Bugs item #1397671, was opened at 2006-01-05 12:14 Message generated for change (Comment added) made by rkindl You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=402868&aid=1397671&group_id=31650
Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Core Group: 0.85 Status: Open Resolution: Fixed Priority: 5 Submitted By: Robert Kindl (rkindl) Assigned to: Gert Driesen (drieseng) Summary: Memory leak Initial Comment: This is minimized sample that reproduces "memory leak" problem. dotest.build - sample file simulating some "real" work Test\default.build - using <nant> task - leaking (using about 200MB of memory at end of 1000th cycle) - reproduced on WXP, W2003 with .NET 1.1 SP1 - not reproduced on WXP with .NET 1.1 SP0 - not reproduced on W2000 with .NET 1.1 SP1 - reproduced on Nant RC2, Nant RC3 and Latest build from 20051213 TestCall\default.build - using <call> task is not leaking - much faster - good workaround, but still it's not same functionality for "real life" Conclusion: - heavy use of <nant> task is causing memory leaks and is much slower (possibly due to memory consumption) - strange is that it's not reproducible on "all platforms" PS: - I know that in .NET there are no "real" memory leaks - only "referenced objects" - So there probably are some "still referenced objects" that eats all that memory :-) ---------------------------------------------------------------------- >Comment By: Robert Kindl (rkindl) Date: 2006-03-19 16:56 Message: Logged In: YES user_id=1240287 I downloaded nightly build 20060319 which contains your changes. It is definitely better, thanks. I can confirm that new version is: - about 20x faster using <nant> task on "synthetic test" - It saves about 30 seconds from total 930 seconds on our "real life" build (we already replaced a lot of <nant> tasks with <call>) But there is still something wrong with memory consumption: - Nant definitely is not leaking memory so havily - before your fix it takes about 200MB for 1000 cycles - after the fix it takes about 25MB for 1000 cycles (or 250MB for 10000 cycles) - this leaks has same symptoms that I already described - they are not "reproducible on all platforms" (W2000 is OK while WXP and W2003 are affected) - this consumption of memory does not depend on "available free memory". I tried to run in Virtual PC with 128MB and 512MB and allocation was the same. I think that there is still something to investigate about this "leaks". ---------------------------------------------------------------------- Comment By: Gert Driesen (drieseng) Date: 2006-03-19 00:54 Message: Logged In: YES user_id=707851 You see much better release with the next nightly build. Also, speed of the <nant> task have been improved about 20x (times, not percentage LOL). The 1000 cycles now finish in about 12 seconds on my system (which isn't really state of the art). Let me know if you see similar results, so we can close this bug report. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=402868&aid=1397671&group_id=31650 ------------------------------------------------------- This SF.Net email is sponsored by xPML, a groundbreaking scripting language that extends applications into web and mobile media. Attend the live webcast and join the prime developer group breaking into this new coding territory! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642 _______________________________________________ nant-developers mailing list nant-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nant-developers
