On 04/10/2013 03:29 PM, Todd Tannenbaum wrote:
On 4/10/2013 1:45 PM, Todd Tannenbaum wrote:
On 3/27/2013 8:49 PM, Matthew Farrellee wrote:
Certainly looks broken to me in 7.8 --
$ for f in /var/log/condor/NegotiatorLog /var/log/condor/MatchLog; do
tail -n1000 $f | md5sum ; done
c6b4e83a66a2cebf6c55b6454ad617d7 -
c6b4e83a66a2cebf6c55b6454ad617d7 -
Re the above -
Indeed this functionality appears broken in v7.8. However, it does work
properly in v7.9.4+, probably because this part of the logging code was
substantially re-written for enhancement reasons in the v7.9 series.
Given that v7.9 is very soon to become the next stable series, and that
very few people currently rely on this behavior, and other priorities, I
am inclined to spend resources fixing this bug in v7.8 since I know it
works properly moving forward. Folks who strongly feel otherwise, speak
up!
Doh! I meant to say "_not_ inclined to spend resource" above.
Thanks to the careful readers out there,
Todd
Does the enhancement effort add any tests, especially for this case, so
that we'll quickly identify if functionality regresses in the future?
Best,
matt
|
![[Computer Systems Lab]](http://www.cs.wisc.edu/pics/csl_logo.gif){kind=logo}