Hello,
On penktadienis 04 Gruodis 2009 21:58:06 Joey Hess wrote:
> Modestas Vainius wrote:
> > Hello,
> >
> > On ketvirtadienis 03 Gruodis 2009 20:50:29 Joey Hess wrote:
> > > Modestas Vainius wrote:
> > > > ++test_is_parallel( do_parallel_mk("--parallel"),
> > > > ++ "DEB_BUILD_OPTIONS=paral
Modestas Vainius wrote:
> Hello,
>
> On ketvirtadienis 03 Gruodis 2009 20:50:29 Joey Hess wrote:
> > Modestas Vainius wrote:
> > > ++test_is_parallel( do_parallel_mk("--parallel"),
> > > ++"DEB_BUILD_OPTIONS=parallel=5 with --parallel" );
> >
> > Surely that code cannot work?
>
> Huh, bu
Hello,
On ketvirtadienis 03 Gruodis 2009 20:50:29 Joey Hess wrote:
> Modestas Vainius wrote:
> > ++test_is_parallel( do_parallel_mk("--parallel"),
> > ++ "DEB_BUILD_OPTIONS=parallel=5 with --parallel" );
>
> Surely that code cannot work?
Huh, but isn't the whole point of tests to check that it
Modestas Vainius wrote:
> ++test_is_parallel( do_parallel_mk("--parallel"),
> ++"DEB_BUILD_OPTIONS=parallel=5 with --parallel" );
Surely that code cannot work?
Also, I seem to see lots of changes before this that I remember
removing before, which means I will have to go back through the entir
Hello,
On trečiadienis 02 Gruodis 2009 02:44:10 Joey Hess wrote:
> --no-max-parallel is another option, but --parallel is ok by me.
The patch for --parallel is attached. It also brings support for legacy
punctuation hacks back (backwards compatibility with previous <= 7.3 series
when no paralle
--no-max-parallel is another option, but --parallel is ok by me.
--
see shy jo
signature.asc
Description: Digital signature
Hello,
On trečiadienis 02 Gruodis 2009 01:11:18 Modestas Vainius wrote:
> What about just --parallel (the name used in parallel branch) as an alias
> to --max-parallel and allowing optional arg? I know you don't like aliases
> but... in documentation, it could something among the lines of:
>
>
Hello,
On antradienis 01 Gruodis 2009 23:59:48 Joey Hess wrote:
> Modestas Vainius wrote:
> > Then how do you suggest me to allow unlimited jobs by not imposing some
> > artificial big number on jobs? I fail to see how --max-parallel=9
> > looks better?
>
> I was thinking about something like
Modestas Vainius wrote:
> Then how do you suggest me to allow unlimited jobs by not imposing some
> artificial big number on jobs? I fail to see how --max-parallel=9 looks
> better?
I was thinking about something like perhaps --max-parallel=unlimited.
--
see shy jo
signature.asc
Descrip
Hello,
On pirmadienis 30 Lapkritis 2009 21:50:09 you wrote:
> > On antradienis 24 Lapkritis 2009 00:32:09 Modestas Vainius wrote:
> > > What is more, now optional argument to --max-parallel makes sense (i.e.
> > > "unlimited parallel") again. I don't want to specify --max-parallel=-1
> > > anywher
Modestas Vainius wrote:
> Hello,
>
> On antradienis 24 Lapkritis 2009 00:32:09 Modestas Vainius wrote:
> > What is more, now optional argument to --max-parallel makes sense (i.e.
> > "unlimited parallel") again. I don't want to specify --max-parallel=-1
> > anywhere, --max-parallel looks better. T
Hello,
On antradienis 24 Lapkritis 2009 00:32:09 Modestas Vainius wrote:
> What is more, now optional argument to --max-parallel makes sense (i.e.
> "unlimited parallel") again. I don't want to specify --max-parallel=-1
> anywhere, --max-parallel looks better. The attached patch also includes re-
Hi Joey,
Joey Hess wrote:
> FWIW, re-enabling it in v8 is a possibility, but it puts an annoying
> burden on the package maintainer who wants to update to v8 for some
> other reason. They would have to test parallel builds of their package.
If it is well documented (i.e. in the upgrade checklist)
Hello,
On pirmadienis 23 Lapkritis 2009 21:23:47 Joey Hess wrote:
> FWIW, re-enabling it in v8 is a possibility, but it puts an annoying
> burden on the package maintainer who wants to update to v8 for some
> other reason. They would have to test parallel builds of their package.
> Which can take
I've spoken with the buildd maintainers, and they were entirely
unsympathetic to the idea that parallel building on their systems might
be a problem, or that it might be a good idea for debhelper to support
parallel building by default as a way to get more than a small fraction
of Debian packages t
Modestas Vainius wrote:
> Probably compat bump was worthwhile to default --max-parallel to -1 as people
> are likely to bump compat rather than set --max-parallel (imho). Joey, my
> last
> patch had --max-parallel=0 "do nothing" (or current --max-parallel=1 mode)
> which you didn't apply. Proba
Micha Lenk wrote:
> This is a very intrusive change that apparently broke a build of libaqbanking,
> which is not at all ready for parallel builds:
> https://buildd.debian.org/fetch.cgi?pkg=libaqbanking&arch=powerpc&ver=4.2.0-2&stamp=1258742157&file=log&as=raw
Why is an autobuilder enabling parall
Hello,
On pirmadienis 23 Lapkritis 2009 11:46:23 Micha Lenk wrote:
> In debhelper 7.4.4 parallel builds were enabled by default:
> > Changes:
> > debhelper (7.4.4) unstable; urgency=low
> > .
> >* The makefile buildsystem (and derived buildsystems cmake, autoconf,
> > etc) now supports paral
Hi Joey,
In debhelper 7.4.4 parallel builds were enabled by default:
> Changes:
> debhelper (7.4.4) unstable; urgency=low
> .
>* The makefile buildsystem (and derived buildsystems cmake, autoconf, etc)
> now supports parallel building by default, as specified via
> DEB_BUILD_OPTIO
19 matches
Mail list logo
