[LLVMbugs] [Bug 20672] New: test/CodeGen/X86/asm-invalid-register-class-crasher.ll doesn't test what it attempts to test

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=20672

            Bug ID: 20672
           Summary: test/CodeGen/X86/asm-invalid-register-class-crasher.ll
                    doesn't test what it attempts to test
           Product: libraries
           Version: trunk
          Hardware: Other
                OS: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: X86
          Assignee: unassignedbugs at nondot.org
          Reporter: wschmidt at linux.vnet.ibm.com
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified

I recently debugged a problem on PowerPC exposed by the subject test case.  In
the process, I discovered that the test doesn't do what it's intended to do.

During make check-all, the test generates the following script in
test/CodeGen/X86/Output/asm-invalid-register-class-crasher.ll.script:

set -o pipefail;{ not
/home/wschmidt/llvm/build/llvm-test/Release+Asserts/bin/l\
lc <
/home/wschmidt/llvm/llvm-test/test/CodeGen/X86/asm-invalid-register-class-\
crasher.ll -mtriple=i386-apple-darwin 2>&1
/home/wschmidt/llvm/build/llvm-test/\
test/CodeGen/X86/Output/asm-invalid-register-class-crasher.ll.tmp; }

This ends up actually attempting to read from
asm-invalid-register-class-crasher.ll.tmp, which doesn't exist.  So the test
fails because of the non-existence of the file, which "not" then reports as
success.  The IR in the test case is ignored.

The RUN line in the test is:

; RUN: not llc < %s -mtriple=i386-apple-darwin 2>&1 %t

I tried changing this to 

; RUN: not llc < %s -mtriple=i386-apple-darwin >%t 2>&1

but "not" returns 1 in this case.  Not sure how you want to fix this up but
it's not particulary useful as is (except for helping expose an elusive
self-hosting clang bug for PowerPC when opening a missing file). ;)

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20140815/8d0d7121/attachment.html>