Re: [FFmpeg-devel] [PATCH v2] Let clang-FORTIFY build; NFC.

2016-09-04 Thread Carl Eugen Hoyos
Hi! 2016-08-31 0:49 GMT+02:00 George Burgess IV : > +++ b/libavfilter/avf_aphasemeter.c > @@ -190,7 +190,7 @@ static int filter_frame(AVFilterLink *inlink, AVFrame *in) > > metadata = avpriv_frame_get_metadatap(out); > if (metadata) { > -uint8_t value[128]; > +char value[

Re: [FFmpeg-devel] [PATCH v2] Let clang-FORTIFY build; NFC.

2016-08-31 Thread George Burgess
> I'm blindly assuming the signedness isn't relevant here That's actually what matters here, sadly. The only way we've been able to get FORTIFY on clang to work nearly as well as FORTIFY on gcc is with overloading (via clang's __attribute__((overloadable)) ). This attribute is specced to use C++'s

Re: [FFmpeg-devel] [PATCH v2] Let clang-FORTIFY build; NFC.

2016-08-31 Thread Michael Niedermayer
On Wed, Aug 31, 2016 at 04:24:36AM -0400, Ronald S. Bultje wrote: > Hi George, > > On Tue, Aug 30, 2016 at 8:47 PM, George Burgess wrote: > > > Thanks for the feedback! I agree the casts aren't pretty. :) > > > > > Isn't it easier to change your fortify-clang and add a compiler option > > to dis

Re: [FFmpeg-devel] [PATCH v2] Let clang-FORTIFY build; NFC.

2016-08-31 Thread Ronald S. Bultje
Hi George, On Tue, Aug 30, 2016 at 8:47 PM, George Burgess wrote: > Thanks for the feedback! I agree the casts aren't pretty. :) > > > Isn't it easier to change your fortify-clang and add a compiler option > to disable this specific error for specific targets? > > The short answer is "in some ca

Re: [FFmpeg-devel] [PATCH v2] Let clang-FORTIFY build; NFC.

2016-08-30 Thread George Burgess
Thanks for the feedback! I agree the casts aren't pretty. :) > Isn't it easier to change your fortify-clang and add a compiler option to disable this specific error for specific targets? The short answer is "in some cases, yes. Sadly, this doesn't seem to be one of those cases." The longer answe

Re: [FFmpeg-devel] [PATCH v2] Let clang-FORTIFY build; NFC.

2016-08-30 Thread Ronald S. Bultje
Hi, On Tue, Aug 30, 2016 at 6:49 PM, George Burgess IV wrote: > ChromeOS is adopting a new FORTIFY implementation tailored for clang. As > an artifact of how this new FORTIFY is implemented, a handful of > implicit conversion warnings get turned into errors. This patch fixes > the implicit conve

[FFmpeg-devel] [PATCH v2] Let clang-FORTIFY build; NFC.

2016-08-30 Thread George Burgess IV
ChromeOS is adopting a new FORTIFY implementation tailored for clang. As an artifact of how this new FORTIFY is implemented, a handful of implicit conversion warnings get turned into errors. This patch fixes the implicit conversions in ffmpeg that clang-FORTIFY has an issue with. Signed-off-by: Ge