Re: [ecasound] Ecasound build process stuck after tests

From: Kai Vehmanen <kvehmanen@email-addr-hidden>
Date: Sun Apr 03 2011 - 15:11:59 EEST


On Sat, 2 Apr 2011, Alessandro Ghedini wrote:

>> Ok, here's a snapshot with some changes:
> Unfortunately the problem is still there :/

d'oh, then this is getting even more interesting. :P

> Maybe a quick and dirty solution would be to simply not create the watchdog
> thread when it is not needed (e.g. for the various --version, --help and
> similar options). Not sure how much work would be needed (or if it will
> actually resolve the issue) though.

I'd like to first understand what's causing this. I mean, it's very likely
this could happen with other use-cases as well, so just whitelisting a few
trivial things like '--version', would just hide the bug.

For a dirty solution, I'd replace pthread_join() with pthread_cancel().
Thread cancelling has been a can of worms (at least in the past, with
legacy pthread implementations). But if basic signal blocking and delivery
mechanisms are not working for reason, things can't get much worse. But
hopefully this is simply be a bug in ecasound code.

>> PS Can you reproduce this only with stresstest.rb or perhaps
>> just by running "ecasound --version"...?
> Just running "ecasound --version" (or "--help") works as well. What I do is
> simply running the command several times until it gets stuck.

ok, great. Could you then patch ecasound/ecasound.cpp with:


... recompile and retest? You should now get traces on what is happening
when terminating.

Create and publish websites with WebMatrix
Use the most popular FREE web apps or write code yourself;
WebMatrix provides all the features you need to develop and
publish your website.
Ecasound-list mailing list
Received on Sun Apr 3 16:15:01 2011

This archive was generated by hypermail 2.1.8 : Sun Apr 03 2011 - 16:15:01 EEST