On 02.02.2017 12:04, Nick Kew wrote: > On Wed, 2017-02-01 at 00:23 +0100, Stefan wrote: >> Hi, >> >> the issue was discovered as part of tracing down a deadlock condition in >> an SVN test [1]. >> >> As far as I understand it, the problem lies in the C-standard not >> explicitly defining when a function registered with atexit() is called >> in the context of thread termination [2]. > I had no idea atexit() was called on thread termination.
It's not. > I guess the > manpage must be misleading in telling us it happens at process exit? > >> atexit(apr_terminate); >> apr_initialize(); > That whole widely-used idiom looks deeply suspect if multiple > thread exits will each invoke the atexit! Are you sure this > isn't just a platform bug? In which case, a patch should perhaps > live in platform-specific code? That answer in SO is nether here nor there. In C (and C++), atexit() handlers are called when the main() function terminates; either duet to a call to exit(), or due to an implicit or explicit return from main(). And the "Linux" function _exit() is something else than the standard C exit() and is closer in semantics to standard C _Exit(). I don't have a recent edition of the C standard available; I do know that C90 has nothing to say about threads. The C++11 standard does describe what happens on thread termination and says nothing about atexit handlers being called. I'd be surprised if C++ behaviour was different from C in this case. -- Brane
