Article View: comp.lang.java.advocacy
Article #150Re: Java memory
From: Jordi Guitart
Date: Fri, 23 Feb 2001 17:08
Date: Fri, 23 Feb 2001 17:08
40 lines
3385 bytes
3385 bytes
Hi, You are right when you talk about all the problems I can found when replacing threads library. But now I'm in the final phase of the replacement (I think so), and only with a special Java application I sometimes have memory problems (some constants of the program become 0 without any reason). "Joe Seigh" <jseigh@genuity.com> wrote in message news:3A967BA3.C506833B@genuity.com... > Jordi Guitart wrote: > > > > I'm only modifying a little part of the JVM (in fact, I'm not "modifying" > > anything, I'm only replacing threads library with another one, and > > unfortunately I have lateral memory problems, so I'm trying to discover > > why?) > > > > JGF > > That's probably not a good idea as you've found out. Java has rather > specific semantics. The various thread libraries have slightly different > semantics, even if they're POSIX pthreads. There's a number of places you > could have problems. One or more of the following could apply. > > 1. The jvm implementor's don't understand java/jvm semantics correctly. > 2. The jvm implementor's did not implement java/jvm semantics correctly. > 3. The two thread libraries have different semantics and so they're not > really interchangeable. > 4. The jvm implementor's are relying on implementation side effects of > the first library that aren't in the second library. e.g. the > first library has FIFO locks and the second does not. Another way > of saying this is they're relying on undocumented behavior in > the threads library version they are using. > > So there's a lot of possibilities here, very subtle ones. Are you > sure you want to get into this? > > Joe Seigh > > Joe Seigh
Technical Details
Message-ID:
<9765hi$942$1@defalla.upc.es>Path:
arc1-ca.pugleaf.net!supernews.google.com!sn-xit-03!supernews.com!logbridge.uoregon.edu!newsfeed.skycache.com!Cidera!news!news.rediris.es!news.cesca.es!news.upc.es!not-for-mailReferences:
<96trij$4ns$1@defalla.upc.es> <3A93B440.F8232537@evtechnology.com> <970ehj$337$1@defalla.upc.es> <3A93C9D3.6060706@evtechnology.com> <970j55$6af$1@defalla.upc.es> <3A967BA3.C506833B@genuity.com>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 10568c,3df73e04f851a6c7 X-Google-Attributes: gid10568c,public X-Google-Thread: fdb77,3df73e04f851a6c7 X-Google-Attributes: gidfdb77,public X-Google-Thread: 10bb9b,3df73e04f851a6c7 X-Google-Attributes: gid10bb9b,public X-Google-Thread: 113ae9,3df73e04f851a6c7 X-Google-Attributes: gid113ae9,public X-Google-Thread: 10a146,3df73e04f851a6c7 X-Google-Attributes: gid10a146,public X-Google-Thread: 1116f3,3df73e04f851a6c7 X-Google-Attributes: gid1116f3,public X-Google-ArrivalTime: 2001-02-23 09:10:10 PST Path: arc1-ca.pugleaf.net!supernews.google.com!sn-xit-03!supernews.com!logbridge.uoregon.edu!newsfeed.skycache.com!Cidera!news!news.rediris.es!news.cesca.es!news.upc.es!not-for-mail From: "Jordi Guitart" <jguitart@ac.upc.es> Newsgroups: comp.lang.java.advocacy,comp.lang.java.help,comp.lang.java.machine,comp.lang.java.programmer,de.comp.lang.java,es.comp.lenguajes.java Subject: Re: Java memory Date: Fri, 23 Feb 2001 18:08:52 +0100 Organization: Universitat Politecnica de Catalunya Lines: 42 Distribution: es Message-ID: <9765hi$942$1@defalla.upc.es> References: <96trij$4ns$1@defalla.upc.es> <3A93B440.F8232537@evtechnology.com> <970ehj$337$1@defalla.upc.es> <3A93C9D3.6060706@evtechnology.com> <970j55$6af$1@defalla.upc.es> <3A967BA3.C506833B@genuity.com> NNTP-Posting-Host: amat.ac.upc.es X-Trace: defalla.upc.es 982948210 9346 147.83.31.105 (23 Feb 2001 17:10:10 GMT) X-Complaints-To: newsmngr@upc.es NNTP-Posting-Date: 23 Feb 2001 17:10:10 GMT X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400