Hi,
I've just committed two very simple patches to copy.c and
trigger.c which caused backend to grow until transaction end.
trigger.c didn't expected that trigger function could have
returned another heap tuple that was built inside of trigger
with SPI_copytuple().
In copy.c I'n not absolutely sure why it was as it was. In
CopyFrom() the array for the values was palloc()'d once
before entering the copy loop, and then again at the top of
the loop. But there was only one pfree() after loop exited.
I've removed the palloc() inside the loop. Seems to work for
the regression test. Telling here only for the case someone
encounters problems on COPY FROM.
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#======================================== jwieck@debis.com (Jan Wieck) #
