Re: Issues with gendocs and images

[Akim Demaille]

Hi Karl,

Le 25 oct. 2012 à 23:49, Karl Berry a écrit :

>    Actually, why would we rename it instead of simply passing to -o the
>    wanted directory, instead of creating this directory?
> 
> No, this has nothing to do with -o.  gendocs runs makeinfo to generate
> the split HTML, which is created in a build subdirectory PACKAGE.html/.
> This happens inside the source/build tree.  Then we want to move the
> results to the "target" $outdir directory ("manual/" by default), just
> as we move the .pdf and the rest.

So the point is really to have one work area, and then move
"finished" products to the right place, right?  So basically,
that is why I had a release of Bison with wrong info
files: the work area is the regular builddir of the project,
yet gendocs.sh does not use the same setup as the hosting
project…

Really, wouldn't it be saner to build in the output zone, and
kill it on error instead?

>    My questions are related to changes I am making to import the needed
>    images
> 
> TIA :).

YAWIA :)

Here is my proposal.  It works for Bison.  I'm surprised that the
tarballs do not expand in a single directory, but in many files.
Also, it might not be so great to simply gzip the monolithic HTML
file, as it prevents providing more files (say images for instance :).
A tarball would make sense too.

0001-gendocs-copy-images-need-by-HTML-output.patch
Description: Binary data