Outstanding Stuff

[Boehm, Hans]

In my opinion, the configure.in part of this patch does the right thing.

As far as GC_DLL is concerned, your patch looks fine in the short term.  In the slightly longer term, the macro test for DLL-ness in the gcc are already a mess, and this makes it slightly worse, by not being entirely consistent with the current, excessively baroque, usage.

If you want to check this in in the short term, that's OK with me.  I'll make some more changes in my tree, which should eventually make it fully correct.

What I would like to do:

1) Consistently use GC_DLL to test whether the GC is in its own dll.

2) Implicitly set GC_DLL in gc_priv.h if _DLL is defined and GC_NO_DLL is not defined.  (I read the MS documentation as saying that _DLL should be defined when building either a dll or its client, and the result is thread safe.  Under non-gcc compilers, it seems like a reasonable choice for the GC_DLL default, but it can probably be wrong in both directions.  For gcj, this is irrelevant.)

Hans
> -----Original Message-----
> From: Mohan Embar [mailto:gnustuff@thisiscool.com]
> Sent: Tuesday, April 01, 2003 6:57 PM
> To: java@gcc.gnu.org
> Subject: Outstanding Stuff
> 
> 
> Ranjit> Which reminds me (OT for this thread, sorry): 
> whatever happened 
> Ranjit> to Mohan's strict-case-open in the front-end patch for Win32? 
> Ranjit> Was it accepted?
> 
> And while we're on (off) the subject of pending stuff, here's my
> current list (in addition to the case-sensitivity stuff):
> 
> GC_DLL: http://gcc.gnu.org/ml/java-patches/2003-q1/msg00669.html
> 
> (
> Even Hans blessed the approach, though not necessarily my 
> implementation:
> http://gcc.gnu.org/ml/java/2003-02/msg00291.html
> )
> 
> "What do we do about the broken MingW 3.3 build"?:
> http://gcc.gnu.org/ml/java-patches/2003-q1/msg00843.html
> 
> -- Mohan
> http://www.thisiscool.com/
> http://www.animalsong.org/
> 
> 
> 
>