[opensuse-factory] GCC 4.7 status

older
[opensuse-factory] /etc/pristine -...

Dominique Leuenberger a.k.a DimStar

10 Apr 2012 10 Apr '12

08:58

Hi all, as it has become custom to send out regular updates (be it only to keep people aware of the topic), here another round for the "GCC 4.7 as default in Factory" - Update: Not much has happened in the last week. A critical component failing to build with Gcc 4.7 is currently libzypp, which, understandably, blocks the switch (well, I think it could always be decided to switch to 4.7, but keep zypp build with 4.6?) Anyway, let's not paint it too dark: libzypp folks are helpful and looked into the matter. The first result was a regression in gcc which had been identified for libzypp. There are currently 22 packages failing in my branch (home:dimstar:gcc47): libzypp, which causes at least PackageKit to fail as a result. blender (this might be to do with the openCOLLADA changes... could be trickier to find the cause) ekiga/opal: nice bundle.. 'stable' series has been stale for a while. unstable series has it fixed, but backporting 'just that' is close to, well, a lot of work :) gsd: not looked at that one yet jack: likely not gcc 4.7 related, as it fails the same way in Factory.. it hangs the build bot. kdebase3/kdepim3: presumably not too hard to patch... maybe I'll find time. kvm: error: call to '__warn_memset_zero_len' declared with attribute warning: memset used with constant zero length (sounds like just a nice bug :) ) libQtWebKit4: I know AJ looked at that one already, but more work is needed. limal*: not really sure: looks like it just does not like the new boost libs. llvm: honestly, I have no clue, especially as it fails in stage 2... so the integration to the new headers seems to cause griev. mariadb: conversion errors... openvas-administrator: not really gcc4.7 related.. it's actually a glib 2.32 fallout.. rubygem-passenger: likely it does not like new boost libs seamonkey: big beast... anybody? sleuthkit: just a buffer overflow in ext2fs... it actually builds; just brp check raises it. Wondering why it's not triggered with gcc 4.6 though. virtualbox: nasty... an update to the latest version should fix the gcc 4.7 issues (tried in my branch... raises a can of new worms) If you feel like working on any of those, don't hesitate! The more people looking at them, the higher chance we get them fixed... and the more likely we can actually switch to Gcc 4.7 in time for 12.2 (mind it: we're passing M3... so there is not much time left if we want to do that switch for 12.2). Looking forward for all your help! Dominique -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Show replies by date

Robert Kaiser

10 Apr 10 Apr

12:48

Dominique Leuenberger a.k.a DimStar schrieb:

...

seamonkey: big beast... anybody?

Almost all of its code is identical to Firefox and/or Thunderbird, does it have a different problem than those? If so, please point me to it and I'll bring it up with the upstream team - after all, I'm the former project lead of SeaMonkey. ;-) Robert Kaiser -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dominique Leuenberger

13:34

Quoting Robert Kaiser :

...

Dominique Leuenberger a.k.a DimStar schrieb:

...
seamonkey: big beast... anybody?

Almost all of its code is identical to Firefox and/or Thunderbird, does it have a different problem than those? If so, please point me to it and I'll bring it up with the upstream team - after all, I'm the former project lead of SeaMonkey. ;-)

For FF and ThunderBird I agreed with Wolfgang to not take any action, as 'the next update' he prepared had it solved. I think it's worthy to verify if there is also an update for SM in the pipeline taking care of that. @Wolfgang: can you comment on that? Before Robert (or anybody else) starts wasting resources on that one. Best regards, Dominique -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Wolfgang Rosenauer

11 Apr 11 Apr

07:13

Hi, Am 10.04.2012 15:34, schrieb Dominique Leuenberger:

...

Quoting Robert Kaiser :

...
Dominique Leuenberger a.k.a DimStar schrieb:

...
seamonkey: big beast... anybody?

Almost all of its code is identical to Firefox and/or Thunderbird, does it have a different problem than those? If so, please point me to it and I'll bring it up with the upstream team - after all, I'm the former project lead of SeaMonkey. ;-)

For FF and ThunderBird I agreed with Wolfgang to not take any action, as 'the next update' he prepared had it solved.

I think it's worthy to verify if there is also an update for SM in the pipeline taking care of that.

@Wolfgang: can you comment on that? Before Robert (or anybody else) starts wasting resources on that one.

I've replied earlier but something in our list setup refused to send the mail. No idea why as it was just a simple reply from K9-Mail :-( and I have very limited internet access atm. So basically a new SeaMonkey version is released together with Firefox and Thunderbird to pick up the new Gecko release. I have not checked SeaMonkey specifically against gcc 4.7 but I know that Firefox should be fixed and most stuff is shared anyway. Therefore my expectation is that Seamonkey-next is already fixed and I will take care of it in case it's not. Is it urgent to verify? Releases are scheduled for 24th so I can verify around that date (probably a bit earlier but I'm travelling atm). Wolfgang -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dominique Leuenberger a.k.a DimStar

09:03

Quoting Wolfgang Rosenauer :

...

Hi, I've replied earlier but something in our list setup refused to send the mail. No idea why as it was just a simple reply from K9-Mail :-( and I have very limited internet access atm.

So basically a new SeaMonkey version is released together with Firefox and Thunderbird to pick up the new Gecko release. I have not checked SeaMonkey specifically against gcc 4.7 but I know that Firefox should be fixed and most stuff is shared anyway. Therefore my expectation is that Seamonkey-next is already fixed and I will take care of it in case it's not. Is it urgent to verify? Releases are scheduled for 24th so I can verify around that date (probably a bit earlier but I'm travelling atm).

Wolfgang, Thanks for your reply.... that's pretty much what I expected to happen with Seamonkey, so never touched it. I'm not sure about how 'urgent' it is. There has not been any commitment or statement IF we actually switch to Gcc 4.7 as default compiler; Of course with all the time I invested on fixing packages I would like to see this happen now too... Also, we are at M3, not sure how much longer we can/wait if we want to do the switch... I think post M3 release should be a good time... some more packages might fail which I did not spot (I generally ignored packages also failing in openSUSE:Factory and rated them as not-gcc related failures.. which can be wrong of course). Dominique -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Claudio Freire

10 Apr 10 Apr

14:34

On Tue, Apr 10, 2012 at 5:58 AM, Dominique Leuenberger a.k.a DimStar wrote:

...

jack: likely not gcc 4.7 related, as it fails the same way in Factory.. it hangs the build bot.

I've seen this happen in big projects at link time, since linking uses a lot of memory (and makes the build host thrash). Does it hang when linking? Can we see a link to check it out? If it's the case, it could be fixed by passing an option to ld that makes it use a lot less memory while linking, instead using more CPU - which is fine in this case, it's better than thrashing. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dominique Leuenberger a.k.a DimStar

14:42

Quoting Claudio Freire :

...

On Tue, Apr 10, 2012 at 5:58 AM, Dominique Leuenberger a.k.a DimStar wrote:

...
jack: likely not gcc 4.7 related, as it fails the same way in Factory.. it hangs the build bot.

I've seen this happen in big projects at link time, since linking uses a lot of memory (and makes the build host thrash).

Does it hang when linking? Can we see a link to check it out?

If it's the case, it could be fixed by passing an option to ld that makes it use a lot less memory while linking, instead using more CPU - which is fine in this case, it's better than thrashing.

Sure, can be seen in the project openSUSE:Factory (package jack) and also in home:dimstar:gcc47 (package jack). Every once in a while it works just fine... (might depend on which worker you end up on). Dominique -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dominique Leuenberger a.k.a DimStar

11 Apr 11 Apr

08:58

Quoting Claudio Freire :

...

On Tue, Apr 10, 2012 at 5:58 AM, Dominique Leuenberger a.k.a DimStar wrote:

...
jack: likely not gcc 4.7 related, as it fails the same way in Factory.. it hangs the build bot.

I've seen this happen in big projects at link time, since linking uses a lot of memory (and makes the build host thrash).

Does it hang when linking? Can we see a link to check it out?

If it's the case, it could be fixed by passing an option to ld that makes it use a lot less memory while linking, instead using more CPU - which is fine in this case, it's better than thrashing.

Thanks for the works done on this! A fixed jack has been submitted to multimedia:libs (seems waf -j<n>, where n > 1 is not a good idea for the obs builders). Thanks a lot Claudio! Dominique -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Cristian Rodríguez

10 Apr 10 Apr

15:21

El 10/04/12 05:58, Dominique Leuenberger a.k.a DimStar escribió:

...

sleuthkit: just a buffer overflow in ext2fs... it actually builds; just brp check raises it. Wondering why it's not triggered with gcc 4.6 though.

See section "General Optimizer Improvements" in http://gcc.gnu.org/gcc-4.7/changes.html in the paragraph "A string length optimization pass has been added...." -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dominique Leuenberger a.k.a DimStar

11 Apr 11 Apr

12:23

Quoting Cristian Rodríguez :

...

El 10/04/12 05:58, Dominique Leuenberger a.k.a DimStar escribió:

...
sleuthkit: just a buffer overflow in ext2fs... it actually builds; just brp check raises it. Wondering why it's not triggered with gcc 4.6 though.

See section "General Optimizer Improvements" in http://gcc.gnu.org/gcc-4.7/changes.html in the paragraph "A string length optimization pass has been added...."

Thanks Christian! Read through the code, identified the error, agreed with the compiler and fixed it :) Patch has been submitted to the devel project (no upstream contact just yet.. will likely do that in the evening). Best regards, Dominique -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Carl-Daniel Hailfinger

10 Apr 10 Apr

21:10

Hi Dominique, Am 10.04.2012 10:58 schrieb Dominique Leuenberger a.k.a DimStar:

...

as it has become custom to send out regular updates (be it only to keep people aware of the topic), here another round for the "GCC 4.7 as default in Factory" - Update:

There are currently 22 packages failing in my branch (home:dimstar:gcc47): llvm: honestly, I have no clue, especially as it fails in stage 2... so the integration to the new headers seems to cause griev.

http://llvm.org/bugs/show_bug.cgi?id=11916 http://llvm.org/viewvc/llvm-project?view=rev&revision=153999 That might be enough to fix it. Regards, Carl-Daniel -- http://www.hailfinger.org/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dominique Leuenberger a.k.a DimStar

11 Apr 11 Apr

09:00

Quoting Carl-Daniel Hailfinger :

...

Hi Dominique,

Am 10.04.2012 10:58 schrieb Dominique Leuenberger a.k.a DimStar:

...
as it has become custom to send out regular updates (be it only to keep people aware of the topic), here another round for the "GCC 4.7 as default in Factory" - Update:

There are currently 22 packages failing in my branch (home:dimstar:gcc47): llvm: honestly, I have no clue, especially as it fails in stage 2... so the integration to the new headers seems to cause griev.

http://llvm.org/bugs/show_bug.cgi?id=11916 http://llvm.org/viewvc/llvm-project?view=rev&revision=153999

Possibly enogugh, sadly we can't just take that patch set on top of our current llvm package. Will need to look a bit longer at that package if I want to fix it... preferably though this should be done by somebody actually understanding that package; compilers are always a little bit 'nastier to break'. Dominique -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Ismail Dönmez

09:31

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 04/11/2012 11:00 AM, Dominique Leuenberger a.k.a DimStar wrote:

...

Quoting Carl-Daniel Hailfinger :

...
Hi Dominique,

Am 10.04.2012 10:58 schrieb Dominique Leuenberger a.k.a DimStar:

...
as it has become custom to send out regular updates (be it only to keep people aware of the topic), here another round for the "GCC 4.7 as default in Factory" - Update:

There are currently 22 packages failing in my branch (home:dimstar:gcc47): llvm: honestly, I have no clue, especially as it fails in stage 2... so the integration to the new headers seems to cause griev.

http://llvm.org/bugs/show_bug.cgi?id=11916 http://llvm.org/viewvc/llvm-project?view=rev&revision=153999

Possibly enogugh, sadly we can't just take that patch set on top of our current llvm package.

Will need to look a bit longer at that package if I want to fix it... preferably though this should be done by somebody actually understanding that package; compilers are always a little bit 'nastier to break'.

llvm 3.1 will fix this and will be out in a month or so. Regards. - -- Ismail Dönmez - openSUSE Booster SUSE LINUX Products GmbH Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.18 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJPhU9oAAoJEJrs5hT7LFEcCuoH/3qXIfDy+sDpEQQqUSyuq9Op fnDRGKcittNzpDGjSoYgcBK1f0unQwCbQTiHkbaeJn9Sa4+D5YE+E+RKJqhIAGya CW2nQf/0rSwG051PK1oEMSELs0PcUX8OTPrbTQthov4n5Uk4vNIRHQ88eF/ro/Wj /fny9bILTVyUNVVQ1LHtCp2WDGxjoXTc4f1XR0A+taYhIeb02eby3G1NeVEPPEU6 WGbwPCK2+oPZ3pPdYO3/QeWAkmPgFxZk21vC/FAzKD2fXo4wIFycTMyh9XM7s/zm wVwPS2aapSnwHXZ3kO5EHzzb+seQAzMjC0v2wgIAmHuPRItC7AN/AvH6UKlxjqU= =N14u -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Alin M Elena

09:37

so let me summarise to see if I understood correctly... at the moment the only serious issue is libzypp&friends all the other packages that Dominique pointed out can be solved by updating them to the new versions.... Maybe we shall have a shout to ask people to push the latest versions from the devel repos into factory at least to know the latest and shiniest stable versions fail... on libzypp... if I remember correctly there were other issues too... missing headers and wrong order of functions in the files... for sure that cannot be blamed on the gcc team... Is there a bug report on the suggested regression in the compiler? regards, Alin On Wed 11 Apr 2012 11:31:20 Ismail Dönmez wrote:

...

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1

On 04/11/2012 11:00 AM, Dominique Leuenberger a.k.a DimStar wrote:

...
Quoting Carl-Daniel Hailfinger

:

...
Hi Dominique,

Am 10.04.2012 10:58 schrieb Dominique Leuenberger a.k.a DimStar:

...
as it has become custom to send out regular updates (be it only to keep people aware of the topic), here another round for the "GCC 4.7 as default in Factory" - Update:

There are currently 22 packages failing in my branch (home:dimstar:gcc47): llvm: honestly, I have no clue, especially as it fails in stage 2... so the integration to the new headers seems to cause griev.

http://llvm.org/bugs/show_bug.cgi?id=11916 http://llvm.org/viewvc/llvm-project?view=rev&revision=153999

Possibly enogugh, sadly we can't just take that patch set on top of our current llvm package.

Will need to look a bit longer at that package if I want to fix it... preferably though this should be done by somebody actually understanding that package; compilers are always a little bit 'nastier to break'.

llvm 3.1 will fix this and will be out in a month or so.

Regards.

- -- Ismail Dönmez - openSUSE Booster SUSE LINUX Products GmbH Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg)

-----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.18 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJPhU9oAAoJEJrs5hT7LFEcCuoH/3qXIfDy+sDpEQQqUSyuq9Op fnDRGKcittNzpDGjSoYgcBK1f0unQwCbQTiHkbaeJn9Sa4+D5YE+E+RKJqhIAGya CW2nQf/0rSwG051PK1oEMSELs0PcUX8OTPrbTQthov4n5Uk4vNIRHQ88eF/ro/Wj /fny9bILTVyUNVVQ1LHtCp2WDGxjoXTc4f1XR0A+taYhIeb02eby3G1NeVEPPEU6 WGbwPCK2+oPZ3pPdYO3/QeWAkmPgFxZk21vC/FAzKD2fXo4wIFycTMyh9XM7s/zm wVwPS2aapSnwHXZ3kO5EHzzb+seQAzMjC0v2wgIAmHuPRItC7AN/AvH6UKlxjqU= =N14u -----END PGP SIGNATURE----- -- Without Questions there are no Answers!

Alin Marin ELENA Advanced Molecular Simulation Research Laboratory School of Physics, University College Dublin ---- Ardionsamblú Móilíneach Saotharlann Taighde Scoil na Fisice, An Coláiste Ollscoile, Baile Átha Cliath ----------------------------------------------------------------------------------- http://alin.elenaworld.net ______________________________________________________________________ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dominique Leuenberger a.k.a DimStar

09:43

Quoting Alin M Elena :

...

so let me summarise to see if I understood correctly... at the moment the only serious issue is libzypp&friends all the other packages that Dominique pointed out can be solved by updating them to the new versions....

Maybe we shall have a shout to ask people to push the latest versions from the devel repos into factory at least to know the latest and shiniest stable versions fail...

on libzypp... if I remember correctly there were other issues too... missing headers and wrong order of functions in the files... for sure that cannot be blamed on the gcc team... Is there a bug report on the suggested regression in the compiler?

I think that pretty much sums it up :) For libzypp, there is this meta bug: https://bugzilla.novell.com/show_bug.cgi?id=755324 It links to the gcc regression mentioned. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Alin M Elena

10:06

...

For libzypp, there is this meta bug: https://bugzilla.novell.com/show_bug.cgi?id=755324

It links to the gcc regression mentioned.

I looked a little bit on the regression issue... it is marked as a serious issue with a fix targeting 4.7.1 that means looking on their history of releases some time in june... in this situation probably someone, if we want gcc 4.7 in 12.2... has to create a temporary patch as indicated in the workarounds from the bug. http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52841 Alin -- Without Questions there are no Answers! ______________________________________________________________________ Alin Marin ELENA Advanced Molecular Simulation Research Laboratory School of Physics, University College Dublin ---- Ardionsamblú Móilíneach Saotharlann Taighde Scoil na Fisice, An Coláiste Ollscoile, Baile Átha Cliath ----------------------------------------------------------------------------------- http://alin.elenaworld.net ______________________________________________________________________ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Cristian Rodríguez

22:08

El 10/04/12 05:58, Dominique Leuenberger a.k.a DimStar escribió:

...

Hi all,

as it has become custom to send out regular updates (be it only to keep people aware of the topic), here another round for the "GCC 4.7 as default in Factory" - Update:

IMHO our gcc 4.7 should default to -grecord-gcc-switches when building with debuginfo. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Richard Guenther

12 Apr 12 Apr

08:28

On Wed, 11 Apr 2012, Cristian Rodríguez wrote:

...

El 10/04/12 05:58, Dominique Leuenberger a.k.a DimStar escribió:

...
Hi all,

as it has become custom to send out regular updates (be it only to keep people aware of the topic), here another round for the "GCC 4.7 as default in Factory" - Update:

IMHO our gcc 4.7 should default to -grecord-gcc-switches when building with debuginfo.

It's as simple as adding that flag to RPM_OPT_FLAGS (or its debug variant). Btw, we already have our own internal switches recording which we use in rpmlint checks (can we drop that if we do the above?). Richard. -- Richard Guenther SUSE / SUSE Labs SUSE LINUX Products GmbH - Nuernberg - AG Nuernberg - HRB 16746 GF: Jeff Hawn, Jennifer Guild, Felix ImendÃ¶rffer

Cristian Rodríguez

11 May 11 May

18:56

El 12/04/12 05:28, Richard Guenther escribió:

...

On Wed, 11 Apr 2012, Cristian Rodríguez wrote:

...
El 10/04/12 05:58, Dominique Leuenberger a.k.a DimStar escribió:

...
Hi all,

as it has become custom to send out regular updates (be it only to keep people aware of the topic), here another round for the "GCC 4.7 as default in Factory" - Update:

IMHO our gcc 4.7 should default to -grecord-gcc-switches when building with debuginfo.

It's as simple as adding that flag to RPM_OPT_FLAGS (or its debug variant). Btw, we already have our own internal switches recording which we use in rpmlint checks (can we drop that if we do the above?).

Richard.

What about backporting this patch http://gcc.gnu.org/git/?p=gcc.git;a=commit;h=052166fd4a8051c7dc4c87d408be091... from GCC HEAD where this option is the default and dwarf_version is 4 ? and then drop the custom gcc hack ? Cheers ! -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Richard Guenther

14 May 14 May

08:09

On Fri, 11 May 2012, Cristian Rodríguez wrote:

...

El 12/04/12 05:28, Richard Guenther escribió:

...
On Wed, 11 Apr 2012, Cristian Rodríguez wrote:

...
El 10/04/12 05:58, Dominique Leuenberger a.k.a DimStar escribió:

...
Hi all,

as it has become custom to send out regular updates (be it only to keep people aware of the topic), here another round for the "GCC 4.7 as default in Factory" - Update:

IMHO our gcc 4.7 should default to -grecord-gcc-switches when building with debuginfo.

It's as simple as adding that flag to RPM_OPT_FLAGS (or its debug variant). Btw, we already have our own internal switches recording which we use in rpmlint checks (can we drop that if we do the above?).

Richard.

What about backporting this patch

http://gcc.gnu.org/git/?p=gcc.git;a=commit;h=052166fd4a8051c7dc4c87d408be091...

from GCC HEAD where this option is the default and dwarf_version is 4 ? and then drop the custom gcc hack ?

It doesn't backport easily (I tried it when seeing that patch). Thus, -ENOTIME. Richard. -- Richard Guenther SUSE / SUSE Labs SUSE LINUX Products GmbH - Nuernberg - AG Nuernberg - HRB 16746 GF: Jeff Hawn, Jennifer Guild, Felix ImendÃ¶rffer

Dominique Leuenberger a.k.a DimStar

12 Apr 12 Apr

11:22

Quoting "Dominique Leuenberger a.k.a DimStar" :

...

Hi all,

...

virtualbox: nasty... an update to the latest version should fix the gcc 4.7 issues (tried in my branch... raises a can of new worms)

Latest virtualbox has been submitted to Factory, which is GREAT! Now virtualbox is depending on this gcc 4.7 bug: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52862 Dominique -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Ismail Dönmez

13 Apr 13 Apr

09:27

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 04/12/2012 01:22 PM, Dominique Leuenberger a.k.a DimStar wrote:

...

Quoting "Dominique Leuenberger a.k.a DimStar" :

...
Hi all,

...
virtualbox: nasty... an update to the latest version should fix the gcc 4.7 issues (tried in my branch... raises a can of new worms)

Latest virtualbox has been submitted to Factory, which is GREAT! Now virtualbox is depending on this gcc 4.7 bug: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52862

richi fixed that now: http://gcc.gnu.org/viewcvs?view=revision&revision=186407 Regards. - -- Ismail Dönmez - openSUSE Booster SUSE LINUX Products GmbH Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.18 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJPh/FtAAoJEJrs5hT7LFEcwLoH/jE4Sr7FxGwa7aM6MP/IRJ/F 7KZ+TRM7a4KIn44Dxwjy3jzJCnU4ccIibSn165EzZTRJgDQnrGUXeb1fC2g0nlKR NZ+8JUvh1M/yMvbGYaSub+hZw0m+BzIpWiFLREz5m745ZubtLkfOAJvp5Ovj8TMZ uoQjCSCZaznJsFcBPgIxVeDrwlieqxA/Y5E5vK9dfqlgvP5+N/dgssts+qWGlzCO bIfBLhkP5wZ+m10VvFNuWTFUkvAgLOs/3OYjigZ3si2fTHsaErgs6usw7M2P49og GAgjLD396J3aZ/gGtHwwUeTtcjd2OCcFRGw2u0x7ZK7u1fgGo6+8Xv+T35YjRn8= =rpta -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dominique Leuenberger a.k.a DimStar

14:38

Quoting Ismail Dönmez :

...

-----BEGIN PGP SIGNED MESSAGE-----

...
Latest virtualbox has been submitted to Factory, which is GREAT! Now virtualbox is depending on this gcc 4.7 bug: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52862

richi fixed that now: http://gcc.gnu.org/viewcvs?view=revision&revision=186407

Great news! I just updated the linkrev of gcc47 in openSUSE:Factory:Staging:Gcc47 to point to the latest and greatest gcc 4.7 in devel:gcc, which Richard updated today already to contain this very fix! Thanks for that Richard! So in a few 'days' we should get an updated overview of packages failing with 'this version of gcc 4.7'... presumably at least virtualbox is getting fixed now :) so again: one package closer to the goal. Dominique -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Ismail Dönmez

18 Apr 18 Apr

09:06

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 04/13/2012 04:38 PM, Dominique Leuenberger a.k.a DimStar wrote:

...

Quoting Ismail Dönmez :

...
-----BEGIN PGP SIGNED MESSAGE-----

...
Latest virtualbox has been submitted to Factory, which is GREAT! Now virtualbox is depending on this gcc 4.7 bug: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52862

richi fixed that now: http://gcc.gnu.org/viewcvs?view=revision&revision=186407

Great news! I just updated the linkrev of gcc47 in openSUSE:Factory:Staging:Gcc47 to point to the latest and greatest gcc 4.7 in devel:gcc, which Richard updated today already to contain this very fix! Thanks for that Richard!

So in a few 'days' we should get an updated overview of packages failing with 'this version of gcc 4.7'... presumably at least virtualbox is getting fixed now :) so again: one package closer to the goal.

I submitted an llvm update which should fix gcc 4.7 problems. Regards. - -- Ismail Dönmez - openSUSE Booster SUSE LINUX Products GmbH Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.18 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQEcBAEBAgAGBQJPjoQlAAoJEJrs5hT7LFEcYREH/3OEW7k+ODIGQMJgT6zVSl/p UULoqm+RqW65I8aDzwZiQSOxscd4g2Bq6o3pBPWFTSRw6LSGY919H2kB51pVhvlo WoOrD6hgBUaQPiLDGQJn0g+AtwU1AXBFql+RQzYtKXasX/Pgutj+bu4UB3o89Kam 28wHGQj9JJY4tduI9WrhgwqC3fUxhxcAiN00Ua5lp6YSOblC9meA1HINbx3Qy+3u kGNcJfL/oc4s0OF99gHZwzkAl8QAHHfJwz9qjVvnLB21usB55JSBMi864z/FrXzs RWQypNIl6s4Bv64KRfce6+P2s76oJiyFEfZLkVdV7emzUNwuT/5qmi17zJ+b8gg= =ggVb -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Michal Šebeň

21 Apr 21 Apr

21:37

virtualbox failed again :-( this time the issue occurs when kbuild package is builded using gcc4.7 ( and afterwards kbuild is used for building virtualbox ) issue already reported to kbuild upstream by fedora guy: http://svn.netlabs.org/kbuild/ticket/112

...

So in a few 'days' we should get an updated overview of packages failing with 'this version of gcc 4.7'... presumably at least virtualbox is getting fixed now :) so again: one package closer to the goal.

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Andreas Jaeger

12 Apr 12 Apr

12:53

On Tuesday, April 10, 2012 10:58:27 Dominique Leuenberger a.k.a DimStar wrote:

...

Hi all,

as it has become custom to send out regular updates (be it only to keep people aware of the topic), here another round for the "GCC 4.7 as default in Factory" - Update:

Not much has happened in the last week. A critical component failing to build with Gcc 4.7 is currently libzypp, which, understandably, blocks the switch (well, I think it could always be decided to switch to 4.7, but keep zypp build with 4.6?)

Anyway, let's not paint it too dark: libzypp folks are helpful and looked into the matter. The first result was a regression in gcc which had been identified for libzypp.

There are currently 22 packages failing in my branch (home:dimstar:gcc47): libzypp, which causes at least PackageKit to fail as a result.

For libzypp and friends: It's not clear whether we have a bug in libzypp or a bug in GCC, might be that the old compiler accepts invalid code... We can always BuildRequire gcc46 and build using CXX="g++-46" etc, so this is no show stopper. Michael spend already a couple of days trying to figure out how to workaround the bug, Andreas -- Andreas Jaeger aj@{suse.com,opensuse.org} Twitter/Identica: jaegerandi SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Jeff Hawn,Jennifer Guild,Felix Imendörffer,HRB16746 (AG Nürnberg) GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dominique Leuenberger a.k.a DimStar

13:04

Quoting Andreas Jaeger :

...

On Tuesday, April 10, 2012 10:58:27 Dominique Leuenberger a.k.a DimStar

...
There are currently 22 packages failing in my branch (home:dimstar:gcc47): libzypp, which causes at least PackageKit to fail as a result.

For libzypp and friends: It's not clear whether we have a bug in libzypp or a bug in GCC, might be that the old compiler accepts invalid code...

We can always BuildRequire gcc46 and build using CXX="g++-46" etc, so this is no show stopper.

Michael spend already a couple of days trying to figure out how to workaround the bug,

Great if this is accepted as a 'solution' in short term... We should make sure though that 'only an accepted list of exceptions' is allowed to do that, as most packages wre rather simple to fix.... so don't just take this as the even easier way out. Dominique... -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Andreas Jaeger

13:10

On Thursday, April 12, 2012 15:04:07 Dominique Leuenberger a.k.a DimStar wrote:

...

Quoting Andreas Jaeger :

...
On Tuesday, April 10, 2012 10:58:27 Dominique Leuenberger a.k.a DimStar

...
There are currently 22 packages failing in my branch (home:dimstar:gcc47): libzypp, which causes at least PackageKit to fail as a result.

For libzypp and friends: It's not clear whether we have a bug in libzypp or a bug in GCC, might be that the old compiler accepts invalid code...

We can always BuildRequire gcc46 and build using CXX="g++-46" etc, so this is no show stopper.

Michael spend already a couple of days trying to figure out how to workaround the bug,

Great if this is accepted as a 'solution' in short term...

I hope ;) since it's something we did in the past.

...

We should make sure though that 'only an accepted list of exceptions' is allowed to do that, as most packages wre rather simple to fix.... so don't just take this as the even easier way out.

Indeed, Andreas -- Andreas Jaeger aj@{suse.com,opensuse.org} Twitter/Identica: jaegerandi SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany GF: Jeff Hawn,Jennifer Guild,Felix Imendörffer,HRB16746 (AG Nürnberg) GPG fingerprint = 93A3 365E CE47 B889 DF7F FED1 389A 563C C272 A126 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

4371

Age (days ago)

4405

Last active (days ago)

List overview

Download

27 comments

12 participants

participants (12)

Alin M Elena
Andreas Jaeger
Carl-Daniel Hailfinger
Claudio Freire
Cristian Rodríguez
Dominique Leuenberger
Dominique Leuenberger a.k.a DimStar
Ismail Dönmez
Michal Šebeň
Richard Guenther
Robert Kaiser
Wolfgang Rosenauer