[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: compiler version checks
From: |
Ralf Wildenhues |
Subject: |
Re: compiler version checks |
Date: |
Fri, 1 May 2009 11:03:02 +0200 |
User-agent: |
Mutt/1.5.18 (2008-05-17) |
Hi Eric,
* Eric Blake wrote on Thu, Apr 30, 2009 at 02:38:00PM CEST:
> According to Ralf Wildenhues on 4/28/2009 12:02 AM:
> >> for ac_option in --version -v -V -qversion ; do
> >> _AC_DO_LIMIT([$ac_compiler $ac_option >&AS_MESSAGE_LOG_FD])
> >> done
> >
> > Sure. Unfortunately, it wouldn't work to check the exit status of
> > the command, as some don't reliably fail with an unknown argument.
>
> Agreed, so no blind use of break within the loop, and no way to reduce the
> number of forks in general. On the other hand, maybe it is worth adding
> an exception that if the first iteration, using --version, detects gcc,
> then we don't need the other iterations (it is okay to make life on free
> systems optimal, as long as it doesn't preclude life on other systems);
> but special-casing gcc does not have to be part of this patch. Overall,
> in case I wasn't clear earlier, I like the intent of this patch, so please
> commit once you've converted all of four of the original linear compiler
> checks into loops.
You were clear enough, I just ran out of steam the other night.
I've commited the patch now after conversion to loops.
For gcc, it's actually helpful to see both --version and -v output;
also, it typically produces short and concise error messages for options
it does not understand, so I think special-casing it is unneeded here.
Thanks,
Ralf
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- Re: compiler version checks,
Ralf Wildenhues <=