[vlc-devel] [PATCH 2/3] configure: add -Werror=incompatible-pointer-types option to the compiler
Rémi Denis-Courmont
remi at remlab.net
Wed Dec 18 11:00:29 CET 2019
As was pointed out earlier, missing function prototypes always lead to bugs, as parameters are down-converted to int. Now it's even worse because it also breaks CFI.
But missing initialisers are not always or almost always a bug. Sometimes they are and sometimes, they're not.
Le 18 décembre 2019 11:27:57 GMT+02:00, Steve Lhomme <robux4 at ycbcr.xyz> a écrit :
>On 2019-12-18 10:21, Rémi Denis-Courmont wrote:
>> We've already had this discussion, and this kind of change was
>already
>> rejected.
>>
>> Warnings as errors are screwing anybody using a configuration
>different
>> from CI's. Packagers have consistently rejected warning as errors.
>
>That's why an option to reject dirty code during development could be
>useful. I guess we could also add the defines in each target of the CI,
>
>it would have the same effect of rejecting dubious code.
>
>> By your argument, basically all warnings should be errors. You're
>> basically denying the concept of warnings.
>
>You're generalizing my proposal to say I'm against warnings in general.
>
>We have -Werror-implicit-function-declaration already in configure.ac.
>Others could be added. It doesn't mean we want to treat all warnings as
>
>errors. I certainly don't.
>
>> Le 18 décembre 2019 11:16:24 GMT+02:00, Steve Lhomme
><robux4 at ycbcr.xyz>
>> a écrit :
>>
>> On 2019-12-18 10:03, Rémi Denis-Courmont wrote:
>>
>> That's exactly what's wrong. They're warnings and they should
>> stay that way.
>>
>>
>> I disagree.
>>
>> You're welcome to attend to them.
>>
>>
>> I can't compile on macOS, iOS or Android. Should I check the full
>logs
>> of each MR I'm doing on Gitlab to see if somewhere there's a
>change I
>> did that is doing something bad ? Will you ?
>>
>> Le 18 décembre 2019 10:13:42 GMT+02:00, Steve Lhomme
>> <robux4 at ycbcr.xyz>
>> a écrit :
>>
>> On 2019-12-17 18:20, Rémi Denis-Courmont wrote:
>>
>> Le tiistaina 17. joulukuuta 2019, 16.09.28 EET Steve Lhomme a
>> écrit :
>>
>> This avoid using pointers of different types. It may lead to
>> serious issues
>> in the code so rather than a warning, it should be
>> avoided/fixed.
>>
>> This will raise warnings in harmless cases of implicit
>> conversions. Some of
>>
>>
>> It's not adding any warning. It's turning existing
>(unattended)
>> warnings
>> into errors.
>>
>> those might not be easily fixable - coming from external
>headers.
>>
>>
>> True, but it seems we have none of these on Windows, Linux
>(and
>> macOS
>> I'm told). I'd rather not allow new code using dirty pointer
>> casts, even
>> from external library headers.
>>
>------------------------------------------------------------------------
>> vlc-devel mailing list
>> To unsubscribe or modify your subscription options:
>> https://mailman.videolan.org/listinfo/vlc-devel
>>
>>
>> --
>> Envoyé de mon appareil Android avec Courriel K-9 Mail.
>Veuillez
>> excuser
>> ma brièveté.
>>
>>
>------------------------------------------------------------------------
>> vlc-devel mailing list
>> To unsubscribe or modify your subscription options:
>> https://mailman.videolan.org/listinfo/vlc-devel
>>
>>
>> --
>> Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez
>excuser
>> ma brièveté.
>>
>> _______________________________________________
>> vlc-devel mailing list
>> To unsubscribe or modify your subscription options:
>> https://mailman.videolan.org/listinfo/vlc-devel
>>
>_______________________________________________
>vlc-devel mailing list
>To unsubscribe or modify your subscription options:
>https://mailman.videolan.org/listinfo/vlc-devel
--
Envoyé de mon appareil Android avec Courriel K-9 Mail. Veuillez excuser ma brièveté.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/vlc-devel/attachments/20191218/bf833ee1/attachment.html>
More information about the vlc-devel
mailing list