Re: Dropping valgrind from armel?
On lun, ott 28, 2013 at 09:36:22 -0400, Jerry Stuckle wrote: On 10/28/2013 6:02 PM, Alessandro Ghedini wrote: On Thu, Oct 24, 2013 at 08:24:28PM +0200, Alessandro Ghedini wrote: Hi all, long story short, a couple years ago armel builds for valgrind were enabled (despite the fact that valgrind only supports ARMv7) by building the package in cross-compile mode and forcing the -march=armv7-a option on buildds that didn't support ARMv7 natively. This was done so that ARMv7 armel systems could use valgrind (see #592614). This has sort of worked for a while, until a couple months ago when valgrind started FTBFS on armel (#720409). This was a simple routine rebuild for the openmpi transition, so I'm inclined to think that I did not broke anything myself. My next upload 1:3.8.1-5 (a month later) still failed to build, making me think that this is not a transient failure. Hence the idea: what about dropping valgrind from armel? Or alternatively, is there anyone who cares about valgrind on armel and wants to debug and try to fix this (possibly without making the original kinda ugly hack any worse)? So, no one? In the next few days I'm going to upload a new version disabling armel builds and ask the release team to drop it as well. If it turns out that many people actually used valgrind on armel, I guess I can re-enable it later (once it works again). Cheers Are you disabling all armel builds? Or just valgrind on armel? Not sure if I understood the question correctly, but I was referring only to valgrind's armel build (I don't quite have the power to eliminate a whole Debian port I'm afraid ;) Also, please CC me since I'm not subscibed to the list (I forgot to mention that before). Cheers -- perl -E '$_=q;$/= @{[@_]};and s;\S+;inidehG ordnasselA;eg;say~~reverse' signature.asc Description: Digital signature
Re: Dropping valgrind from armel?
On Thu, Oct 24, 2013 at 08:24:28PM +0200, Alessandro Ghedini wrote: Hi all, long story short, a couple years ago armel builds for valgrind were enabled (despite the fact that valgrind only supports ARMv7) by building the package in cross-compile mode and forcing the -march=armv7-a option on buildds that didn't support ARMv7 natively. This was done so that ARMv7 armel systems could use valgrind (see #592614). This has sort of worked for a while, until a couple months ago when valgrind started FTBFS on armel (#720409). This was a simple routine rebuild for the openmpi transition, so I'm inclined to think that I did not broke anything myself. My next upload 1:3.8.1-5 (a month later) still failed to build, making me think that this is not a transient failure. Hence the idea: what about dropping valgrind from armel? Or alternatively, is there anyone who cares about valgrind on armel and wants to debug and try to fix this (possibly without making the original kinda ugly hack any worse)? So, no one? In the next few days I'm going to upload a new version disabling armel builds and ask the release team to drop it as well. If it turns out that many people actually used valgrind on armel, I guess I can re-enable it later (once it works again). Cheers -- perl -E '$_=q;$/= @{[@_]};and s;\S+;inidehG ordnasselA;eg;say~~reverse' signature.asc Description: Digital signature
Re: Dropping valgrind from armel?
On 10/28/2013 6:02 PM, Alessandro Ghedini wrote: On Thu, Oct 24, 2013 at 08:24:28PM +0200, Alessandro Ghedini wrote: Hi all, long story short, a couple years ago armel builds for valgrind were enabled (despite the fact that valgrind only supports ARMv7) by building the package in cross-compile mode and forcing the -march=armv7-a option on buildds that didn't support ARMv7 natively. This was done so that ARMv7 armel systems could use valgrind (see #592614). This has sort of worked for a while, until a couple months ago when valgrind started FTBFS on armel (#720409). This was a simple routine rebuild for the openmpi transition, so I'm inclined to think that I did not broke anything myself. My next upload 1:3.8.1-5 (a month later) still failed to build, making me think that this is not a transient failure. Hence the idea: what about dropping valgrind from armel? Or alternatively, is there anyone who cares about valgrind on armel and wants to debug and try to fix this (possibly without making the original kinda ugly hack any worse)? So, no one? In the next few days I'm going to upload a new version disabling armel builds and ask the release team to drop it as well. If it turns out that many people actually used valgrind on armel, I guess I can re-enable it later (once it works again). Cheers Are you disabling all armel builds? Or just valgrind on armel? Jerry -- To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/526f1116.4020...@attglobal.net
Re: Dropping valgrind from armel?
On gio, ott 24, 2013 at 11:11:52 +0200, Hector Oron wrote: Hello, 2013/10/24 Alessandro Ghedini gh...@debian.org: Hence the idea: what about dropping valgrind from armel? Or alternatively, is there anyone who cares about valgrind on armel and wants to debug and try to fix this (possibly without making the original kinda ugly hack any worse)? If interested, also check https://bugs.kde.org/show_bug.cgi?id=248998 I'm aware of those patches (see #701691), they don't work though (even if you make them apply and build). Cheers -- perl -E '$_=q;$/= @{[@_]};and s;\S+;inidehG ordnasselA;eg;say~~reverse' signature.asc Description: Digital signature
Re: Dropping valgrind from armel?
On Thu, Oct 24, 2013 at 08:24:28PM +0200, Alessandro Ghedini wrote: Hi all, long story short, a couple years ago armel builds for valgrind were enabled (despite the fact that valgrind only supports ARMv7) by building the package in cross-compile mode and forcing the -march=armv7-a option on buildds that didn't support ARMv7 natively. This was done so that ARMv7 armel systems could use valgrind (see #592614). This has sort of worked for a while, until a couple months ago when valgrind started FTBFS on armel (#720409). This was a simple routine rebuild for the openmpi transition, so I'm inclined to think that I did not broke anything myself. My next upload 1:3.8.1-5 (a month later) still failed to build, making me think that this is not a transient failure. To add a little more context: * 1:3.8.1-4 built on ancina on 2013-06-01 * 1:3.8.1-4+b1 failed twice on ancina on 2013-08-21/22 * 1:3.8.1-5 failed on antheil on 2013-09-24 The fact that it built fine and then 2 months later failed on ancina looks weird to me, and I can't think of any change that could have triggered this. Also, please CC me since I'm not subscibed to the list. Cheers -- perl -E '$_=q;$/= @{[@_]};and s;\S+;inidehG ordnasselA;eg;say~~reverse' signature.asc Description: Digital signature
Dropping valgrind from armel?
Hi all, long story short, a couple years ago armel builds for valgrind were enabled (despite the fact that valgrind only supports ARMv7) by building the package in cross-compile mode and forcing the -march=armv7-a option on buildds that didn't support ARMv7 natively. This was done so that ARMv7 armel systems could use valgrind (see #592614). This has sort of worked for a while, until a couple months ago when valgrind started FTBFS on armel (#720409). This was a simple routine rebuild for the openmpi transition, so I'm inclined to think that I did not broke anything myself. My next upload 1:3.8.1-5 (a month later) still failed to build, making me think that this is not a transient failure. Hence the idea: what about dropping valgrind from armel? Or alternatively, is there anyone who cares about valgrind on armel and wants to debug and try to fix this (possibly without making the original kinda ugly hack any worse)? Cheers -- perl -E '$_=q;$/= @{[@_]};and s;\S+;inidehG ordnasselA;eg;say~~reverse' signature.asc Description: Digital signature
Re: Dropping valgrind from armel?
Hello, 2013/10/24 Alessandro Ghedini gh...@debian.org: Hence the idea: what about dropping valgrind from armel? Or alternatively, is there anyone who cares about valgrind on armel and wants to debug and try to fix this (possibly without making the original kinda ugly hack any worse)? If interested, also check https://bugs.kde.org/show_bug.cgi?id=248998 Cheers -- perl -E '$_=q;$/= @{[@_]};and s;\S+;inidehG ordnasselA;eg;say~~reverse' -- Héctor Orón -.. . -... .. .- -. -.. . ...- . .-.. --- .--. . .-. -- To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/caodfwehaqeuxtvxa+zb8emdgnh_dzfkfxqfhkzenawccchz...@mail.gmail.com