The big Ada merge (mainline only) wrongly reverted some patches that had been made to the FSF GCC tree. These need to be unreverted, and a proper check made for whether any other patches were reverted that shouldn't have been. A list, which may be incomplete, is at <URL:http://gcc.gnu.org/ml/gcc-patches/2002-03/msg00402.html>. Release: 3.2 20020504 (experimental) Environment: System: Linux digraph 2.2.20 #2 Sat Nov 10 16:44:22 UTC 2001 i686 unknown Architecture: i686 host: i686-pc-linux-gnu build: i686-pc-linux-gnu target: i686-pc-linux-gnu configured with: ../gcc-cvs/configure --prefix=/opt/gcc/mainline --disable-shared --enable-threads=posix --with-system-zlib
Fix: Unrevert the patches, taking care about whether they need updating in consequence of other changes to the Ada front end. Check carefully for any other reverted patches, and unrevert them.
Responsible-Changed-From-To: unassigned->bosch Responsible-Changed-Why: It's the responsibility of the person who reverted the patches by mistake in the merge to check carefully for what was reverted and unrevert it.
State-Changed-From-To: open->analyzed State-Changed-Why: Patch reversion still present in the tree. Incorrect reversion forms a regression.
From: Zack Weinberg <zack@codesourcery.com> To: Nathanael Nerode <neroden@twcny.rr.com> Cc: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, bosch@gcc.gnu.org, jsm28@cam.ac.uk, gcc-bugs@gcc.gnu.org Subject: Re: ada/6558: [mainline regression] Wrongly reverted patches Date: Mon, 28 Oct 2002 23:19:19 -0800 On Tue, Oct 29, 2002 at 02:05:06AM -0500, Nathanael Nerode wrote: > What sort of progress can we make on this? > > Clearly what needs to be checked are the patches to the FSF tree from > 2001 and earlier. But how much earlier ?!? Is there an earlier point > at which we know that the ACT (?) tree had been merged *from* the FSF > tree, so that we don't need to look earlier than that? Otherwise this > is potentially an endless problem. My understanding is that the only patches at issue were those lost when this change occurred: 2002-03-07 Geert Bosch <bosch@gnat.com> * 41intnam.ads, 42intnam.ads, 4aintnam.ads, 4cintnam.ads, 4dintnam.ads, 4gintnam.ads, 4hintnam.ads, 4lintnam.ads, 4mintnam.ads, 4pintnam.ads, 4rintnam.ads, 4sintnam.ads, 4uintnam.ads, 4vcalend.adb, 4zintnam.ads, 52system.ads, 5amastop.adb, 5asystem.ads, 5ataprop.adb, 5atpopsp.adb, 5avxwork.ads, 5bosinte.adb, 5bsystem.ads, 5esystem.ads, [...] utils.c, utils2.c, validsw.adb, xnmake.adb, xr_tabls.adb, xr_tabls.ads, xref_lib.adb, xref_lib.ads : Merge in ACT changes. * 1ssecsta.adb, 1ssecsta.ads, a-chlat9.ads, a-cwila9.ads, g-enblsp.adb, g-md5.adb, g-md5.ads, gnatname.adb, gnatname.ads, mkdir.c, osint-b.adb, osint-b.ads, osint-c.adb, osint-c.ads, osint-l.adb, osint-l.ads, osint-m.adb, osint-m.ads : New files * 3lsoccon.ads, 5qparame.ads, 5qvxwork.ads, 5smastop.adb, 5zparame.ads, gnatmain.adb, gnatmain.ads, gnatpsys.adb : Removed * mdllfile.adb, mdllfile.ads, mdlltool.adb, mdlltool.ads : Renamed to mdll-fil.ad[bs] and mdll-util.ad[bs] * mdll-fil.adb, mdll-fil.ads, mdll-utl.adb, mdll-utl.ads : Renamed from mdllfile.ad[bs] and mdlltool.ad[bs] I would like to point out that it is just shy of eight months since this happened, and there has been no indication that Mr. Bosch has lifted a finger to address the problem that he created (beyond claiming that it would be fixed soon). I am tempted to invoke the patch reversion procedure and demand that this entire merge be backed out. I will certainly do so if the issue is not resolved by the time 3.3 branches. zw
From: Nathanael Nerode <neroden@twcny.rr.com> To: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, bosch@gcc.gnu.org, jsm28@cam.ac.uk, gcc-bugs@gcc.gnu.org Cc: Subject: Re: ada/6558: [mainline regression] Wrongly reverted patches Date: Tue, 29 Oct 2002 02:05:06 -0500 What sort of progress can we make on this? Clearly what needs to be checked are the patches to the FSF tree from 2001 and earlier. But how much earlier ?!? Is there an earlier point at which we know that the ACT (?) tree had been merged *from* the FSF tree, so that we don't need to look earlier than that? Otherwise this is potentially an endless problem. http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6558
From: "Joseph S. Myers" <jsm28@cam.ac.uk> To: Zack Weinberg <zack@codesourcery.com> Cc: Nathanael Nerode <neroden@twcny.rr.com>, <gcc-gnats@gcc.gnu.org>, <bosch@gcc.gnu.org>, <gcc-bugs@gcc.gnu.org> Subject: Re: ada/6558: [mainline regression] Wrongly reverted patches Date: Tue, 29 Oct 2002 10:27:44 +0000 (GMT) On Mon, 28 Oct 2002, Zack Weinberg wrote: > My understanding is that the only patches at issue were those lost > when this change occurred: > > 2002-03-07 Geert Bosch <bosch@gnat.com> Yes. It's all patches from non-ACT people to the FSF tree before then that need to be checked to make sure they are still in the tree or have been superseded by subsequent changes. I think someone found an example (which may or may not now have been unreverted individually) that wasn't among the examples I gave of patches wrongly reverted by this change, so it is necessary to do that check rather than relying on my list. > I would like to point out that it is just shy of eight months since > this happened, and there has been no indication that Mr. Bosch has > lifted a finger to address the problem that he created (beyond > claiming that it would be fixed soon). -- Joseph S. Myers jsm28@cam.ac.uk
From: Nathanael Nerode <neroden@twcny.rr.com> To: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, bosch@gcc.gnu.org, jsm28@cam.ac.uk, gcc-bugs@gcc.gnu.org Cc: Subject: Re: ada/6558: [mainline regression] Wrongly reverted patches Date: Tue, 29 Oct 2002 15:37:19 -0500 Hmm. I think I see a way of cutting down on the number of patches to check. Looking at cvsweb, it seems that most of the commits to the Ada ChangeLog are by Mr. Bosch himself. :-) If I assume that he didn't accidentally revert any of his own patches, then I have a more reasonable number to go through. http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6558
From: Nathanael Nerode <neroden@twcny.rr.com> To: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, bosch@gcc.gnu.org, jsm28@cam.ac.uk, gcc-bugs@gcc.gnu.org Cc: Subject: Re: ada/6558: [mainline regression] Wrongly reverted patches Date: Tue, 29 Oct 2002 15:52:51 -0500 All patches appear OK up through 1.102. This is a good start. http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6558
From: Geert Bosch <bosch@gnat.com> To: Nathanael Nerode <neroden@twcny.rr.com> Cc: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, bosch@gcc.gnu.org, jsm28@cam.ac.uk, gcc-bugs@gcc.gnu.org Subject: Re: ada/6558: [mainline regression] Wrongly reverted patches Date: Tue, 29 Oct 2002 15:59:58 -0500 On Tuesday, Oct 29, 2002, at 15:37 America/New_York, Nathanael Nerode wrote: > Hmm. I think I see a way of cutting down on the number of patches to > check. Looking at cvsweb, it seems that most of the commits to the > Ada ChangeLog are by Mr. Bosch himself. :-) If I assume that he > didn't accidentally revert any of his own patches, then I have a more > reasonable number to go through. The same is true for all patches from @gnat.com addresses. The main error I made was to try to merge ACT->FSF without first doing FSF->ACT. I tried to go through the entire diff and make sure no changes were undone, but I missed some cases. I will not make this mistake again. Nathanael, thank you for your help. -Geert
From: Nathanael Nerode <neroden@twcny.rr.com> To: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, bosch@gcc.gnu.org, jsm28@cam.ac.uk, gcc-bugs@gcc.gnu.org Cc: Subject: Re: ada/6558: [mainline regression] Wrongly reverted patches Date: Tue, 29 Oct 2002 16:04:08 -0500 Found the first bogus reversion, from ChangeLog 1.105 by neil. ada: * misc.c (gnat_decode_option, gnat_init_options): Make definitions static too. (gnat_init): Don't return NULL. http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6558 So the ChangeLog only has to be checked from here to Geert's fatal patch at 1.201. I can and will do this; I'd like someone else to help me by unreverting the patches as I find them. I think we can get this done pretty efficiently, actually. --Nathanael
From: Geert Bosch <bosch@gnat.com> To: Nathanael Nerode <neroden@twcny.rr.com> Cc: gcc-gnats@gcc.gnu.org, gcc-prs@gcc.gnu.org, bosch@gcc.gnu.org, jsm28@cam.ac.uk, gcc-bugs@gcc.gnu.org Subject: Re: ada/6558: [mainline regression] Wrongly reverted patches Date: Tue, 29 Oct 2002 16:08:31 -0500 On Tuesday, Oct 29, 2002, at 16:04 America/New_York, Nathanael Nerode wrote: > So the ChangeLog only has to be checked from here to Geert's fatal > patch at 1.201. I can and will do this; I'd like someone else to help > me by unreverting the patches as I find them. I think we can get this > done pretty efficiently, actually. I'll unrevert any reverted patches you'll find. -Geert
From: bosch@gcc.gnu.org To: gcc-gnats@gcc.gnu.org Cc: Subject: ada/6558 Date: 29 Oct 2002 22:01:19 -0000 CVSROOT: /cvs/gcc Module name: gcc Changes by: bosch@gcc.gnu.org 2002-10-29 14:01:19 Modified files: gcc/ada : misc.c ChangeLog Log message: PR ada/6558 * misc.c: Unrevert misc.c (1.13) Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/misc.c.diff?cvsroot=gcc&r1=1.42&r2=1.43 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/ChangeLog.diff?cvsroot=gcc&r1=1.262&r2=1.263
From: bosch@gcc.gnu.org To: gcc-gnats@gcc.gnu.org Cc: Subject: ada/6558 Date: 29 Oct 2002 22:18:08 -0000 CVSROOT: /cvs/gcc Module name: gcc Changes by: bosch@gcc.gnu.org 2002-10-29 14:18:08 Modified files: gcc/ada : Make-lang.in ChangeLog Log message: PR ada/6558 * Make-lang.in (gnatbind): Depend on CONFIG_H Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/Make-lang.in.diff?cvsroot=gcc&r1=1.16&r2=1.17 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/ChangeLog.diff?cvsroot=gcc&r1=1.263&r2=1.264
From: bosch@gcc.gnu.org To: gcc-gnats@gcc.gnu.org Cc: Subject: ada/6558 Date: 30 Oct 2002 17:42:55 -0000 CVSROOT: /cvs/gcc Module name: gcc Changes by: bosch@gcc.gnu.org 2002-10-30 09:42:54 Modified files: gcc/ada : misc.c Make-lang.in ChangeLog Log message: PR ada/6558 * misc.c : Include optabs.h * Make-lang.in (misc.o): Add dependency on optabs.h Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/misc.c.diff?cvsroot=gcc&r1=1.43&r2=1.44 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/Make-lang.in.diff?cvsroot=gcc&r1=1.17&r2=1.18 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/ChangeLog.diff?cvsroot=gcc&r1=1.264&r2=1.265
State-Changed-From-To: analyzed->closed State-Changed-Why: Fixed.
From: neroden@gcc.gnu.org To: gcc-gnats@gcc.gnu.org Cc: Subject: ada/6558 Date: 13 Nov 2002 21:19:51 -0000 CVSROOT: /cvs/gcc Module name: gcc Changes by: neroden@gcc.gnu.org 2002-11-13 13:19:51 Modified files: gcc/ada : ChangeLog config-lang.in Log message: 2002-11-13 Nathanael Nerode <neroden@gcc.gnu.org> PR ada/6558 * config-lang.in: Remove diff_excludes. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/ChangeLog.diff?cvsroot=gcc&r1=1.266&r2=1.267 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/ada/config-lang.in.diff?cvsroot=gcc&r1=1.6&r2=1.7