Re: [llvm-commits] Regression tests exist for a reason
On Apr 28, 2007, at 9:07 PM, Reid Spencer wrote: Hi Chris, On Sat, 28 Apr 2007 20:09:14 -0500 Christopher Lamb <[EMAIL PROTECTED]> wrote: As part of the SVN migration, perhaps we could get an autobuilder that builds and runs at least the llvm/test suite for each changeset. That'd at least make identifying the offending changeset possible. That's a pretty good idea. It could even auto-email the llvm- commits list saying "this change set broke the build or caused regressions" Could you please file an enhancement bug asking for this feature? That way it won't get forgotten. Done. http://llvm.org/bugs/show_bug.cgi?id=1368 -- Christopher Lamb ___ llvm-commits mailing list llvm-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
Re: [llvm-commits] Regression tests exist for a reason
Hi Chris, On Sat, 28 Apr 2007 20:09:14 -0500 Christopher Lamb <[EMAIL PROTECTED]> wrote: >As part of the SVN migration, perhaps we could get an autobuilder that builds >and runs at least the llvm/test suite for each changeset. That'd at least >make identifying the offending changeset possible. That's a pretty good idea. It could even auto-email the llvm-commits list saying "this change set broke the build or caused regressions" Could you please file an enhancement bug asking for this feature? That way it won't get forgotten. Thanks, Reid. > >-- >Christopher Lamb > > > ___ llvm-commits mailing list llvm-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
Re: [llvm-commits] Regression tests exist for a reason
As part of the SVN migration, perhaps we could get an autobuilder that builds and runs at least the llvm/test suite for each changeset. That'd at least make identifying the offending changeset possible. -- Christopher Lamb ___ llvm-commits mailing list llvm-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
Re: [llvm-commits] Regression tests exist for a reason
On Sat, 28 Apr 2007 09:45:19 -0700 Jeff Cohen <[EMAIL PROTECTED]> wrote: >That leaves us with six failures: > >FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmStruct.c >FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmUnion.c >FAIL: >/usr/home/jeffc/llvm/test/CodeGen/Generic/2007-04-08-MultipleFrameIndices.ll >for PR1308 >FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-29-lbrx-asm.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-31-InlineAsmAddrMode.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/2006-07-12-InlineAsmQConstraint.ll >for PR828 > >Which commit is responsible for these? We wouldn't have to ask the question >if everyone ran the regression tests before committing, much less waste time >answering it. > >Yes, I'm ranting here, but this happens too frequently. These look to be regressions from the inline assembly stuff that Chris was working on last night. Well, at least 4 of them. Not sure about the other two. > >Reid Spencer wrote: >> Jeff, >> >> On Sat, 28 Apr 2007 08:53:34 -0700 >> Jeff Cohen <[EMAIL PROTECTED]> wrote: >> >>> Recent commits have broken 28 tests. Regression tests exist for a >>> reason. Run them. Before committing. >>> >> >> Sorry. I thought I was committing a trivial change (compile breakage for >> jump across case labels) and accidentally committed my PR1146 changes too. >> Working over a hotel connection is less than satisfying. >> >> This should be fixed now. I'm running regressions as I type this. >> >> Reid. >> >> >>> FAIL: /usr/home/jeffc/llvm/test/Assembler/2006-05-26-VarargsCallEncode.ll >>> FAIL: >>> /usr/home/jeffc/llvm/test/C++Frontend/2003-11-25-ReturningOpaqueByValue.cpp >>> FAIL: /usr/home/jeffc/llvm/test/C++Frontend/2004-09-27-CompilerCrash.cpp >>> FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-04-10-StructParameters.c >>> FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-07-14-MiscTests.c >>> FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-07-14-MiscTests3.c >>> FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmStruct.c >>> FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmUnion.c >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/2007-03-06-AddR7.ll >>> FAIL: >>> /usr/home/jeffc/llvm/test/CodeGen/ARM/2007-03-26-RegScavengerAssert.ll >>> for PR1266 >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/sxt_rot.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/uxt_rot.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/add.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/i32_sub_1.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/zapnot.ll >>> FAIL: >>> /usr/home/jeffc/llvm/test/CodeGen/Generic/2007-04-08-MultipleFrameIndices.ll >>> >>> for PR1308 >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-29-lbrx-asm.ll >>> FAIL: >>> /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-31-InlineAsmAddrMode.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/and-elim.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/and_sext.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/small-arguments.ll >>> FAIL: >>> /usr/home/jeffc/llvm/test/CodeGen/X86/2006-07-12-InlineAsmQConstraint.ll >>> for PR828 >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/2006-11-12-CSRetCC.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/mingw-alloca.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/trunc-to-bool.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/x86-64-arg.ll >>> FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/x86-64-shortint.ll >>> FAIL: /usr/home/jeffc/llvm/test/Feature/paramattrs.ll >>> FAIL: /usr/home/jeffc/llvm/test/Integer/paramattrs_bt.ll >>> >>> >>> ___ >>> llvm-commits mailing list >>> llvm-commits@cs.uiuc.edu >>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits >>> >> >> >> >> >> > ___ llvm-commits mailing list llvm-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
Re: [llvm-commits] Regression tests exist for a reason
That leaves us with six failures: FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmStruct.c FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmUnion.c FAIL: /usr/home/jeffc/llvm/test/CodeGen/Generic/2007-04-08-MultipleFrameIndices.ll for PR1308 FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-29-lbrx-asm.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-31-InlineAsmAddrMode.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/2006-07-12-InlineAsmQConstraint.ll for PR828 Which commit is responsible for these? We wouldn't have to ask the question if everyone ran the regression tests before committing, much less waste time answering it. Yes, I'm ranting here, but this happens too frequently. Reid Spencer wrote: > Jeff, > > On Sat, 28 Apr 2007 08:53:34 -0700 > Jeff Cohen <[EMAIL PROTECTED]> wrote: > >> Recent commits have broken 28 tests. Regression tests exist for a >> reason. Run them. Before committing. >> > > Sorry. I thought I was committing a trivial change (compile breakage for jump > across case labels) and accidentally committed my PR1146 changes too. Working > over a hotel connection is less than satisfying. > > This should be fixed now. I'm running regressions as I type this. > > Reid. > > >> FAIL: /usr/home/jeffc/llvm/test/Assembler/2006-05-26-VarargsCallEncode.ll >> FAIL: >> /usr/home/jeffc/llvm/test/C++Frontend/2003-11-25-ReturningOpaqueByValue.cpp >> FAIL: /usr/home/jeffc/llvm/test/C++Frontend/2004-09-27-CompilerCrash.cpp >> FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-04-10-StructParameters.c >> FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-07-14-MiscTests.c >> FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-07-14-MiscTests3.c >> FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmStruct.c >> FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmUnion.c >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/2007-03-06-AddR7.ll >> FAIL: >> /usr/home/jeffc/llvm/test/CodeGen/ARM/2007-03-26-RegScavengerAssert.ll >> for PR1266 >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/sxt_rot.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/uxt_rot.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/add.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/i32_sub_1.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/zapnot.ll >> FAIL: >> /usr/home/jeffc/llvm/test/CodeGen/Generic/2007-04-08-MultipleFrameIndices.ll >> for PR1308 >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-29-lbrx-asm.ll >> FAIL: >> /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-31-InlineAsmAddrMode.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/and-elim.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/and_sext.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/small-arguments.ll >> FAIL: >> /usr/home/jeffc/llvm/test/CodeGen/X86/2006-07-12-InlineAsmQConstraint.ll >> for PR828 >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/2006-11-12-CSRetCC.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/mingw-alloca.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/trunc-to-bool.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/x86-64-arg.ll >> FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/x86-64-shortint.ll >> FAIL: /usr/home/jeffc/llvm/test/Feature/paramattrs.ll >> FAIL: /usr/home/jeffc/llvm/test/Integer/paramattrs_bt.ll >> >> >> ___ >> llvm-commits mailing list >> llvm-commits@cs.uiuc.edu >> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits >> > > > > > ___ llvm-commits mailing list llvm-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
Re: [llvm-commits] Regression tests exist for a reason
Jeff, On Sat, 28 Apr 2007 08:53:34 -0700 Jeff Cohen <[EMAIL PROTECTED]> wrote: >Recent commits have broken 28 tests. Regression tests exist for a >reason. Run them. Before committing. Sorry. I thought I was committing a trivial change (compile breakage for jump across case labels) and accidentally committed my PR1146 changes too. Working over a hotel connection is less than satisfying. This should be fixed now. I'm running regressions as I type this. Reid. > >FAIL: /usr/home/jeffc/llvm/test/Assembler/2006-05-26-VarargsCallEncode.ll >FAIL: >/usr/home/jeffc/llvm/test/C++Frontend/2003-11-25-ReturningOpaqueByValue.cpp >FAIL: /usr/home/jeffc/llvm/test/C++Frontend/2004-09-27-CompilerCrash.cpp >FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-04-10-StructParameters.c >FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-07-14-MiscTests.c >FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-07-14-MiscTests3.c >FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmStruct.c >FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmUnion.c >FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/2007-03-06-AddR7.ll >FAIL: >/usr/home/jeffc/llvm/test/CodeGen/ARM/2007-03-26-RegScavengerAssert.ll >for PR1266 >FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/sxt_rot.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/uxt_rot.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/add.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/i32_sub_1.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/zapnot.ll >FAIL: >/usr/home/jeffc/llvm/test/CodeGen/Generic/2007-04-08-MultipleFrameIndices.ll >for PR1308 >FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-29-lbrx-asm.ll >FAIL: >/usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-31-InlineAsmAddrMode.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/and-elim.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/and_sext.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/small-arguments.ll >FAIL: >/usr/home/jeffc/llvm/test/CodeGen/X86/2006-07-12-InlineAsmQConstraint.ll >for PR828 >FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/2006-11-12-CSRetCC.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/mingw-alloca.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/trunc-to-bool.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/x86-64-arg.ll >FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/x86-64-shortint.ll >FAIL: /usr/home/jeffc/llvm/test/Feature/paramattrs.ll >FAIL: /usr/home/jeffc/llvm/test/Integer/paramattrs_bt.ll > > >___ >llvm-commits mailing list >llvm-commits@cs.uiuc.edu >http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits ___ llvm-commits mailing list llvm-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
[llvm-commits] Regression tests exist for a reason
Recent commits have broken 28 tests. Regression tests exist for a reason. Run them. Before committing. FAIL: /usr/home/jeffc/llvm/test/Assembler/2006-05-26-VarargsCallEncode.ll FAIL: /usr/home/jeffc/llvm/test/C++Frontend/2003-11-25-ReturningOpaqueByValue.cpp FAIL: /usr/home/jeffc/llvm/test/C++Frontend/2004-09-27-CompilerCrash.cpp FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-04-10-StructParameters.c FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-07-14-MiscTests.c FAIL: /usr/home/jeffc/llvm/test/CFrontend/2002-07-14-MiscTests3.c FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmStruct.c FAIL: /usr/home/jeffc/llvm/test/CFrontend/2007-04-11-InlineAsmUnion.c FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/2007-03-06-AddR7.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/2007-03-26-RegScavengerAssert.ll for PR1266 FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/sxt_rot.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/ARM/uxt_rot.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/add.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/i32_sub_1.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/Alpha/zapnot.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/Generic/2007-04-08-MultipleFrameIndices.ll for PR1308 FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-29-lbrx-asm.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/2007-01-31-InlineAsmAddrMode.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/and-elim.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/and_sext.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/PowerPC/small-arguments.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/2006-07-12-InlineAsmQConstraint.ll for PR828 FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/2006-11-12-CSRetCC.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/mingw-alloca.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/trunc-to-bool.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/x86-64-arg.ll FAIL: /usr/home/jeffc/llvm/test/CodeGen/X86/x86-64-shortint.ll FAIL: /usr/home/jeffc/llvm/test/Feature/paramattrs.ll FAIL: /usr/home/jeffc/llvm/test/Integer/paramattrs_bt.ll ___ llvm-commits mailing list llvm-commits@cs.uiuc.edu http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
