[
https://jira.codehaus.org/browse/DOXIA-467?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=323004#comment-323004
]
Robert Scholte commented on DOXIA-467:
--------------------------------------
The patches contain fixes for several different issue.
I've started by isolating the single trailing backward slash at the end of a
line, which has been fixed in [r1463173|http://svn.apache.org/r1463173]
The backwardslashes within a monospaced block is fixed in
[r1463228|http://svn.apache.org/r1463228]
The rest hasn't been patched, because I can't figure out which tests belong to
which Parser adjustment.
If you could split them up in smaller patches, I'm happy to apply them.
> Confluence: Inconsistent handling of '\' escape characters, incorrect
> handling of \\ inside {{monospace}} blocks.
> -----------------------------------------------------------------------------------------------------------------
>
> Key: DOXIA-467
> URL: https://jira.codehaus.org/browse/DOXIA-467
> Project: Maven Doxia
> Issue Type: Bug
> Components: Module - Confluence
> Affects Versions: 1.3
> Reporter: Valters Vingolds
> Attachments: confluence-escaping-hg.patch,
> confluence-escaping-patch-files.zip, monospace-escapes.patch
>
>
> I noticed that our generated pages are mostly nonsense where we have
> \\unc\path\strings and also some general DOS-style paths (c:\dos\run). The
> issue is that \ are "eaten" because they are treated as escape characters and
> get suppressed from output. This is incorrect, per Confluence rules, they
> should appear almost always, and only function as escapes when directly
> preceding _ and * characters with special meaning (note, in Confluence they
> also have + and - and ^ and ~ with special meaning, it would be used to
> escape them too, but we don't have that yet).
> The other issue is with \\, confluence actually does not support line breaks
> within the line (it seems to me), and only forces linebreak if \\ appears at
> the end of line. This makes sense, because otherwise writing \\unc\path is
> impossible (as you can't quite escape the '\\' with another '\'). I haven't
> yet looked how feasible it is to make \\ only work as line break if it
> appears at the end of line, but my requirement was simpler: to make \\ work
> inside {{monospace}} blocks, i.e., make it appear verbatim, and not produce
> linebreak. So that {{\\unc\path}} can be written and appear so.
> I want to also propose to make handling of italic less zealous, as we happen
> to have a lot of code variables that use underscore (variables_like_this),
> and currently the underscores get eaten, and (like) part gets italicized.
> (Escaping all underscores is not practical.)
> I wish to propose that italic modifier _ should only work if preceded by
> whitespace or * (+ some other edge cases). So _this_ should get italicized
> but something_like_this would not.
> (You can actually see the above work right here in this comment, as JIRA
> parses this text according to Confluence rules. We should strive to replicate
> this behaviour, as I see one use-case of this module is to transfer wiki
> pages created in Confluence into source tree as documentation. That does not
> work if our implementation does not faithfully render markup as Confluence
> wiki would).
> I am attaching a proposed patch.
> (Sorry this patch also includes lines for DOXIA-466, I don't know how to
> remove them. So if you already applied that patch, you will get some rejected
> hunks).
> Another part is that * and _ are always produced verbatim in monospace blocks.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
