On 7 Dec 2011, at 15:54, David Chisnall wrote:

> On 7 Dec 2011, at 15:48, Richard Frith-Macdonald wrote:
> 
>> I suggest you re-read the test code and comments because there's nothing 
>> random about using latin1.
> 
> Okay, I just looked at the test.  Apparently the LANG environment variable is 
> set, so it is not doing something random (is LANG set inside GNUstep Make for 
> subsequent compiles though?)
> 
>> Of course there may be errors ... but you don't seem to have found any, and 
>> just calling it random is not helpful.
>> If I'm not testing what I describe, please provide a report of why not and a 
>> fix.
> 
> I still don't understand WHY it is testing it.

To check whether the compiler supports non-ascii string literals.

> It seems that you have broken the build for lots of people (or, at least, 
> made them explicitly add a flag to remove a scary warning)

I think 'lots' is a small minority ... but how about suggesting a non-scary 
warning?

> to provide a compiler flag that is never needed in any of our existing code 
> and should never be provided for sane code.

1. The check is NOT to provide a compiler flag, it's to check compiler 
behavior.  The fact that checking behavior allows us to add some flags for 
safety/clarity is an incidental benefit.
2. The reason is that we now support non-ascii string literals for 
compatibility with OSX.  Of course we don't use it ... because we didn't 
support it before.
_______________________________________________
Gnustep-dev mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/gnustep-dev

Reply via email to