forgejo/modules/setting/markup.go

191 lines
5.7 KiB
Go
Raw Normal View History

// Copyright 2019 The Gitea Authors. All rights reserved.
// SPDX-License-Identifier: MIT
package setting
import (
"regexp"
"strings"
"code.gitea.io/gitea/modules/log"
)
// ExternalMarkupRenderers represents the external markup renderers
var (
ExternalMarkupRenderers []*MarkupRenderer
ExternalSanitizerRules []MarkupSanitizerRule
MermaidMaxSourceCharacters int
)
const (
RenderContentModeSanitized = "sanitized"
RenderContentModeNoSanitizer = "no-sanitizer"
RenderContentModeIframe = "iframe"
)
// Markdown settings
var Markdown = struct {
EnableHardLineBreakInComments bool
EnableHardLineBreakInDocuments bool
CustomURLSchemes []string `ini:"CUSTOM_URL_SCHEMES"`
FileExtensions []string
EnableMath bool
}{
EnableHardLineBreakInComments: true,
EnableHardLineBreakInDocuments: false,
Add .livemd as a markdown extension (#22730) ## Needs and benefits [Livebook](https://livebook.dev/) notebooks are used for code documentation and for deep dives and note-taking in the elixir ecosystem. Rendering these in these as Markdown on frogejo has many benefits, since livemd is a subset of markdown. Some of the benefits are: - New users of elixir and livebook are scared by unformated .livemd files, but are shown what they expect - Sharing a notebook is as easy as sharing a link, no need to install the software in order to see the results. [goldmark-meraid ](https://github.com/abhinav/goldmark-mermaid) is a mermaid-js parser already included in gitea. This makes the .livemd rendering integration feature complete. With this PR class diagrams, ER Diagrams, flow charts and much more will be rendered perfectly. With the additional functionality gitea will be an ideal tool for sharing resources with fellow software engineers working in the elixir ecosystem. Allowing the git forge to be used without needing to install any software. ## Feature Description This issue requests the .livemd extension to be added as a Markdown language extension. - `.livemd` is the extension of Livebook which is an Elixir version of Jupyter Notebook. - `.livemd` is` a subset of Markdown. This would require the .livemd to be recognized as a markdown file. The Goldmark the markdown parser should handle the parsing and rendering automatically. Here is the corresponding commit for GitHub linguist: https://github.com/github/linguist/pull/5672 Here is a sample page of a livemd file: https://github.com/github/linguist/blob/master/samples/Markdown/livebook.livemd ## Screenshots The first screenshot shows how github shows the sample .livemd in the browser. The second screenshot shows how mermaid js, renders my development notebook and its corresponding ER Diagram. The source code can be found here: https://codeberg.org/lgh/Termi/src/commit/79615f74281789a1f2967b57bad0c67c356cef1f/termiNotes.livemd ## Testing I just changed the file extension from `.livemd`to `.md`and the document already renders perfectly on codeberg. Check you can it out [here](https://codeberg.org/lgh/Termi/src/branch/livemd2md/termiNotes.md) --------- Co-authored-by: techknowlogick <techknowlogick@gitea.io>
2023-04-26 17:22:54 +02:00
FileExtensions: strings.Split(".md,.markdown,.mdown,.mkd,.livemd", ","),
EnableMath: true,
}
// MarkupRenderer defines the external parser configured in ini
type MarkupRenderer struct {
Enabled bool
MarkupName string
Command string
FileExtensions []string
IsInputFile bool
NeedPostProcess bool
MarkupSanitizerRules []MarkupSanitizerRule
RenderContentMode string
}
// MarkupSanitizerRule defines the policy for whitelisting attributes on
// certain elements.
type MarkupSanitizerRule struct {
Element string
AllowAttr string
Regexp *regexp.Regexp
AllowDataURIImages bool
}
func loadMarkupFrom(rootCfg ConfigProvider) {
mustMapSetting(rootCfg, "markdown", &Markdown)
MermaidMaxSourceCharacters = rootCfg.Section("markup").Key("MERMAID_MAX_SOURCE_CHARACTERS").MustInt(5000)
ExternalMarkupRenderers = make([]*MarkupRenderer, 0, 10)
ExternalSanitizerRules = make([]MarkupSanitizerRule, 0, 10)
for _, sec := range rootCfg.Section("markup").ChildSections() {
name := strings.TrimPrefix(sec.Name(), "markup.")
if name == "" {
log.Warn("name is empty, markup " + sec.Name() + "ignored")
continue
}
if name == "sanitizer" || strings.HasPrefix(name, "sanitizer.") {
newMarkupSanitizer(name, sec)
} else {
newMarkupRenderer(name, sec)
}
}
}
func newMarkupSanitizer(name string, sec ConfigSection) {
rule, ok := createMarkupSanitizerRule(name, sec)
if ok {
if strings.HasPrefix(name, "sanitizer.") {
names := strings.SplitN(strings.TrimPrefix(name, "sanitizer."), ".", 2)
name = names[0]
}
for _, renderer := range ExternalMarkupRenderers {
if name == renderer.MarkupName {
renderer.MarkupSanitizerRules = append(renderer.MarkupSanitizerRules, rule)
return
}
}
ExternalSanitizerRules = append(ExternalSanitizerRules, rule)
}
}
func createMarkupSanitizerRule(name string, sec ConfigSection) (MarkupSanitizerRule, bool) {
var rule MarkupSanitizerRule
ok := false
if sec.HasKey("ALLOW_DATA_URI_IMAGES") {
rule.AllowDataURIImages = sec.Key("ALLOW_DATA_URI_IMAGES").MustBool(false)
ok = true
}
if sec.HasKey("ELEMENT") || sec.HasKey("ALLOW_ATTR") {
rule.Element = sec.Key("ELEMENT").Value()
rule.AllowAttr = sec.Key("ALLOW_ATTR").Value()
if rule.Element == "" || rule.AllowAttr == "" {
log.Error("Missing required values from markup.%s. Must have ELEMENT and ALLOW_ATTR defined!", name)
return rule, false
}
regexpStr := sec.Key("REGEXP").Value()
if regexpStr != "" {
// Validate when parsing the config that this is a valid regular
// expression. Then we can use regexp.MustCompile(...) later.
compiled, err := regexp.Compile(regexpStr)
if err != nil {
log.Error("In markup.%s: REGEXP (%s) failed to compile: %v", name, regexpStr, err)
return rule, false
}
rule.Regexp = compiled
}
ok = true
}
if !ok {
log.Error("Missing required keys from markup.%s. Must have ELEMENT and ALLOW_ATTR or ALLOW_DATA_URI_IMAGES defined!", name)
return rule, false
}
return rule, true
}
func newMarkupRenderer(name string, sec ConfigSection) {
extensionReg := regexp.MustCompile(`\.\w`)
extensions := sec.Key("FILE_EXTENSIONS").Strings(",")
exts := make([]string, 0, len(extensions))
for _, extension := range extensions {
if !extensionReg.MatchString(extension) {
log.Warn(sec.Name() + " file extension " + extension + " is invalid. Extension ignored")
} else {
exts = append(exts, extension)
}
}
if len(exts) == 0 {
log.Warn(sec.Name() + " file extension is empty, markup " + name + " ignored")
return
}
command := sec.Key("RENDER_COMMAND").MustString("")
if command == "" {
log.Warn(" RENDER_COMMAND is empty, markup " + name + " ignored")
return
}
if sec.HasKey("DISABLE_SANITIZER") {
log.Error("Deprecated setting `[markup.*]` `DISABLE_SANITIZER` present. This fallback will be removed in v1.18.0")
}
renderContentMode := sec.Key("RENDER_CONTENT_MODE").MustString(RenderContentModeSanitized)
if !sec.HasKey("RENDER_CONTENT_MODE") && sec.Key("DISABLE_SANITIZER").MustBool(false) {
renderContentMode = RenderContentModeNoSanitizer // if only the legacy DISABLE_SANITIZER exists, use it
}
if renderContentMode != RenderContentModeSanitized &&
renderContentMode != RenderContentModeNoSanitizer &&
renderContentMode != RenderContentModeIframe {
log.Error("invalid RENDER_CONTENT_MODE: %q, default to %q", renderContentMode, RenderContentModeSanitized)
renderContentMode = RenderContentModeSanitized
}
ExternalMarkupRenderers = append(ExternalMarkupRenderers, &MarkupRenderer{
Enabled: sec.Key("ENABLED").MustBool(false),
MarkupName: name,
FileExtensions: exts,
Command: command,
IsInputFile: sec.Key("IS_INPUT_FILE").MustBool(false),
NeedPostProcess: sec.Key("NEED_POSTPROCESS").MustBool(true),
RenderContentMode: renderContentMode,
})
}