final thoughts - bug tracking system

David Cuthbert dacut at
Sat Sep 17 16:27:28 PDT 2005

Matthew Dillon wrote:
    I am certainly not going to run something that
    takes a lot of effort to upgrade... and believe me, ANYTHING based on
    java takes a lot of effort to ugprade if the vendor starts relying on
    anything other then basic java constructs.  I have enough problems
    just using Java, because of SUN's braindamaged policies.  I do not
    want java in our core, production infrastructure.
I have to agree here... I've been burned by Java so many times that I 
just won't touch it.  Trying to get component X, which requires 
1.4.2_05, to run with component Y, which requires 1.4.2_04, can be a 
nightmare.  If it's not that, it's some bug or performance issue in the 
JRE, which "will be fixed in the next version, due out RSN."

One bit that I do think we'd be missing out, though, by going the DIY 
way is timeliness.  I think that DragonFly needs a bug tracking system 
yesterday, even if it's just someone updating a text file every night 
and posting it on the Wiki.  (That said, I'd hope that this isn't the 
approach taken, and I'm definitely not volunteering for the job. :-)

If you're serious about going the DIY route (and I have no doubt that 
you are), Matt, then I'd recommend putting something simple in place now 
with the idea of migrating the data over and/or maintaining it in 
parallel.  Polish this off, and this would give MSBTS (Matt's Spiffy Bug 
Tracking System) the added feature of importing data from foreign 
sources (which, having been through three bug tracking migrations at my 
last job, can be a Big Deal).

More information about the Kernel mailing list