cvs commit: src/sys/dev/sound/pcm channel.c
Alexander Leidinger
Alexander at Leidinger.net
Fri Nov 18 09:10:47 GMT 2005
Stefan Farfeleder <stefanf at FreeBSD.org> wrote:
> On Thu, Nov 17, 2005 at 02:51:39PM +0100, Alexander Leidinger wrote:
>> Stefan Farfeleder <stefanf at FreeBSD.org> wrote:
>>
>> >On Tue, Nov 15, 2005 at 08:23:06PM +0800, Ariff Abdullah wrote:
>> >>On Tue, 15 Nov 2005 09:58:52 +0000 (GMT)
>> >>Robert Watson <rwatson at FreeBSD.org> wrote:
>>
>> >>> Probably a combination of -Wall and -Werror, which generally
>> >>> generates a warning along the following lines:
>> >>>
>> >>> warning: suggest parentheses around comparison in operand of &
>> >>>
>> >>That means, warnings caused by excessive inlining while compiling
>> >>maestro.c should also break the build if WARNS=1, right? Has anybody
>> >>encounter and getting annoyed by this?
>> >
>> >No, function inlining doesn't change the semantics of expressions.
>
> I misread Ariff's question. Please disregard.
Ok.
>> So -Wall and -Werror doesn' result in issuing an error for each warning but
>> instead in an error for warnings, which are issued because the code in
>> question may change the semantic depending on the change which needs to be
>> done to get rid of the warning?
>
> Whether GCC prints 'warning:' or 'error:' is not changed by -Werror.
But the behavior of gcc is changed by -Werror: gcc aborts instead of
generating an object file.
> I'm not sure understand the second part of your question. The warning
> is issued because a & b == c is parsed as a & (b == c).
The second part was how gcc behaves, I think it behaves like:
if -Werror && any_warning_generated:
bail_out
Your answer to the (misread) message from Ariff suggested:
if -Werror && warning_because_of_possible_semantic_change:
bail_out
Since you misread the message, no need to answer (except I'm wrong).
Bye,
Alexander.
--
http://www.Leidinger.net Alexander @ Leidinger.net: PGP ID = B0063FE7
http://www.FreeBSD.org netchild @ FreeBSD.org : PGP ID = 72077137
Until Eve arrived, this was a man's world.
-- Richard Amour
More information about the cvs-src
mailing list