Quoting Lynne (2023-05-28 04:48:00) > May 28, 2023, 03:07 by jamr...@gmail.com: > > > On 5/24/2023 9:32 PM, Lynne wrote: > > > >> May 24, 2023, 23:24 by leo.i...@gmail.com: > >> > >>> On 5/24/23 16:35, Lynne wrote: > >>> > >>>> Patch attached. > >>>> > >>> > >>> + av_log((void *)&tx_class, AV_LOG_DEBUG, "%s\n", bp.str); > >>> > >>> The type of the first argument to av_log should be AVClass **, but this > >>> only appears to be AVClass *. See libavutil/log.c line 428. > >>> > >>> - Leo Izen > >>> > >> > >> Right, thanks, changed to: > >> > >>> static const AVClass tx_class = { > >>> .class_name = "tx", > >>> .item_name = av_default_item_name, > >>> .version = LIBAVUTIL_VERSION_INT, > >>> }; > >>> > >>> static const struct { > >>> const AVClass *tx_class; > >>> } tx_log = { > >>> &tx_class, > >>> }; > >>> > >> Will push this tomorrow. > >> > > > > Can't add an AVClass* field to AVTXContext and set it to &tx_class during > > init? > > > > The struct is accessed from asm, didn't really want to fix all the loads > for something which only runs at init, and only if !CONFIG_SMALL.
That's largely missing the point of having a non-NULL logging context, which is that the log callback has access to the context. It's private in this case, but AVOptions could still be used to retrieve things like per-context opaque. There have also been discussions recently about a per-context logging callback being highly useful for some cases. -- Anton Khirnov _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".