Docs: makelsr.py: normalize snippet titles into more manageable node names

Commas are forbidden in Texinfo node names, and @q and @qq cause
problems to scripts/build/extract_texi_filenames.py, which is not able
to normalize node names as well as Texinfo official formatters.

diff --git a/scripts/auxiliar/makelsr.py b/scripts/auxiliar/makelsr.py
index 593925b3bf456c17bb8d5fba1e932d2edc2712d2..001106d32753384dc2aaa73aab2be8033d98578a 100755 (executable)
--- a/scripts/auxiliar/makelsr.py
+++ b/scripts/auxiliar/makelsr.py
@@ -66,6 +66,16 @@ notags_files = []
 # mark the section that will be printed verbatim by lilypond-book
 end_header_re = re.compile ('(\\header {.+?doctitle = ".+?})\n', re.M | re.S)
 
+doctitle_re = re.compile (r'(doctitle[a-zA-Z_]{0,6}\s*=\s*")((?:\\"|[^"\n])*)"')
+texinfo_q_re = re.compile (r'@q{(.*?)}')
+texinfo_qq_re = re.compile (r'@qq{(.*?)}')
+def doctitle_sub (title_match):
+    # Comma forbidden in Texinfo node name
+    title = title_match.group (2).replace (',', '')
+    title = texinfo_q_re.sub (r"`\1'", title)
+    title = texinfo_qq_re.sub (r'"\1"', title)
+    return title_match.group (1) + title + '"'
+
 def mark_verbatim_section (ly_code):
     return end_header_re.sub ('\\1 % begin verbatim\n\n', ly_code, 1)
 
@@ -118,6 +128,7 @@ def copy_ly (srcdir, name, tags):
             texidoc_translation = texidoc_translation.replace ('\\', '\\\\')
             s = begin_header_re.sub ('\\g<0>\n' + texidoc_translation, s, 1)
 
+    s = doctitle_re.sub (doctitle_sub, s)
     if in_dir and in_dir in srcdir:
         s = LY_HEADER_LSR + add_tags (s, tags)
     else: