[llvm-dev] [RFC] A proposal for byval in a world with opaque pointers

[David Blaikie via llvm-dev]

On Tue, Jan 19, 2016 at 3:30 PM, Eddy B. <edy.burt at gmail.com> wrote:

>
>
> 2016-01-20 1:18 GMT+02:00 David Blaikie <dblaikie at gmail.com>:
>
>>
>>
>> On Tue, Jan 19, 2016 at 3:16 PM, Eddy B. <edy.burt at gmail.com> wrote:
>>
>>> > Do all tests need it? Align is just for non-default alignment (>
>>> align(1)) so it may be OK for the tests to omit it & it's probably not
>>> right to /require/ an align attribute on a byval parameter.
>>>
>>> I've had to fix some of the tests which checked that the right
>>> alignment was applied,
>>>
>>
>> Rough order of magnitude number of tests that needed to be fixed? (10s?
>> 100s?)
>>
>
> Relevant commit:
> https://github.com/eddyb/llvm/commit/c76e3d6bd6b7008795e16f4422a014d2ff7d48dd
> It's just 4 tests, but I had to split each one because they were getting
> ran against
> multiple targets, and those needed different alignment attributes.
>

Only casually glanced at that, but do we need to test it with different
alignments on different targets? Or can we just pick one alignment to test
on all of them, do you think? (if you're not sure, that's OK - I/we can
look into it & see what the tests are trying to do, etc)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20160119/1024a52e/attachment.html>