Thanks for this information!
This could be a Java implementation issue, or a Java implementation
issue specific to Solaris, or it could even have something to do with
the window manager you're using.
Minimize/maximize has no significant effect on memory usage here,
running Linux 2.5.44 with Sawfish and Blackdown 1.4.1 beta.
I just had an active j session running for about two days. Just before
I finally exited, top reported 237 MB VIRT, 72 MB RES, 50 MB SHR. My
current session has been up for about 2 hours, and top reports 237 MB
VIRT, 59 MB RES, 50 MB SHR.
This suggests that the resident size does tend to increase over time,
but I think in my case this is mostly because I've been using and
killing a lot of Lisp shells inside j that may not be cleaning
themselves up exactly right (this is a new feature isn't fully debugged
yet).
If you're just using j to edit normal files, some data structures that
keep track of history do tend to get bigger as time goes by, but that
shouldn't amount to very much. At some point I'd like to audit all of
that code and put quotas on everything, so that j could run for months
at a time with constant memory usage, but even now, just from a memory-
leak point of view, I think a j session could plausibly last from one
release to the next, if you don't get too frisky.
You're more likely to restart because of some other bug than run out of
memory... ;)
-Peter
> Muy intersante. Not sure if you know about this --
> minimize/maximize window affects memory usage. This may be
> a Java2 implemenation issue.
>
> I have the J session up for an hour, memory incr from 60mb
> to 85mb. I minimize/maximize J window three times in a
> minute. "top" reports memory usage as 100mb / 70mb (size /
> res ).
>
> Cmd line: ./j2re1.4.0_01/bin/java -Xmx16m -jar
> ./j0.17/j.jar
>
> Sounds like a "repaint" or "paint" issue.
>
> ----------------------------------------------------------------------
>
> Comment By: Marc Ubaldino (ubaldino)
> Date: 2002-12-19 13:57
>
> Message:
> Logged In: YES
> user_id=389485
>
> Fyi, More data points. Running J for about 45 minutes. Even
> with limiting heap to 64m, size still grows... session was
> idle until I actually started to use the session about 40
> minutes after invoking it. Again X11 session, remote side
> sol-7, locally sol-8.
>
> PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU
> COMMAND
> 16905 ubaldino 19 58 0 218M 191M sleep 0:42 2.59%
> java
|