1. The work to remove Arduino is progressing. I ended up making three patches directly for Arduino stuff, two are merged and the third is still in review. There's a 4th patch (not submitted yet) which is for the single-threaded support which is not directly Arduino, but was only used by Arduino. There are a few other areas which look like they exist only because of Arduino, I have not poked at whether anything could benefit by dropping that support, and don't intend to in the near term.
2. I didn't hear any interest in removing the currently unused (and unmaintained) remote-access stuff, so leaving that topic for now. One should not expect to actually use this code. 3. Mac (and cousin iOS) remains unmaintained, with an ongoing low-level interest in keeping and improving it. Patches here welcome; I'll try to help walk them through if I can (there's a simple one pending that turns on some iOS stuff). Mac/iOS remains very much non-mainstream for IoTivity at this time (just helped convince the OCF tools group that their proposed onboarding tool should not be required to run on iOS for this reason). 4. There is build support for MinGW and the msys target environment. I am not aware of whether anyone is actually using this, or whether it is even wanted. I have never attempted to build this, and we don't have a Jenkins builder for it, so it's possible (or rather: likely) that the build environment has rotted. Is this setup still of interest? If so we should build it occasionally to keep it working. 5. Cross-compiling doesn't really work. You can build for a target which is not the same as the host under certain specific circumstances: using the Android IDE, you can target Android in various configurations; using the Mac Xcode SDK you can target MacOS and iOS, subject to limitations already mentioned. These work, mostly (some features that would fail, such as regenerating the dat files because that happens using a tool generated during the build, don't run for those targets anyway). There are dodgy areas in some of the extlibs builds that I know of. Cross-building from say, linux-x86_64 to linux-arm, fail pretty convincingly. It's currently better to get an ARM box that runs Linux and is vaguley capable, and do a Linux build so host==target, then you can move the binaries to a smaller box. A 2gb-RAM Odroid works pretty well; a 1gb Raspberry Pi takes the better part of a day which is pretty painful. I'm not sure if this situation is actually a problem. If people think it is, I can put it on the list to think about - I'm trying to restrain myself and not work on everything "just because", then nothing actually gets done. _______________________________________________ iotivity-dev mailing list [email protected] https://lists.iotivity.org/mailman/listinfo/iotivity-dev
