-----BEGIN PGP SIGNED MESSAGE----- Hash: RIPEMD160
Just for completeness. From the post Ben made:What you describe above sounds like the "normal" way compilers are bootstrapped. The question is, how did they obtain the original Darwin/OSX M3 compiler then? Via a cross compiler? That's about the only possibility I see, except maybe that they have a minimal non-M3 written compiler that can bootstrap M3, but if that was the case, then why don't they distribute that by default... seems unlikely.
New Target Platform PPC_DARWIN (Darwin, MacOS X on PowerPC Hardware)
- --------------------------------------------------------------------
CM3 5.2.0 is the first Modula-3 release in history that will run on
some of Apple Computer's operating systems, i.e. those that are based
on Darwin. Cross-compilation was done from a FreeBSD4 system using the
new gcc back-end; the target host was a Darwin 6.3 machine identifying
itself as
Darwin lamancha.opendarwin.org 6.3 Darwin Kernel Version 6.3: \
Sat Dec 14 03:11:25 PST 2002; root:xnu/xnu-344.23.obj~4/RELEASE_PPC \
Power Macintosh powerpc
All the runtime and Unix implementation modules were initially copied
from the FreeBSD3 platform; in hindsight, this was only a mediocre
choice, as FreeBSD3 and PPC_DARWIN differ in endianness.
The following list gives detailed information about the status of
PPC_DARWIN support:
o Compiler front end (cm3) and code generator (cm3cg) work without
warnings or failures for all core and standard packages
(do-cm3-core.sh (core system compilation), do-cm3-std.sh (standard
packages installation)).
o The runtime is still very basic, it only supports the necessary
minimum, i.e.
- multi threading based on setjmp/longjmp and signals
- simple copy-and-sweep garbage collection (no incremental,
generational collector)
- access to all standard operating system services via m3core and
libm3
- no virtual memory protection, no system call wrappers
- no thread stack protection via vm
- exception handling based on setjmp/longjmp (no table based
exception handling (impossible with new gcc))
Several of the basic Unix interfaces in m3core have been carefully
checked; others haven't been reviewed at all, so that there will be
room for improvements ;-)
o The system configuration (cm3.cfg) supports static and dynamic
linking of M3 libraries (but not of external (system) libraries,
those are always linked dynamically); debugging support via stabs
is on by default; profiling hasn't been tested yet; all generated
code is position independent (-fPIC).
o Graphical user interface programming is supported by Trestle on X
and seems to work fine; there is of course no native GUI support
yet (Trestle on Aqua?).
o The following applications have been tested on PPC_DARWIN and seem
to work as expected: m3tohtml, formsedit, solitaire, badbricks,
columns, fisheye, obliq, Juno-2, CVSup.
http://groups.google.com/groups?dq=&hl=en&lr=&ie=UTF- 8&threadm=rlc2a.1046%248Z5.16958%40sea-read.news.verio.net&prev=/ groups%3Fq%3Dcomp.lang.modula3
Max
- - we may race and we may run, but we can not undo what has been done. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (Darwin) iD8DBQE+U7L7iW/Ta/pxHPQRAxqrAKDdoXU5sU4q5CiA7NAfybdw5XO0pgCg4frn wnjBK++fNapb6aNPjkTepvk= =Snn1 -----END PGP SIGNATURE----- ------------------------------------------------------- This SF.net email is sponsored by: SlickEdit Inc. Develop an edge. The most comprehensive and flexible code editor you can use. Code faster. C/C++, C#, Java, HTML, XML, many more. FREE 30-Day Trial. www.slickedit.com/sourceforge _______________________________________________ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
