Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
So to further update on this, I think with the fixes I've merged, OE-Core should be clean for all the QA tests except desktop file warnings and ldflags. ldflags was a problem as the code wasn't triggering at all in Yocto due to a bug in bitbake to do with the way ??= vs ?= was being handled. I'll have to run some further builds to see how bad the ldflags problem is and it may take a little longer to address those issues but its a step in the right direction at least and we can go ahead and make all the other areas apart from desktop and ldflags become errors whenever we want IMO. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
On Friday 01 July 2011 18:12:47 Richard Purdie wrote: Any volunteers for qt4-x11-free-4.7.3? I'll take a look at it. Cheers, Paul -- Paul Eggleton Intel Open Source Technology Centre ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
On Thu, 2011-06-30 at 17:49 +0200, Koen Kooi wrote: It's a white list, so: # 0 - non dev contains .so # 5 - .la contains installed=yes or reference to the workdir # 7 - the desktop file is not valid # 8 - .la contains reference to the workdir # 9 - LDFLAGS ignored Are warnings and # 1 - package contains a dangerous RPATH # 2 - package depends on debug package # 3 - non dbg contains .so # 4 - wrong architecture # 6 - .pc contains reference to /usr/include or workdir # 10 - Build paths in binaries # 11 - package depends on devel package Are fatal errors. The splits seems arbitrary to me, but it that's how it was last year before RP disabled all fatal errors. I guess the split does make some sense as it is, although I can't see any reason for #8 not to be in the fatal set. #5 also seems like it would belong there except that, as far as I can tell, that test doesn't actually exist in the code so it's a bit academic how the results are treated. #7 is, in the scheme of things, a relatively minor infringement (and usually an upstream bug anyway) so probably oughtn't to make a package unshippable. #9 is potentially a nuisance but in most cases doesn't cause any actual problems, so again I think it's fair for this to be a warning. Incidentally, it seems that the description for #6 is a bit wrong: it doesn't actually do any checking for /usr/include. And #3 should obviously be talking about .debug not .so. p. ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
On Thu, 2011-06-30 at 16:59 +0100, Paul Eggleton wrote: OK, comprehension fail on my part, sorry. Good to have them listed out explicitly though, thanks for that. The fatal errors are definitely not good things to be allowing through if we can avoid it. Without checking, do you have a handle on how many of these we would have to fix in OE-core to make it build without fatal errors? For a world build of OE-Core/Poky on qemux86 (on a 64 bit machine so it is cross) and ignoring desktop errors and things I've posted fixes for: db-5.1.19, non dev contains .so, db, /work/i586-poky-linux/db-5.1.19-r3/packages-split/db/usr/lib/libdb-5.so perl-5.12.3, package contains RPATH, perl-module-compress, /work/i586-poky-linux/perl-5.12.3-r1/packages-split/perl-module-compress/usr/lib/perl/5.12.3/auto/Compress/Raw/Zlib/Zlib.so libxml-parser-perl-2.40, package contains RPATH, libxml-parser-perl, /work/i586-poky-linux/libxml-parser-perl-2.40-r2/packages-split/libxml-parser-perl/usr/lib/perl/vendor_perl/5.12.3/auto/XML/Parser/Expat/Expat.so oprofile-0.9.6, non dev contains .so, oprofile, /work/i586-poky-linux/oprofile-0.9.6-r1/packages-split/oprofile/usr/lib/oprofile/libopagent.so lttng-viewer-0.12.38, non dev contains .so, lttng-viewer, /work/i586-poky-linux/lttng-viewer-0.12.38-r1/packages-split/lttng-viewer/usr/lib/liblttvtraceread.so gcc-4.6.0+svnr175150, non dev contains .so, gcc, /work/i586-poky-linux/gcc-4.6.0+svnr175150-r4/packages-split/gcc/usr/libexec/gcc/i586-poky-linux/4.6.1/liblto_plugin.so libgsmd-0.1+svnr5394, non dev contains .so, gsmd-plugin-machine-generic, /work/i586-poky-linux/libgsmd-0.1+svnr5394-r2/packages-split/gsmd-plugin-machine-generic/usr/lib/gsmd/libgsmd-machine_generic.so libgsmd-0.1+svnr5394, non dev contains .so, gsmd-plugin-machine-tihtc, /work/i586-poky-linux/libgsmd-0.1+svnr5394-r2/packages-split/gsmd-plugin-machine-tihtc/usr/lib/gsmd/libgsmd-machine_tihtc.so libgsmd-0.1+svnr5394, non dev contains .so, gsmd-plugin-machine-gta01, /work/i586-poky-linux/libgsmd-0.1+svnr5394-r2/packages-split/gsmd-plugin-machine-gta01/usr/lib/gsmd/libgsmd-machine_gta01.so libgsmd-0.1+svnr5394, non dev contains .so, gsmd-plugin-vendor-bcm, /work/i586-poky-linux/libgsmd-0.1+svnr5394-r2/packages-split/gsmd-plugin-vendor-bcm/usr/lib/gsmd/libgsmd-vendor_bcm.so libgsmd-0.1+svnr5394, non dev contains .so, gsmd-plugin-vendor-qc, /work/i586-poky-linux/libgsmd-0.1+svnr5394-r2/packages-split/gsmd-plugin-vendor-qc/usr/lib/gsmd/libgsmd-vendor_qc.so libgsmd-0.1+svnr5394, non dev contains .so, gsmd-plugin-vendor-ti, /work/i586-poky-linux/libgsmd-0.1+svnr5394-r2/packages-split/gsmd-plugin-vendor-ti/usr/lib/gsmd/libgsmd-vendor_ti.so libgsmd-0.1+svnr5394, non dev contains .so, gsmd-plugin-vendor-tihtc, /work/i586-poky-linux/libgsmd-0.1+svnr5394-r2/packages-split/gsmd-plugin-vendor-tihtc/usr/lib/gsmd/libgsmd-vendor_tihtc.so qemu-0.14.0, wrong architecture, qemu, /work/i586-poky-linux/qemu-0.14.0-r2/packages-split/qemu/usr/share/qemu/openbios-ppc qemu-0.14.0, wrong architecture, qemu, /work/i586-poky-linux/qemu-0.14.0-r2/packages-split/qemu/usr/share/qemu/openbios-sparc32 qemu-0.14.0, wrong architecture, qemu, /work/i586-poky-linux/qemu-0.14.0-r2/packages-split/qemu/usr/share/qemu/openbios-sparc64 qt4-x11-free-4.7.3, wrong architecture, qt4-tools-dbg, /work/i586-poky-linux/qt4-x11-free-4.7.3-r23.1/packages-split/qt4-tools-dbg/usr/bin/.debug/uic qt4-x11-free-4.7.3, wrong architecture, qt4-tools-dbg, /work/i586-poky-linux/qt4-x11-free-4.7.3-r23.1/packages-split/qt4-tools-dbg/usr/bin/.debug/rcc qt4-x11-free-4.7.3, wrong architecture, qt4-tools-dbg, /work/i586-poky-linux/qt4-x11-free-4.7.3-r23.1/packages-split/qt4-tools-dbg/usr/bin/.debug/moc qt4-x11-free-4.7.3, wrong architecture, qt4-tools, /work/i586-poky-linux/qt4-x11-free-4.7.3-r23.1/packages-split/qt4-tools/usr/bin/uic qt4-x11-free-4.7.3, wrong architecture, qt4-tools, /work/i586-poky-linux/qt4-x11-free-4.7.3-r23.1/packages-split/qt4-tools/usr/bin/rcc qt4-x11-free-4.7.3, wrong architecture, qt4-tools, /work/i586-poky-linux/qt4-x11-free-4.7.3-r23.1/packages-split/qt4-tools/usr/bin/moc qt4-x11-free-4.7.3, wrong architecture, qt4-linguist, /work/i586-poky-linux/qt4-x11-free-4.7.3-r23.1/packages-split/qt4-linguist/usr/bin/lrelease The qemu ones are false positives, the qt ones are nasty as are the libxml-parser-perl and perl ones. The .so errors are hard to get excited about but should get looked at and resolved. In the meantime I'll run a world build for mips or ppc and see if that shows anything different up. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
On 7/1/11 12:16 PM, Khem Raj wrote: On 07/01/2011 10:12 AM, Richard Purdie wrote: gcc-4.6.0+svnr175150, non dev contains .so, gcc, /work/i586-poky-linux/gcc-4.6.0+svnr175150-r4/packages-split/gcc/usr/libexec/gcc/i586-poky-linux/4.6.1/liblto_plugin.so This should be packages with gcc itself. May be adding to FILES will get rid of this warning The warning above indicates that it -is- being packaged in the 'gcc' package. The check verified that no files of the name .so exit and they are not symlinks: if not name.endswith(-dev) and not name.endswith(-dbg) and path.endswith(.so) and os.path.islink(path): In this case, I believe that the file is likely supposed to be there. Is there a way to selectively disable insane checks on a per-recipe basis when we know they are wrong? --Mark ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
Op 1 jul 2011, om 19:25 heeft Mark Hatle het volgende geschreven: On 7/1/11 12:16 PM, Khem Raj wrote: On 07/01/2011 10:12 AM, Richard Purdie wrote: gcc-4.6.0+svnr175150, non dev contains .so, gcc, /work/i586-poky-linux/gcc-4.6.0+svnr175150-r4/packages-split/gcc/usr/libexec/gcc/i586-poky-linux/4.6.1/liblto_plugin.so This should be packages with gcc itself. May be adding to FILES will get rid of this warning The warning above indicates that it -is- being packaged in the 'gcc' package. The check verified that no files of the name .so exit and they are not symlinks: if not name.endswith(-dev) and not name.endswith(-dbg) and path.endswith(.so) and os.path.islink(path): In this case, I believe that the file is likely supposed to be there. Is there a way to selectively disable insane checks on a per-recipe basis when we know they are wrong? We only have the big INSANE_SKIP_gcc = True hammer ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
On Fri, 2011-07-01 at 19:33 +0200, Koen Kooi wrote: Op 1 jul 2011, om 19:25 heeft Mark Hatle het volgende geschreven: On 7/1/11 12:16 PM, Khem Raj wrote: On 07/01/2011 10:12 AM, Richard Purdie wrote: gcc-4.6.0+svnr175150, non dev contains .so, gcc, /work/i586-poky-linux/gcc-4.6.0+svnr175150-r4/packages-split/gcc/usr/libexec/gcc/i586-poky-linux/4.6.1/liblto_plugin.so This should be packages with gcc itself. May be adding to FILES will get rid of this warning The warning above indicates that it -is- being packaged in the 'gcc' package. The check verified that no files of the name .so exit and they are not symlinks: if not name.endswith(-dev) and not name.endswith(-dbg) and path.endswith(.so) and os.path.islink(path): In this case, I believe that the file is likely supposed to be there. Is there a way to selectively disable insane checks on a per-recipe basis when we know they are wrong? We only have the big INSANE_SKIP_gcc = True hammer Although as I hint at in my insane.bbclass patch, I just made it much easier to implement skipping specific checks. Its not there yet but shouldn't be too difficult from where the code it at now... Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
[OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
These were turned off by: commit fae8d5e985e9b05ce90f1eca434ad4dbf2259725 Author: Richard Purdie rpur...@linux.intel.com Date: Thu Jul 8 23:51:06 2010 +0100 insane.bbclass: Relax fatal errors for now until we get have time to work through the backlog The current metadata triggers so many of these that they need to be made fatal so people will actually fix them. Signed-off-by: Koen Kooi k...@dominion.thruhere.net --- meta/classes/insane.bbclass |1 - 1 files changed, 0 insertions(+), 1 deletions(-) diff --git a/meta/classes/insane.bbclass b/meta/classes/insane.bbclass index f7e5524..e06d2b2 100644 --- a/meta/classes/insane.bbclass +++ b/meta/classes/insane.bbclass @@ -115,7 +115,6 @@ def package_qa_make_fatal_error(error_class, name, path,d): TODO: Load a whitelist of known errors -return False return not error_class in [0, 5, 7, 8, 9] def package_qa_write_error(error_class, name, path, d): -- 1.6.6.1 ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
On Thursday 30 June 2011 16:11:29 Koen Kooi wrote: These were turned off by: commit fae8d5e985e9b05ce90f1eca434ad4dbf2259725 Author: Richard Purdie rpur...@linux.intel.com Date: Thu Jul 8 23:51:06 2010 +0100 insane.bbclass: Relax fatal errors for now until we get have time to work through the backlog The current metadata triggers so many of these that they need to be made fatal so people will actually fix them. ... return not error_class in [0, 5, 7, 8, 9] So in principle I can agree that making these fatal again will make people sort out the problems that they are flagging up. However, why is class 7 - .desktop files being invalid - a fatal error, considering there are many sub- classes of invalidity being tested for with varying levels of significance? (This has probably been discussed on the OE list before but IMHO it merits revisiting if so.) Cheers, Paul -- Paul Eggleton Intel Open Source Technology Centre ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
Op 30 jun 2011, om 17:33 heeft Paul Eggleton het volgende geschreven: On Thursday 30 June 2011 16:11:29 Koen Kooi wrote: These were turned off by: commit fae8d5e985e9b05ce90f1eca434ad4dbf2259725 Author: Richard Purdie rpur...@linux.intel.com Date: Thu Jul 8 23:51:06 2010 +0100 insane.bbclass: Relax fatal errors for now until we get have time to work through the backlog The current metadata triggers so many of these that they need to be made fatal so people will actually fix them. ... return not error_class in [0, 5, 7, 8, 9] So in principle I can agree that making these fatal again will make people sort out the problems that they are flagging up. However, why is class 7 - .desktop files being invalid - a fatal error, considering there are many sub- classes of invalidity being tested for with varying levels of significance? (This has probably been discussed on the OE list before but IMHO it merits revisiting if so.) It's a white list, so: # 0 - non dev contains .so # 5 - .la contains installed=yes or reference to the workdir # 7 - the desktop file is not valid # 8 - .la contains reference to the workdir # 9 - LDFLAGS ignored Are warnings and # 1 - package contains a dangerous RPATH # 2 - package depends on debug package # 3 - non dbg contains .so # 4 - wrong architecture # 6 - .pc contains reference to /usr/include or workdir # 10 - Build paths in binaries # 11 - package depends on devel package Are fatal errors. The splits seems arbitrary to me, but it that's how it was last year before RP disabled all fatal errors. ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
On 06/30/2011 08:11 AM, Koen Kooi wrote: These were turned off by: commit fae8d5e985e9b05ce90f1eca434ad4dbf2259725 Author: Richard Purdie rpur...@linux.intel.com Date: Thu Jul 8 23:51:06 2010 +0100 insane.bbclass: Relax fatal errors for now until we get have time to work through the backlog The current metadata triggers so many of these that they need to be made fatal so people will actually fix them. Signed-off-by: Koen Kooi k...@dominion.thruhere.net Given that a number of them can probably just be pulled from oe.dev as fixes too: Acked-by: Tom Rini tom_r...@mentor.com --- meta/classes/insane.bbclass |1 - 1 files changed, 0 insertions(+), 1 deletions(-) diff --git a/meta/classes/insane.bbclass b/meta/classes/insane.bbclass index f7e5524..e06d2b2 100644 --- a/meta/classes/insane.bbclass +++ b/meta/classes/insane.bbclass @@ -115,7 +115,6 @@ def package_qa_make_fatal_error(error_class, name, path,d): TODO: Load a whitelist of known errors -return False return not error_class in [0, 5, 7, 8, 9] def package_qa_write_error(error_class, name, path, d): -- Tom Rini Mentor Graphics Corporation ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
On Thursday 30 June 2011 16:49:57 Koen Kooi wrote: It's a white list, so: # 0 - non dev contains .so # 5 - .la contains installed=yes or reference to the workdir # 7 - the desktop file is not valid # 8 - .la contains reference to the workdir # 9 - LDFLAGS ignored Are warnings and # 1 - package contains a dangerous RPATH # 2 - package depends on debug package # 3 - non dbg contains .so # 4 - wrong architecture # 6 - .pc contains reference to /usr/include or workdir # 10 - Build paths in binaries # 11 - package depends on devel package Are fatal errors. The splits seems arbitrary to me, but it that's how it was last year before RP disabled all fatal errors. OK, comprehension fail on my part, sorry. Good to have them listed out explicitly though, thanks for that. The fatal errors are definitely not good things to be allowing through if we can avoid it. Without checking, do you have a handle on how many of these we would have to fix in OE-core to make it build without fatal errors? Cheers, Paul -- Paul Eggleton Intel Open Source Technology Centre ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
On 6/30/11 10:58 AM, Tom Rini wrote: On 06/30/2011 08:11 AM, Koen Kooi wrote: These were turned off by: commit fae8d5e985e9b05ce90f1eca434ad4dbf2259725 Author: Richard Purdie rpur...@linux.intel.com Date: Thu Jul 8 23:51:06 2010 +0100 insane.bbclass: Relax fatal errors for now until we get have time to work through the backlog The current metadata triggers so many of these that they need to be made fatal so people will actually fix them. Signed-off-by: Koen Kooi k...@dominion.thruhere.net Given that a number of them can probably just be pulled from oe.dev as fixes too: Acked-by: Tom Rini tom_r...@mentor.com I'm really worried about the current warnings/QA issues that keep coming up.. so I agree we really need to do this.. (note, we also really need to tell people to expect some breakage as we go and fix the mess!) Acked-by: Mark Hatle mark.ha...@windriver.com --- meta/classes/insane.bbclass |1 - 1 files changed, 0 insertions(+), 1 deletions(-) diff --git a/meta/classes/insane.bbclass b/meta/classes/insane.bbclass index f7e5524..e06d2b2 100644 --- a/meta/classes/insane.bbclass +++ b/meta/classes/insane.bbclass @@ -115,7 +115,6 @@ def package_qa_make_fatal_error(error_class, name, path,d): TODO: Load a whitelist of known errors -return False return not error_class in [0, 5, 7, 8, 9] def package_qa_write_error(error_class, name, path, d): ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
Op 30 jun 2011, om 18:08 heeft Mark Hatle het volgende geschreven: On 6/30/11 10:58 AM, Tom Rini wrote: On 06/30/2011 08:11 AM, Koen Kooi wrote: These were turned off by: commit fae8d5e985e9b05ce90f1eca434ad4dbf2259725 Author: Richard Purdie rpur...@linux.intel.com Date: Thu Jul 8 23:51:06 2010 +0100 insane.bbclass: Relax fatal errors for now until we get have time to work through the backlog The current metadata triggers so many of these that they need to be made fatal so people will actually fix them. Signed-off-by: Koen Kooi k...@dominion.thruhere.net Given that a number of them can probably just be pulled from oe.dev as fixes too: Acked-by: Tom Rini tom_r...@mentor.com I'm really worried about the current warnings/QA issues that keep coming up.. so I agree we really need to do this.. (note, we also really need to tell people to expect some breakage as we go and fix the mess!) During todays TSC meeting the proposed date to switch was the 11th and encourage people to start fixing things *NOW*. What are peoples thought on that? regards, Koen ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH] insane bbclass: turn fatal errors back into fatal errors
On 06/30/2011 01:35 PM, Koen Kooi wrote: Op 30 jun 2011, om 18:08 heeft Mark Hatle het volgende geschreven: On 6/30/11 10:58 AM, Tom Rini wrote: On 06/30/2011 08:11 AM, Koen Kooi wrote: These were turned off by: commit fae8d5e985e9b05ce90f1eca434ad4dbf2259725 Author: Richard Purdierpur...@linux.intel.com Date: Thu Jul 8 23:51:06 2010 +0100 insane.bbclass: Relax fatal errors for now until we get have time to work through the backlog The current metadata triggers so many of these that they need to be made fatal so people will actually fix them. Signed-off-by: Koen Kooik...@dominion.thruhere.net Given that a number of them can probably just be pulled from oe.dev as fixes too: Acked-by: Tom Rinitom_r...@mentor.com I'm really worried about the current warnings/QA issues that keep coming up.. so I agree we really need to do this.. (note, we also really need to tell people to expect some breakage as we go and fix the mess!) During todays TSC meeting the proposed date to switch was the 11th and encourage people to start fixing things *NOW*. What are peoples thought on that? Sounds like an excellent compromise! Scott -- Scott Garman Embedded Linux Engineer - Yocto Project Intel Open Source Technology Center ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core