[llvm] r223774 - [x86] Fix the test to actually test things for the CPU names, add the

[Justin Bogner]

Chandler Carruth <chandlerc at gmail.com> writes:
> On Fri, Dec 19, 2014 at 3:27 PM, Steven Wu <stevenwu at apple.com> wrote:
>
>     Hi Chandler
>    
>     The testcases you committed is breaking our internal bots running with
>     gmalloc. When gmalloc is enable, it will print version numbers and etc. to
>     stderr.
>     Here is an example:
>     GuardMalloc[llc-23060]: Allocations will be placed on 16 byte boundaries.
>     GuardMalloc[llc-23060]:  - Some buffer overruns may not be noticed.
>     GuardMalloc[llc-23060]:  - Applications using vector instructions (e.g.,
>     SSE) should work.
>     GuardMalloc[llc-23060]: version 104
>    
>     Can you modify the testcase so that it won’t be triggered by gmalloc (like
>     grep for error?)
>
> I would much rather not.
>
> I don't know why it is reasonable to support arbitrary output to stderr in
> test cases. Lots of our test cases examine stderr for specific output... Why
> do those work?

Because they check for the output they want (or don't, as the case may
be), rather than checking for no output. Please use FileCheck for this:

    ; CHECK-NOT: not a recognized processor for this target
    ; RUN: llc < %s -o /dev/null -mtriple=x86_64-unknown-unknown -mcpu=nocona 2>&1 | FileCheck --allow-empty