Patch - Comment Parsing

Issue

A bug identified in http://moinmo.in/MoinMoinBugs/CommentBreaksListNumbering has been an issue for awhile.

When a comment line (viz. ## comment is placed with a list, the list is broken. MoinMoin stops this break occurring in tables but not in lists.

These comment lines only appear in the raw text of a page, and cannot be revealed by toggling comments as with /* comment */ (if enabled by user preferences). So it would make sense to just ignore them and prevent parsing in all instances, not just with tables.

Patch

A simple patch to the parser text_moin_wiki.py can resolve this issue.

Copy the parser text_moin_wiki.py from MoinMoin/parser to .../data/plugin/parser, and add the following patch.

   1              # ignore processing instructions
   2              if self.in_processing_instructions:
   3                  found = False
   4                  for pi in ("##", "#format", "#refresh", "#redirect", "#deprecated",
   5                             "#pragma", "#form", "#acl", "#language"):
   6                      if line.lower().startswith(pi):
   7                          self.request.write(self.formatter.comment(line))
   8                          found = True
   9                          break
  10                  if not found:
  11                      self.in_processing_instructions = 0
  12                  else:
  13                      continue # do not parse this line
  14                     
  15 +            # ignore comment lines beyond processing instructions
  16 +            if line.startswith("##") and not self.in_pre:
  17 +                continue # do not parse this line
  18 + 

Within a parser section (self.in_pre == true), ##comment lines are not ignored, unless the parser section is {{{#!wiki....

The patch makes the checking for lines starting with ## after this point redundant, but this does not seem to create a problem.

This patch has been checked against slideshow_wiki.py parser, which uses ##slide and ##notes as tokens for delimiting slides and slide notes, and it does not prevent the slideshow parser working. However, if a list is split across two slides, the numbering for the second slides with need to be restarted explicitly (e.g. 1.#7). So little has been gained by the patch in this context.

Hits

367