Big-endian platforms (usually running OpenWrt) have a known issue in RTMP code. ARUT has started to deal with it in branch "big-endian" but never finished it. His changes never got into master. There are at least 3 open issues in his repo (#324 #576 #739) that describe the problem. I can confirm that package nginx-all-module in Openwrt 19.07.2 uses your source code (ut0mt8/nginx-rtmp-module) and still has the problem on MIPS. Log from nginx:
I create this Issue to document current status. Suggested course of action is to review rtmp code and make it endian-agnostic. Unfortunately I have no time to do that now. It's a pity ARUT hasn't started using NTOH/HTON from the beginning.
Is the DASH implementation more robust (eg. is it tested on big-endian machines)?
Big-endian platforms (usually running OpenWrt) have a known issue in RTMP code. ARUT has started to deal with it in branch "big-endian" but never finished it. His changes never got into master. There are at least 3 open issues in his repo (#324 #576 #739) that describe the problem. I can confirm that package nginx-all-module in Openwrt 19.07.2 uses your source code (ut0mt8/nginx-rtmp-module) and still has the problem on MIPS. Log from nginx:
I create this Issue to document current status. Suggested course of action is to review rtmp code and make it endian-agnostic. Unfortunately I have no time to do that now. It's a pity ARUT hasn't started using NTOH/HTON from the beginning.
Is the DASH implementation more robust (eg. is it tested on big-endian machines)?