> I'd suggest tagging CVS and making it clear that java gump may deviate > from documentation, etc but permit changes to the code.
We could take the opportunity to: 1) Move Gump (Python) core to SVN 2) Move Gump Metadata to a separate CVS repository [to move to SVN one day in the future, when all ASF committer are "comfortable" w/ SVN]. We could leave the current repository as is, tagged but even CVS HEAD stable, for automated Gumps out there that depend upon things there. > My main concern > at the moment is that python gump seems a /lot/ slower than java gump on > my under-powered, over-worked Sun Ultra 5. On that machine a Java gump > run would take in the order of 19 hours. I started a Python gump run on > the 6th June at 2.33pm local time and it still isn't finished (as of the > 8th June at 5.05pm). Performance has become my bugbear w/ this stuff. I just don't get the language well enough to be able to crack it. I've done profiling, GC inspection, reference inspection, and I don't get it. Some things do grow (although I can see no leak) but I can't imagine that too much memory is consumed. I have work to do here. [BTW: The branch I currently have has a lot of __del__ work done, to try to ensure no circular references. Python seems to really need help with those.] FWIIW: I find small (50 project) workspaces zip by, but large ones (500+) seem to creep. BTW: How many projects do you have in this workspace? What is the purpose of it? Do you "trim/tailor" the project sequence by passing "myproj*" project masks? BTW: Do you use forrest batch, or not forrest xdocs? regards Adam --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
