Re: Patch for review: java.lang.Class - signers issue

[Michael Koch <konqueror at gmx dot de>]

On Wed, Dec 03, 2003 at 02:54:27PM +0000, Andrew Haley wrote:
> Michael Koch writes:
>  > On Wed, Dec 03, 2003 at 02:34:46PM +0000, Andrew Haley wrote:
>  > > Michael Koch writes:
>  > >  > On Wed, Dec 03, 2003 at 02:17:28PM +0000, Andrew Haley wrote:
>  > >  > > Michael Koch writes:
>  > >  > >  > -----BEGIN PGP SIGNED MESSAGE-----
>  > >  > >  > Hash: SHA1
>  > >  > >  > 
>  > >  > >  > 
>  > >  > >  > I wrote the attached two patches to make it possible to compile 
>  > >  > >  > classpath with gcj again. I see these patches as start for a 
>  > >  > >  > discussion of this problem. These patches fix PR/12768 which is a 
>  > >  > >  > regression from 3.3.
>  > >  > > 
>  > >  > > Tom wanted to call the field hack_signers.  I don't know why.  I would
>  > >  > > have thought _Jv_signers a better choice.  However, this should be
>  > >  > > good enough for 3.4.
>  > >  > 
>  > >  > Tom said that the renaming to hack_signers is only a temporary solution
>  > >  > for 3.4.
>  > > 
>  > > I read that.  I still don't know why hack_signers is a suitable name.
>  > > 
>  > >  > In fact someone should implement a better solution for this
>  > >  > problem. My fixes are just a workaround.
>  > > 
>  > > OK.  You'll need to post the compiler part to gcc-patches.
>  > 
>  > Will do when we have agreed on a name with Tom. ;-)
>  > These patches were only written to start something. I never thought to
>  > apply them as they are.
> 
> I don't see why not.  It's a trivial problem with a trivial workaround.

Okay, if Tom agrees I will send the gcj patch to gcc-patches and commit
them.


Michael