This is the mail archive of the
mailing list for the Java project.
Re: Allocate Class instancs on heap; not conservatively scan DSOs
- From: Andrew Haley <aph at redhat dot com>
- To: Bryce McKinlay <mckinlay at redhat dot com>
- Cc: gcc-patches at gcc dot gnu dot org, java-patches at gcc dot gnu dot org
- Date: Wed, 26 Apr 2006 18:15:56 +0100
- Subject: Re: Allocate Class instancs on heap; not conservatively scan DSOs
- References: <email@example.com> <444FA867.firstname.lastname@example.org>
Bryce McKinlay writes:
> For a long time we've talked about putting the class data in a
> pointer-free format, instead of storing them as static jclasses. It
> looks like this patch has taken care of much of the hard work and will
> make changing the class data format easier.
Indeed it will, yes.
> One suggestion - I think it would be good to make the comments a bit
> more clear on which libgcj functions are called by the compiler to
> register classes for the -findirect-classes case, and which are for the
> non-indirect case. Its obvious from reading the patch why there is both
> a _Jv_RegisterClasses and _Jv_RegisterNewClasses, but it may not be for
> someone browsing the code later.
That seems reasonable.
> Also, the comment on emit_indirect_register_classes() doesn't look right
> - it doesn't call _Jv_NewClassFromInitailizer directly.
Oops. TVM for looking at this.