docs: add derived-id section reference lint

.github/workflows/ci.yml

@@ -187,6 +187,8 @@ jobs:
         .github/scripts/build-doc.sh
     - name: Check manpage names match filenames
       run: scripts/manpage-name-check.py --enforce docs/build/man
+    - name: Check for derived-id section references
+      run: scripts/docs-anchor-check.py --enforce
     - name: Verify no untracked or modified files after build
       run: |
         #*.po and documentation.pot are modifyed by build. Ignore them for now.

docs/src/man/man1/linuxcncrsh.1.adoc

@@ -178,6 +178,7 @@ The supported commands are as follows:
   Help will respond regardless of whether a 'HELLO' has been
   successfully negotiated.
 
+[[_subcommands]]
 == SUBCOMMANDS
 
 Commands and parameters are not case sensitive, except for the

docs/src/man/man1/mesambccc.1.adoc

@@ -28,6 +28,7 @@ below.
   Output a verbose list of configuration parameters, devices, Modbus messages
   and HAL pins.
 
+[[_mbccs_file_format]]
 == MBCCS FILE FORMAT
 The Modbus command control source file (mbccs file) is an XML formatted
 document describing the communication parameters, connected devices, HAL pins
@@ -59,6 +60,7 @@ All values are checked to be within an acceptable range. Errors and warnings
 are emitted when values are out of bounds. In case of a warning they may be
 clamped to the acceptable range.
 
+[[_modbus_functions]]
 === MODBUS FUNCTIONS
 A subset of the Modbus functions is supported (with function number in parentheses):
 
@@ -82,6 +84,7 @@ A subset of the Modbus functions is supported (with function number in parenthes
 You can use the function's symbolic name or numerical value in '<command>' in
 the 'function' attribute (as in `function="W_REGISTERS"` or `function="16"`).
 
+[[_modbus_types]]
 === MODBUS TYPES
 The 'modbustype' attribute declares the interpretation of values to or from a
 Modbus device's registers. They can be a signed integer (S), unsigned integer
@@ -387,6 +390,7 @@ Recognized '<initlist>/<command>/<data>' attributes:
   16-bit integer but depends on the 'modbustype' attribute and the range of
   acceptable values depends on the Modbus function.
 
+[[_hal_types]]
 === HAL TYPES
 A '<command>' in the '<commands>' section maps to one or more HAL pins with
 specific type using the 'haltype' attribute. Recognized are:

scripts/docs-anchor-check.py

@@ -0,0 +1,84 @@
+#!/usr/bin/env python3
+# Flag section references relying on auto-generated AsciiDoc ids:
+# <<_derived_id>> xrefs and link:...html#_derived_id URLs. Derived ids come
+# from the section title, so the reference breaks when the title is
+# translated or retitled. Pin the target with an explicit [[anchor]] instead.
+# Pass 1 collects explicit anchor definitions in docs/src, pass 2 reports
+# derived-id references with no explicit target.
+# Warn-only unless --enforce or DOCS_ANCHOR_CHECK_ENFORCE is set.
+
+import os
+import re
+import sys
+import glob
+
+# The script lives in scripts/, the docs are one level up at ../docs, so it runs from anywhere.
+HERE = os.path.dirname(os.path.realpath(__file__, strict=True))
+DOCS = os.path.normpath(os.path.join(HERE, '..', 'docs'))
+SRC = os.path.join(DOCS, 'src')
+
+# Explicit anchor definition forms, mirroring ANCHOR_DEF in
+# docs/src/extensions/xref_resolver.rb, plus the anchor: macro.
+ANCHOR_DEF = re.compile(r"""
+    \[\[ ([A-Za-z_][\w:.-]*) (?:,[^\]]*)? \]\]     |   # [[id]] or [[id,reftext]]
+    \[\# ([A-Za-z_][\w:.-]*) (?:[.%][^\]]*)? \]    |   # [#id]
+    \[ (?:[^,\]]*,\s*)* id\s*=\s*["']? ([A-Za-z_][\w:.-]*) ["']? [,\]] |   # [id="foo"]
+    ^anchor: ([A-Za-z_][\w:.-]*) \[\]              |   # anchor:id[]
+    ^:id:\s* ([A-Za-z_][\w:.-]*)                       # :id: foo
+""", re.X | re.M)
+
+# Derived-id references: <<_foo>> / <<_foo,Title>> xrefs and
+# link:...html#_foo[...] URL fragments.  Namespaced anchors carry a ':', so a
+# target that starts with '_' and has no ':' is a derived id by convention.
+XREF = re.compile(r'<<(_[A-Za-z0-9][\w.-]*)(?:,.*?)?>>')
+LINKURL = re.compile(r'link:[^\s\[]*#(_[A-Za-z0-9][\w.-]*)\[')
+
+def collect(adoc_files):
+    defined = {}  # anchor -> file
+    for path in adoc_files:
+        text = open(path, encoding='utf-8', errors='replace').read()
+        for m in ANCHOR_DEF.finditer(text):
+            anchor = next(g for g in m.groups() if g)
+            defined.setdefault(anchor, path)
+    return defined
+
+def find_derived_refs(adoc_files, defined):
+    problems = []
+    for path in adoc_files:
+        lines = open(path, encoding='utf-8', errors='replace').read().splitlines()
+        for lineno, line in enumerate(lines, 1):
+            for regex in (XREF, LINKURL):
+                for m in regex.finditer(line):
+                    target = m.group(1)
+                    if target not in defined:
+                        rel = os.path.relpath(path, DOCS)
+                        problems.append((rel, lineno, m.group(0)[:60], target))
+    return problems
+
+def main():
+    enforce = '--enforce' in sys.argv or os.environ.get('DOCS_ANCHOR_CHECK_ENFORCE')
+    adoc_files = sorted(glob.glob(f'{SRC}/**/*.adoc', recursive=True))
+    defined = collect(adoc_files)
+    problems = find_derived_refs(adoc_files, defined)
+    if not problems:
+        return 0
+    out = ['Derived-id section references with no explicit anchor target',
+           '(these break in translated docs when the target title is translated;',
+           'pin the target section with an explicit [[anchor]] and reference that):',
+           '']
+    for rel, lineno, ref, target in problems:
+        out.append(f'{rel}:{lineno}: {ref}')
+        out.append(f'    target `[[{target}]]` is not explicitly defined anywhere in docs/src')
+    text = '\n'.join(out)
+    print(text)
+    if os.environ.get('GITHUB_ACTIONS'):
+        summary = os.environ.get('GITHUB_STEP_SUMMARY')
+        if summary:
+            with open(summary, 'a', encoding='utf-8') as f:
+                f.write('## Derived-id anchor check\n\n```\n' + text + '\n```\n')
+        level = 'error' if enforce else 'warning'
+        print(f'::{level} title=Derived-id anchor check::{len(problems)} reference(s) rely on auto-generated section ids, see job summary')
+    return 1 if enforce else 0
+
+if __name__ == '__main__':
+    sys.exit(main())