xloem

No, I expect I might not return to this project, so I wanted to leave the string code somewhere people could find it if needed. Personally it took a lot...

I looked through this first commit and it looks harmless and organised to my error-prone eye. It appears backwards-compatible regarding varargs use of options and use of no_device_discovery on android....

@harbulot do you have interest in verifying that this doesn't break older android code?

- i just tested this branch on android with https://github.com/harbulot/LibUsbAndroidIssue942 and it passed great - note that no_enumeration is still a static global in this branch, so it doesn't alone...

> @xloem, do you agree that although the situation for NO_DISCOVERY and other options is not satisfactory in current git without this API change (and #1003), it is not worse...

hi, android-dev is out of date, use the android branch but yes logs + debugging help please, it's all a work in progress

additional possible approaches: - passing options in a list structure to libusb_init - separating allocation of the usb context from initialisation so early options can be set with data -...

yeah. and I guess by initialisation I did mean just initial device discovery and likely hotplug setup I suppose; I don't remember the internal design well.

58ac16bd1e0f5f5f01a26c40720501d5f2974a30 is part of #874 . I have trouble thinking about the PR, and am unlikely to resolve that before the release. There may be other changes in the branch...

Thanks for so much thought you've both put into this quite minor issue. https://github.com/pothosware/SoapySDR/blob/4d46096c49071bb18e135036e9383d51ed2cd5fd/python3/CMakeLists.txt#L10 I'm guessing that `find_package(SWIG 2.0.11)` and also in the python2 subfolder, would prevent it recurring, and...