[AIX] XFAIL tests because of no big archive writer operation support

[Jake-Egan]

Summary

Big archive writer operation is not currently supported so mark these tests XFAIL for now.

[jsji]

LGTM. Thanks Jake.

[MaskRay]

The problem is whether there are other tests using "big archive" and needing XFAIL: system-aix (FWIW I think UNSUPPORTED is more appropriate).
llvm/test/Object and llvm/test/tools/llvm-ar have a dozen archive writer tests. If they all need UNSUPPORTED: system-aix, that's too much of a churn.

[Jake-Egan]

The patch D122746 XFAILed the archive writer tests in llvm/test/Object and llvm/test/tools/llvm-ar, but it missed the 5 tests in this patch. There aren't any other tests that need to be dealt with at the moment. I agree UNSUPPORTED probably makes more sense, but I was trying to be consistent with the other patch.

[jhenderson]

In D122949#3424498, @Jake-Egan wrote:

The patch D122746 XFAILed the archive writer tests in llvm/test/Object and llvm/test/tools/llvm-ar, but it missed the 5 tests in this patch. There aren't any other tests that need to be dealt with at the moment. I agree UNSUPPORTED probably makes more sense, but I was trying to be consistent with the other patch.

I considered this when reviewing, but I actually think XFAIL is better, because there are reasonably near term plans to implement the functionality. Marking as UNSUPPORTED would mean the tests won't start XPASSing at that point, meaning they may never be run for the new format, even though they should be.

That being said, for tests not testing the archiver, but rather testing a different tool, you should instead use the explicit --format option, so that they continue working on AIX. Note that this is what the original patch did.

[Jake-Egan]

In D122949#3425633, @jhenderson wrote:

In D122949#3424498, @Jake-Egan wrote:

The patch D122746 XFAILed the archive writer tests in llvm/test/Object and llvm/test/tools/llvm-ar, but it missed the 5 tests in this patch. There aren't any other tests that need to be dealt with at the moment. I agree UNSUPPORTED probably makes more sense, but I was trying to be consistent with the other patch.

I considered this when reviewing, but I actually think XFAIL is better, because there are reasonably near term plans to implement the functionality. Marking as UNSUPPORTED would mean the tests won't start XPASSing at that point, meaning they may never be run for the new format, even though they should be.

That being said, for tests not testing the archiver, but rather testing a different tool, you should instead use the explicit --format option, so that they continue working on AIX. Note that this is what the original patch did.

In that case, I can post a patch to use --format instead for the tests not testing the archiver.

llvm/test/tools/llvm-link/archive.ll
llvm/test/tools/llvm-lipo/create-archive-input.test
llvm/test/tools/llvm-objcopy/MachO/universal-object.test

llvm/test/tools/llvm-ar/mri1.test

+; XFAIL: system-aix
+
 ; RUN: echo create %t.a > %t.mri
 ; RUN: echo save >> %t.mri
 ; RUN: echo end >> %t.mri
 
 ; RUN: llvm-ar -M  < %t.mri
 ; RUN: llvm-ar t %t.a

llvm/test/tools/llvm-ar/response.test

+# XFAIL: system-aix
 ## llvm-ar should be able to consume response files.
 
 # RUN: echo 'contents' > %t.txt
 # RUN: echo 'rc %t1.a %t.txt' > %t.response1.txt
 # RUN: llvm-ar @%t.response1.txt
 # RUN: llvm-ar p %t1.a | FileCheck %s --check-prefix=CONTENTS
 
 ## Quotes and Spaces.

llvm/test/tools/llvm-link/archive.ll

+# XFAIL: system-aix
+
 # RUN: llvm-as %S/Inputs/f.ll -o %t.f.bc
 # RUN: llvm-as %S/Inputs/g.ll -o %t.g.bc
 # RUN: llvm-ar cr %t.fg.a %t.f.bc %t.g.bc
 # RUN: llvm-ar cr %t.empty.lib
 # RUN: llvm-link %S/Inputs/h.ll %t.fg.a %t.empty.lib -o %t.linked.bc
 
 # RUN: llvm-nm %t.linked.bc | FileCheck %s
 

llvm/test/tools/llvm-lipo/create-archive-input.test

+# XFAIL: system-aix
+
 # RUN: yaml2obj %p/Inputs/i386-slice.yaml -o %t-i386.o
 # RUN: yaml2obj %p/Inputs/x86_64-slice.yaml -o %t-x86_64.o
 # RUN: yaml2obj %p/Inputs/i386-x86_64-universal.yaml -o %t-universal.o
 # RUN: llvm-as %p/Inputs/armv7-ios.ll -o %t-ir-armv7.o
 # RUN: llvm-as %p/Inputs/x64-osx.ll -o %t-ir-x86_64.o
 
 # RUN: llvm-ar cr %t.empty.a
 # RUN: not llvm-lipo %t.empty.a -create -output /dev/null 2>&1 | FileCheck --check-prefix=EMPTY-ARCHIVE %s

llvm/test/tools/llvm-objcopy/MachO/universal-object.test

+# XFAIL: system-aix
 # This test verifies that llvm-objcopy copies a univeral Mach-O object file properly.
 
 # RUN: yaml2obj %p/Inputs/i386.yaml -o %t.i386
 # RUN: yaml2obj %p/Inputs/x86_64.yaml -o %t.x86_64
 
 ## Case 1: copy a universal object containing regular Mach-O objects.
 # RUN: llvm-lipo %t.i386 %t.x86_64 -create -output %t.universal
 # RUN: llvm-objcopy %t.universal %t.universal.copy