From b6ce2d6dc9db16227c523b2d0a39a231e5d38945 Mon Sep 17 00:00:00 2001 From: wxiaoguang Date: Wed, 27 Nov 2024 00:46:02 +0800 Subject: [PATCH] Refactor markup render system (#32645) This PR mainly removes some global variables, moves some code and renames some functions to make code clearer. This PR also removes a testing-only option ForceHardLineBreak during refactoring since the behavior is clear now. --- modules/markup/html.go | 121 +++++------ modules/markup/html_internal_test.go | 4 +- modules/markup/html_test.go | 14 +- modules/markup/markdown/goldmark.go | 4 +- modules/markup/markdown/markdown_test.go | 258 +++++++++++------------ modules/markup/render.go | 10 +- tests/fuzz/fuzz_test.go | 2 +- 7 files changed, 188 insertions(+), 225 deletions(-) diff --git a/modules/markup/html.go b/modules/markup/html.go index 0b1e9b32242..04b768bb8ec 100644 --- a/modules/markup/html.go +++ b/modules/markup/html.go @@ -5,9 +5,9 @@ package markup import ( "bytes" + "fmt" "io" "regexp" - "slices" "strings" "sync" @@ -133,75 +133,49 @@ func CustomLinkURLSchemes(schemes []string) { common.GlobalVars().LinkRegex, _ = xurls.StrictMatchingScheme(strings.Join(withAuth, "|")) } -type postProcessError struct { - context string - err error -} - -func (p *postProcessError) Error() string { - return "PostProcess: " + p.context + ", " + p.err.Error() -} - type processor func(ctx *RenderContext, node *html.Node) -var defaultProcessors = []processor{ - fullIssuePatternProcessor, - comparePatternProcessor, - codePreviewPatternProcessor, - fullHashPatternProcessor, - shortLinkProcessor, - linkProcessor, - mentionProcessor, - issueIndexPatternProcessor, - commitCrossReferencePatternProcessor, - hashCurrentPatternProcessor, - emailAddressProcessor, - emojiProcessor, - emojiShortCodeProcessor, -} - -// PostProcess does the final required transformations to the passed raw HTML +// PostProcessDefault does the final required transformations to the passed raw HTML // data, and ensures its validity. Transformations include: replacing links and // emails with HTML links, parsing shortlinks in the format of [[Link]], like // MediaWiki, linking issues in the format #ID, and mentions in the format // @user, and others. -func PostProcess(ctx *RenderContext, input io.Reader, output io.Writer) error { - return postProcess(ctx, defaultProcessors, input, output) -} - -var commitMessageProcessors = []processor{ - fullIssuePatternProcessor, - comparePatternProcessor, - fullHashPatternProcessor, - linkProcessor, - mentionProcessor, - issueIndexPatternProcessor, - commitCrossReferencePatternProcessor, - hashCurrentPatternProcessor, - emailAddressProcessor, - emojiProcessor, - emojiShortCodeProcessor, +func PostProcessDefault(ctx *RenderContext, input io.Reader, output io.Writer) error { + procs := []processor{ + fullIssuePatternProcessor, + comparePatternProcessor, + codePreviewPatternProcessor, + fullHashPatternProcessor, + shortLinkProcessor, + linkProcessor, + mentionProcessor, + issueIndexPatternProcessor, + commitCrossReferencePatternProcessor, + hashCurrentPatternProcessor, + emailAddressProcessor, + emojiProcessor, + emojiShortCodeProcessor, + } + return postProcess(ctx, procs, input, output) } // RenderCommitMessage will use the same logic as PostProcess, but will disable -// the shortLinkProcessor and will add a defaultLinkProcessor if defaultLink is -// set, which changes every text node into a link to the passed default link. +// the shortLinkProcessor. func RenderCommitMessage(ctx *RenderContext, content string) (string, error) { - procs := commitMessageProcessors - return renderProcessString(ctx, procs, content) -} - -var commitMessageSubjectProcessors = []processor{ - fullIssuePatternProcessor, - comparePatternProcessor, - fullHashPatternProcessor, - linkProcessor, - mentionProcessor, - issueIndexPatternProcessor, - commitCrossReferencePatternProcessor, - hashCurrentPatternProcessor, - emojiShortCodeProcessor, - emojiProcessor, + procs := []processor{ + fullIssuePatternProcessor, + comparePatternProcessor, + fullHashPatternProcessor, + linkProcessor, + mentionProcessor, + issueIndexPatternProcessor, + commitCrossReferencePatternProcessor, + hashCurrentPatternProcessor, + emailAddressProcessor, + emojiProcessor, + emojiShortCodeProcessor, + } + return postProcessString(ctx, procs, content) } var emojiProcessors = []processor{ @@ -214,7 +188,18 @@ var emojiProcessors = []processor{ // emailAddressProcessor, will add a defaultLinkProcessor if defaultLink is set, // which changes every text node into a link to the passed default link. func RenderCommitMessageSubject(ctx *RenderContext, defaultLink, content string) (string, error) { - procs := slices.Clone(commitMessageSubjectProcessors) + procs := []processor{ + fullIssuePatternProcessor, + comparePatternProcessor, + fullHashPatternProcessor, + linkProcessor, + mentionProcessor, + issueIndexPatternProcessor, + commitCrossReferencePatternProcessor, + hashCurrentPatternProcessor, + emojiShortCodeProcessor, + emojiProcessor, + } procs = append(procs, func(ctx *RenderContext, node *html.Node) { ch := &html.Node{Parent: node, Type: html.TextNode, Data: node.Data} node.Type = html.ElementNode @@ -223,19 +208,19 @@ func RenderCommitMessageSubject(ctx *RenderContext, defaultLink, content string) node.Attr = []html.Attribute{{Key: "href", Val: defaultLink}, {Key: "class", Val: "muted"}} node.FirstChild, node.LastChild = ch, ch }) - return renderProcessString(ctx, procs, content) + return postProcessString(ctx, procs, content) } // RenderIssueTitle to process title on individual issue/pull page func RenderIssueTitle(ctx *RenderContext, title string) (string, error) { // do not render other issue/commit links in an issue's title - which in most cases is already a link. - return renderProcessString(ctx, []processor{ + return postProcessString(ctx, []processor{ emojiShortCodeProcessor, emojiProcessor, }, title) } -func renderProcessString(ctx *RenderContext, procs []processor, content string) (string, error) { +func postProcessString(ctx *RenderContext, procs []processor, content string) (string, error) { var buf strings.Builder if err := postProcess(ctx, procs, strings.NewReader(content), &buf); err != nil { return "", err @@ -246,7 +231,7 @@ func renderProcessString(ctx *RenderContext, procs []processor, content string) // RenderDescriptionHTML will use similar logic as PostProcess, but will // use a single special linkProcessor. func RenderDescriptionHTML(ctx *RenderContext, content string) (string, error) { - return renderProcessString(ctx, []processor{ + return postProcessString(ctx, []processor{ descriptionLinkProcessor, emojiShortCodeProcessor, emojiProcessor, @@ -256,7 +241,7 @@ func RenderDescriptionHTML(ctx *RenderContext, content string) (string, error) { // RenderEmoji for when we want to just process emoji and shortcodes // in various places it isn't already run through the normal markdown processor func RenderEmoji(ctx *RenderContext, content string) (string, error) { - return renderProcessString(ctx, emojiProcessors, content) + return postProcessString(ctx, emojiProcessors, content) } func postProcess(ctx *RenderContext, procs []processor, input io.Reader, output io.Writer) error { @@ -276,7 +261,7 @@ func postProcess(ctx *RenderContext, procs []processor, input io.Reader, output strings.NewReader(""), )) if err != nil { - return &postProcessError{"invalid HTML", err} + return fmt.Errorf("markup.postProcess: invalid HTML: %w", err) } if node.Type == html.DocumentNode { @@ -308,7 +293,7 @@ func postProcess(ctx *RenderContext, procs []processor, input io.Reader, output // Render everything to buf. for _, node := range newNodes { if err := html.Render(output, node); err != nil { - return &postProcessError{"error rendering processed HTML", err} + return fmt.Errorf("markup.postProcess: html.Render: %w", err) } } return nil diff --git a/modules/markup/html_internal_test.go b/modules/markup/html_internal_test.go index 7f2057a3438..651e674108c 100644 --- a/modules/markup/html_internal_test.go +++ b/modules/markup/html_internal_test.go @@ -277,12 +277,12 @@ func TestRender_AutoLink(t *testing.T) { test := func(input, expected string) { var buffer strings.Builder - err := PostProcess(NewTestRenderContext(localMetas), strings.NewReader(input), &buffer) + err := PostProcessDefault(NewTestRenderContext(localMetas), strings.NewReader(input), &buffer) assert.Equal(t, err, nil) assert.Equal(t, strings.TrimSpace(expected), strings.TrimSpace(buffer.String())) buffer.Reset() - err = PostProcess(NewTestRenderContext(localMetas), strings.NewReader(input), &buffer) + err = PostProcessDefault(NewTestRenderContext(localMetas), strings.NewReader(input), &buffer) assert.Equal(t, err, nil) assert.Equal(t, strings.TrimSpace(expected), strings.TrimSpace(buffer.String())) } diff --git a/modules/markup/html_test.go b/modules/markup/html_test.go index f806f66d112..54bd91f3b3b 100644 --- a/modules/markup/html_test.go +++ b/modules/markup/html_test.go @@ -445,14 +445,14 @@ func Test_ParseClusterFuzz(t *testing.T) { data := "
` var res strings.Builder - err := markup.PostProcess(markup.NewTestRenderContext(localMetas), strings.NewReader(data), &res) + err := markup.PostProcessDefault(markup.NewTestRenderContext(localMetas), strings.NewReader(data), &res) assert.NoError(t, err) assert.Equal(t, data, res.String()) } @@ -514,7 +514,7 @@ func BenchmarkEmojiPostprocess(b *testing.B) { b.ResetTimer() for i := 0; i < b.N; i++ { var res strings.Builder - err := markup.PostProcess(markup.NewTestRenderContext(localMetas), strings.NewReader(data), &res) + err := markup.PostProcessDefault(markup.NewTestRenderContext(localMetas), strings.NewReader(data), &res) assert.NoError(b, err) } } @@ -522,7 +522,7 @@ func BenchmarkEmojiPostprocess(b *testing.B) { func TestFuzz(t *testing.T) { s := "t/l/issues/8#/../../a" renderContext := markup.NewTestRenderContext() - err := markup.PostProcess(renderContext, strings.NewReader(s), io.Discard) + err := markup.PostProcessDefault(renderContext, strings.NewReader(s), io.Discard) assert.NoError(t, err) } @@ -530,7 +530,7 @@ func TestIssue18471(t *testing.T) { data := `http://domain/org/repo/compare/783b039...da951ce` var res strings.Builder - err := markup.PostProcess(markup.NewTestRenderContext(localMetas), strings.NewReader(data), &res) + err := markup.PostProcessDefault(markup.NewTestRenderContext(localMetas), strings.NewReader(data), &res) assert.NoError(t, err) assert.Equal(t, `783b039...da951ce`, res.String()) diff --git a/modules/markup/markdown/goldmark.go b/modules/markup/markdown/goldmark.go index ed95cecf8b0..620a39ebfde 100644 --- a/modules/markup/markdown/goldmark.go +++ b/modules/markup/markdown/goldmark.go @@ -80,9 +80,7 @@ func (g *ASTTransformer) Transform(node *ast.Document, reader text.Reader, pc pa // many places render non-comment contents with no mode=document, then these contents also use comment's hard line break setting // especially in many tests. markdownLineBreakStyle := ctx.RenderOptions.Metas["markdownLineBreakStyle"] - if markup.RenderBehaviorForTesting.ForceHardLineBreak { - v.SetHardLineBreak(true) - } else if markdownLineBreakStyle == "comment" { + if markdownLineBreakStyle == "comment" { v.SetHardLineBreak(setting.Markdown.EnableHardLineBreakInComments) } else if markdownLineBreakStyle == "document" { v.SetHardLineBreak(setting.Markdown.EnableHardLineBreakInDocuments) diff --git a/modules/markup/markdown/markdown_test.go b/modules/markup/markdown/markdown_test.go index aac5ccbb1f8..22ab39ebfa9 100644 --- a/modules/markup/markdown/markdown_test.go +++ b/modules/markup/markdown/markdown_test.go @@ -85,94 +85,13 @@ func TestRender_Images(t *testing.T) { `

`+title+`

`) } -func testAnswers(baseURL string) []string { - return []string{ - `

Wiki! Enjoy :)

- -

See commit 65f1bf27bc

-

Ideas and codes

- -`, - `

What is Wine Staging?

-

Wine Staging on website wine-staging.com.

-

Quick Links

-

Here are some links to the most important topics. You can find the full list of pages at the sidebar.

- - - - - - - - - - - - - -
images/icon-install.pngInstallation
images/icon-usage.pngUsage
-`, - `

Excelsior JET allows you to create native executables for Windows, Linux and Mac OS X.

-
    -
  1. Package your libGDX application
    -images/1.png
    2. -
  2. Perform a test run by hitting the Run! button.
    -images/2.png
    3. -
-

More tests

-

(from https://www.markdownguide.org/extended-syntax/)

-

Checkboxes

-
    -
  • unchecked
    • -
  • checked
    • -
  • still unchecked
    • -
-

Definition list

-
-
First Term
-
This is the definition of the first term.
-
Second Term
-
This is one definition of the second term.
-
This is another definition of the second term.
-
-

Footnotes

-

Here is a simple footnote,1 and here is a longer one.2

-
-
-
    -
  1. -

    This is the first footnote. ↩︎

    -
    2. -
  2. -

    Here is one with multiple paragraphs and code.

    -

    Indent paragraphs to include them in the footnote.

    -

    { my code }

    -

    Add as many paragraphs as you like. ↩︎

    -
    3. -
-
-`, `
    -
  • If you want to rebase/retry this PR, click this checkbox.
    • -
-
-

This PR has been generated by Renovate Bot.

-`, - } -} +func TestTotal_RenderString(t *testing.T) { + defer test.MockVariableValue(&markup.RenderBehaviorForTesting.DisableAdditionalAttributes, true)() -// Test cases without ambiguous links -var sameCases = []string{ - // dear imgui wiki markdown extract: special wiki syntax - `Wiki! Enjoy :) + // Test cases without ambiguous links (It is not right to copy a whole file here, instead it should clearly test what is being tested) + sameCases := []string{ + // dear imgui wiki markdown extract: special wiki syntax + `Wiki! Enjoy :) - [[Links, Language bindings, Engine bindings|Links]] - [[Tips]] @@ -185,8 +104,8 @@ Ideas and codes - Node graph editors https://github.com/ocornut/imgui/issues/306 - [[Memory Editor|memory_editor_example]] - [[Plot var helper|plot_var_example]]`, - // wine-staging wiki home extract: tables, special wiki syntax, images - `## What is Wine Staging? + // wine-staging wiki home extract: tables, special wiki syntax, images + `## What is Wine Staging? **Wine Staging** on website [wine-staging.com](http://wine-staging.com). ## Quick Links @@ -196,8 +115,8 @@ Here are some links to the most important topics. You can find the full list of |--------------------------------|----------------------------------------------------------| | [[images/icon-usage.png]] | [[Usage]] | `, - // libgdx wiki page: inline images with special syntax - `[Excelsior JET](http://www.excelsiorjet.com/) allows you to create native executables for Windows, Linux and Mac OS X. + // libgdx wiki page: inline images with special syntax + `[Excelsior JET](http://www.excelsiorjet.com/) allows you to create native executables for Windows, Linux and Mac OS X. 1. [Package your libGDX application](https://github.com/libgdx/libgdx/wiki/Gradle-on-the-Commandline#packaging-for-the-desktop) [[images/1.png]] @@ -237,7 +156,7 @@ Here is a simple footnote,[^1] and here is a longer one.[^bignote] Add as many paragraphs as you like. `, - ` + ` - [ ] If you want to rebase/retry this PR, click this checkbox. --- @@ -245,21 +164,101 @@ Here is a simple footnote,[^1] and here is a longer one.[^bignote] This PR has been generated by [Renovate Bot](https://github.com/renovatebot/renovate). `, -} + } + + baseURL := "" + testAnswers := []string{ + `

Wiki! Enjoy :)

+ +

See commit 65f1bf27bc

+

Ideas and codes

+ +`, + `

What is Wine Staging?

+

Wine Staging on website wine-staging.com.

+

Quick Links

+

Here are some links to the most important topics. You can find the full list of pages at the sidebar.

+ + + + + + + + + + + + + +
images/icon-install.pngInstallation
images/icon-usage.pngUsage
+`, + `

Excelsior JET allows you to create native executables for Windows, Linux and Mac OS X.

+
    +
  1. Package your libGDX application +images/1.png
    2. +
  2. Perform a test run by hitting the Run! button. +images/2.png
    3. +
+

More tests

+

(from https://www.markdownguide.org/extended-syntax/)

+

Checkboxes

+
    +
  • unchecked
    • +
  • checked
    • +
  • still unchecked
    • +
+

Definition list

+
+
First Term
+
This is the definition of the first term.
+
Second Term
+
This is one definition of the second term.
+
This is another definition of the second term.
+
+

Footnotes

+

Here is a simple footnote,1 and here is a longer one.2

+
+
+
    +
  1. +

    This is the first footnote. ↩︎

    +
    2. +
  2. +

    Here is one with multiple paragraphs and code.

    +

    Indent paragraphs to include them in the footnote.

    +

    { my code }

    +

    Add as many paragraphs as you like. ↩︎

    +
    3. +
+
+`, + `
    +
  • If you want to rebase/retry this PR, click this checkbox.
    • +
+
+

This PR has been generated by Renovate Bot.

+`, + } -func TestTotal_RenderString(t *testing.T) { - defer test.MockVariableValue(&markup.RenderBehaviorForTesting.ForceHardLineBreak, true)() - defer test.MockVariableValue(&markup.RenderBehaviorForTesting.DisableAdditionalAttributes, true)() markup.Init(&markup.RenderHelperFuncs{ IsUsernameMentionable: func(ctx context.Context, username string) bool { return username == "r-lyeh" }, }) - answers := testAnswers("") for i := 0; i < len(sameCases); i++ { line, err := markdown.RenderString(markup.NewTestRenderContext(localMetas), sameCases[i]) assert.NoError(t, err) - assert.Equal(t, answers[i], string(line)) + assert.Equal(t, testAnswers[i], string(line)) } } @@ -312,10 +311,9 @@ func TestRenderSiblingImages_Issue12925(t *testing.T) { testcase := `![image1](/image1) ![image2](/image2) ` - expected := `

image1
+ expected := `

image1 image2

` - defer test.MockVariableValue(&markup.RenderBehaviorForTesting.ForceHardLineBreak, true)() res, err := markdown.RenderRawString(markup.NewTestRenderContext(), testcase) assert.NoError(t, err) assert.Equal(t, expected, res) @@ -525,43 +523,33 @@ mail@domain.com space${SPACE}${SPACE} ` input = strings.ReplaceAll(input, "${SPACE}", " ") // replace ${SPACE} with " ", to avoid some editor's auto-trimming - cases := []struct { - Expected string - }{ - { - Expected: `

space @mention-user
-/just/a/path.bin
-https://example.com/file.bin
-local link
-remote link
-local link
-remote link
-local image
-local image
-local image
-remote image
-local image
-remote link
-88fc37a3c0...12fc37a3c0 (hash)
-com 88fc37a3c0a4dda553bdcfc80c178a58247f42fb...12fc37a3c0a4dda553bdcfc80c178a58247f42fb pare
-88fc37a3c0
-com 88fc37a3c0a4dda553bdcfc80c178a58247f42fb mit
-👍
-mail@domain.com
-@mention-user test
-#123
+ expected := `

space @mention-user
+/just/a/path.bin +https://example.com/file.bin +local link +remote link +local link +remote link +local image +local image +local image +remote image +local image +remote link +88fc37a3c0...12fc37a3c0 (hash) +com 88fc37a3c0a4dda553bdcfc80c178a58247f42fb...12fc37a3c0a4dda553bdcfc80c178a58247f42fb pare +88fc37a3c0 +com 88fc37a3c0a4dda553bdcfc80c178a58247f42fb mit +👍 +mail@domain.com +@mention-user test +#123 space

-`, - }, - } - - defer test.MockVariableValue(&markup.RenderBehaviorForTesting.ForceHardLineBreak, true)() +` defer test.MockVariableValue(&markup.RenderBehaviorForTesting.DisableAdditionalAttributes, true)() - for i, c := range cases { - result, err := markdown.RenderString(markup.NewTestRenderContext(localMetas), input) - assert.NoError(t, err, "Unexpected error in testcase: %v", i) - assert.Equal(t, c.Expected, string(result), "Unexpected result in testcase %v", i) - } + result, err := markdown.RenderString(markup.NewTestRenderContext(localMetas), input) + assert.NoError(t, err) + assert.Equal(t, expected, string(result)) } func TestAttention(t *testing.T) { diff --git a/modules/markup/render.go b/modules/markup/render.go index be75d08c8c3..3b112b1a142 100644 --- a/modules/markup/render.go +++ b/modules/markup/render.go @@ -28,14 +28,6 @@ const ( ) var RenderBehaviorForTesting struct { - // Markdown line break rendering has 2 default behaviors: - // * Use hard: replace "\n" with "
" for comments, setting.Markdown.EnableHardLineBreakInComments=true - // * Keep soft: "\n" for non-comments (a.k.a. documents), setting.Markdown.EnableHardLineBreakInDocuments=false - // In history, there was a mess: - // * The behavior was controlled by `Metas["mode"] != "document", - // * However, many places render the content without setting "mode" in Metas, all these places used comment line break setting incorrectly - ForceHardLineBreak bool - // Gitea will emit some additional attributes for various purposes, these attributes don't affect rendering. // But there are too many hard-coded test cases, to avoid changing all of them again and again, we can disable emitting these internal attributes. DisableAdditionalAttributes bool @@ -218,7 +210,7 @@ func render(ctx *RenderContext, renderer Renderer, input io.Reader, output io.Wr eg.Go(func() (err error) { if r, ok := renderer.(PostProcessRenderer); ok && r.NeedPostProcess() { - err = PostProcess(ctx, pr1, pw2) + err = PostProcessDefault(ctx, pr1, pw2) } else { _, err = io.Copy(pw2, pr1) } diff --git a/tests/fuzz/fuzz_test.go b/tests/fuzz/fuzz_test.go index 946f7c46f11..01d562d995c 100644 --- a/tests/fuzz/fuzz_test.go +++ b/tests/fuzz/fuzz_test.go @@ -27,6 +27,6 @@ func FuzzMarkdownRenderRaw(f *testing.F) { func FuzzMarkupPostProcess(f *testing.F) { f.Fuzz(func(t *testing.T, data []byte) { setting.AppURL = "http://localhost:3000/" - markup.PostProcess(newFuzzRenderContext(), bytes.NewReader(data), io.Discard) + markup.PostProcessDefault(newFuzzRenderContext(), bytes.NewReader(data), io.Discard) }) }