Gabor Hojtsy wrote:
[xref on para is not supported]
The id xref is linking to, is missing in the de-tree, but present in the
missing-ids.xml file:
<para id="ini.com.allow-dcom"></para>
In the english file in question (faq/com.xml) xref is used like:
<xref linkend="ini.com.allow-dcom"/>
As this points to <para id="ini.com.allow-dcom"></para> in
missing-ids.xml for the de tree, obviously no link text can be produced.
This should not affect the building prozess,(it's only a warning) but
renders not very nice:
http://www.php.net/manual/de/faq.com.php#faq.com.q8
The link text ist empty.
Anything we can do about?
Possibilities (other than updating the german manual ;-)):
1.) use xreflabel
2.) produce content for the para's in missing-ids.xml
3.) split the faq
No splitting will help this error message popping up, since it occurs in
different places in different translations. It can happen anytime you
have an outdated file in your translation.
Yeah, I know.
It was the subtle attempt to suggest splitting of large files.
(but that's another story)
The "problem" is the
increased usage of <xref>. Previously authors were encouraged to use
<link> to add internal links, which requires text content, so this was
no problem. The adoption of <xref> lead to this problem.
Should we avoid using xref?
1. What would we put into <para xreflabel="..."> in place of the dots in
the missing-ids.xml file?
The proper link-test?
Sure, this depends on the content of the id=".."
For the example <para id="ini.com.allow-dcom"> this would become
<para id="ini.com.allow-dcom" xreflabel="com.allow_dcom">
2. How would this be better then the previous option?
In no case, just another option to generate link text via xref.
<para id="ini.com.allow-dcom">com.allow_dcom</para>
Either way, if we choose option 1 or 2 it increases processing time,
maybe we can place just some generic text inside? Not nice also...
3. This is not just a faq problem.
I agree, it was just an example of one missing-id.
Currently the missing-ids.xml file is a bit pointless; it "only" serves
a technichal purpose: building the manual, and this not too clever as it
produces unneccesary warnings and no content for the readers (I don't
blame anyone for this!)
Either we don't use xref or we should fix the generating of missing id's.
In general, I really do not care much about this "problem" and if we can
come up with a solution - fine, if not life goes on.
However, the build process for translated manuals should not be hindered
by this issue.
Friedhelm
p.s.:
True, there are some big files in the
FAQ folder..
See above, we might want to split them?
