Unix specific (was Re: OOC Garbage Collector Proposal)

[Brian Hawley <brian.hawley@bigfoot.com>]

>From: Michael Griebling <mgriebling@inspired.to>
>> >> need a working prototype and some experimental results with OOC.  For
>> >> my part I volunteer to provide the necessary compiler support in oo2c
>> >> for such a prototype.
>> >
>> >I was hoping you would volunteer for the oo2c bits.  Thanks. ;-)
>>
>> I was hoping that someone else would volunteer. I would love to rewrite
>
>Sorry, I wasn't referring to you, Brian.  I believe Michael van Acken, the
>originator of the compiler wrote that last message.  I was thanking him.
>Of course, the more people we have volunteering on this project, the
>better the final result will be.

I figured that one out :). I would like to help though. I have studied
various GC techniques for a few years, and would like to apply them to
Oberon.

>> the RScheme collector in Oberon but I have had a few problems with OOC
>> so far that have made it difficult to use. For one thing, I now have
>> little access to Unix boxes, just Win32 boxes, so the configure scripts,
>> the libraries and RCS are rather awkward for me to use. I am going to
>
>If you are still interested in helping, I suggest you get yourself a version
>of Linux.  It is practically free and runs on just about any PC.  Although
>it's a bit difficult to install, it does run X-windows and the OO2C compiler,
>and VisualOberon quite well.  You needed worry about loosing your
>Windows stuff since Linux comes with a multiple OS boot selector.You will
>need a separate hard disk or partition though.

I wish I could, but these aren't really my computers. I'm afraid I'm stuck
with Win32 unless I use DJGPP for Dos. I would like to port oo2c to LCC-Win32,
a very good, light-weight Ansi-C compiler (it fits on one floppy, with the
source to the compiler included, and an IDE). It uses the standard C runtime
that comes with Win32, just like Mingw32. Cygnus Gnu-Win32 is really too much
to install as a backend, not to mention their licensing restrictions.

I could cross-compile from an HP-UX 10.20 box, if necessary :(

>> buy CS-RCS so the last one shouldn't be difficult, but rewriting the
>> libraries for native Win32 support should be an interesting task. Can
>> we split the Unix-specific aspects of the C backend and libs into their
>> own subdirectories, so I have a better idea of how much I need to port?
>
>What you are proposing is a really BIG task.  Just thought I would warn
>you.  I'm not familiar with CS-RCS.  Is this a compiler?

Component Software RCS, based on Gnu RCS, but not GPL'ed itself. The free
version performs all of the same functions as Gnu RCS, diff, etc., but is
not compatible. The Unix-compatible version you pay for. It seems a shame
to have to get it when I already got Intersolv PVCS bundled with Delphi C/S,
but that's not compatible either.

Unfortunately, I have a good idea of how big the task is. It still needs
doing though. We'll never get the native backends to be compatible until we 
set the groundwork. Unless we start out portable, we'll have to go through
the nightmare that Cygnus is with GCC. And I'll have to stick with Delphi
and Java (AAACK).

So, what would be the best use of my time? GC or portability cleanup and
Win32?

Brian Hawley