This is the mail archive of the
newlib@sourceware.org
mailing list for the newlib project.
Re: [help-texinfo] makeinfo --html doesn't split properly if a node is named index
- From: Eli Zaretskii <eliz at gnu dot org>
- To: Jon Beniston <jon at beniston dot com>
- Cc: help-texinfo at gnu dot org, newlib at sourceware dot org
- Date: Fri, 17 Oct 2014 18:36:28 +0300
- Subject: Re: [help-texinfo] makeinfo --html doesn't split properly if a node is named index
- Authentication-results: sourceware.org; auth=none
- References: <009e01cfe9fc$2fc95db0$8f5c1910$ at beniston dot com>
- Reply-to: Eli Zaretskii <eliz at gnu dot org>
> From: "Jon Beniston" <jon@beniston.com>
> Date: Fri, 17 Oct 2014 12:19:14 +0100
>
> When using any of the split options (--split=chapter|node|section or
> --split-size=5000000), the output is corrupted. In the index.html file, what
> should be one of the sub-nodes, appears at the top, and many of the
> navigation footers on random pages are missing.
>
> I suspect this relates to the node name "index" being used. The node names
> are generated based on C function names, and there is an index function.
> Changing this name to something else, seems to fix the problem. However, all
> of the other doc formats (info/pdf etc) seem ok.
>
> Is this an issue with makeinfo --html, or should "@node index" not be used
> in the newlib tex files?
Long time ago, when the split-HTML options were implemented in
makeinfo, the design was that if several nodes map to the same file
name, they all appear in the file by that name, one after the other,
but the URLs that point to each node on that file have a #node-name
appended to them, so that the links still work correctly.
Since then makeinfo was reimplemented in Perl, and it's possible that
this feature has a bug of some sort. I suggest to post this to
bug-texinfo list, where the maintainers will see it and act on it.