Mailinglist Archive: opensuse-mingw (33 mails)
| < Previous | Next > |
Re: [opensuse-mingw] Invocation of make in spec file
- From: Fridrich Strba <fridrich.strba@xxxxxxxxxx>
- Date: Fri, 15 Apr 2011 17:33:30 +0200
- Message-id: <4DA8654A.50804@bluewin.ch>
On 15/04/11 16:46, Maarten Bosmans wrote:
Here, I don't agree: Some packages break with -j3 but don't with -j4 or -j2. We never know how many processes the builder worker will allow. So, this is just a precaution I took after real experience with failures. It was better then to have to run several time "osc rebuild"
Also disagree. For instance if I push a webkit update from the web-interface using my windows laptop, where I could not build the package before, but could do the diff according to the error, the multiprocess build will first finish all the processes that were not broken with their warnings and then fail. Looking for error in this situation is just a PITA, the second make being invoked as single process one, the log will have the error clearly at the end.
I really don't see the benefit of this one.
F.
--
Please avoid sending me Word, Excel or PowerPoint attachments.
See http://www.gnu.org/philosophy/no-word-attachments.html
--
To unsubscribe, e-mail: opensuse-mingw+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse-mingw+help@xxxxxxxxxxxx
-%{_mingwXX_make} %{?_smp_mflags} || %{_mingwXX_make}
+%{_mingwXX_make}
or just
-make %{?_smp_mflags} || make
+make
Why is the first style (with ||) used? I would think that a package
maintainer should know whether a package is buildable in parallel or
not and just give the right invocation, without the retry on fail. The
Here, I don't agree: Some packages break with -j3 but don't with -j4 or -j2. We never know how many processes the builder worker will allow. So, this is just a precaution I took after real experience with failures. It was better then to have to run several time "osc rebuild"
extra make also makes errors harder to spot in the build log.
Also disagree. For instance if I push a webkit update from the web-interface using my windows laptop, where I could not build the package before, but could do the diff according to the error, the multiprocess build will first finish all the processes that were not broken with their warnings and then fail. Looking for error in this situation is just a PITA, the second make being invoked as single process one, the log will have the error clearly at the end.
I'd propose to just do a batch convert to make %{?_smp_mflags} (at
least for all autotools-based packages) and if any packages fail to
build, replace that line with a simple make.
I really don't see the benefit of this one.
F.
--
Please avoid sending me Word, Excel or PowerPoint attachments.
See http://www.gnu.org/philosophy/no-word-attachments.html
--
To unsubscribe, e-mail: opensuse-mingw+unsubscribe@xxxxxxxxxxxx
For additional commands, e-mail: opensuse-mingw+help@xxxxxxxxxxxx
| < Previous | Next > |