Follow qTox/qTox#5127.
I previously patched golang to let my tox bot work months ago, and now it's qtox. They all crashed due to the small stack size limit of musl-libc(80K).
Since there're already two applications having crashed, i think we should have a discussion: should the caller of the setstacksize() be applications, or the toxcore library?
I prefer patching library, since i just have 28 friends, plus the nodes info and so, it's so small a scale, but it crashed. It may be very easy to touch the stack size limit of musl, and patching library will make things much more convenient.
As for the size, 512K should be big enough, as i said in qTox/qTox#5127. But i think we need to think more about it, estimating the stack size generally cost by the library. Anybody have idea on how to do this?
UPDATE: qTox/qTox#5127 (comment). sudden6 pointed out that this maybe a bug or an oversight in toxcore or qtox, because toxcore should not take so much stack in his opinion.
Follow qTox/qTox#5127.
I previously patched golang to let my tox bot work months ago, and now it's qtox. They all crashed due to the small stack size limit of musl-libc(80K).
Since there're already two applications having crashed, i think we should have a discussion: should the caller of the
setstacksize()be applications, or the toxcore library?I prefer patching library, since i just have 28 friends, plus the nodes info and so, it's so small a scale, but it crashed. It may be very easy to touch the stack size limit of musl, and patching library will make things much more convenient.
As for the size, 512K should be big enough, as i said in qTox/qTox#5127. But i think we need to think more about it, estimating the stack size generally cost by the library. Anybody have idea on how to do this?
UPDATE: qTox/qTox#5127 (comment). sudden6 pointed out that this maybe a bug or an oversight in toxcore or qtox, because toxcore should not take so much stack in his opinion.