[llvm] r281289 - [lib/LTO] Expose getModule() in lto::InputFile.

Davide Italiano via llvm-commits llvm-commits at lists.llvm.org
Mon Sep 12 23:38:26 PDT 2016


On Mon, Sep 12, 2016 at 10:07 PM, Mehdi Amini <mehdi.amini at apple.com> wrote:
>
> On Sep 12, 2016, at 9:03 PM, Davide Italiano <davide at freebsd.org> wrote:
>
> On Sep 12, 2016 8:51 PM, "Mehdi Amini" <mehdi.amini at apple.com> wrote:
>>
>>
>>
>> Sent from my iPhone
>>
>> On Sep 12, 2016, at 8:13 PM, Davide Italiano <davide at freebsd.org> wrote:
>>
>>> On Mon, Sep 12, 2016 at 8:05 PM, Mehdi Amini <mehdi.amini at apple.com>
>>> wrote:
>>>>
>>>> Sneaky ping on https://reviews.llvm.org/D23132 :)
>>>>
>>>>
>>>> (In the meantime: +1 to Peter, please revert)
>>>>
>>>>
>>>
>>> This patch is not breaking any buildbot
>>
>>
>> That's not a criteria.
>>
>>> and it's very unlikely that's
>>> is gonna be used by anything outside of lld
>>
>>
>> As long as I delete this API without updating your client code, that's
>> fine with me.
>
> hmm. We already had this discussion when Peter tried to implement ThinLTO in
> lld. I don't think it's acceptable because we should all try to keep the
> bots green and happy. I think, instead, you might want to consider and
> comment on my idea of moving the check inside the library, checking if that
> would break any of your use cases.
>
> Right, we can discuss whatever you want. But right now you need to revert
> this API, it is not acceptable in the first place.
>

r281296 should end the whole story.


-- 
Davide

"There are no solved problems; there are only problems that are more
or less solved" -- Henri Poincare


More information about the llvm-commits mailing list