Re: [FFmpeg-devel] [PATCH] avformat/libsrt: Decode URL parameter strings

2023-08-11 Thread Armin Hasitzka
I think accepting URL params in first place isn't inherently bad (depending on the use case, it might make certain interactions easier). I expected regressions to be brought up and just want to point out that, while absolutely possible, actual regressions will likely be fairly rare: - any "un

Re: [FFmpeg-devel] [PATCH] avformat/libsrt: Decode URL parameter strings

2023-08-10 Thread zhilizhao(赵志立)
> On Aug 11, 2023, at 00:13, Armin Hasitzka wrote: > > Hi again, > > we found this the other day when using a stream ID containing "%2F", > expecting this to be resolved to "/". While "%2F" could obviously be sent > decoded, "&" (decoded) would currently end the value and not be used, "+" > (d

[FFmpeg-devel] [PATCH] avformat/libsrt: Decode URL parameter strings

2023-08-10 Thread Armin Hasitzka
Hi again, we found this the other day when using a stream ID containing "%2F", expecting this to be resolved to "/". While "%2F" could obviously be sent decoded, "&" (decoded) would currently end the value and not be used, "+" (decoded) would be overwritten with " ", and "=" (decoded) could cause