BTW, I have been able to deduce from my own usage patterns that most of the leak seems to take place in the jelly ant task wrapper tag...for whatever it's worth, this is always where the memory disappears never to reappear again...
-j On Fri, 2003-12-12 at 17:39, John Casey wrote: > I have scanned the bugs listed in jira, looking for the memory leak I'm > experiencing when I run a reactor-ized multiproject:xxx build. > > What I find are several bugs which talk about various memory leaks, and > are marked fixed (most in the relatively distant past), and some more > which describe the "fix" as simply increasing the memory allocated to > the JVM. While I'm sure this is a valid workaround for most, I don't see > it as any sort of fix; it's like the old reaction of throwing > money/hardware at a problem in the hopes of [buying time] making it go > away. > > I've seen some things float by on the dev list about whether or not to > fix "the" memory leak before 1.0 releases (I see this as vital to > distinguishing maven as more than a novelty project), and I'm wondering > if there is somewhere I can go for a summation of the leak's symptoms > and current diagnoses. You see, my company uses maven extensively, and > has ~30 projects which currently will not build in a single sweep with > multiproject, due to lack of memory...regardless of my adjustment up to > -Xmx512M. > > I'm willing to donate another set of eyes to this problem, in the > interest of making my life easier at work, but I don't want to spend too > much time retracing others' steps - I feel this is a waste of time. > > Can anyone offer any advice? > > Thanks, > john > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
