hugo/hugolib/shortcode.go

771 lines
21 KiB
Go
Raw Permalink Normal View History

// Copyright 2019 The Hugo Authors. All rights reserved.
2013-07-04 15:32:55 +00:00
//
2015-11-24 03:16:36 +00:00
// Licensed under the Apache License, Version 2.0 (the "License");
2013-07-04 15:32:55 +00:00
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
2015-11-24 03:16:36 +00:00
// http://www.apache.org/licenses/LICENSE-2.0
2013-07-04 15:32:55 +00:00
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package hugolib
import (
"bytes"
"context"
"errors"
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
"fmt"
"html/template"
"path"
"reflect"
"regexp"
"sort"
"strconv"
"strings"
"sync"
"github.com/gohugoio/hugo/common/herrors"
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
"github.com/gohugoio/hugo/common/types"
"github.com/gohugoio/hugo/parser/pageparser"
"github.com/gohugoio/hugo/resources/page"
"github.com/gohugoio/hugo/common/maps"
"github.com/gohugoio/hugo/common/text"
"github.com/gohugoio/hugo/common/urls"
"github.com/gohugoio/hugo/output"
bp "github.com/gohugoio/hugo/bufferpool"
"github.com/gohugoio/hugo/tpl"
)
2013-07-04 15:32:55 +00:00
var (
_ urls.RefLinker = (*ShortcodeWithPage)(nil)
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
_ types.Unwrapper = (*ShortcodeWithPage)(nil)
_ text.Positioner = (*ShortcodeWithPage)(nil)
)
2016-03-24 13:11:04 +00:00
// ShortcodeWithPage is the "." context in a shortcode template.
2013-07-04 15:32:55 +00:00
type ShortcodeWithPage struct {
Params any
Inner template.HTML
Page page.Page
Parent *ShortcodeWithPage
Name string
IsNamedParams bool
// Zero-based ordinal in relation to its parent. If the parent is the page itself,
// this ordinal will represent the position of this shortcode in the page content.
Ordinal int
// Indentation before the opening shortcode in the source.
indentation string
innerDeindentInit sync.Once
innerDeindent template.HTML
// pos is the position in bytes in the source file. Used for error logging.
posInit sync.Once
posOffset int
pos text.Position
scratch *maps.Scratch
2013-07-04 15:32:55 +00:00
}
// InnerDeindent returns the (potentially de-indented) inner content of the shortcode.
func (scp *ShortcodeWithPage) InnerDeindent() template.HTML {
if scp.indentation == "" {
return scp.Inner
}
scp.innerDeindentInit.Do(func() {
b := bp.GetBuffer()
text.VisitLinesAfter(string(scp.Inner), func(s string) {
if strings.HasPrefix(s, scp.indentation) {
b.WriteString(strings.TrimPrefix(s, scp.indentation))
} else {
b.WriteString(s)
}
})
scp.innerDeindent = template.HTML(b.String())
bp.PutBuffer(b)
})
return scp.innerDeindent
}
// Position returns this shortcode's detailed position. Note that this information
// may be expensive to calculate, so only use this in error situations.
func (scp *ShortcodeWithPage) Position() text.Position {
scp.posInit.Do(func() {
if p, ok := mustUnwrapPage(scp.Page).(pageContext); ok {
scp.pos = p.posOffset(scp.posOffset)
}
})
return scp.pos
}
2016-03-24 13:11:04 +00:00
// Site returns information about the current site.
func (scp *ShortcodeWithPage) Site() page.Site {
return scp.Page.Site()
}
// Ref is a shortcut to the Ref method on Page. It passes itself as a context
// to get better error messages.
func (scp *ShortcodeWithPage) Ref(args map[string]any) (string, error) {
return scp.Page.RefFrom(args, scp)
Provide (relative) reference funcs & shortcodes. - `.Ref` and `.RelRef` take a reference (the logical filename for a page, including extension and/or a document fragment ID) and return a permalink (or relative permalink) to the referenced document. - If the reference is a page name (such as `about.md`), the page will be discovered and the permalink will be returned: `/about/` - If the reference is a page name with a fragment (such as `about.md#who`), the page will be discovered and used to add the `page.UniqueID()` to the resulting fragment and permalink: `/about/#who:deadbeef`. - If the reference is a fragment and `.*Ref` has been called from a `Node` or `SiteInfo`, it will be returned as is: `#who`. - If the reference is a fragment and `.*Ref` has been called from a `Page`, it will be returned with the page’s unique ID: `#who:deadbeef`. - `.*Ref` can be called from either `Node`, `SiteInfo` (e.g., `Node.Site`), `Page` objects, or `ShortcodeWithPage` objects in templates. - `.*Ref` cannot be used in content, so two shortcodes have been created to provide the functionality to content: `ref` and `relref`. These are intended to be used within markup, like `[Who]({{% ref about.md#who %}})` or `<a href="{{% ref about.md#who %}}">Who</a>`. - There are also `ref` and `relref` template functions (used to create the shortcodes) that expect a `Page` or `Node` object and the reference string (e.g., `{{ relref . "about.md" }}` or `{{ "about.md" | ref . }}`). It actually looks for `.*Ref` as defined on `Node` or `Page` objects. - Shortcode handling had to use a *differently unique* wrapper in `createShortcodePlaceholder` because of the way that the `ref` and `relref` are intended to be used in content.
2014-11-24 06:15:34 +00:00
}
// RelRef is a shortcut to the RelRef method on Page. It passes itself as a context
// to get better error messages.
func (scp *ShortcodeWithPage) RelRef(args map[string]any) (string, error) {
return scp.Page.RelRefFrom(args, scp)
Provide (relative) reference funcs & shortcodes. - `.Ref` and `.RelRef` take a reference (the logical filename for a page, including extension and/or a document fragment ID) and return a permalink (or relative permalink) to the referenced document. - If the reference is a page name (such as `about.md`), the page will be discovered and the permalink will be returned: `/about/` - If the reference is a page name with a fragment (such as `about.md#who`), the page will be discovered and used to add the `page.UniqueID()` to the resulting fragment and permalink: `/about/#who:deadbeef`. - If the reference is a fragment and `.*Ref` has been called from a `Node` or `SiteInfo`, it will be returned as is: `#who`. - If the reference is a fragment and `.*Ref` has been called from a `Page`, it will be returned with the page’s unique ID: `#who:deadbeef`. - `.*Ref` can be called from either `Node`, `SiteInfo` (e.g., `Node.Site`), `Page` objects, or `ShortcodeWithPage` objects in templates. - `.*Ref` cannot be used in content, so two shortcodes have been created to provide the functionality to content: `ref` and `relref`. These are intended to be used within markup, like `[Who]({{% ref about.md#who %}})` or `<a href="{{% ref about.md#who %}}">Who</a>`. - There are also `ref` and `relref` template functions (used to create the shortcodes) that expect a `Page` or `Node` object and the reference string (e.g., `{{ relref . "about.md" }}` or `{{ "about.md" | ref . }}`). It actually looks for `.*Ref` as defined on `Node` or `Page` objects. - Shortcode handling had to use a *differently unique* wrapper in `createShortcodePlaceholder` because of the way that the `ref` and `relref` are intended to be used in content.
2014-11-24 06:15:34 +00:00
}
2016-03-24 13:11:04 +00:00
// Scratch returns a scratch-pad scoped for this shortcode. This can be used
// as a temporary storage for variables, counters etc.
func (scp *ShortcodeWithPage) Scratch() *maps.Scratch {
2016-03-21 10:10:57 +00:00
if scp.scratch == nil {
scp.scratch = maps.NewScratch()
2016-03-21 10:10:57 +00:00
}
return scp.scratch
2015-04-19 12:50:24 +00:00
}
2016-03-24 13:11:04 +00:00
// Get is a convenience method to look up shortcode parameters by its key.
func (scp *ShortcodeWithPage) Get(key any) any {
if scp.Params == nil {
return nil
}
if reflect.ValueOf(scp.Params).Len() == 0 {
return nil
}
var x reflect.Value
switch key.(type) {
case int64, int32, int16, int8, int:
if reflect.TypeOf(scp.Params).Kind() == reflect.Map {
// We treat this as a non error, so people can do similar to
// {{ $myParam := .Get "myParam" | default .Get 0 }}
// Without having to do additional checks.
return nil
} else if reflect.TypeOf(scp.Params).Kind() == reflect.Slice {
idx := int(reflect.ValueOf(key).Int())
ln := reflect.ValueOf(scp.Params).Len()
if idx > ln-1 {
return ""
}
x = reflect.ValueOf(scp.Params).Index(idx)
}
case string:
if reflect.TypeOf(scp.Params).Kind() == reflect.Map {
x = reflect.ValueOf(scp.Params).MapIndex(reflect.ValueOf(key))
if !x.IsValid() {
return ""
}
} else if reflect.TypeOf(scp.Params).Kind() == reflect.Slice {
// We treat this as a non error, so people can do similar to
// {{ $myParam := .Get "myParam" | default .Get 0 }}
// Without having to do additional checks.
return nil
}
}
return x.Interface()
}
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
// For internal use only.
func (scp *ShortcodeWithPage) Unwrapv() any {
return scp.Page
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
// Note - this value must not contain any markup syntax
const shortcodePlaceholderPrefix = "HAHAHUGOSHORTCODE"
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
func createShortcodePlaceholder(sid string, id uint64, ordinal int) string {
return shortcodePlaceholderPrefix + strconv.FormatUint(id, 10) + sid + strconv.Itoa(ordinal) + "HBHB"
}
2013-07-04 15:32:55 +00:00
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
type shortcode struct {
name string
isInline bool // inline shortcode. Any inner will be a Go template.
isClosing bool // whether a closing tag was provided
inner []any // string or nested shortcode
params any // map or array
ordinal int
indentation string // indentation from source.
info tpl.Info // One of the output formats (arbitrary)
templs []tpl.Template // All output formats
// If set, the rendered shortcode is sent as part of the surrounding content
// to Goldmark and similar.
// Before Hug0 0.55 we didn't send any shortcode output to the markup
// renderer, and this flag told Hugo to process the {{ .Inner }} content
// separately.
2024-02-18 11:16:30 +00:00
// The old behavior can be had by starting your shortcode template with:
// {{ $_hugo_config := `{ "version": 1 }`}}
doMarkup bool
// the placeholder in the source when passed to Goldmark etc.
// This also identifies the rendered shortcode.
placeholder string
pos int // the position in bytes in the source file
length int // the length in bytes in the source file
2013-07-04 15:32:55 +00:00
}
func (s shortcode) insertPlaceholder() bool {
return !s.doMarkup || s.configVersion() == 1
}
func (s shortcode) needsInner() bool {
return s.info != nil && s.info.ParseInfo().IsInner
}
func (s shortcode) configVersion() int {
if s.info == nil {
// Not set for inline shortcodes.
return 2
}
return s.info.ParseInfo().Config.Version
}
func (s shortcode) innerString() string {
var sb strings.Builder
for _, inner := range s.inner {
sb.WriteString(inner.(string))
}
return sb.String()
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
func (sc shortcode) String() string {
// for testing (mostly), so any change here will break tests!
var params any
switch v := sc.params.(type) {
case map[string]any:
// sort the keys so test assertions won't fail
var keys []string
for k := range v {
keys = append(keys, k)
}
sort.Strings(keys)
tmp := make(map[string]any)
for _, k := range keys {
tmp[k] = v[k]
}
params = tmp
default:
// use it as is
params = sc.params
}
return fmt.Sprintf("%s(%q, %t){%s}", sc.name, params, sc.doMarkup, sc.inner)
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
}
type shortcodeHandler struct {
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
filename string
s *Site
// Ordered list of shortcodes for a page.
shortcodes []*shortcode
// All the shortcode names in this set.
nameSet map[string]bool
nameSetMu sync.RWMutex
// Configuration
enableInlineShortcodes bool
}
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
func newShortcodeHandler(filename string, s *Site) *shortcodeHandler {
sh := &shortcodeHandler{
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
filename: filename,
s: s,
enableInlineShortcodes: s.ExecHelper.Sec().EnableInlineShortcodes,
shortcodes: make([]*shortcode, 0, 4),
nameSet: make(map[string]bool),
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
return sh
}
const (
innerNewlineRegexp = "\n"
innerCleanupRegexp = `\A<p>(.*)</p>\n\z`
innerCleanupExpand = "$1"
)
func prepareShortcode(
ctx context.Context,
level int,
s *Site,
tplVariants tpl.TemplateVariants,
sc *shortcode,
parent *ShortcodeWithPage,
p *pageState,
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
isRenderString bool,
) (shortcodeRenderer, error) {
toParseErr := func(err error) error {
source := p.m.content.mustSource()
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
return p.parseError(fmt.Errorf("failed to render shortcode %q: %w", sc.name, err), source, sc.pos)
}
// Allow the caller to delay the rendering of the shortcode if needed.
var fn shortcodeRenderFunc = func(ctx context.Context) ([]byte, bool, error) {
if p.m.pageConfig.IsGoldmark && sc.doMarkup {
// Signal downwards that the content rendered will be
// parsed and rendered by Goldmark.
ctx = tpl.Context.IsInGoldmark.Set(ctx, true)
}
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
r, err := doRenderShortcode(ctx, level, s, tplVariants, sc, parent, p, isRenderString)
if err != nil {
return nil, false, toParseErr(err)
}
b, hasVariants, err := r.renderShortcode(ctx)
if err != nil {
return nil, false, toParseErr(err)
}
return b, hasVariants, nil
}
return fn, nil
}
func doRenderShortcode(
ctx context.Context,
level int,
s *Site,
tplVariants tpl.TemplateVariants,
sc *shortcode,
parent *ShortcodeWithPage,
p *pageState,
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
isRenderString bool,
) (shortcodeRenderer, error) {
var tmpl tpl.Template
// Tracks whether this shortcode or any of its children has template variations
// in other languages or output formats. We are currently only interested in
// the output formats, so we may get some false positives -- we
// should improve on that.
var hasVariants bool
if sc.isInline {
if !p.s.ExecHelper.Sec().EnableInlineShortcodes {
return zeroShortcode, nil
}
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
templName := path.Join("_inline_shortcode", p.Path(), sc.name)
if sc.isClosing {
templStr := sc.innerString()
var err error
tpl/tplimpl: Rework template management to get rid of concurrency issues This more or less completes the simplification of the template handling code in Hugo started in v0.62. The main motivation was to fix a long lasting issue about a crash in HTML content files without front matter. But this commit also comes with a big functional improvement. As we now have moved the base template evaluation to the build stage we now use the same lookup rules for `baseof` as for `list` etc. type of templates. This means that in this simple example you can have a `baseof` template for the `blog` section without having to duplicate the others: ``` layouts ├── _default │   ├── baseof.html │   ├── list.html │   └── single.html └── blog └── baseof.html ``` Also, when simplifying code, you often get rid of some double work, as shown in the "site building" benchmarks below. These benchmarks looks suspiciously good, but I have repeated the below with ca. the same result. Compared to master: ``` name old time/op new time/op delta SiteNew/Bundle_with_image-16 13.1ms ± 1% 10.5ms ± 1% -19.34% (p=0.029 n=4+4) SiteNew/Bundle_with_JSON_file-16 13.0ms ± 0% 10.7ms ± 1% -18.05% (p=0.029 n=4+4) SiteNew/Tags_and_categories-16 46.4ms ± 2% 43.1ms ± 1% -7.15% (p=0.029 n=4+4) SiteNew/Canonify_URLs-16 52.2ms ± 2% 47.8ms ± 1% -8.30% (p=0.029 n=4+4) SiteNew/Deep_content_tree-16 77.9ms ± 1% 70.9ms ± 1% -9.01% (p=0.029 n=4+4) SiteNew/Many_HTML_templates-16 43.0ms ± 0% 37.2ms ± 1% -13.54% (p=0.029 n=4+4) SiteNew/Page_collections-16 58.2ms ± 1% 52.4ms ± 1% -9.95% (p=0.029 n=4+4) name old alloc/op new alloc/op delta SiteNew/Bundle_with_image-16 3.81MB ± 0% 2.22MB ± 0% -41.70% (p=0.029 n=4+4) SiteNew/Bundle_with_JSON_file-16 3.60MB ± 0% 2.01MB ± 0% -44.20% (p=0.029 n=4+4) SiteNew/Tags_and_categories-16 19.3MB ± 1% 14.1MB ± 0% -26.91% (p=0.029 n=4+4) SiteNew/Canonify_URLs-16 70.7MB ± 0% 69.0MB ± 0% -2.40% (p=0.029 n=4+4) SiteNew/Deep_content_tree-16 37.1MB ± 0% 31.2MB ± 0% -15.94% (p=0.029 n=4+4) SiteNew/Many_HTML_templates-16 17.6MB ± 0% 10.6MB ± 0% -39.92% (p=0.029 n=4+4) SiteNew/Page_collections-16 25.9MB ± 0% 21.2MB ± 0% -17.99% (p=0.029 n=4+4) name old allocs/op new allocs/op delta SiteNew/Bundle_with_image-16 52.3k ± 0% 26.1k ± 0% -50.18% (p=0.029 n=4+4) SiteNew/Bundle_with_JSON_file-16 52.3k ± 0% 26.1k ± 0% -50.16% (p=0.029 n=4+4) SiteNew/Tags_and_categories-16 336k ± 1% 269k ± 0% -19.90% (p=0.029 n=4+4) SiteNew/Canonify_URLs-16 422k ± 0% 395k ± 0% -6.43% (p=0.029 n=4+4) SiteNew/Deep_content_tree-16 401k ± 0% 313k ± 0% -21.79% (p=0.029 n=4+4) SiteNew/Many_HTML_templates-16 247k ± 0% 143k ± 0% -42.17% (p=0.029 n=4+4) SiteNew/Page_collections-16 282k ± 0% 207k ± 0% -26.55% (p=0.029 n=4+4) ``` Fixes #6716 Fixes #6760 Fixes #6768 Fixes #6778
2020-01-15 14:59:56 +00:00
tmpl, err = s.TextTmpl().Parse(templName, templStr)
if err != nil {
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
if isRenderString {
return zeroShortcode, p.wrapError(err)
}
2022-05-15 09:40:34 +00:00
fe := herrors.NewFileErrorFromName(err, p.File().Filename())
pos := fe.Position()
pos.LineNumber += p.posOffset(sc.pos).LineNumber
fe = fe.UpdatePosition(pos)
return zeroShortcode, p.wrapError(fe)
}
} else {
// Re-use of shortcode defined earlier in the same page.
var found bool
tpl/tplimpl: Rework template management to get rid of concurrency issues This more or less completes the simplification of the template handling code in Hugo started in v0.62. The main motivation was to fix a long lasting issue about a crash in HTML content files without front matter. But this commit also comes with a big functional improvement. As we now have moved the base template evaluation to the build stage we now use the same lookup rules for `baseof` as for `list` etc. type of templates. This means that in this simple example you can have a `baseof` template for the `blog` section without having to duplicate the others: ``` layouts ├── _default │   ├── baseof.html │   ├── list.html │   └── single.html └── blog └── baseof.html ``` Also, when simplifying code, you often get rid of some double work, as shown in the "site building" benchmarks below. These benchmarks looks suspiciously good, but I have repeated the below with ca. the same result. Compared to master: ``` name old time/op new time/op delta SiteNew/Bundle_with_image-16 13.1ms ± 1% 10.5ms ± 1% -19.34% (p=0.029 n=4+4) SiteNew/Bundle_with_JSON_file-16 13.0ms ± 0% 10.7ms ± 1% -18.05% (p=0.029 n=4+4) SiteNew/Tags_and_categories-16 46.4ms ± 2% 43.1ms ± 1% -7.15% (p=0.029 n=4+4) SiteNew/Canonify_URLs-16 52.2ms ± 2% 47.8ms ± 1% -8.30% (p=0.029 n=4+4) SiteNew/Deep_content_tree-16 77.9ms ± 1% 70.9ms ± 1% -9.01% (p=0.029 n=4+4) SiteNew/Many_HTML_templates-16 43.0ms ± 0% 37.2ms ± 1% -13.54% (p=0.029 n=4+4) SiteNew/Page_collections-16 58.2ms ± 1% 52.4ms ± 1% -9.95% (p=0.029 n=4+4) name old alloc/op new alloc/op delta SiteNew/Bundle_with_image-16 3.81MB ± 0% 2.22MB ± 0% -41.70% (p=0.029 n=4+4) SiteNew/Bundle_with_JSON_file-16 3.60MB ± 0% 2.01MB ± 0% -44.20% (p=0.029 n=4+4) SiteNew/Tags_and_categories-16 19.3MB ± 1% 14.1MB ± 0% -26.91% (p=0.029 n=4+4) SiteNew/Canonify_URLs-16 70.7MB ± 0% 69.0MB ± 0% -2.40% (p=0.029 n=4+4) SiteNew/Deep_content_tree-16 37.1MB ± 0% 31.2MB ± 0% -15.94% (p=0.029 n=4+4) SiteNew/Many_HTML_templates-16 17.6MB ± 0% 10.6MB ± 0% -39.92% (p=0.029 n=4+4) SiteNew/Page_collections-16 25.9MB ± 0% 21.2MB ± 0% -17.99% (p=0.029 n=4+4) name old allocs/op new allocs/op delta SiteNew/Bundle_with_image-16 52.3k ± 0% 26.1k ± 0% -50.18% (p=0.029 n=4+4) SiteNew/Bundle_with_JSON_file-16 52.3k ± 0% 26.1k ± 0% -50.16% (p=0.029 n=4+4) SiteNew/Tags_and_categories-16 336k ± 1% 269k ± 0% -19.90% (p=0.029 n=4+4) SiteNew/Canonify_URLs-16 422k ± 0% 395k ± 0% -6.43% (p=0.029 n=4+4) SiteNew/Deep_content_tree-16 401k ± 0% 313k ± 0% -21.79% (p=0.029 n=4+4) SiteNew/Many_HTML_templates-16 247k ± 0% 143k ± 0% -42.17% (p=0.029 n=4+4) SiteNew/Page_collections-16 282k ± 0% 207k ± 0% -26.55% (p=0.029 n=4+4) ``` Fixes #6716 Fixes #6760 Fixes #6768 Fixes #6778
2020-01-15 14:59:56 +00:00
tmpl, found = s.TextTmpl().Lookup(templName)
if !found {
return zeroShortcode, fmt.Errorf("no earlier definition of shortcode %q found", sc.name)
}
}
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
tmpl = tpl.AddIdentity(tmpl)
} else {
var found, more bool
tpl/tplimpl: Rework template management to get rid of concurrency issues This more or less completes the simplification of the template handling code in Hugo started in v0.62. The main motivation was to fix a long lasting issue about a crash in HTML content files without front matter. But this commit also comes with a big functional improvement. As we now have moved the base template evaluation to the build stage we now use the same lookup rules for `baseof` as for `list` etc. type of templates. This means that in this simple example you can have a `baseof` template for the `blog` section without having to duplicate the others: ``` layouts ├── _default │   ├── baseof.html │   ├── list.html │   └── single.html └── blog └── baseof.html ``` Also, when simplifying code, you often get rid of some double work, as shown in the "site building" benchmarks below. These benchmarks looks suspiciously good, but I have repeated the below with ca. the same result. Compared to master: ``` name old time/op new time/op delta SiteNew/Bundle_with_image-16 13.1ms ± 1% 10.5ms ± 1% -19.34% (p=0.029 n=4+4) SiteNew/Bundle_with_JSON_file-16 13.0ms ± 0% 10.7ms ± 1% -18.05% (p=0.029 n=4+4) SiteNew/Tags_and_categories-16 46.4ms ± 2% 43.1ms ± 1% -7.15% (p=0.029 n=4+4) SiteNew/Canonify_URLs-16 52.2ms ± 2% 47.8ms ± 1% -8.30% (p=0.029 n=4+4) SiteNew/Deep_content_tree-16 77.9ms ± 1% 70.9ms ± 1% -9.01% (p=0.029 n=4+4) SiteNew/Many_HTML_templates-16 43.0ms ± 0% 37.2ms ± 1% -13.54% (p=0.029 n=4+4) SiteNew/Page_collections-16 58.2ms ± 1% 52.4ms ± 1% -9.95% (p=0.029 n=4+4) name old alloc/op new alloc/op delta SiteNew/Bundle_with_image-16 3.81MB ± 0% 2.22MB ± 0% -41.70% (p=0.029 n=4+4) SiteNew/Bundle_with_JSON_file-16 3.60MB ± 0% 2.01MB ± 0% -44.20% (p=0.029 n=4+4) SiteNew/Tags_and_categories-16 19.3MB ± 1% 14.1MB ± 0% -26.91% (p=0.029 n=4+4) SiteNew/Canonify_URLs-16 70.7MB ± 0% 69.0MB ± 0% -2.40% (p=0.029 n=4+4) SiteNew/Deep_content_tree-16 37.1MB ± 0% 31.2MB ± 0% -15.94% (p=0.029 n=4+4) SiteNew/Many_HTML_templates-16 17.6MB ± 0% 10.6MB ± 0% -39.92% (p=0.029 n=4+4) SiteNew/Page_collections-16 25.9MB ± 0% 21.2MB ± 0% -17.99% (p=0.029 n=4+4) name old allocs/op new allocs/op delta SiteNew/Bundle_with_image-16 52.3k ± 0% 26.1k ± 0% -50.18% (p=0.029 n=4+4) SiteNew/Bundle_with_JSON_file-16 52.3k ± 0% 26.1k ± 0% -50.16% (p=0.029 n=4+4) SiteNew/Tags_and_categories-16 336k ± 1% 269k ± 0% -19.90% (p=0.029 n=4+4) SiteNew/Canonify_URLs-16 422k ± 0% 395k ± 0% -6.43% (p=0.029 n=4+4) SiteNew/Deep_content_tree-16 401k ± 0% 313k ± 0% -21.79% (p=0.029 n=4+4) SiteNew/Many_HTML_templates-16 247k ± 0% 143k ± 0% -42.17% (p=0.029 n=4+4) SiteNew/Page_collections-16 282k ± 0% 207k ± 0% -26.55% (p=0.029 n=4+4) ``` Fixes #6716 Fixes #6760 Fixes #6768 Fixes #6778
2020-01-15 14:59:56 +00:00
tmpl, found, more = s.Tmpl().LookupVariant(sc.name, tplVariants)
if !found {
s.Log.Errorf("Unable to locate template for shortcode %q in page %q", sc.name, p.File().Path())
return zeroShortcode, nil
}
hasVariants = hasVariants || more
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
}
data := &ShortcodeWithPage{Ordinal: sc.ordinal, posOffset: sc.pos, indentation: sc.indentation, Params: sc.params, Page: newPageForShortcode(p), Parent: parent, Name: sc.name}
if sc.params != nil {
data.IsNamedParams = reflect.TypeOf(sc.params).Kind() == reflect.Map
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
if len(sc.inner) > 0 {
var inner string
for _, innerData := range sc.inner {
switch innerData := innerData.(type) {
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
case string:
inner += innerData
case *shortcode:
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
s, err := prepareShortcode(ctx, level+1, s, tplVariants, innerData, data, p, isRenderString)
if err != nil {
return zeroShortcode, err
}
ss, more, err := s.renderShortcodeString(ctx)
hasVariants = hasVariants || more
if err != nil {
return zeroShortcode, err
}
inner += ss
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
default:
s.Log.Errorf("Illegal state on shortcode rendering of %q in page %q. Illegal type in inner data: %s ",
sc.name, p.File().Path(), reflect.TypeOf(innerData))
return zeroShortcode, nil
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
}
2024-02-18 11:16:30 +00:00
// Pre Hugo 0.55 this was the behavior even for the outer-most
// shortcode.
if sc.doMarkup && (level > 0 || sc.configVersion() == 1) {
var err error
b, err := p.pageOutput.contentRenderer.ParseAndRenderContent(ctx, []byte(inner), false)
if err != nil {
return zeroShortcode, err
}
newInner := b.Bytes()
Provide (relative) reference funcs & shortcodes. - `.Ref` and `.RelRef` take a reference (the logical filename for a page, including extension and/or a document fragment ID) and return a permalink (or relative permalink) to the referenced document. - If the reference is a page name (such as `about.md`), the page will be discovered and the permalink will be returned: `/about/` - If the reference is a page name with a fragment (such as `about.md#who`), the page will be discovered and used to add the `page.UniqueID()` to the resulting fragment and permalink: `/about/#who:deadbeef`. - If the reference is a fragment and `.*Ref` has been called from a `Node` or `SiteInfo`, it will be returned as is: `#who`. - If the reference is a fragment and `.*Ref` has been called from a `Page`, it will be returned with the page’s unique ID: `#who:deadbeef`. - `.*Ref` can be called from either `Node`, `SiteInfo` (e.g., `Node.Site`), `Page` objects, or `ShortcodeWithPage` objects in templates. - `.*Ref` cannot be used in content, so two shortcodes have been created to provide the functionality to content: `ref` and `relref`. These are intended to be used within markup, like `[Who]({{% ref about.md#who %}})` or `<a href="{{% ref about.md#who %}}">Who</a>`. - There are also `ref` and `relref` template functions (used to create the shortcodes) that expect a `Page` or `Node` object and the reference string (e.g., `{{ relref . "about.md" }}` or `{{ "about.md" | ref . }}`). It actually looks for `.*Ref` as defined on `Node` or `Page` objects. - Shortcode handling had to use a *differently unique* wrapper in `createShortcodePlaceholder` because of the way that the `ref` and `relref` are intended to be used in content.
2014-11-24 06:15:34 +00:00
// If the type is “” (unknown) or “markdown”, we assume the markdown
Provide (relative) reference funcs & shortcodes. - `.Ref` and `.RelRef` take a reference (the logical filename for a page, including extension and/or a document fragment ID) and return a permalink (or relative permalink) to the referenced document. - If the reference is a page name (such as `about.md`), the page will be discovered and the permalink will be returned: `/about/` - If the reference is a page name with a fragment (such as `about.md#who`), the page will be discovered and used to add the `page.UniqueID()` to the resulting fragment and permalink: `/about/#who:deadbeef`. - If the reference is a fragment and `.*Ref` has been called from a `Node` or `SiteInfo`, it will be returned as is: `#who`. - If the reference is a fragment and `.*Ref` has been called from a `Page`, it will be returned with the page’s unique ID: `#who:deadbeef`. - `.*Ref` can be called from either `Node`, `SiteInfo` (e.g., `Node.Site`), `Page` objects, or `ShortcodeWithPage` objects in templates. - `.*Ref` cannot be used in content, so two shortcodes have been created to provide the functionality to content: `ref` and `relref`. These are intended to be used within markup, like `[Who]({{% ref about.md#who %}})` or `<a href="{{% ref about.md#who %}}">Who</a>`. - There are also `ref` and `relref` template functions (used to create the shortcodes) that expect a `Page` or `Node` object and the reference string (e.g., `{{ relref . "about.md" }}` or `{{ "about.md" | ref . }}`). It actually looks for `.*Ref` as defined on `Node` or `Page` objects. - Shortcode handling had to use a *differently unique* wrapper in `createShortcodePlaceholder` because of the way that the `ref` and `relref` are intended to be used in content.
2014-11-24 06:15:34 +00:00
// generation has been performed. Given the input: `a line`, markdown
// specifies the HTML `<p>a line</p>\n`. When dealing with documents as a
// whole, this is OK. When dealing with an `{{ .Inner }}` block in Hugo,
// this is not so good. This code does two things:
//
// 1. Check to see if inner has a newline in it. If so, the Inner data is
// unchanged.
// 2 If inner does not have a newline, strip the wrapping <p> block and
// the newline.
switch p.m.pageConfig.Markup {
case "", "markdown":
Provide (relative) reference funcs & shortcodes. - `.Ref` and `.RelRef` take a reference (the logical filename for a page, including extension and/or a document fragment ID) and return a permalink (or relative permalink) to the referenced document. - If the reference is a page name (such as `about.md`), the page will be discovered and the permalink will be returned: `/about/` - If the reference is a page name with a fragment (such as `about.md#who`), the page will be discovered and used to add the `page.UniqueID()` to the resulting fragment and permalink: `/about/#who:deadbeef`. - If the reference is a fragment and `.*Ref` has been called from a `Node` or `SiteInfo`, it will be returned as is: `#who`. - If the reference is a fragment and `.*Ref` has been called from a `Page`, it will be returned with the page’s unique ID: `#who:deadbeef`. - `.*Ref` can be called from either `Node`, `SiteInfo` (e.g., `Node.Site`), `Page` objects, or `ShortcodeWithPage` objects in templates. - `.*Ref` cannot be used in content, so two shortcodes have been created to provide the functionality to content: `ref` and `relref`. These are intended to be used within markup, like `[Who]({{% ref about.md#who %}})` or `<a href="{{% ref about.md#who %}}">Who</a>`. - There are also `ref` and `relref` template functions (used to create the shortcodes) that expect a `Page` or `Node` object and the reference string (e.g., `{{ relref . "about.md" }}` or `{{ "about.md" | ref . }}`). It actually looks for `.*Ref` as defined on `Node` or `Page` objects. - Shortcode handling had to use a *differently unique* wrapper in `createShortcodePlaceholder` because of the way that the `ref` and `relref` are intended to be used in content.
2014-11-24 06:15:34 +00:00
if match, _ := regexp.MatchString(innerNewlineRegexp, inner); !match {
cleaner, err := regexp.Compile(innerCleanupRegexp)
if err == nil {
newInner = cleaner.ReplaceAll(newInner, []byte(innerCleanupExpand))
}
}
}
// TODO(bep) we may have plain text inner templates.
Provide (relative) reference funcs & shortcodes. - `.Ref` and `.RelRef` take a reference (the logical filename for a page, including extension and/or a document fragment ID) and return a permalink (or relative permalink) to the referenced document. - If the reference is a page name (such as `about.md`), the page will be discovered and the permalink will be returned: `/about/` - If the reference is a page name with a fragment (such as `about.md#who`), the page will be discovered and used to add the `page.UniqueID()` to the resulting fragment and permalink: `/about/#who:deadbeef`. - If the reference is a fragment and `.*Ref` has been called from a `Node` or `SiteInfo`, it will be returned as is: `#who`. - If the reference is a fragment and `.*Ref` has been called from a `Page`, it will be returned with the page’s unique ID: `#who:deadbeef`. - `.*Ref` can be called from either `Node`, `SiteInfo` (e.g., `Node.Site`), `Page` objects, or `ShortcodeWithPage` objects in templates. - `.*Ref` cannot be used in content, so two shortcodes have been created to provide the functionality to content: `ref` and `relref`. These are intended to be used within markup, like `[Who]({{% ref about.md#who %}})` or `<a href="{{% ref about.md#who %}}">Who</a>`. - There are also `ref` and `relref` template functions (used to create the shortcodes) that expect a `Page` or `Node` object and the reference string (e.g., `{{ relref . "about.md" }}` or `{{ "about.md" | ref . }}`). It actually looks for `.*Ref` as defined on `Node` or `Page` objects. - Shortcode handling had to use a *differently unique* wrapper in `createShortcodePlaceholder` because of the way that the `ref` and `relref` are intended to be used in content.
2014-11-24 06:15:34 +00:00
data.Inner = template.HTML(newInner)
} else {
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
data.Inner = template.HTML(inner)
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
}
result, err := renderShortcodeWithPage(ctx, s.Tmpl(), tmpl, data)
if err != nil && sc.isInline {
2022-05-15 09:40:34 +00:00
fe := herrors.NewFileErrorFromName(err, p.File().Filename())
pos := fe.Position()
pos.LineNumber += p.posOffset(sc.pos).LineNumber
fe = fe.UpdatePosition(pos)
return zeroShortcode, fe
}
if len(sc.inner) == 0 && len(sc.indentation) > 0 {
b := bp.GetBuffer()
i := 0
text.VisitLinesAfter(result, func(line string) {
// The first line is correctly indented.
if i > 0 {
b.WriteString(sc.indentation)
}
i++
b.WriteString(line)
})
result = b.String()
bp.PutBuffer(b)
}
return prerenderedShortcode{s: result, hasVariants: hasVariants}, err
}
func (s *shortcodeHandler) addName(name string) {
s.nameSetMu.Lock()
defer s.nameSetMu.Unlock()
s.nameSet[name] = true
}
func (s *shortcodeHandler) transferNames(in *shortcodeHandler) {
s.nameSetMu.Lock()
defer s.nameSetMu.Unlock()
for k := range in.nameSet {
s.nameSet[k] = true
}
}
func (s *shortcodeHandler) hasName(name string) bool {
s.nameSetMu.RLock()
defer s.nameSetMu.RUnlock()
_, ok := s.nameSet[name]
return ok
}
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
func (s *shortcodeHandler) prepareShortcodesForPage(ctx context.Context, p *pageState, f output.Format, isRenderString bool) (map[string]shortcodeRenderer, error) {
rendered := make(map[string]shortcodeRenderer)
tplVariants := tpl.TemplateVariants{
Language: p.Language().Lang,
OutputFormat: f,
}
for _, v := range s.shortcodes {
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
s, err := prepareShortcode(ctx, 0, s.s, tplVariants, v, nil, p, isRenderString)
if err != nil {
return nil, err
}
rendered[v.placeholder] = s
}
return rendered, nil
}
all: Rework page store, add a dynacache, improve partial rebuilds, and some general spring cleaning There are some breaking changes in this commit, see #11455. Closes #11455 Closes #11549 This fixes a set of bugs (see issue list) and it is also paying some technical debt accumulated over the years. We now build with Staticcheck enabled in the CI build. The performance should be about the same as before for regular sized Hugo sites, but it should perform and scale much better to larger data sets, as objects that uses lots of memory (e.g. rendered Markdown, big JSON files read into maps with transform.Unmarshal etc.) will now get automatically garbage collected if needed. Performance on partial rebuilds when running the server in fast render mode should be the same, but the change detection should be much more accurate. A list of the notable new features: * A new dependency tracker that covers (almost) all of Hugo's API and is used to do fine grained partial rebuilds when running the server. * A new and simpler tree document store which allows fast lookups and prefix-walking in all dimensions (e.g. language) concurrently. * You can now configure an upper memory limit allowing for much larger data sets and/or running on lower specced PCs. We have lifted the "no resources in sub folders" restriction for branch bundles (e.g. sections). Memory Limit * Hugos will, by default, set aside a quarter of the total system memory, but you can set this via the OS environment variable HUGO_MEMORYLIMIT (in gigabytes). This is backed by a partitioned LRU cache used throughout Hugo. A cache that gets dynamically resized in low memory situations, allowing Go's Garbage Collector to free the memory. New Dependency Tracker: Hugo has had a rule based coarse grained approach to server rebuilds that has worked mostly pretty well, but there have been some surprises (e.g. stale content). This is now revamped with a new dependency tracker that can quickly calculate the delta given a changed resource (e.g. a content file, template, JS file etc.). This handles transitive relations, e.g. $page -> js.Build -> JS import, or $page1.Content -> render hook -> site.GetPage -> $page2.Title, or $page1.Content -> shortcode -> partial -> site.RegularPages -> $page2.Content -> shortcode ..., and should also handle changes to aggregated values (e.g. site.Lastmod) effectively. This covers all of Hugo's API with 2 known exceptions (a list that may not be fully exhaustive): Changes to files loaded with template func os.ReadFile may not be handled correctly. We recommend loading resources with resources.Get Changes to Hugo objects (e.g. Page) passed in the template context to lang.Translate may not be detected correctly. We recommend having simple i18n templates without too much data context passed in other than simple types such as strings and numbers. Note that the cachebuster configuration (when A changes then rebuild B) works well with the above, but we recommend that you revise that configuration, as it in most situations should not be needed. One example where it is still needed is with TailwindCSS and using changes to hugo_stats.json to trigger new CSS rebuilds. Document Store: Previously, a little simplified, we split the document store (where we store pages and resources) in a tree per language. This worked pretty well, but the structure made some operations harder than they needed to be. We have now restructured it into one Radix tree for all languages. Internally the language is considered to be a dimension of that tree, and the tree can be viewed in all dimensions concurrently. This makes some operations re. language simpler (e.g. finding translations is just a slice range), but the idea is that it should also be relatively inexpensive to add more dimensions if needed (e.g. role). Fixes #10169 Fixes #10364 Fixes #10482 Fixes #10630 Fixes #10656 Fixes #10694 Fixes #10918 Fixes #11262 Fixes #11439 Fixes #11453 Fixes #11457 Fixes #11466 Fixes #11540 Fixes #11551 Fixes #11556 Fixes #11654 Fixes #11661 Fixes #11663 Fixes #11664 Fixes #11669 Fixes #11671 Fixes #11807 Fixes #11808 Fixes #11809 Fixes #11815 Fixes #11840 Fixes #11853 Fixes #11860 Fixes #11883 Fixes #11904 Fixes #7388 Fixes #7425 Fixes #7436 Fixes #7544 Fixes #7882 Fixes #7960 Fixes #8255 Fixes #8307 Fixes #8863 Fixes #8927 Fixes #9192 Fixes #9324
2023-12-24 18:11:05 +00:00
func posFromInput(filename string, input []byte, offset int) text.Position {
if offset < 0 {
return text.Position{
Filename: filename,
}
}
lf := []byte("\n")
input = input[:offset]
lineNumber := bytes.Count(input, lf) + 1
endOfLastLine := bytes.LastIndex(input, lf)
return text.Position{
Filename: filename,
LineNumber: lineNumber,
ColumnNumber: offset - endOfLastLine,
Offset: offset,
}
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
// pageTokens state:
// - before: positioned just before the shortcode start
// - after: shortcode(s) consumed (plural when they are nested)
func (s *shortcodeHandler) extractShortcode(ordinal, level int, source []byte, pt *pageparser.Iterator) (*shortcode, error) {
if s == nil {
panic("handler nil")
}
sc := &shortcode{ordinal: ordinal}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
// Back up one to identify any indentation.
if pt.Pos() > 0 {
pt.Backup()
item := pt.Next()
if item.IsIndentation() {
sc.indentation = item.ValStr(source)
}
}
cnt := 0
nestedOrdinal := 0
nextLevel := level + 1
closed := false
const errorPrefix = "failed to extract shortcode"
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
Loop:
for {
currItem := pt.Next()
switch {
case currItem.IsLeftShortcodeDelim():
next := pt.Peek()
if next.IsRightShortcodeDelim() {
// no name: {{< >}} or {{% %}}
return sc, errors.New("shortcode has no name")
}
if next.IsShortcodeClose() {
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
continue
2014-01-29 22:50:31 +00:00
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
if cnt > 0 {
// nested shortcode; append it to inner content
pt.Backup()
nested, err := s.extractShortcode(nestedOrdinal, nextLevel, source, pt)
nestedOrdinal++
if nested != nil && nested.name != "" {
s.addName(nested.name)
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
if err == nil {
sc.inner = append(sc.inner, nested)
} else {
return sc, err
}
2014-01-29 22:50:31 +00:00
} else {
sc.doMarkup = currItem.IsShortcodeMarkupDelimiter()
2014-01-29 22:50:31 +00:00
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
cnt++
case currItem.IsRightShortcodeDelim():
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
// we trust the template on this:
// if there's no inner, we're done
if !sc.isInline {
if !sc.info.ParseInfo().IsInner {
return sc, nil
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
}
case currItem.IsShortcodeClose():
closed = true
next := pt.Peek()
if !sc.isInline {
if !sc.needsInner() {
if next.IsError() {
// return that error, more specific
continue
}
return nil, fmt.Errorf("%s: shortcode %q does not evaluate .Inner or .InnerDeindent, yet a closing tag was provided", errorPrefix, next.ValStr(source))
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
}
}
if next.IsRightShortcodeDelim() {
// self-closing
pt.Consume(1)
} else {
sc.isClosing = true
pt.Consume(2)
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
return sc, nil
case currItem.IsText():
sc.inner = append(sc.inner, currItem.ValStr(source))
case currItem.IsShortcodeName():
sc.name = currItem.ValStr(source)
// Used to check if the template expects inner content.
templs := s.s.Tmpl().LookupVariants(sc.name)
if templs == nil {
return nil, fmt.Errorf("%s: template for shortcode %q not found", errorPrefix, sc.name)
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
sc.info = templs[0].(tpl.Info)
sc.templs = templs
case currItem.IsInlineShortcodeName():
sc.name = currItem.ValStr(source)
sc.isInline = true
case currItem.IsShortcodeParam():
if !pt.IsValueNext() {
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
continue
} else if pt.Peek().IsShortcodeParamVal() {
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
// named params
if sc.params == nil {
params := make(map[string]any)
params[currItem.ValStr(source)] = pt.Next().ValTyped(source)
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
sc.params = params
2014-01-29 22:50:31 +00:00
} else {
if params, ok := sc.params.(map[string]any); ok {
params[currItem.ValStr(source)] = pt.Next().ValTyped(source)
} else {
return sc, fmt.Errorf("%s: invalid state: invalid param type %T for shortcode %q, expected a map", errorPrefix, params, sc.name)
}
2014-01-29 22:50:31 +00:00
}
} else {
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
// positional params
if sc.params == nil {
var params []any
params = append(params, currItem.ValTyped(source))
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
sc.params = params
2014-01-29 22:50:31 +00:00
} else {
if params, ok := sc.params.([]any); ok {
params = append(params, currItem.ValTyped(source))
sc.params = params
} else {
return sc, fmt.Errorf("%s: invalid state: invalid param type %T for shortcode %q, expected a slice", errorPrefix, params, sc.name)
}
2014-01-29 22:50:31 +00:00
}
}
case currItem.IsDone():
if !currItem.IsError() {
if !closed && sc.needsInner() {
return sc, fmt.Errorf("%s: shortcode %q must be closed or self-closed", errorPrefix, sc.name)
}
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
// handled by caller
pt.Backup()
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
break Loop
2014-01-29 22:50:31 +00:00
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
}
return sc, nil
}
// Replace prefixed shortcode tokens with the real content.
// Note: This function will rewrite the input slice.
func expandShortcodeTokens(
ctx context.Context,
source []byte,
tokenHandler func(ctx context.Context, token string) ([]byte, error),
) ([]byte, error) {
start := 0
pre := []byte(shortcodePlaceholderPrefix)
post := []byte("HBHB")
pStart := []byte("<p>")
pEnd := []byte("</p>")
k := bytes.Index(source[start:], pre)
for k != -1 {
j := start + k
postIdx := bytes.Index(source[j:], post)
if postIdx < 0 {
// this should never happen, but let the caller decide to panic or not
return nil, errors.New("illegal state in content; shortcode token missing end delim")
}
end := j + postIdx + 4
key := string(source[j:end])
newVal, err := tokenHandler(ctx, key)
if err != nil {
return nil, err
}
// Issue #1148: Check for wrapping p-tags <p>
if j >= 3 && bytes.Equal(source[j-3:j], pStart) {
if (k+4) < len(source) && bytes.Equal(source[end:end+4], pEnd) {
j -= 3
end += 4
}
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
}
Provide (relative) reference funcs & shortcodes. - `.Ref` and `.RelRef` take a reference (the logical filename for a page, including extension and/or a document fragment ID) and return a permalink (or relative permalink) to the referenced document. - If the reference is a page name (such as `about.md`), the page will be discovered and the permalink will be returned: `/about/` - If the reference is a page name with a fragment (such as `about.md#who`), the page will be discovered and used to add the `page.UniqueID()` to the resulting fragment and permalink: `/about/#who:deadbeef`. - If the reference is a fragment and `.*Ref` has been called from a `Node` or `SiteInfo`, it will be returned as is: `#who`. - If the reference is a fragment and `.*Ref` has been called from a `Page`, it will be returned with the page’s unique ID: `#who:deadbeef`. - `.*Ref` can be called from either `Node`, `SiteInfo` (e.g., `Node.Site`), `Page` objects, or `ShortcodeWithPage` objects in templates. - `.*Ref` cannot be used in content, so two shortcodes have been created to provide the functionality to content: `ref` and `relref`. These are intended to be used within markup, like `[Who]({{% ref about.md#who %}})` or `<a href="{{% ref about.md#who %}}">Who</a>`. - There are also `ref` and `relref` template functions (used to create the shortcodes) that expect a `Page` or `Node` object and the reference string (e.g., `{{ relref . "about.md" }}` or `{{ "about.md" | ref . }}`). It actually looks for `.*Ref` as defined on `Node` or `Page` objects. - Shortcode handling had to use a *differently unique* wrapper in `createShortcodePlaceholder` because of the way that the `ref` and `relref` are intended to be used in content.
2014-11-24 06:15:34 +00:00
// This and other cool slice tricks: https://github.com/golang/go/wiki/SliceTricks
source = append(source[:j], append(newVal, source[end:]...)...)
start = j
k = bytes.Index(source[start:], pre)
}
return source, nil
Shortcode rewrite, take 2 This commit contains a restructuring and partial rewrite of the shortcode handling. Prior to this commit rendering of the page content was mingled with handling of the shortcodes. This led to several oddities. The new flow is: 1. Shortcodes are extracted from page and replaced with placeholders. 2. Shortcodes are processed and rendered 3. Page is processed 4. The placeholders are replaced with the rendered shortcodes The handling of summaries is also made simpler by this. This commit also introduces some other chenges: 1. distinction between shortcodes that need further processing and those who do not: * `{{< >}}`: Typically raw HTML. Will not be processed. * `{{% %}}`: Will be processed by the page's markup engine (Markdown or (infuture) Asciidoctor) The above also involves a new shortcode-parser, with lexical scanning inspired by Rob Pike's talk called "Lexical Scanning in Go", which should be easier to understand, give better error messages and perform better. 2. If you want to exclude a shortcode from being processed (for documentation etc.), the inner part of the shorcode must be commented out, i.e. `{{%/* movie 47238zzb */%}}`. See the updated shortcode section in the documentation for further examples. The new parser supports nested shortcodes. This isn't new, but has two related design choices worth mentioning: * The shortcodes will be rendered individually, so If both `{{< >}}` and `{{% %}}` are used in the nested hierarchy, one will be passed through the page's markdown processor, the other not. * To avoid potential costly overhead of always looking far ahead for a possible closing tag, this implementation looks at the template itself, and is branded as a container with inner content if it contains a reference to `.Inner` Fixes #565 Fixes #480 Fixes #461 And probably some others.
2014-10-27 20:48:30 +00:00
}
2013-07-04 15:32:55 +00:00
func renderShortcodeWithPage(ctx context.Context, h tpl.TemplateHandler, tmpl tpl.Template, data *ShortcodeWithPage) (string, error) {
2015-01-30 19:19:46 +00:00
buffer := bp.GetBuffer()
defer bp.PutBuffer(buffer)
err := h.ExecuteWithContext(ctx, tmpl, buffer, data)
2014-01-29 22:50:31 +00:00
if err != nil {
return "", fmt.Errorf("failed to process shortcode: %w", err)
2014-01-29 22:50:31 +00:00
}
return buffer.String(), nil
2013-07-04 15:32:55 +00:00
}