gnucash master: Disable regeneration of POTFILES.in

[Geert Janssens]

On maandag 11 september 2017 23:47:44 CEST John Ralls wrote:
> > On Sep 11, 2017, at 2:00 PM, Geert Janssens <geert.gnucash at kobaltwit.be>
> > wrote:> 
> > On maandag 11 september 2017 21:24:30 CEST John Ralls wrote:
> >> Updated	 via  https://github.com/Gnucash/gnucash/commit/ac86ca94 
(commit)
> >> 
> >> 	from  https://github.com/Gnucash/gnucash/commit/3f84f484 (commit)
> >> 
> >> commit ac86ca949f1e76c0ce55e74c2ec8e094a6bfb550
> >> Author: John Ralls <jralls at ceridwen.us>
> >> Date:   Mon Sep 11 12:18:57 2017 -0700
> >> 
> >>    Disable regeneration of POTFILES.in
> >>    
> >>    On Windows CMake converts "\n" to CRLF, which causes xgettext to
> >>    mangle paths with [type: gettext/gsettings] in front of them and
> >>    also dirties the source directory with uncommittable changes.
> >>    The autotools build doesn't regenerate POTFILES.in and we've
> >> 
> >> historically maintained it by hand.
> >> 
> >> 
> >> 
> >> Summary of changes:
> >> po/CMakeLists.txt | 2 +-
> >> 1 file changed, 1 insertion(+), 1 deletion(-)
> >> 
> >> _______________________________________________
> >> gnucash-patches mailing list
> >> gnucash-patches at gnucash.org
> >> https://lists.gnucash.org/mailman/listinfo/gnucash-patches
> > 
> > Would it be ok for you to disable it only on Windows ? I've come to
> > appreciate it reminds me when I add new files or rename files on linux.
> 
> We could, but that's not what's reminding you.  check-po runs
> 'intltool-update -m' which tests for files in the tree that aren't in one
> of POTFILES.in, POTFILES.skip, or POTFILES.ignore. It's what puts up the
> "POTFILES.ignore is deprecated" warning as well.
> 
> The POTFILES.in regeneration recreates a new POTFILES.in with every run of
> CMake and silently includes any file that's not already in POTFILES.skip or
> POTFILES.ignore, defeating the test.
> 

That's true. What I really meant with 'reminds me' is that the file changes 
show up immediately in my git status. I like it that the such file changes are 
committed together with the commit that introduces them. In a way it makes the 
'intltool-update -m' test redundant.

I understand we don't want this automatic regeneration to mess up the jhbuild 
change detection. Perhaps the function MAKE_GNUCAH_POTFILES can be refined to 
only write to POTFILES.in if the content is different before and after. And we 
can add a Windows only bit that converts CRLF back to "\n". Perhaps 
configure_file() can help here ?

Geert