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

Re: Bug#948120: libreoffice: Incorrect conditional test to enable BUILD_NOGUI_PACKAGES



On 1/4/20 2:57 PM, Rene Engelhard wrote:
> Yes, because BUILD_NOGUI_PACKAGES=y is set because of findstring vs. filter.
> 
> There you *are* right, I didn't say anything else.

This bug report was just about this particular issue
of using the wrong keyword.

>> What? I'm not sure why this would be a workaround.
> 
> But that just disables -nogui there which is not what we want
> -> workaround.
> Non-workaround would be to fix OOO_NOGUI_ARCHS to have ppc64.

Well, whether you want ppc64 to build the "nogui" packages
is a different story. The point is, currently it's not part of
OOO_NOGUI_ARCHS, so it should be able to build without
the "nogui" packages.

Whether you want "nogui" or not is a matter of opinion, I
don't have a strong preference in this case. If you want
to build the nogui packages on ppc64, knock yourself out.

Feel free to do the same for powerpc as the build machine
for ppc64 and powerpc is the same and the new one for
both architectures will be even faster.

>> You didn't put ppc64 in the list of NOGUI architectures.
> 
> Yes, because I didn't have it in radar when I did this.

Sure. But that was not my point.

>> That means the nogui code itself wasn't build.
> 
> Erm, no. Not at the stage of the failure.
> 
> (...)
> 
> And a build log clearly shows 1,1a,2,5 done.

I did not go through the whole rules script to check for
the build order. I assumed that you would install the
packages at the end of the build, not inbetween. But
anyway, it's not really important for this bug report.

>> You want to use "filter" when looking for architectures, not "findstring"
>> which will match "ppc64" in "ppc64el".
> 
> Maybe, yes. But in fact ppc64 should be added to OOO_NOGUI_ARCHS anyways, which is
> the real issue here.

Again, whether ppc64 is in OOO_NOGUI_ARCHS or not, is a matter
of taste. Whether you should use "filter" or "findstring" is
not, because the latter matches substrings and hence will match
architectures that you didn't mean to match. That's my point.

And I think it's probably a good idea to check all occurrences
of "findstring" in debian/rules to make sure that it really
matches the strings that you intended to match.

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaubitz@debian.org
`. `'   Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913


Reply to: