Re: findSystemClass

Date view Thread view Subject view Author view

From: Tim Wilkinson (tim@transvirtual.com)
Date: Thu Dec 10 1998 - 18:32:08 EST


Godmar,

Completely right - I should stop fixing the symptoms shouldn't I.

I've attached the install shield program so you can try it.

Cheers
Tim

> First of all, this is not the root of the problem.
> Second, if you remove it, you fail a test case: findSystemClass
> (which Sun's JDK passes.)
>
> Why don't you point me at InstallShield (and what you're trying to do
> with it) and I check whether it works with my tree?
>
> As an aside, it will still be wrong to hold the verification lock
> when calling out to loadClass, because a classloader thread may
> just go in an infinite loop. We'll attend to that problem later,
> though as it's not as critical.
>
> - Godmar
>
> >
> > If we no longer run the processClass method at the end of
> > java_lang_ClassLoader_findSystemClass0 then we do away with the verifier
> > re-entry (possible not for all apps but certinaly for such shings as
> > InstallSield). Now the question is - is this legal? I've checked the
> > docs on this and it's decidedly unhelpful as to whether findSystemClass
> > should run the classes clinit method (it says the class is loaded and
> > linked but that doesn't mean clinit is called does it?).
> >
> > Anyway, if you make this simple change InstallShield works fine - I've
> > not committed it yet though until people say what they think.
> >
> > Cheers
> > Tim
> >
> > --
> > Tim Wilkinson Tel: +1 510 704 1660
> > Transvirtual Technologies, Inc., Fax: +1 510 704 1893
> > Berkeley, CA, USA. Email: tim@transvirtual.com
> >
> >
> >
> >

--
  Tim Wilkinson                         Tel:     +1 510 704 1660
  Transvirtual Technologies, Inc.,      Fax:     +1 510 704 1893
  Berkeley, CA, USA.                    Email:   tim@transvirtual.com



Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Sat Sep 23 2000 - 19:57:11 EDT