I gathered up some courage and ran the regression tests with the
Electric Fence library.
A quick overview of the queries that failed follows. (In the following a
is a test where the backend dies.) This usually means electric fence
detected some overrun.
one fancy select statement
If we are lucky these are all cause by the same bug -:). Anybody want
to find out for sure? If you are interested in the regression.diff file I
post it to the list.
pgsql-hackers by date
|Next:||From: Andrew Martin||Date: 1998-03-26 18:30:36|
|Subject: Re: [HACKERS] URGENT PROBLEM with the 6.3->6.3.1 patch|
|Previous:||From: Maurice Gittens||Date: 1998-03-26 16:46:43|
|Subject: Re: [HACKERS] patch for memory overrun on Linux(i386)|