| Michael Witten: | | > -const char *avio_enum_protocols(void **opaque, int output) | > +const char *avio_enum_protocols(void **const opaque, const int output) | | Andreas Rheinhardt: | | > This thing makes nothing more const-correct at all; C uses call be | > value, so we only deal with our own copy of opaque which we may modify | > as we please. | > The reason for the const-incorrectness lies in the fact that this | > function makes *opaque point to something const (namely an entry in a | > const list of pointers (each pointing to a const protocol, but that is | > irrelevant)), so that the user now has a pointer whose pointed to-type | > is not const-qualified, despite the actual target being const. | > See here: | > https://patchwork.ffmpeg.org/project/ffmpeg/patch/20190821090438.10260-2-andreas.rheinha...@gmail.com/ | > for how to fix the const-correctness if one wants to keep the general | > design of the functions. See libavformat/allformats.c, | > libavcodec/allcodecs.c, libavfilter/allfilters.c for the other approach | > (that involves storing an index in the pointer). | > | > - Andreas
* The general design of the functions has not been altered by this series. * The previous patch fixed the const incorrectness as warned by the compiler. This patch merely improves the const correctness in general. * We don't want to modify those parameters as we please, and so we'd like to ask the compiler to remind us of that fact as necessary. It's const-correct in the sense that it expresses the following fact: Those variables are not going to be changed; if there is no intention to modify a variable, then it's good practice to mark that variable as const. Because this function is being edited, it's a good time to sprinkle the 'const' keyword around, as a matter of mental hygiene. Sincerely, Michael Witten _______________________________________________ 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".
