You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by Hans Gilde <hg...@earthlink.net> on 2004/12/19 23:26:23 UTC
RE: Memory-leak test seems to go through
Paul, your activemath.org account isn't working.
This looks good for the memory leak but... if we're going to change the API,
why not just store the cache in the JellyContext and be done with it?
-----Original Message-----
From: Paul Libbrecht [mailto:paul@activemath.org]
Sent: Friday, December 17, 2004 4:11 PM
To: Hans Gilde
Subject: Memory-leak test seems to go through
Hey Hans,
This mail will arrive probably too late as I am offline now but...
anyways... just to let you know that my TagHolderMap thread-local seems
to succeed with the memory-leak test that you have proposed. I'd like
it if you can confirm it.
Careful, I had to explicitly add
org.apache.commons.jelly.impl.TagScript.clearTagHolderMap(); at line
118 of BaseMemoryLeak until I find the good places to put this
clearing.
I don't like to put it even in the parser's returned ScriptBlock's run
as I anticipate this reference will play the role of the session within
some more elaborate setting (like my reloadable tags)... advice is
welcome there.
(in a sense, an approach would be to have a trace of when the
tagReferenceMap was created so that it becomes the place when it's
removed...).
paul
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org
Re: Memory-leak test seems to go through
Posted by Paul Libbrecht <pa...@dfki.de>.
Le 19 déc. 04, à 23:26, Hans Gilde a écrit :
> This looks good for the memory leak but... if we're going to change
> the API, why not just store the cache in the JellyContext and be done
> with it?
I would have nothing against it... I think there's no way to avoid
changing the API as things are really pulled from everywhere.
Is there a patch for the jelly-context already ?
(and do I understand correctly that, anyways, the modification would
only be enrichment of the API ?).
paul
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org