On Sep 1, 2005, at 2:32 PM, Minty wrote:
# Without this line, undef gets passed to write_vint.
# The output is the same, but this thwarts an uninitialized
# warning and allows me to turn all warnings back on.
$ti->doc_freq(0) unless $ti->doc_freq;
I seem to recall coming across this. Might it be caused by a case of
if ($value) where $value = 0
I guess what I'm getting at, is why are undef's being passed in in the
first place? Correcting lower down the chain is a great idea and all,
but maybe it's worth looking further up the chain as well...
I'm shocked I pulled such a dumb maneuver, but... yeah, I think I was
about to hide a bug with that line. It's happens when the $self->
{last_ti} gets add()ed, since that's not initialized to anything.
I'm having trouble figuring it out, but if you want to nail it down
-- that would be appreciated. I have a gnarly buffering bug in
Convert::Lucene that's been biting me for hours...
Marvin Humphrey
Rectangular Research
http://www.rectangular.com/
|
Try Searching:
servers, voip, java, networking, microsoft ...
|
|
|
| 
