[FFmpeg-devel] Warnings (Was: Re: [PATCH 1/2] avutil/attributes: extend av_uninit to clang)

Ganesh Ajjanagadde gajjanag at mit.edu
Wed Oct 14 01:51:46 CEST 2015


On Tue, Oct 13, 2015 at 12:54 PM, Moritz Barsnick <barsnick at gmx.net> wrote:
> On Wed, Oct 07, 2015 at 23:35:08 +0200, Moritz Barsnick wrote:
>> On Wed, Oct 07, 2015 at 16:07:53 -0400, Ronald S. Bultje wrote:
>> > > BTW, are the "remaining" (meaning: more than from gcc) warnings emitted
>> > > by the Intel C(++) Compiler redundant or false? The icc configuration
>> > > should then probably silence them.
>> >
>> > I doubt anyone has ever investigated them...
>>
>> configure creates a quite explicit list of warning suppressions for icc
>> already. Somebody did investigate at one point. My icc is a bit old
>> (14.0.3.174 is the newest I have a license for), but I can take a stab
>> sometime, or ask here what the point of the warnings could be, if noone
>> minds.
>
> I don't even need to run my icc, fate has logs from icc builds with
> plenty of versions, such as:
> http://fate.ffmpeg.org/log.cgi?time=20151012035217&log=compile&slot=x86_64-linux-gnu-icc-2013_sp1.2.144
>
> Apart from the intentional "declared deprecated" warnings:
> - quite a few "deprecated conversion of string literal to char* (should
>   be const char*)", which are classified as warning #3179 and could be
>   suppressed (I think I once figured out why they are there - but can't
>   remember);
> - only two other warnings:
>   - /home/fate/fate/slots/x86_64-linux-gnu-icc-2013_sp1.2.144/src/libavcodec/pngenc.c(273): warning #2332: a value of type "const uint8_t={unsigned char} *" cannot be assigned to an entity of type "Bytef={Byte={unsigned char}} *" (dropping qualifiers)
>           s->zstream.next_in  = data;
>                               ^
>   - /home/fate/fate/slots/x86_64-linux-gnu-icc-2013_sp1.2.144/src/libavcodec/tscc.c(82): warning #2332: a value of type "const uint8_t={unsigned char} *" cannot be assigned to an entity of type "Bytef={Byte={unsigned char}} *" (dropping qualifiers)
>           c->zstream.next_in   = buf;
>                                ^
> - and these presumably new and as-yet unresolved unused return value warnings:
>   - /home/fate/fate/slots/x86_64-linux-gnu-icc-2013_sp1.2.144/src/ffmpeg.c(220): warning #2218: result of call is not used
>           av_buffersrc_add_frame_flags(ist->filters[i]->filter, frame,
>           ^
>
>     /home/fate/fate/slots/x86_64-linux-gnu-icc-2013_sp1.2.144/src/ffmpeg.c(295): warning #2218: result of call is not used
>           av_buffersrc_add_frame(ist->filters[i]->filter, NULL);
>           ^

the last two have been added by me via the av_warn_unused_result
attribute. I may fix it in a week if no one addresses it before then.

>
> So, nothing much in special. May need a short check.
> Moritz
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel


More information about the ffmpeg-devel mailing list