On Fri, 6 Mar 2026 at 09:58, Scott Kostyshak <[email protected]> wrote:

> On Thu, Mar 05, 2026 at 09:10:26PM +0100, Thibaut Cuvelier wrote:
> > I ended up pushing f13d9199, a variant of the patch that doesn't have
> side
> > effects. (I had forgotten the DocBookSection parameter!)
>
> Great! I uninverted the test at b2e6ab84 thanks to your fix.
>

Just to be clear: if the test passes, it's just because the testing
procedure is not very demanding. The output is valid XML (syntactically
correct) but not valid DocBook (semantically correct: the figures have no
content apart from comments).


> The following two tests fail for me:
>
>
> INVERTED.DOCBOOK_export/examples/Language_Support/Localization_Test_%28for_Developers%29_docbook5
> export/examples/ja/Modules/Linguistics_docbook5
>
> It seems we can invert the "INVERTED" test above. i.e., that export used
> to fail but now passes. I'm not sure if thanks to your fix today, or
> thanks to a previous fix. If it's OK with you, I can "uninvert" this
> test.
>

It might very well be related, as I changed a few elements in standard
definitions (7d5b9d78).


> I'm not sure about the Linguistics_docbook5. Maybe it started failing
> due to a change in the Linguistics manual. If so, we might want to
> invert this one (since it would not be a regression in DocBook).


 I just had a look at that document and it made LyX crash… I reworked a bit
the code to output a comment instead of crashing LyX (especially as the
error scenario doesn't seem too rare). Locally, this document is valid
DocBook 5 (as of 5f92d566).
-- 
lyx-devel mailing list
[email protected]
https://lists.lyx.org/mailman/listinfo/lyx-devel

Reply via email to