[Python-Dev] Worse news

Tim Peters tim.one@home.com
Mon, 22 Jan 2001 16:07:27 -0500

[Jeremy Hylton]
> Funny (strange or haha?) that test_extcall is failing since the two
> pieces of code I've modified most recently are compile.c and the
> section of ceval.c that handles extended call syntax.

Ya, I knew that, but I avoided wagging a Finger of Shame in your direction
because coincidence isn't proof <wink>.

> ...
> As for the test_sax failure,

There is no test_sax failure anywhere anymore that I know of (Martin found a
dead-wrong array decl in contributed pyexpat.c code and repaired it).

And I believe my "rt -x test_sax" failure in test_extcall almost certainly
has nothing to do with test_sax -- far more likely the connection to
test_sax is an accident, and that if I spend umpteen hours trying other
things at random I'll provoke the same memory accident leading to a bad
pointer via excluding some other test.  I just picked test_sax because that
*was* broken and I wanted to get thru the rest of the tests.

BTW, delighted(?) to hear that test_cpickle fails for you too!  I'm sure
test_extcall is going to blow up for other people eventually too -- but it
is sooooo hard to provoke even for me.  I've dropped the effort pending news
from someone running Insure++ or efence or whatever.