Files
summercms/internal/docsite/check_policy.go
Jakub Zych 73c72af244 fix(11.1-07): check go fences, README src= and shell fences from the AST
- goLang follows the highlighter's chroma lookup, so golang and main.go need src=
- a src= fence in a module README is refused and never captioned
- shell fences inside callouts and lists are command-checked
- a fence with four or more leading spaces is an indented code block
2026-10-01 08:30:19 +02:00

116 lines
3.5 KiB
Go

package docsite
import (
"fmt"
"slices"
"strings"
gast "github.com/yuin/goldmark/ast"
)
// calloutTypes are the only `> [!TYPE]` callouts the theme renders.
var calloutTypes = []string{"NOTE", "TIP", "WARNING"}
// checkPolicy enforces the page content rules:
// - every Go-lexer fence in a docs/ page carries src=, at any depth
// (module READMEs are rendered as written, D-18);
// - callouts are NOTE, TIP or WARNING;
// - docs/ headings are plain ASCII text without links or code spans.
func (s *site) checkPolicy(docs []parsedDoc) ([]Problem, error) {
var problems []Problem
for _, d := range docs {
if d.page == nil {
continue
}
guide := d.page.Module == ""
if guide {
fences, err := collectFences(d.doc, d.body)
if err != nil {
return nil, err
}
for _, f := range fences {
if _, hasSrc := ParseSrc(f.info); goLang(f.lang) && !hasSrc {
problems = append(problems, Problem{File: d.file, Line: d.line + f.line, Rule: "snippet",
Message: f.lang + " code block has no src= reference"})
}
}
}
problems = append(problems, calloutProblems(d)...)
if guide {
problems = append(problems, headingProblems(d)...)
}
}
return problems, nil
}
// calloutProblems reports blockquotes whose marker is not NOTE, TIP or
// WARNING. calloutTransformer has already turned the known types into
// callout nodes, so a remaining blockquote with a marker is unknown.
// The marker is read from the AST, so a copy of it inside a code fence
// is not a callout.
func calloutProblems(d parsedDoc) []Problem {
var problems []Problem
_ = gast.Walk(d.doc, func(n gast.Node, entering bool) (gast.WalkStatus, error) {
bq, ok := n.(*gast.Blockquote)
if !entering || !ok {
return gast.WalkContinue, nil
}
para, ok := bq.FirstChild().(*gast.Paragraph)
if !ok || para.Lines().Len() == 0 {
return gast.WalkContinue, nil
}
first := para.Lines().At(0)
m := calloutLine.FindStringSubmatch(strings.TrimSpace(string(first.Value(d.body))))
if m == nil || slices.Contains(calloutTypes, m[1]) {
return gast.WalkContinue, nil
}
problems = append(problems, Problem{File: d.file, Line: lineOf(d.body, first.Start, d.line), Rule: "callout",
Message: fmt.Sprintf("unknown type %s (use NOTE, TIP or WARNING)", m[1])})
return gast.WalkContinue, nil
})
return problems
}
const headingMessage = "headings must be plain ASCII text without links or code"
// headingProblems reports headings with a link, image, code span or raw
// HTML, or with non-ASCII text: their IDs would not be stable across the
// site, Gitea and GitHub.
func headingProblems(d parsedDoc) []Problem {
var problems []Problem
_ = gast.Walk(d.doc, func(n gast.Node, entering bool) (gast.WalkStatus, error) {
h, ok := n.(*gast.Heading)
if !entering || !ok {
return gast.WalkContinue, nil
}
bad := false
_ = gast.Walk(h, func(c gast.Node, entering bool) (gast.WalkStatus, error) {
if !entering {
return gast.WalkContinue, nil
}
switch t := c.(type) {
case *gast.Link, *gast.Image, *gast.CodeSpan, *gast.RawHTML, *gast.AutoLink:
bad = true
return gast.WalkStop, nil
case *gast.Text:
for _, b := range t.Segment.Value(d.body) {
if b >= 0x80 {
bad = true
return gast.WalkStop, nil
}
}
}
return gast.WalkContinue, nil
})
if bad {
off := -1
if h.Lines().Len() > 0 {
off = h.Lines().At(0).Start
}
problems = append(problems, Problem{File: d.file, Line: lineOf(d.body, off, d.line), Rule: "heading", Message: headingMessage})
}
return gast.WalkSkipChildren, nil
})
return problems
}