[llvm] r179111 - Revert r176408 and r176407 to address PR15540.

[Shuxin Yang]

Hi, Nuno:

   Let me go through  the events about your changes.

   1.  Your change originally triggered a bug about LTO bootstrap build.
        Bill revert your change and file a bug.

   2.  My manager asked me to figure out the root cause. '
       I fixed the problem and reiterate (in many places) that your 
problematic should be applied with my change.
       Unfortunately, you silently let your code creep in without even a 
single word addressing my concerns.

   3. Yi Jiang later on realized that your change increase GVN 
compile-time by 10%.
       You try to "reproduce" with debug-built compiler and claim you 
cannot reproduce the problem,
       and you claim you are too busy to dig into the problem. (Why the 
hack you have time to write
       the code in the 1st place?).

    4. Nadav asked you to revert your change,  you seems to be quite 
reluctant, and ask Nadav to
        "judiciously" revert "some code", "not too much".

        This is really ridiculous! I have never heard before that a 
original author of the problematic
        patch asked other people to "judiciously" revert part of patch.

    5. Also you promise you bring the comment to "object" back, you 
never did!

   It seems it is very difficult to work w you. go back to this mail.

  > You should not remove  getUnderlyingObjectSize(), since it's 
required for correctness of BasicAA.
Which bug did you fixed in the 176407?
And, why we cannot remove getUnderlyingObjectSize()?

In 176407, you just redo I made previously in order to enable your 2nd 
change.
Check your emails, 176407 was similar to my original proposal, after 
discuss with Arnold,
I ditch this proposal, and go for his proposal.

If you want to revert 176407, you first need to reproduce the defect Yi 
report, and convince us with
data you collect.



On 4/13/13 7:55 PM, Nadav Rotem wrote:
>
> On Apr 13, 2013, at 6:22 PM, Nuno Lopes <nunoplopes at sapo.pt 
> <mailto:nunoplopes at sapo.pt>> wrote:
>
>> Hi Nadav,
>>
>> I'm sorry to say but you reverted too much. In particular:
>> - You should not remove  getUnderlyingObjectSize(), since it's 
>> required for correctness of BasicAA.
>> - BasicAA changes shouldn't be reverted for the same reason.
>> - Tests 18 and 19 have nothing to do with this, and should stay.
>>
>> I can perform these changes myself if you prefer.
>>
>
> Hi Nuno,
>
> Thanks for looking at this. Are you going to re-apply r176407 ?
>
> Thanks,
> Nadav
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20130413/2c567146/attachment.html>