unofficial mirror of guix-devel@gnu.org 
 help / color / mirror / code / Atom feed
* 80 new build failures for MIPS on the 'dbus-update' branch
@ 2015-10-21 23:50 Mark H Weaver
  2015-10-22 12:56 ` Andreas Enge
  0 siblings, 1 reply; 9+ messages in thread
From: Mark H Weaver @ 2015-10-21 23:50 UTC (permalink / raw)
  To: guix-devel

The following link shows the builds that fail on the 'dbus-update'
branch but are successful on 'master':

  http://hydra.gnu.org:3000/eval/107477?filter=&compare=107476#tabs-now-fail

Among them are about 80 new dependency failures on MIPS, including key
packages like emacs, xfce, evince, and gnome-terminal.  The reason is
that these packages now depend on 'p11-kit', which has never
successfully built on MIPS.

      Mark

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: 80 new build failures for MIPS on the 'dbus-update' branch
  2015-10-21 23:50 80 new build failures for MIPS on the 'dbus-update' branch Mark H Weaver
@ 2015-10-22 12:56 ` Andreas Enge
  2015-10-22 14:09   ` 宋文武
  0 siblings, 1 reply; 9+ messages in thread
From: Andreas Enge @ 2015-10-22 12:56 UTC (permalink / raw)
  To: Mark H Weaver; +Cc: guix-devel

On Wed, Oct 21, 2015 at 07:50:00PM -0400, Mark H Weaver wrote:
> Among them are about 80 new dependency failures on MIPS, including key
> packages like emacs, xfce, evince, and gnome-terminal.  The reason is
> that these packages now depend on 'p11-kit', which has never
> successfully built on MIPS.

Do you see where this comes in? The git commits in dbus-update do not
mention p11-kit, and
   $ guix graph gnome-terminal | grep p11
returns nothing.

Andreas

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: 80 new build failures for MIPS on the 'dbus-update' branch
  2015-10-22 12:56 ` Andreas Enge
@ 2015-10-22 14:09   ` 宋文武
  2015-10-22 14:43     ` Andreas Enge
  0 siblings, 1 reply; 9+ messages in thread
From: 宋文武 @ 2015-10-22 14:09 UTC (permalink / raw)
  To: Andreas Enge; +Cc: guix-devel

Andreas Enge <andreas@enge.fr> writes:

> On Wed, Oct 21, 2015 at 07:50:00PM -0400, Mark H Weaver wrote:
>> Among them are about 80 new dependency failures on MIPS, including key
>> packages like emacs, xfce, evince, and gnome-terminal.  The reason is
>> that these packages now depend on 'p11-kit', which has never
>> successfully built on MIPS.
>
> Do you see where this comes in? The git commits in dbus-update do not
> mention p11-kit, and
>    $ guix graph gnome-terminal | grep p11
> returns nothing.
oops, It's introduced by:
  gtk+ -> rest -> libsoup -> glib-networking -> p11-kit

if we can't fix p11-kit on mips, then rest can be removed from gtk+.
it's used for “cloud printing”.

maybe not releated, the build log of p11-kit
<http://hydra.gnu.org/build/736459/log/raw> have:

  ffi.h:224:5: warning: "HAVE_LONG_DOUBLE_VARIANT" is not defined

which is fixed in the newer version (3.2.1) of libffi.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: 80 new build failures for MIPS on the 'dbus-update' branch
  2015-10-22 14:09   ` 宋文武
@ 2015-10-22 14:43     ` Andreas Enge
  2015-10-23  9:30       ` Andreas Enge
  2015-10-25 21:33       ` Ludovic Courtès
  0 siblings, 2 replies; 9+ messages in thread
From: Andreas Enge @ 2015-10-22 14:43 UTC (permalink / raw)
  To: 宋文武; +Cc: guix-devel

On Thu, Oct 22, 2015 at 10:09:07PM +0800, 宋文武 wrote:
> Andreas Enge <andreas@enge.fr> writes:
> > Do you see where this comes in? The git commits in dbus-update do not
> > mention p11-kit, and
> >    $ guix graph gnome-terminal | grep p11
> > returns nothing.

Sorry, I was confused here and did not check with the dbus-update branch.

> oops, It's introduced by:
>   gtk+ -> rest -> libsoup -> glib-networking -> p11-kit
> if we can't fix p11-kit on mips, then rest can be removed from gtk+.
> it's used for “cloud printing”.

This looks like a good temporary fix, to disentangle the dbus-update branch.
My impression is that we should not have changed so much at the same time,
anyway, and instead have concentrated on the dbus-update and closely related
modifications, then merged back, then done other modifications and updates
like gtk+.

> maybe not releated, the build log of p11-kit
> <http://hydra.gnu.org/build/736459/log/raw> have:
>   ffi.h:224:5: warning: "HAVE_LONG_DOUBLE_VARIANT" is not defined
> which is fixed in the newer version (3.2.1) of libffi.

Interesting, so maybe the problem will be solved; then we can still add
rest as an input to gtk+ again.

Thanks!

Andreas

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: 80 new build failures for MIPS on the 'dbus-update' branch
  2015-10-22 14:43     ` Andreas Enge
@ 2015-10-23  9:30       ` Andreas Enge
  2015-10-23 10:08         ` Andreas Enge
  2015-10-24  1:48         ` 宋文武
  2015-10-25 21:33       ` Ludovic Courtès
  1 sibling, 2 replies; 9+ messages in thread
From: Andreas Enge @ 2015-10-23  9:30 UTC (permalink / raw)
  To: 宋文武; +Cc: guix-devel

On Thu, Oct 22, 2015 at 04:43:29PM +0200, Andreas Enge wrote:
> On Thu, Oct 22, 2015 at 10:09:07PM +0800, 宋文武 wrote:
> > oops, It's introduced by:
> >   gtk+ -> rest -> libsoup -> glib-networking -> p11-kit
> > if we can't fix p11-kit on mips, then rest can be removed from gtk+.
> > it's used for “cloud printing”.
> This looks like a good temporary fix, to disentangle the dbus-update branch.

I gave it a try on my local machine, but several tests fail since they want
to open my display:
make[3]: Entering directory '/tmp/nix-build-gtk+-3.18.2.drv-0/gtk+-3.18.2/testsuite/gdk'
TEST: cairo... (pid=2676)

(/tmp/nix-build-gtk+-3.18.2.drv-0/gtk+-3.18.2/testsuite/gdk/.libs/cairo:2676): Gdk-WARNING **: cannot open display: 
FAIL: cairo
TEST: display... (pid=2691)
  /display/unset-display:                                              OK
  /display/bad-display:                                                OK
PASS: display
TEST: encoding... (pid=2710)

(/tmp/nix-build-gtk+-3.18.2.drv-0/gtk+-3.18.2/testsuite/gdk/.libs/encoding:2710): Gdk-WARNING **: cannot open display: 
FAIL: encoding
TEST: keysyms... (pid=2724)

(/tmp/nix-build-gtk+-3.18.2.drv-0/gtk+-3.18.2/testsuite/gdk/.libs/keysyms:2724): Gdk-WARNING **: cannot open display: 
FAIL: keysyms
TEST: rgba... (pid=2738)
  /rgba/parse:                                                         OK
  /rgba/to-string:                                                     OK
  /rgba/copy:                                                          OK
  /rgba/parse/nonsense:                                                OK
PASS: rgba
Makefile:952: recipe for target 'test-cwd' failed

What next?

Andreas

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: 80 new build failures for MIPS on the 'dbus-update' branch
  2015-10-23  9:30       ` Andreas Enge
@ 2015-10-23 10:08         ` Andreas Enge
  2015-10-24  1:48         ` 宋文武
  1 sibling, 0 replies; 9+ messages in thread
From: Andreas Enge @ 2015-10-23 10:08 UTC (permalink / raw)
  To: 宋文武; +Cc: guix-devel

Another candidate for fixing in this list:
   http://hydra.gnu.org/eval/107489?compare=master#tabs-now-fail
is libgnomecanvasmm, and a few other packages that now fail everywhere.
For instance bluez, which blocks libreoffice.

Andreas

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: 80 new build failures for MIPS on the 'dbus-update' branch
  2015-10-23  9:30       ` Andreas Enge
  2015-10-23 10:08         ` Andreas Enge
@ 2015-10-24  1:48         ` 宋文武
  2015-10-24  7:48           ` Andreas Enge
  1 sibling, 1 reply; 9+ messages in thread
From: 宋文武 @ 2015-10-24  1:48 UTC (permalink / raw)
  To: Andreas Enge; +Cc: guix-devel

Andreas Enge <andreas@enge.fr> writes:

> On Thu, Oct 22, 2015 at 04:43:29PM +0200, Andreas Enge wrote:
>> On Thu, Oct 22, 2015 at 10:09:07PM +0800, 宋文武 wrote:
>> > oops, It's introduced by:
>> >   gtk+ -> rest -> libsoup -> glib-networking -> p11-kit
>> > if we can't fix p11-kit on mips, then rest can be removed from gtk+.
>> > it's used for “cloud printing”.
>> This looks like a good temporary fix, to disentangle the dbus-update branch.
>
> I gave it a try on my local machine, but several tests fail since they want
> to open my display:
> make[3]: Entering directory '/tmp/nix-build-gtk+-3.18.2.drv-0/gtk+-3.18.2/testsuite/gdk'
> TEST: cairo... (pid=2676)
>
> (/tmp/nix-build-gtk+-3.18.2.drv-0/gtk+-3.18.2/testsuite/gdk/.libs/cairo:2676): Gdk-WARNING **: cannot open display: 
> FAIL: cairo
> TEST: display... (pid=2691)
>   /display/unset-display:                                              OK
>   /display/bad-display:                                                OK
> PASS: display
> TEST: encoding... (pid=2710)
>
> (/tmp/nix-build-gtk+-3.18.2.drv-0/gtk+-3.18.2/testsuite/gdk/.libs/encoding:2710):
> Gdk-WARNING **: cannot open display:
> FAIL: encoding
> TEST: keysyms... (pid=2724)
>
> (/tmp/nix-build-gtk+-3.18.2.drv-0/gtk+-3.18.2/testsuite/gdk/.libs/keysyms:2724):
> Gdk-WARNING **: cannot open display:
> FAIL: keysyms
> TEST: rgba... (pid=2738)
>   /rgba/parse:                                                         OK
>   /rgba/to-string:                                                     OK
>   /rgba/copy:                                                          OK
>   /rgba/parse/nonsense:                                                OK
> PASS: rgba
> Makefile:952: recipe for target 'test-cwd' failed
I remove rest from gtk+, build successful locally (x86_64), and pushed.
Are you fails on mips64el?
>
> What next?
Not sure, but I will fix as much as I can.

Lesson learned, next time, I won't do unreleated updates in a topic branch.

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: 80 new build failures for MIPS on the 'dbus-update' branch
  2015-10-24  1:48         ` 宋文武
@ 2015-10-24  7:48           ` Andreas Enge
  0 siblings, 0 replies; 9+ messages in thread
From: Andreas Enge @ 2015-10-24  7:48 UTC (permalink / raw)
  To: 宋文武; +Cc: guix-devel

On Sat, Oct 24, 2015 at 09:48:00AM +0800, 宋文武 wrote:
> I remove rest from gtk+, build successful locally (x86_64), and pushed.
> Are you fails on mips64el?

No, also on x86_64. I am not sure what happened, maybe I made a mistake
somewhere.

Andreas

^ permalink raw reply	[flat|nested] 9+ messages in thread

* Re: 80 new build failures for MIPS on the 'dbus-update' branch
  2015-10-22 14:43     ` Andreas Enge
  2015-10-23  9:30       ` Andreas Enge
@ 2015-10-25 21:33       ` Ludovic Courtès
  1 sibling, 0 replies; 9+ messages in thread
From: Ludovic Courtès @ 2015-10-25 21:33 UTC (permalink / raw)
  To: Andreas Enge; +Cc: guix-devel

Andreas Enge <andreas@enge.fr> skribis:

> This looks like a good temporary fix, to disentangle the dbus-update branch.
> My impression is that we should not have changed so much at the same time,
> anyway, and instead have concentrated on the dbus-update and closely related
> modifications, then merged back, then done other modifications and updates
> like gtk+.

+1

Ludo'.

^ permalink raw reply	[flat|nested] 9+ messages in thread

end of thread, other threads:[~2015-10-25 21:33 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-10-21 23:50 80 new build failures for MIPS on the 'dbus-update' branch Mark H Weaver
2015-10-22 12:56 ` Andreas Enge
2015-10-22 14:09   ` 宋文武
2015-10-22 14:43     ` Andreas Enge
2015-10-23  9:30       ` Andreas Enge
2015-10-23 10:08         ` Andreas Enge
2015-10-24  1:48         ` 宋文武
2015-10-24  7:48           ` Andreas Enge
2015-10-25 21:33       ` Ludovic Courtès

Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/guix.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).