bug-gnu-utils
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [patch] skip one more glade test


From: Matthew Burgess
Subject: Re: [patch] skip one more glade test
Date: Fri, 21 May 2010 13:06:51 -0600
User-agent: RoundCube Webmail/0.1b

On Fri, 21 May 2010 01:34:42 +0200, Bruno Haible <address@hidden> wrote:
> Hi,
> 
> Matthew Burgess wrote:
>> Attached is a patch that skips it when glade support is unavailable.
> 
> Thanks for the patch. But I think the support for all documented file
> formats in xgettext is important enough that it warrants a test failure.
> 
> Yes I know that some other missing features, such as of Java or Tcl
> support,
> yield a SKIP rather than FAIL. But xgettext on Glade input files is less
> "exotic".
>
> The FAILure notice is meant as a reminder to the distributor to look again
> whether he really wants to distribute binaries with such a limitation.
> It is still the distributor's decision and responsibility to do so if he
> wants. But the FAILure serves as a warning.

OK, understood.  Would you accept the attached patch then, that makes
all glade-related tests consistent by having them all fail in the absence
of glade support in xgettext?  My intention with this and the previous
patch is simply to make all the glade tests produce consistent results,
I don't particularly mind whether they're all skipped or all fail, but
having some behave differently than the others just strikes me as a
little odd.

Whilst on the subject, is there anything else that upstream expect
rather than simply 'recommend' (as per the DEPENDENCIES') file?  Is it
worth perhaps marking such dependencies as 'Mandatory' rather than simply
'Recommended'?

My understanding is that cvs/git are runtime dependencies?  As such,
I would think that adding '--with-git' to LFS would be the right thing
to do so that our users can avoid installing the legacy CVS package, if
they need to run 'autopoint'?
 
> I plan to rewrite that support to use libxml2 instead of expat; this
> should reduce the number of prerequisite packages by 1. (libxml2 is used in
> so many places.)

Whilst that would be welcome, I think LFS can and should install expat
prior to gettext in order to respect upstream's expectations and perhaps
avoid you receiving too many bogus bug reports.
 
Thanks,

Matt.

Attachment: gettext-0.18-glade_support-1.patch
Description: Binary data


reply via email to

[Prev in Thread] Current Thread [Next in Thread]