Re: Broken zh_CN cross-references

[Gavin Smith]

On 2 November 2015 at 12:52, Brendan O'Dea <address@hidden> wrote:
> Hey folks.  I have a translation issue: I'm generating a zh_CN localised
> page using po4a and the word splitting is breaking some of the
> cross-references.
>
> Example .texi and .info files attached.  To reproduce: open the .info file,
> and jump to the node: "调用 help2man".  The first cross reference in the
> section discussing "--name" splits across two lines for me, so instead of
> attempting to lookup the node "包含文字" (which does exist), it tries to lookup
> "包含 文字", which doesn't.
>
> --bod

Once I got over the fact that the file had "@setfilename
help2man.info" in it, so the output file wasn't called
help2man-zh_CN.info, I tried suppressing line breaks with @w, as in

This option overrides an include file @samp{[name]} section (@address@hidden).

but it made no difference.

This didn't work either:

This option overrides an include file @samp{[name]} section(@address@hidden).

Also doing @node @w{包含文字} made no difference.

The only way I could get satisfactory output is by manually specifying
line breaks:

This option overrides an include file @samp{[name]} address@hidden
(@pxref{包含文字}).

The fact that @w didn't work here is possibly a bug. Even if it didn't
work, it wouldn't be a very good solution.

So there are two possible solutions, fix the reader or fix the
generator. In node names multiple spaces or newlines are collapsed to
one space. It would be unpleasant to complicate this by adding special
rules for Chinese characters. Fixing the program that generates the
Info files not to output the line breaks should be possible (i.e. an
implicit @w around the Chinese characters).