Re: thread safe patch for classfile.Utility and generic.Type

Hi Enver

Attached to this mail you will find
- a patch file based on CVS HEAD of tonight for classfile.Utility and
generic.Type
- a zip with complete source of those 2 classes

The patch is rather simple. I would really appreciate if this patch could be
added to the CVS. I don't care for @author or anything, just want it to work
(better) as it should.
Let me know if / when you plan a bcel-5-2 release or if I will have to take
a "cvs build" as soon as you commit it.

Note: the patched is based on CVS HEAD (HEAD differs from 5-1 only as
regards an email adress in @author, but anyway)

Enjoy
Alex



----- Original Message -----
From: "Enver Haase" <[email protected]>
To: <[email protected]>
Cc: "BCEL Developers List" <[email protected]>
Sent: Saturday, July 19, 2003 8:43 PM
Subject: Re: thread safe patch for classfile.Utility and generic.Type


>
> Just send the patch to the list.
>
> regards,
> Enver
>
>
> Alex wrote:
> > Hi
> >
> > I encountered problems in a multithreaded instrumentation layer I made
over BCEL. As soon as I have to thread it generates bad method signatures. I
read the following post
http://www.mail-archive.com/[email protected]/msg00201.html, which
warns about this issue.
> >
> > Indeed there is in classfile.Utility and generic.Type class only static
methods, but several methods are getting/setting the same static field
consumed_chars in several loops. When using BCEL in app server environment,
you always have concurent classloading in different classloader (proxy, skel
/ stub, CMP impl ...).
> >
> > Those 2 classes are really bad designed. Such an issue requires a
minimal effort to provide a patch and ensure thread safety.
> > I have made a patch based on a ThreadLocal wrapping of consumed_chars. I
would really appreciate if it could be incorporated in next release. Let me
know if/where/how I have to provide a diff for it instead of packaging my
own bcel-patch.jar with my app.
> >
> > Alex
> >
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]