r187174 - Fix GNU ObjC ABI for a message returning a struct.
arthur.j.odwyer at gmail.com
Fri Jul 26 18:18:31 PDT 2013
On Fri, Jul 26, 2013 at 5:05 PM, Jonathan Schleifer <js at webkeks.org> wrote:
> Am 27.07.2013 um 01:54 schrieb John McCall:
>> The feature enables users to rely on arbitrary message forwarding
>> if they're willing to live on certain minimum compiler and runtime versions.
>> That’s a pretty simple and comprehensible message for users.
>> The policy of clang is not to introduce __has_feature checks for bug fixes.
> This, however, is different. It will mean that using an old Clang version or
> GCC will generate *crashing code* instead of a feature being unavailable.
> That's a completely different thing!
> Crashing code is *totally* inacceptable! GNU used to be ok with crashing
> code (there was a lot in their runtime that just crashed one some CPU
> architectures), but that doesn't mean we need to be as bad as they were.
> IMHO, code should never crash, period. It should check for the define I
> provide and if it absolutely relies on the feature error out, or of it's
> optional disable it.
Jonathan, I don't think that's a good argument. You've already pointed
out that the test for "does Clang support this feature" is as simple
as "#if defined(__clang__) && (__clang_major__ > 3 || (__clang_major__
== 3 && __clang_minor__ >= 4))". You object that such a test would
leave out bleeding-edge versions of Clang between now and the 3.4
release; but that's a very short-term problem that will go away in a
few months, whereas __has_feature() introduces a problem that will by
definition be around for the rest of everyone's life.
(Besides, this is Objective-C. The Objective-C compiler is a
fast-moving target, and generates silly and/or crashy code on a
regular basis for sufficiently obscure constructs. Forwarding a method
that returns a struct by value, using a particular non-Apple non-GNU
runtime and a particular compiler built from source within a small
time window, seems like the very definition of "sufficiently
More information about the cfe-commits