[ python-Bugs-971395 ] thread.name crashes interpreter

[SourceForge.net]

Bugs item #971395, was opened at 2004-06-11 16:15
Message generated for change (Comment added) made by tim_one
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=105470&aid=971395&group_id=5470

Category: Threads
Group: Python 2.3
>Status: Closed
Resolution: Works For Me
Priority: 5
Submitted By: Jonathan Ellis (ellisj)
Assigned to: Nobody/Anonymous (nobody)
Summary: thread.name crashes interpreter

Initial Comment:
I changed the __repr__ method of the cookbook Future
class --
http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/84317
-- as follows:

    def __repr__(self):
        return '<%s at %s:%s>' % (self.__T.name,
hex(id(self)), self.__status)

this caused obscure crashes with the uninformative message
        Fatal Python error: PyEval_SaveThread: NULL tstate

changing to __T.getName() fixed the crashing.

It seems to me that thread.name should be __name or
_name to help novices not shoot themselves in the foot.

----------------------------------------------------------------------

>Comment By: Tim Peters (tim_one)
Date: 2004-06-13 22:13

Message:
Logged In: YES 
user_id=31435

Well, in the absence of a test case, we can't even be sure 
this specific change was the true cause, so I'm closing this 
for now.  Feel encouraged to reopen it if more evidence 
appears.

FWIW, "NULL tstate" messages are historically almost all due 
to flawed logic in 3rd-party C extension modules.

----------------------------------------------------------------------

Comment By: Jonathan Ellis (ellisj)
Date: 2004-06-12 23:57

Message:
Logged In: YES 
user_id=657828

sorry, if I had been able to fine a simple crash case I
would have given it.  :-|

----------------------------------------------------------------------

Comment By: Tim Peters (tim_one)
Date: 2004-06-12 22:58

Message:
Logged In: YES 
user_id=31435

I've been unable to reproduce any problems with Future 
modified to use your __repr__, under Python 2.3.4 on 
Windows.  Of course any attempt to invoke that method dies 
with

AttributeError: 'Thread' object has no attribute 'name'

but that's not a bug.

Python won't change to add other ways of getting the thread 
name -- the Python philosophy is to supply one clear way to 
do a thing, and is opposed to Tower of Babel approaches that 
try to cater to everything someone might type off the top of 
their head.  There's no end to that (why not .Name 
and .thread_name and ... too?).

Would you kindly add specific code to this report, sufficient 
to reproduce the fatal Python error you mentioned?  If that 
occurs, it's a bug, but I've been unable to provoke it.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=105470&aid=971395&group_id=5470