Bug: Signal processing in 1.5.2 and Python (readline, LinuxThreads etc.)

[Gregor Hoffleit]

Hi,

I just wanted to draw your attention to a bug report I submitted into the
Python BTS (Bug #196; currently available as
http://www.python.org/python-bugs/incoming?id=196;user=guest).

I describe that the combination of Python 1.5.2, LinuxThreads and readline
support has strange effects to the signal handling of the interpreter. I
tested various combinations of 1.5.1, 1.5.2, different glibc's, GNU Pth
instead of LinuxThread's pthreads and tried to find a pattern.

The most obvious observation of this problem certainly is that a Python
1.5.2 interpreter on Debian 2.2 systems will exit if you press Control-C on
the command line, while you would expect a KeyboardInterrupt exception.
Moving the readline module out of the way, *or* compiling the interpreter
without thread support or with Pth threads will cure this a little bit, but
still there are strange effects.

Perhaps you're interested in browsing the report and trying this with
different Linux and different Unix flavors. Please don't hesitate to send me
your results. Maybe somebody can give me a hint where to look for the bad
guy. Candidates seem to be LinuxThreads, GNU readline, the Python readline
module and the Python thread support, but I'm lost where to look for the
bug.

    Gregor