On 15 March 2017 at 23:54, Evgeny Grin <k...@yandex.ru> wrote: > That's why I asked about user code example. I can't test what we trying > to fix.
I understand you wish clearly. I will provide the code tomorrow. > Did you test it or only review? I am always do test the code, because there are might be regressions or new bugs :) > Seems that MHD read data from network after you suspended connection, > which cause overflow buffer overflow and error response. > You issue should be fixed by afb221c32aeeb2731e3b5362cc4ff8949d34edea. I did not tested this commit yet. > Disconnect by timeout for suspended connections should be fixed, so you > can stop playing with timeout values. Good. But I see that the root of the problem is not solved. I told already about this check: timeout <= (MHD_monotonic_sec_counter() - connection->last_activity) There is must be call inside the function MHD_resume_connection() which will do next (or similar): connection->last_activity = MHD_monotonic_sec_counter(); Anyway, I will test latest patches and either raise this suggestion again or do not. > You mentioned that you had some problems in master with epoll and > timeouts. Is it still actual? epoll is not actual. I have solved most (see below) issues with epoll. There are other issues with timeouts raised, I will explain them after testing latest patches (or I will not if all is ok). At least, overall work of MHD is much better than was before. Thanks! -- With Best Regards, Vitaliy V. Tokarev
