[PATCH] use the last passed -munaligned-access / -mno-unaligned-access

[Renato Golin]

r189542.

cheers,
--renato


On 28 August 2013 20:44, Jeroen Hofstee <clang at myspectrum.nl> wrote:

>  On 08/28/2013 09:33 PM, Jeroen Hofstee wrote:
>
> Hello Renato and others,
>
> On 08/27/2013 11:20 AM, Renato Golin wrote:
>
> On 27 August 2013 09:24, Dmitri Gribenko <gribozavr at gmail.com> wrote:
>
>> On Tue, Aug 27, 2013 at 1:11 AM, Renato Golin <renato.golin at linaro.org>
>> wrote:
>> > I don't have that strong an opinion to keep things as they are, but I
>> also
>> > wouldn't change unless there is a good reason (or consensus) to do so.
>>
>>  In my opinion, consistency is a pretty good reason.
>>
>
>  That's where the grey area is... Do we strive to be consistent with the
> previous behaviour, or with other flags?
>
>  The former, how it's implemented now, is that a "strict-align" flag
> would taint the whole argument list, since there was no way to "un-taint".
> The latter, would change how "strict-align" behaves, being more consistent
> with other flags, but less consistent with its former self.
>
>  Jeroen, do you have any input in this? Any special reason for not having
> made like other dual flags?
>
>
> No special reason, I must have been in a good mood I guess.
> I would like it even more to throw an error at the user for passing
> such inconsistent flags, with the question to make up his / her mind
> and filter-out one of them. Or preferably not add them both in the
> first place.
>
> Luckily we don't have to discuss that, for the simple fact gcc seems
> to use the last flag and clang seems to follow accordingly. Patches
> attached, intended to match that behaviour.
>
>
> follow up: Updated the subject and removed personal comment from the test.
>
> Regards,
> Jeroen
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20130829/acc9d3e4/attachment.html>