[llvm-commits] [FIX] for clang-X86_64-freebsd buildbot (very small patch, the only string is changed)

[Jim Grosbach]

Cool. That part's fine w/ me. I haven't followed the rest of the SwitchRange stuff enough to comment beyond that, though.

-Jim

On Jul 11, 2012, at 1:15 PM, Stepan Dyatkovskiy <stpworld at narod.ru> wrote:

> From llvm-diff doc:
> "If llvm-diff finds no differences between the modules, it will exit with 0 and produce no output. Otherwise it will exit with a non-zero value."
> So it llvm-diff is better choice then. Fixed patch is attached.
> -Stepan.
> 
> 12.07.2012, 00:07, "Jim Grosbach" <grosbach at apple.com>:
>> llvm-diff compares bitcode. Perhaps it would work here? Note: I haven't tried it for this; I'm hypothesizing.
>> 
>> -Jim
>> 
>> On Jul 11, 2012, at 1:06 PM, Stepan Dyatkovskiy <STPWORLD at narod.ru> wrote:
>> 
>>>  How to test that BitcodeReader=>Writer=Reader doesn't distorts the data then?
>>>  -Stepan.
>>> 
>>>  12.07.2012, 00:03, "Jim Grosbach" <grosbach at apple.com>:
>>>>  On Jul 11, 2012, at 1:00 PM, Benjamin Kramer <benny.kra at gmail.com> wrote:
>>>>>   On 11.07.2012, at 21:55, Stepan Dyatkovskiy <stpworld at narod.ru> wrote:
>>>>>>   Index: test/Bitcode/2012-05-07-SwitchInstRangesSupport.ll
>>>>>>   ===================================================================
>>>>>>   --- test/Bitcode/2012-05-07-SwitchInstRangesSupport.ll (revision 160066)
>>>>>>   +++ test/Bitcode/2012-05-07-SwitchInstRangesSupport.ll (working copy)
>>>>>>   @@ -6,7 +6,7 @@
>>>>>>   ; RUN: llvm-dis %t.bc -o - | tail -n +2 > %t.ll
>>>>>>   ; RUN: llvm-as %t.ll -o %t2.bc
>>>>>>   ; RUN: llvm-dis %t2.bc -o - | tail -n +2 > %t2.ll
>>>>>>   -; RUN: diff %t.ll %t2.ll | not grep .*
>>>>>>   +; RUN: diff %t.ll %t2.ll | not grep ^.*$
>>>>>   diff returns 1 (error) when there is a difference so you can drop the "| not grep ^.*$" part here.
>>>>  Not all platforms have diff(1). Windows, e.g. Not sure how much we care, though.
>>>> 
>>>>  -j
> <fixbuildbot.patch>