This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH][RFC] Add #pragma message ... to gcc/g++
* Simon Baldwin wrote on Mon, Jul 21, 2008 at 05:34:20PM CEST:
> Ralf Wildenhues wrote:
>>
>> What I meant though was: your warning states that the (malformed) pragma
>> would be ignored, thus not printed. Reading the code tells me that it
>> is printed, however.
[...]
> It's really ignored, not printed. GCC_BAD() is a macro that expands to
> a warning() followed by return, so handle_pragma_message() returns
> without reaching fnotice() on any GCC_BAD(). I'm not wildly fond of
> this style of coding, but it's common in c-pragma.c, perhaps reflecting
> the fact that parts of the module may be rather old. Blending the new
> change in seemed preferable to adding something different or uprooting
> all the others.
Indeed. I didn't notice the 'return' in GCC_BAD, I didn't expect it
there. Stated very diplomatically, I'm not fond of that style either.
Sorry for chasing a wild goose here.
> Dejagnu should, I believe, be checking this by reporting any unexpected
> '#pragma message ...' output as excess errors or warnings -- they'll
> appear on gcc's stdout, but with no dg-message to catch them. I've
> added a bit of extra commentary to the test case to make this more
> apparent -- hopefully this will explain why the test catches all it's
> supposed to.
Ah, thanks.
Cheers,
Ralf