r21677 - gnucash/trunk/src/gnc-module/test - [Testing] Gnc-module: Provide checked error messages to replace noted warnings (c-files only)

[Geert Janssens]

Op woensdag 7 december 2011 06:56:54 schreef John Ralls:
> On Dec 7, 2011, at 3:30 AM, Geert Janssens wrote:
> > Op zondag 4 december 2011 19:08:10 schreef John Ralls:
> >> Author: jralls
> >> Date: 2011-12-04 19:08:10 -0500 (Sun, 04 Dec 2011)
> >> New Revision: 21677
> >> Trac: http://svn.gnucash.org/trac/changeset/21677
> >> 
> >> Modified:
> >>   gnucash/trunk/src/gnc-module/test/misc-mods/Makefile.am
> >>   gnucash/trunk/src/gnc-module/test/test-agedver.c
> >>   gnucash/trunk/src/gnc-module/test/test-dynload.c
> >>   gnucash/trunk/src/gnc-module/test/test-incompatdep.c
> >>   gnucash/trunk/src/gnc-module/test/test-load-c.c
> >>   gnucash/trunk/src/gnc-module/test/test-modsysver.c
> >> 
> >> Log:
> >> [Testing] Gnc-module: Provide checked error messages to replace noted
> >> warnings (c-files only)
> >> 
> >> 
> >> Much better for expected warnings to be tested for than to have a
> >> lead-in message saying that they're really OK.
> >> 
> >> Don't have the test utilities in Scheme yet.
> >> 
> >> Set the phony modules in misc-modules to be libtool modules instead of
> >> shared libraries. Aside from being more correct, this ensures that
> >> they'll have the same name on Linux and OSX.
> >> 
> >> _______________________________________________
> >> gnucash-patches mailing list
> >> gnucash-patches at gnucash.org
> >> https://lists.gnucash.org/mailman/listinfo/gnucash-patches
> > 
> > When I started my fixes on the tests (which you have nicely completed),
> > I
> > considered backporting the changes to the 2.4 branch. I did port most of
> > my changes like that.
> > 
> > Would it make sense to backport your test fixes and if so, before 2.4.9
> > is released this weekend ?
> 
> I'm not done yet. There are still messages that need to be suppressed.
> 
> I don't see any real benefit to backporting it. It wouldn't do any harm, but
> it's not user-facing and it's not likely to have any affect on quality in
> 2.4.
> 
Agreed. I more or less came to the same conclusion after I sent the question.

I was mostly thinking of distribution packagers that currently can't depend on 
make check to see if all is ok. Before your changes, some of the tests simply 
failed or emitted lots of ugly looking warnings you may or may not have been 
told to ignore before. Not really a comfortable situation to evaluate if the 
application is good to go.

But since we started planning towards 2.6 in a not so distant future, it may 
not really matter in 2.4.

Geert