[llvm-commits] [llvm] r167740 - in /llvm/trunk: lib/Transforms/Scalar/Reassociate.cpp test/Transforms/Reassociate/mul_neg.ll test/Transforms/Reassociate/multistep.ll

[Evan Cheng]

On Nov 12, 2012, at 7:19 PM, Shuxin Yang <shuxin.llvm at gmail.com> wrote:

> I hope each pass has more than one owner. In case one owner is busy, the other one could response in
> reasonable short period of time.

FYI. Anyone can review any patch. It doesn't require the owner, especially since many passes don't have owners. 

> 
> During the dev-meeting, I heard quite a few complaint about the pailful "pings".
> 
> If the code review take more than one week, it basically means the author who make the change need
> to re-test from scratch. This cost is reasonable for big change. it is way to high for tiny to medium change.

I agree with this. I think the llvm community is suffering a bit with longer review latency because there are a lot of confusion about the review policy. I think that's part of growing pain. It's something we can deal with as long as the community as a whole remain civil (which we are). 

I personally like to see more review after commit. But I understand it's sometimes hard to draw the line and people have different opinions about it. 

Evan

> 
> It is also make thing hard for incremental change
> 
> On 11/12/2012 07:00 PM, Chandler Carruth wrote:
>> On Mon, Nov 12, 2012 at 6:55 PM, Evan Cheng <evan.cheng at apple.com> wrote:
>>> On Nov 12, 2012, at 3:51 PM, Chris Lattner <clattner at apple.com> wrote:
>>> 
>>>> On Nov 12, 2012, at 1:08 PM, Evan Cheng <evan.cheng at apple.com> wrote:
>>>> 
>>>>> Hi Eli,
>>>>> 
>>>>> The approach seems sound to me. Since no one has raised design concerns at this point and the fact this is a a bug fix for a crash, I think it's appropriate for review after commit.
>>>> This patch isn't "obvious", and Shuxin is still new to this part of the code.  Unless it's been reviewed it shouldn't go in.
>>> The patch was submitted for review back on 11/5 and there were some discussions on 11/6 and 11/7. There were no additional comments since then. Should it continue to sit in review limbo?
>> The standard response, and what everyone else in this position does
>> (and there are *many*...) is to ping, asking for an update.
>> _______________________________________________
>> llvm-commits mailing list
>> llvm-commits at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>