| Message ID | 20190113071850.14337-1-ericbavier@centurylink.net |
|---|---|
| State | Accepted |
| Headers | show |
| Series | [bug#34056] core-updates -- gnu: python2: Fix test flags. | expand |
| Context | Check | Description |
|---|---|---|
| cbaines/package builds | pending | In Cuirass |
| cbaines/applying patch | fail | Apply failed |
On Sun, Jan 13, 2019 at 01:18:49AM -0600, ericbavier@centurylink.net wrote: > From: Eric Bavier <bavier@member.fsf.org> > > * gnu/packages/python.scm (python-2.7)[arguments]: 'EXTRATESTOPTS' -> > 'TESTOPTS'. This overrides the default '-l' argument for memory leak checks > which is not compatible with the -j for parallelism. > --- > gnu/packages/python.scm | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/gnu/packages/python.scm b/gnu/packages/python.scm > index 9b43f465cc..dffded738d 100644 > --- a/gnu/packages/python.scm > +++ b/gnu/packages/python.scm > @@ -194,7 +194,7 @@ > (assoc-ref %outputs "out") "/lib")) > ;; With no -j argument tests use all available cpus, so provide one. > #:make-flags > - (list (format #f "EXTRATESTOPTS=-j~d" (parallel-job-count))) > + (list (format #f "TESTOPTS=-j~d" (parallel-job-count))) > > #:modules ((ice-9 ftw) (ice-9 match) > (guix build utils) (guix build gnu-build-system)) > -- > 2.20.1 > I'm building out python2 now on my aarch64 board but I assume it'll work. While I do prefer the parallelism in the test suite, between the two I'd consider the memory leak checks the more important of the two.
On Sun, 13 Jan 2019 09:48:59 +0200 Efraim Flashner <efraim@flashner.co.il> wrote: > On Sun, Jan 13, 2019 at 01:18:49AM -0600, ericbavier@centurylink.net wrote: > > From: Eric Bavier <bavier@member.fsf.org> > > > > * gnu/packages/python.scm (python-2.7)[arguments]: 'EXTRATESTOPTS' -> > > 'TESTOPTS'. This overrides the default '-l' argument for memory leak checks > > which is not compatible with the -j for parallelism. > > --- > > gnu/packages/python.scm | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/gnu/packages/python.scm b/gnu/packages/python.scm > > index 9b43f465cc..dffded738d 100644 > > --- a/gnu/packages/python.scm > > +++ b/gnu/packages/python.scm > > @@ -194,7 +194,7 @@ > > (assoc-ref %outputs "out") "/lib")) > > ;; With no -j argument tests use all available cpus, so provide one. > > #:make-flags > > - (list (format #f "EXTRATESTOPTS=-j~d" (parallel-job-count))) > > + (list (format #f "TESTOPTS=-j~d" (parallel-job-count))) > > > > #:modules ((ice-9 ftw) (ice-9 match) > > (guix build utils) (guix build gnu-build-system)) > > -- > > 2.20.1 > > > > I'm building out python2 now on my aarch64 board but I assume it'll > work. While I do prefer the parallelism in the test suite, between the > two I'd consider the memory leak checks the more important of the two. > Are the memory-leak checks something that we need to worry about in our packaging of python2? Can memory leaks be introduced in the way we configure and build our python? Otherwise I assume that's something more interesting to someone developing python itself and not as interesting in system CI. IMHO. I see I should probably move the comment about this overriding -l from the commit message to the code comment. `~Eric
On Sun, Jan 13, 2019 at 10:23:20AM -0600, Eric Bavier wrote: > On Sun, 13 Jan 2019 09:48:59 +0200 > Efraim Flashner <efraim@flashner.co.il> wrote: > > > On Sun, Jan 13, 2019 at 01:18:49AM -0600, ericbavier@centurylink.net wrote: > > > From: Eric Bavier <bavier@member.fsf.org> > > > > > > * gnu/packages/python.scm (python-2.7)[arguments]: 'EXTRATESTOPTS' -> > > > 'TESTOPTS'. This overrides the default '-l' argument for memory leak checks > > > which is not compatible with the -j for parallelism. > > > --- > > > gnu/packages/python.scm | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/gnu/packages/python.scm b/gnu/packages/python.scm > > > index 9b43f465cc..dffded738d 100644 > > > --- a/gnu/packages/python.scm > > > +++ b/gnu/packages/python.scm > > > @@ -194,7 +194,7 @@ > > > (assoc-ref %outputs "out") "/lib")) > > > ;; With no -j argument tests use all available cpus, so provide one. > > > #:make-flags > > > - (list (format #f "EXTRATESTOPTS=-j~d" (parallel-job-count))) > > > + (list (format #f "TESTOPTS=-j~d" (parallel-job-count))) > > > > > > #:modules ((ice-9 ftw) (ice-9 match) > > > (guix build utils) (guix build gnu-build-system)) > > > -- > > > 2.20.1 > > > > > > > I'm building out python2 now on my aarch64 board but I assume it'll > > work. While I do prefer the parallelism in the test suite, between the > > two I'd consider the memory leak checks the more important of the two. > > > > Are the memory-leak checks something that we need to worry about in our > packaging of python2? Can memory leaks be introduced in the way we > configure and build our python? Otherwise I assume that's something > more interesting to someone developing python itself and not as > interesting in system CI. IMHO. No idea. > I see I should probably move the comment about this overriding -l from > the commit message to the code comment. > Sounds like a good idea :)
ericbavier@centurylink.net writes: > From: Eric Bavier <bavier@member.fsf.org> > > * gnu/packages/python.scm (python-2.7)[arguments]: 'EXTRATESTOPTS' -> > 'TESTOPTS'. This overrides the default '-l' argument for memory leak checks > which is not compatible with the -j for parallelism. > --- > gnu/packages/python.scm | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/gnu/packages/python.scm b/gnu/packages/python.scm > index 9b43f465cc..dffded738d 100644 > --- a/gnu/packages/python.scm > +++ b/gnu/packages/python.scm > @@ -194,7 +194,7 @@ > (assoc-ref %outputs "out") "/lib")) > ;; With no -j argument tests use all available cpus, so provide one. > #:make-flags > - (list (format #f "EXTRATESTOPTS=-j~d" (parallel-job-count))) > + (list (format #f "TESTOPTS=-j~d" (parallel-job-count))) > > #:modules ((ice-9 ftw) (ice-9 match) > (guix build utils) (guix build gnu-build-system)) Hey Eric, I've also been looking at some Python 2 things on core-updates today, trying to get the linkchecker package building again. I've just pushed a slightly different patch, but to this effect to core-updates, and only just seen this message, sorry about that. Thanks for looking at it though, this approach is probably neater, as I used substitute* instead. Chris
On Sun, 13 Jan 2019 22:56:27 +0000 Christopher Baines <mail@cbaines.net> wrote: > ericbavier@centurylink.net writes: > > > From: Eric Bavier <bavier@member.fsf.org> > > > > * gnu/packages/python.scm (python-2.7)[arguments]: 'EXTRATESTOPTS' -> > > 'TESTOPTS'. This overrides the default '-l' argument for memory leak checks > > which is not compatible with the -j for parallelism. > > --- > > gnu/packages/python.scm | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/gnu/packages/python.scm b/gnu/packages/python.scm > > index 9b43f465cc..dffded738d 100644 > > --- a/gnu/packages/python.scm > > +++ b/gnu/packages/python.scm > > @@ -194,7 +194,7 @@ > > (assoc-ref %outputs "out") "/lib")) > > ;; With no -j argument tests use all available cpus, so provide one. > > #:make-flags > > - (list (format #f "EXTRATESTOPTS=-j~d" (parallel-job-count))) > > + (list (format #f "TESTOPTS=-j~d" (parallel-job-count))) > > > > #:modules ((ice-9 ftw) (ice-9 match) > > (guix build utils) (guix build gnu-build-system)) > > Hey Eric, > > I've also been looking at some Python 2 things on core-updates today, > trying to get the linkchecker package building again. > > I've just pushed a slightly different patch, but to this effect to > core-updates, and only just seen this message, sorry about that. Not a problem. > > Thanks for looking at it though, this approach is probably neater, as I > used substitute* instead. > Would you like to keep your patch, or apply this one instead? `~Eric
Eric Bavier <ericbavier@centurylink.net> writes: > On Sun, 13 Jan 2019 22:56:27 +0000 > Christopher Baines <mail@cbaines.net> wrote: > >> ericbavier@centurylink.net writes: >> >> > From: Eric Bavier <bavier@member.fsf.org> >> > >> > * gnu/packages/python.scm (python-2.7)[arguments]: 'EXTRATESTOPTS' -> >> > 'TESTOPTS'. This overrides the default '-l' argument for memory leak checks >> > which is not compatible with the -j for parallelism. >> > --- >> > gnu/packages/python.scm | 2 +- >> > 1 file changed, 1 insertion(+), 1 deletion(-) >> > >> > diff --git a/gnu/packages/python.scm b/gnu/packages/python.scm >> > index 9b43f465cc..dffded738d 100644 >> > --- a/gnu/packages/python.scm >> > +++ b/gnu/packages/python.scm >> > @@ -194,7 +194,7 @@ >> > (assoc-ref %outputs "out") "/lib")) >> > ;; With no -j argument tests use all available cpus, so provide one. >> > #:make-flags >> > - (list (format #f "EXTRATESTOPTS=-j~d" (parallel-job-count))) >> > + (list (format #f "TESTOPTS=-j~d" (parallel-job-count))) >> > >> > #:modules ((ice-9 ftw) (ice-9 match) >> > (guix build utils) (guix build gnu-build-system)) >> >> Hey Eric, >> >> I've also been looking at some Python 2 things on core-updates today, >> trying to get the linkchecker package building again. >> >> I've just pushed a slightly different patch, but to this effect to >> core-updates, and only just seen this message, sorry about that. > > Not a problem. > >> Thanks for looking at it though, this approach is probably neater, as I >> used substitute* instead. > > Would you like to keep your patch, or apply this one instead? I'm quite happy to switch to this approach. I don't know whether the best way to do that is to switch out the patches by force-pushing to core-updates, or to just apply your patch on top though... Thanks, Chris
On Mon, 21 Jan 2019 19:10:06 +0000 Christopher Baines <mail@cbaines.net> wrote: > Eric Bavier <ericbavier@centurylink.net> writes: > > > On Sun, 13 Jan 2019 22:56:27 +0000 > > Christopher Baines <mail@cbaines.net> wrote: > > > >> ericbavier@centurylink.net writes: > >> > >> > From: Eric Bavier <bavier@member.fsf.org> > >> > > >> > * gnu/packages/python.scm (python-2.7)[arguments]: 'EXTRATESTOPTS' -> > >> > 'TESTOPTS'. This overrides the default '-l' argument for memory leak checks > >> > which is not compatible with the -j for parallelism. > >> > --- > >> > gnu/packages/python.scm | 2 +- > >> > 1 file changed, 1 insertion(+), 1 deletion(-) > >> > > >> > diff --git a/gnu/packages/python.scm b/gnu/packages/python.scm > >> > index 9b43f465cc..dffded738d 100644 > >> > --- a/gnu/packages/python.scm > >> > +++ b/gnu/packages/python.scm > >> > @@ -194,7 +194,7 @@ > >> > (assoc-ref %outputs "out") "/lib")) > >> > ;; With no -j argument tests use all available cpus, so provide one. > >> > #:make-flags > >> > - (list (format #f "EXTRATESTOPTS=-j~d" (parallel-job-count))) > >> > + (list (format #f "TESTOPTS=-j~d" (parallel-job-count))) > >> > > >> > #:modules ((ice-9 ftw) (ice-9 match) > >> > (guix build utils) (guix build gnu-build-system)) > >> > >> Hey Eric, > >> > >> I've also been looking at some Python 2 things on core-updates today, > >> trying to get the linkchecker package building again. > >> > >> I've just pushed a slightly different patch, but to this effect to > >> core-updates, and only just seen this message, sorry about that. > > > > Not a problem. > > > >> Thanks for looking at it though, this approach is probably neater, as I > >> used substitute* instead. > > > > Would you like to keep your patch, or apply this one instead? > > I'm quite happy to switch to this approach. I don't know whether the > best way to do that is to switch out the patches by force-pushing to > core-updates, or to just apply your patch on top though... I can just adjust my patch. I think we don't want to force-push since there are people working on it. Thanks for your attention to core-updates :) `~Eric
On Mon, 21 Jan 2019 19:16:22 -0600 Eric Bavier <ericbavier@centurylink.net> wrote: > On Mon, 21 Jan 2019 19:10:06 +0000 > Christopher Baines <mail@cbaines.net> wrote: > > > Eric Bavier <ericbavier@centurylink.net> writes: > > > > > On Sun, 13 Jan 2019 22:56:27 +0000 > > > Christopher Baines <mail@cbaines.net> wrote: > > > > > >> ericbavier@centurylink.net writes: > > >> > > >> > From: Eric Bavier <bavier@member.fsf.org> > > >> > > > >> > * gnu/packages/python.scm (python-2.7)[arguments]: 'EXTRATESTOPTS' -> > > >> > 'TESTOPTS'. This overrides the default '-l' argument for memory leak checks > > >> > which is not compatible with the -j for parallelism. > > >> > --- > > >> > gnu/packages/python.scm | 2 +- > > >> > 1 file changed, 1 insertion(+), 1 deletion(-) > > >> > > > >> > diff --git a/gnu/packages/python.scm b/gnu/packages/python.scm > > >> > index 9b43f465cc..dffded738d 100644 > > >> > --- a/gnu/packages/python.scm > > >> > +++ b/gnu/packages/python.scm > > >> > @@ -194,7 +194,7 @@ > > >> > (assoc-ref %outputs "out") "/lib")) > > >> > ;; With no -j argument tests use all available cpus, so provide one. > > >> > #:make-flags > > >> > - (list (format #f "EXTRATESTOPTS=-j~d" (parallel-job-count))) > > >> > + (list (format #f "TESTOPTS=-j~d" (parallel-job-count))) > > >> > > > >> > #:modules ((ice-9 ftw) (ice-9 match) > > >> > (guix build utils) (guix build gnu-build-system)) > > >> > > >> Hey Eric, > > >> > > >> I've also been looking at some Python 2 things on core-updates today, > > >> trying to get the linkchecker package building again. > > >> > > >> I've just pushed a slightly different patch, but to this effect to > > >> core-updates, and only just seen this message, sorry about that. > > > > > > Not a problem. > > > > > >> Thanks for looking at it though, this approach is probably neater, as I > > >> used substitute* instead. > > > > > > Would you like to keep your patch, or apply this one instead? > > > > I'm quite happy to switch to this approach. I don't know whether the > > best way to do that is to switch out the patches by force-pushing to > > core-updates, or to just apply your patch on top though... > > I can just adjust my patch. An adjusted patch, which reverts Christopher's patch was pushed to core-updates in commit 32922190977f701694880254a71feb2752222f94 `~Eric
diff --git a/gnu/packages/python.scm b/gnu/packages/python.scm index 9b43f465cc..dffded738d 100644 --- a/gnu/packages/python.scm +++ b/gnu/packages/python.scm @@ -194,7 +194,7 @@ (assoc-ref %outputs "out") "/lib")) ;; With no -j argument tests use all available cpus, so provide one. #:make-flags - (list (format #f "EXTRATESTOPTS=-j~d" (parallel-job-count))) + (list (format #f "TESTOPTS=-j~d" (parallel-job-count))) #:modules ((ice-9 ftw) (ice-9 match) (guix build utils) (guix build gnu-build-system))
From: Eric Bavier <bavier@member.fsf.org> * gnu/packages/python.scm (python-2.7)[arguments]: 'EXTRATESTOPTS' -> 'TESTOPTS'. This overrides the default '-l' argument for memory leak checks which is not compatible with the -j for parallelism. --- gnu/packages/python.scm | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)