Thiago Macieira wrote: Hi,
> of the parent. The fact that it's deadlocking in the code that is supposed to > remove locks so that it won't deadlock is a good irony. Yeah, very funny until it happens to you :) > > Absent further proof to the contrary, this looks like a system bug. You could be right, but how on earth do I end up in tcmalloc code here? In my experiment I'm trying to store ints (index numbers used as the key) as 4- byte strings in QByteArrays which get converted to std::strings for use with leveldb. I half expect that I'm overlooking a place where this goes wrong, but also that that would lead to more variable crash backtraces. I use Chrome as a browser, which must use leveldb internally, so if this were a system bug I'd expect to see deadlocks while surfing. That just doesn't happen, which probably means we can at least not factor Qt out of the equation. R. _______________________________________________ Interest mailing list Interest@qt-project.org http://lists.qt-project.org/mailman/listinfo/interest