Re: Patch: gcj ggc patch

[Richard Henderson <rth at redhat dot com>]

On Wed, Mar 06, 2002 at 10:58:18PM -0800, Per Bothner wrote:
> That seems silly.  Deliberate redundancy is not normally
> how Gcc is engineered (except for sanity checks to detect errors).
> Redundantly marking identifier nodes leads to confusion and bugs,
> because it is not clear whose responsibility it actually
> is to mark the identifier.

I'm thinking more toward geoffk's PCH stuff in which marks
serve the dual purpose of GC and swizzling for export.  In
addition, complete mark coverage opens the door for GC with
compaction.


r~