Hi,
On Thu, 2005-10-20 at 11:11 -0600, Tom Tromey wrote:
> In a case like this the issue is that someone modifying the code won't
> know what application to test. I sometimes add comments about
> particular applications, e.g. see the note about Eclipse in
> ResourceBundle. I don't think there's
> "Mark" == Mark Wielaard <[EMAIL PROTECTED]> writes:
Mark> Yeah, yeah... :) I know that it doesn't beat having real comments for
Mark> code. But I really do try running real known good applications before we
Mark> make any release to check that nothing is broken.
In a case like this the issu
On Wed, 2005-10-19 at 18:09 +0100, Andrew Haley wrote:
> Mark Wielaard writes:
> > On Wed, 2005-10-19 at 17:36 +0100, Andrew Haley wrote:
> > > You really shouldn't consider committing a patch like this without
> > > adding comments. After all, what is to stop someone changing it back?
> >
>
Mark Wielaard writes:
>
> Hi Andrew,
>
> On Wed, 2005-10-19 at 17:36 +0100, Andrew Haley wrote:
> > > 2005-10-19 Mark Wielaard <[EMAIL PROTECTED]>
> > >=20
> > > * java/net/URLClassLoader.java (addURLs): Don't call addURL(), but
> > > call urls.add() and addURLImpl() directly on
Hi Andrew,
On Wed, 2005-10-19 at 17:36 +0100, Andrew Haley wrote:
> > 2005-10-19 Mark Wielaard <[EMAIL PROTECTED]>
> >
> > * java/net/URLClassLoader.java (addURLs): Don't call addURL(), but
> > call urls.add() and addURLImpl() directly on each URL.
> >
> > Andrew, I know you hacked on
Mark Wielaard writes:
> This is the last piece needed to get JBoss (4.0.3) to start up. JBoss
> has some ClassLoaders that subclass URLClassLoader and override the
> protected addURL() method. These ClassLoaders don't expect addURL() to
> be called while they are calling super() in their own constr