* Re: [Qemu-devel] [PATCH] configure: fix ALSA configure test
@ 2012-07-17 19:28 ` Peter Maydell
0 siblings, 0 replies; 20+ messages in thread
From: Peter Maydell @ 2012-07-17 19:28 UTC (permalink / raw)
To: Stefan Weil
Cc: Blue Swirl, qemu-trivial, Igor Mitsyanko, QEMU Developers,
Anthony Liguori
On 17 July 2012 20:24, Stefan Weil <sw@weilnetz.de> wrote:
> The arguments why -Werror is a bad idea for some configure tests
> are reasonable.
>
> Nevertheless the QEMU community was able to produce thousands of
> lines of code which compile without a warning, so we should be able
> to create warning and error free code for a handful of configure
> tests.
The trouble is that the warnings and errors here don't cause the
build to fail noisily; that's a big distinction IMHO.
I suppose we could make compile_prog do something like:
* run the compile test
* if it fails => test failure as now
* if it succeeds (and we're doing a Werror build at all),
rerun the same test with -Werror
* if that fails, abort configure with an error message
Then we would have the same "make the problem obvious" effect
that plain -Werror provides for our main compilation.
> The 4 patches above are valid and can be applied with or without
> -Werror, therefore qemu-trivial or whoever does not have to wait for
> Peter's patch.
Yes, I agree we might as well fix these errors since we've now
noticed them, regardless of whether or not we apply my patch
(which I've just sent).
-- PMM
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-trivial] [Qemu-devel] [PATCH] configure: fix ALSA configure test
2012-07-17 19:28 ` Peter Maydell
@ 2012-07-17 19:42 ` Stefan Weil
-1 siblings, 0 replies; 20+ messages in thread
From: Stefan Weil @ 2012-07-17 19:42 UTC (permalink / raw)
To: Peter Maydell; +Cc: Blue Swirl, qemu-trivial, QEMU Developers, Anthony Liguori
Am 17.07.2012 21:28, schrieb Peter Maydell:
> On 17 July 2012 20:24, Stefan Weil <sw@weilnetz.de> wrote:
>> The arguments why -Werror is a bad idea for some configure tests
>> are reasonable.
>>
>> Nevertheless the QEMU community was able to produce thousands of
>> lines of code which compile without a warning, so we should be able
>> to create warning and error free code for a handful of configure
>> tests.
> The trouble is that the warnings and errors here don't cause the
> build to fail noisily; that's a big distinction IMHO.
> I suppose we could make compile_prog do something like:
> * run the compile test
> * if it fails => test failure as now
> * if it succeeds (and we're doing a Werror build at all),
> rerun the same test with -Werror
> * if that fails, abort configure with an error message
> Then we would have the same "make the problem obvious" effect
> that plain -Werror provides for our main compilation.
Good idea. Of course it will increase the time needed for
running the default configure, but I think that's acceptable
if we don't use it for the tests of the compiler warning options.
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [PATCH] configure: fix ALSA configure test
@ 2012-07-17 19:42 ` Stefan Weil
0 siblings, 0 replies; 20+ messages in thread
From: Stefan Weil @ 2012-07-17 19:42 UTC (permalink / raw)
To: Peter Maydell; +Cc: Blue Swirl, qemu-trivial, QEMU Developers, Anthony Liguori
Am 17.07.2012 21:28, schrieb Peter Maydell:
> On 17 July 2012 20:24, Stefan Weil <sw@weilnetz.de> wrote:
>> The arguments why -Werror is a bad idea for some configure tests
>> are reasonable.
>>
>> Nevertheless the QEMU community was able to produce thousands of
>> lines of code which compile without a warning, so we should be able
>> to create warning and error free code for a handful of configure
>> tests.
> The trouble is that the warnings and errors here don't cause the
> build to fail noisily; that's a big distinction IMHO.
> I suppose we could make compile_prog do something like:
> * run the compile test
> * if it fails => test failure as now
> * if it succeeds (and we're doing a Werror build at all),
> rerun the same test with -Werror
> * if that fails, abort configure with an error message
> Then we would have the same "make the problem obvious" effect
> that plain -Werror provides for our main compilation.
Good idea. Of course it will increase the time needed for
running the default configure, but I think that's acceptable
if we don't use it for the tests of the compiler warning options.
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-trivial] [Qemu-devel] [PATCH] configure: fix ALSA configure test
2012-07-17 19:28 ` Peter Maydell
@ 2012-07-21 9:37 ` Stefan Hajnoczi
-1 siblings, 0 replies; 20+ messages in thread
From: Stefan Hajnoczi @ 2012-07-21 9:37 UTC (permalink / raw)
To: Peter Maydell
Cc: Igor Mitsyanko, qemu-trivial, Stefan Weil, QEMU Developers,
Blue Swirl, Anthony Liguori
On Tue, Jul 17, 2012 at 08:28:40PM +0100, Peter Maydell wrote:
> On 17 July 2012 20:24, Stefan Weil <sw@weilnetz.de> wrote:
> > The arguments why -Werror is a bad idea for some configure tests
> > are reasonable.
> >
> > Nevertheless the QEMU community was able to produce thousands of
> > lines of code which compile without a warning, so we should be able
> > to create warning and error free code for a handful of configure
> > tests.
>
> The trouble is that the warnings and errors here don't cause the
> build to fail noisily; that's a big distinction IMHO.
> I suppose we could make compile_prog do something like:
> * run the compile test
> * if it fails => test failure as now
> * if it succeeds (and we're doing a Werror build at all),
> rerun the same test with -Werror
> * if that fails, abort configure with an error message
> Then we would have the same "make the problem obvious" effect
> that plain -Werror provides for our main compilation.
>
> > The 4 patches above are valid and can be applied with or without
> > -Werror, therefore qemu-trivial or whoever does not have to wait for
> > Peter's patch.
>
> Yes, I agree we might as well fix these errors since we've now
> noticed them, regardless of whether or not we apply my patch
> (which I've just sent).
Build/configure fixes should go straight into qemu.git.
Stefan
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [Qemu-trivial] [PATCH] configure: fix ALSA configure test
@ 2012-07-21 9:37 ` Stefan Hajnoczi
0 siblings, 0 replies; 20+ messages in thread
From: Stefan Hajnoczi @ 2012-07-21 9:37 UTC (permalink / raw)
To: Peter Maydell
Cc: Igor Mitsyanko, qemu-trivial, Stefan Weil, QEMU Developers,
Blue Swirl, Anthony Liguori
On Tue, Jul 17, 2012 at 08:28:40PM +0100, Peter Maydell wrote:
> On 17 July 2012 20:24, Stefan Weil <sw@weilnetz.de> wrote:
> > The arguments why -Werror is a bad idea for some configure tests
> > are reasonable.
> >
> > Nevertheless the QEMU community was able to produce thousands of
> > lines of code which compile without a warning, so we should be able
> > to create warning and error free code for a handful of configure
> > tests.
>
> The trouble is that the warnings and errors here don't cause the
> build to fail noisily; that's a big distinction IMHO.
> I suppose we could make compile_prog do something like:
> * run the compile test
> * if it fails => test failure as now
> * if it succeeds (and we're doing a Werror build at all),
> rerun the same test with -Werror
> * if that fails, abort configure with an error message
> Then we would have the same "make the problem obvious" effect
> that plain -Werror provides for our main compilation.
>
> > The 4 patches above are valid and can be applied with or without
> > -Werror, therefore qemu-trivial or whoever does not have to wait for
> > Peter's patch.
>
> Yes, I agree we might as well fix these errors since we've now
> noticed them, regardless of whether or not we apply my patch
> (which I've just sent).
Build/configure fixes should go straight into qemu.git.
Stefan
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-trivial] [Qemu-devel] [PATCH] configure: fix ALSA configure test
2012-07-17 19:28 ` Peter Maydell
@ 2012-07-23 17:40 ` Blue Swirl
-1 siblings, 0 replies; 20+ messages in thread
From: Blue Swirl @ 2012-07-23 17:40 UTC (permalink / raw)
To: Peter Maydell
Cc: qemu-trivial, Stefan Weil, Igor Mitsyanko, QEMU Developers,
Anthony Liguori
On Tue, Jul 17, 2012 at 7:28 PM, Peter Maydell <peter.maydell@linaro.org> wrote:
> On 17 July 2012 20:24, Stefan Weil <sw@weilnetz.de> wrote:
>> The arguments why -Werror is a bad idea for some configure tests
>> are reasonable.
>>
>> Nevertheless the QEMU community was able to produce thousands of
>> lines of code which compile without a warning, so we should be able
>> to create warning and error free code for a handful of configure
>> tests.
>
> The trouble is that the warnings and errors here don't cause the
> build to fail noisily; that's a big distinction IMHO.
> I suppose we could make compile_prog do something like:
> * run the compile test
Unfortunately that would break cross compiling.
> * if it fails => test failure as now
> * if it succeeds (and we're doing a Werror build at all),
> rerun the same test with -Werror
> * if that fails, abort configure with an error message
> Then we would have the same "make the problem obvious" effect
> that plain -Werror provides for our main compilation.
>
>> The 4 patches above are valid and can be applied with or without
>> -Werror, therefore qemu-trivial or whoever does not have to wait for
>> Peter's patch.
>
> Yes, I agree we might as well fix these errors since we've now
> noticed them, regardless of whether or not we apply my patch
> (which I've just sent).
>
> -- PMM
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [PATCH] configure: fix ALSA configure test
@ 2012-07-23 17:40 ` Blue Swirl
0 siblings, 0 replies; 20+ messages in thread
From: Blue Swirl @ 2012-07-23 17:40 UTC (permalink / raw)
To: Peter Maydell
Cc: qemu-trivial, Stefan Weil, Igor Mitsyanko, QEMU Developers,
Anthony Liguori
On Tue, Jul 17, 2012 at 7:28 PM, Peter Maydell <peter.maydell@linaro.org> wrote:
> On 17 July 2012 20:24, Stefan Weil <sw@weilnetz.de> wrote:
>> The arguments why -Werror is a bad idea for some configure tests
>> are reasonable.
>>
>> Nevertheless the QEMU community was able to produce thousands of
>> lines of code which compile without a warning, so we should be able
>> to create warning and error free code for a handful of configure
>> tests.
>
> The trouble is that the warnings and errors here don't cause the
> build to fail noisily; that's a big distinction IMHO.
> I suppose we could make compile_prog do something like:
> * run the compile test
Unfortunately that would break cross compiling.
> * if it fails => test failure as now
> * if it succeeds (and we're doing a Werror build at all),
> rerun the same test with -Werror
> * if that fails, abort configure with an error message
> Then we would have the same "make the problem obvious" effect
> that plain -Werror provides for our main compilation.
>
>> The 4 patches above are valid and can be applied with or without
>> -Werror, therefore qemu-trivial or whoever does not have to wait for
>> Peter's patch.
>
> Yes, I agree we might as well fix these errors since we've now
> noticed them, regardless of whether or not we apply my patch
> (which I've just sent).
>
> -- PMM
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-trivial] [Qemu-devel] [PATCH] configure: fix ALSA configure test
2012-07-23 17:40 ` Blue Swirl
@ 2012-07-23 17:45 ` Peter Maydell
-1 siblings, 0 replies; 20+ messages in thread
From: Peter Maydell @ 2012-07-23 17:45 UTC (permalink / raw)
To: Blue Swirl
Cc: qemu-trivial, Stefan Weil, Igor Mitsyanko, QEMU Developers,
Anthony Liguori
On 23 July 2012 18:40, Blue Swirl <blauwirbel@gmail.com> wrote:
> On Tue, Jul 17, 2012 at 7:28 PM, Peter Maydell <peter.maydell@linaro.org> wrote:
>> The trouble is that the warnings and errors here don't cause the
>> build to fail noisily; that's a big distinction IMHO.
>> I suppose we could make compile_prog do something like:
>> * run the compile test
>
> Unfortunately that would break cross compiling.
Sorry, that was slightly unclear phrasing. By "run the test"
I meant "do the check that we can compile the test code"
"run the binary produced". This series won't break cross
compiling.
-- PMM
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [PATCH] configure: fix ALSA configure test
@ 2012-07-23 17:45 ` Peter Maydell
0 siblings, 0 replies; 20+ messages in thread
From: Peter Maydell @ 2012-07-23 17:45 UTC (permalink / raw)
To: Blue Swirl
Cc: qemu-trivial, Stefan Weil, Igor Mitsyanko, QEMU Developers,
Anthony Liguori
On 23 July 2012 18:40, Blue Swirl <blauwirbel@gmail.com> wrote:
> On Tue, Jul 17, 2012 at 7:28 PM, Peter Maydell <peter.maydell@linaro.org> wrote:
>> The trouble is that the warnings and errors here don't cause the
>> build to fail noisily; that's a big distinction IMHO.
>> I suppose we could make compile_prog do something like:
>> * run the compile test
>
> Unfortunately that would break cross compiling.
Sorry, that was slightly unclear phrasing. By "run the test"
I meant "do the check that we can compile the test code"
"run the binary produced". This series won't break cross
compiling.
-- PMM
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-trivial] [Qemu-devel] [PATCH] configure: fix ALSA configure test
2012-07-23 17:45 ` Peter Maydell
@ 2012-07-23 19:32 ` Peter Maydell
-1 siblings, 0 replies; 20+ messages in thread
From: Peter Maydell @ 2012-07-23 19:32 UTC (permalink / raw)
To: Blue Swirl
Cc: qemu-trivial, Stefan Weil, Igor Mitsyanko, QEMU Developers,
Anthony Liguori
On 23 July 2012 18:45, Peter Maydell <peter.maydell@linaro.org> wrote:
> On 23 July 2012 18:40, Blue Swirl <blauwirbel@gmail.com> wrote:
>> On Tue, Jul 17, 2012 at 7:28 PM, Peter Maydell <peter.maydell@linaro.org> wrote:
>>> The trouble is that the warnings and errors here don't cause the
>>> build to fail noisily; that's a big distinction IMHO.
>>> I suppose we could make compile_prog do something like:
>>> * run the compile test
>>
>> Unfortunately that would break cross compiling.
>
> Sorry, that was slightly unclear phrasing. By "run the test"
> I meant "do the check that we can compile the test code"
> "run the binary produced".
Gah. Missing 'not', should read:
# By "run the test" I meant
# "do the check that we can compile the test code", not
# "run the binary produced".
-- PMM
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [PATCH] configure: fix ALSA configure test
@ 2012-07-23 19:32 ` Peter Maydell
0 siblings, 0 replies; 20+ messages in thread
From: Peter Maydell @ 2012-07-23 19:32 UTC (permalink / raw)
To: Blue Swirl
Cc: qemu-trivial, Stefan Weil, Igor Mitsyanko, QEMU Developers,
Anthony Liguori
On 23 July 2012 18:45, Peter Maydell <peter.maydell@linaro.org> wrote:
> On 23 July 2012 18:40, Blue Swirl <blauwirbel@gmail.com> wrote:
>> On Tue, Jul 17, 2012 at 7:28 PM, Peter Maydell <peter.maydell@linaro.org> wrote:
>>> The trouble is that the warnings and errors here don't cause the
>>> build to fail noisily; that's a big distinction IMHO.
>>> I suppose we could make compile_prog do something like:
>>> * run the compile test
>>
>> Unfortunately that would break cross compiling.
>
> Sorry, that was slightly unclear phrasing. By "run the test"
> I meant "do the check that we can compile the test code"
> "run the binary produced".
Gah. Missing 'not', should read:
# By "run the test" I meant
# "do the check that we can compile the test code", not
# "run the binary produced".
-- PMM
^ permalink raw reply [flat|nested] 20+ messages in thread