forth linux-mandrake rpm packages available
Guido Draheim
If you have a mandrake (or connectiva) linux system, you might want
to have some forth systems prepackaged. These rpm files will hook
into the (debian-style) menu-interface of these distributions along
with a fancy icon, opening a terminal window for the forth interpreter.
The panel section should be called like "Applications/Interpreters".
available rpms:
gforth-0.5.0
bigforth-2.0.4
ficl-3.0.1
pfe-0.32.38
get them from the download area at http://freespace.sf.net - the direct
link to the page is http://sf.net/project/showfiles.php?group_id=13488
HTH, cheers,
-- guido http://freespace.sf.net/guidod
GCS/E/S/P C++$++++ ULHS L++w- N++@ d(+-) s+a- r+@>+++ y++ 5++X- (geekcode)
Guido Draheim
new available rpms:
kforth-1.0
pforth-21.6
this should now be plenty of forth to check your forth writings with.
Guido Draheim:
Krishna Myneni
>
> new available rpms:
> kforth-1.0
> pforth-21.6
>
> this should now be plenty of forth to check your forth writings with.
>
Thanks for the info Guido. I wanted to have an rpm package
for kforth for sometime now, but there always seem to be more
pressing matters... When I checked the download area on
http://freespace.sf.net the kforth entry appeared
in the list, but there were no files yet. Are you going
to maintain the rpm packages?
Krishna
Guido Draheim
rpm is pretty easy now for me, and it should be possible to just
rebuild with any new tarball you have next time around.
The thing about sourceforge: it uses now a distributed system of
download servers. I have uploaded the rpms and they are already
visible on my (europe) http://prdownloads.sf.net/freespace - but
possibly not yet on yours. The frontpage in the download area will
only get visible when all prdownload-servers have the files
mirrored. (background: for each world region, a different dns-entry
is used for prdownloads.sf.net, so you end up at a different ip-adress).
just wait a few minutes, and the files should be visible from the
frontpage. If you don't want to wait that long, fetch it directly
from the europe server at http://216.136.171.195/freespace/
Guido Draheim
Oh, and btw, I am still waiting for the mandrake guys to put the
forth rpms into the cooker repository, so they might ship next
time officially on a linux-mandrake cd. However, I guess no one
over there knows about forth, so that is more or less just
theoretic. The redhat/rawhide do not have any forth, so that one
would possibly need to put some propaganda over there, so that
it will go into any of the other redhat-like linux distributions.
Sorrily, I do not have much of time to do those communictations,
I just make rpms and put them up at my sourceforge space so far.
-- cheers, guido
p.s. the filemirror is now complete at sfnet :-)
Es schrieb Guido Draheim:
Krishna Myneni
>
> Oh, and btw, I am still waiting for the mandrake guys to put the
> forth rpms into the cooker repository, so they might ship next
> time officially on a linux-mandrake cd. However, I guess no one
> over there knows about forth, so that is more or less just
> theoretic. The redhat/rawhide do not have any forth, so that one
> would possibly need to put some propaganda over there, so that
> it will go into any of the other redhat-like linux distributions.
> Sorrily, I do not have much of time to do those communictations,
> I just make rpms and put them up at my sourceforge space so far.
> -- cheers, guido
>
> p.s. the filemirror is now complete at sfnet :-)
>
Lobbying the redhat folks to put the forth rpms on their
distribution sounds like a worthwhile pursuit for the
Forth community. Having the rpms readily available should
prompt more people to try Forth.
Btw, I now see the kforth rpms on your page. I will put a
link to these on the kforth home page. Thanks.
Regards,
Krishna
Bdale Garbee
> new available rpms:
Just as an aside, I and others have had a number of Forth packages in Debian
GNU/Linux <http://www.debian.org> for several years. At least gforth, kforth
(just noticed I haven't updated this recently, will get around to it shortly),
pforth, and yforth are available. Probably others.
Enjoy.
Bdale
Anton Ertl
Guido Draheim <guidod...@gmx.de> writes:
> The redhat/rawhide do not have any forth, so that one
>would possibly need to put some propaganda over there, so that
>it will go into any of the other redhat-like linux distributions.
I was in contact with someone from Redhat some time ago, and he was
even looking at using Forth as a scripting language for some Red Hat
tool. I don't know what came out of it.
One problem at the time was that gcc-2.96 (RedHat's unofficial
snapshot of gcc) does not compile gforth (my guess is that the
compiler is broken).
Followups to clf.
- anton
--
M. Anton Ertl http://www.complang.tuwien.ac.at/anton/home.html
comp.lang.forth FAQs: http://www.complang.tuwien.ac.at/forth/faq/toc.html
Guido Draheim
>
> In article <3C0C56DC...@gmx.de>,
> Guido Draheim <guidod...@gmx.de> writes:
> > The redhat/rawhide do not have any forth, so that one
> >would possibly need to put some propaganda over there, so that
> >it will go into any of the other redhat-like linux distributions.
>
> I was in contact with someone from Redhat some time ago, and he was
> even looking at using Forth as a scripting language for some Red Hat
> tool. I don't know what came out of it.
>
> One problem at the time was that gcc-2.96 (RedHat's unofficial
> snapshot of gcc) does not compile gforth (my guess is that the
> compiler is broken).
>
that was atleast true for the early versions, i.e. gcc-2.96-1x,
however mandrake ships some gcc-2.96-6x and rawhide has currently
2.96-8x. Well, I have to admit that all later version of gcc did
produce some kind of miscompiled code for pfe - including 2.95.x,
but I was able to fix the places for 2.95.x but not for 2.96. The
rpms you see at my freespace are compiled quite often with gcc-3.0.2,
including gforth and bigforth - I did not look whether you have a
basic check-sequence in gforth (as with "make check" in pfe
to atleast look for this damn compiler problemacy). It does not
seem to be a problem AFAICS.
Making rpms for rawhide would require to follow the redhat guidelines,
while the rpm user does not see them, the rpm writer must watch out
for some differences between the distributions. Therefore it would be
a good idea to have a contact to a packager working with a rawhide
distribution - which is not me, I am using the cooker being the
mandrake variant. The compiled rpms however should be portable to
all newer rpm-based linux distributions, its just the rpm-build that's
different.
Another problem is that places like rpmfind.net will only walk the
rawhide and cooker database - I did contact Daniel Veillard why there
is no latest update of the sourceforge ftp dataspace but he said it
was not quite possible at the moment, so that the newest rpm at my
sf freespace will not show up at about any rpmfinder. IMNSHO, this is
the biggest problem - if users interested in forth would be able to
find a rpm they could install just so, then it would be less a problem
to see (newbie) people just have a look into what forth is...
cheers, guido
/snip/
On Tue, Dec 04, 2001 at 06:13:23AM +0100, Guido Draheim wrote:
>
> hi Daniel,
>
> noticed that rpmfind.net lists quite some of my rpms hosted at sourceforge
> but they are seriously outdated. Is there a chance that there'll be updates?
@ERROR: max connections (10) reached - try again later
that's what I get in my logs ... Nothing I can do !
> secondly, you faq'd that rpms shall go to redhat.contrib, but that's somewhat
> dubious. I have a couple of rpms at http://freespace.sf.net download area
> where each of them implements the Forth language, but it seems no linux
> distro wants to have such a thing. - many people look through rpmfind.net
> to get rpms which their own linux-distro does not ship along, and this
> makes for a perfect double-loose condition.
You understand I cannot maintain mirror for every project out there, right ?
So try to get one of the repository I already mirror to host you,
freshmeat, falsehope or freshrpms for example...
Daniel
Bernd Paysan
> that was atleast true for the early versions, i.e. gcc-2.96-1x,
> however mandrake ships some gcc-2.96-6x and rawhide has currently
> 2.96-8x. Well, I have to admit that all later version of gcc did
> produce some kind of miscompiled code for pfe - including 2.95.x,
> but I was able to fix the places for 2.95.x but not for 2.96. The
> rpms you see at my freespace are compiled quite often with gcc-3.0.2,
> including gforth and bigforth - I did not look whether you have a
> basic check-sequence in gforth (as with "make check" in pfe
> to atleast look for this damn compiler problemacy). It does not
> seem to be a problem AFAICS.
Why didn't you just try "make check"? It's there. BTW: When you make
binaries, please also do a "make bench" and check if the figures are
comparable to those in Benchres. New compilers like gcc-3.0.2 often
require some tweaking with the forced register assignments to produce
good code.
--
Bernd Paysan
"If you want it done right, you have to do it yourself"
http://www.jwdt.com/~paysan/
Guido Draheim
great, I'll keep this is mind for the next release - which I have to
do anyway since I got an infoline in the wrong place and I want to
seperate a doc-rpm from the main-rpm for all packages...
cheers, guido
Anton Ertl
Guido Draheim <guidod...@gmx.de> writes:
>Es schrieb Bernd Paysan:
>> Why didn't you just try "make check"? It's there. BTW: When you make
>> binaries, please also do a "make bench" and check if the figures are
>> comparable to those in Benchres. New compilers like gcc-3.0.2 often
>> require some tweaking with the forced register assignments to produce
>> good code.
>>
>
>great, I'll keep this is mind for the next release - which I have to
>do anyway since I got an infoline in the wrong place and I want to
>seperate a doc-rpm from the main-rpm for all packages...
In order to save you the register tweaking, use the following patch,
and run gcc with -fno-gcse. gcc-2.95.* is nicer to Gforth than
gcc-3.0.*.
Index: machine.h
===================================================================
RCS file: /usr/local/lib/cvs-repository/src-master/gforth/arch/386/machine.h,v
retrieving revision 1.13
retrieving revision 1.15
diff -u -r1.13 -r1.15
--- machine.h 2000/09/23 15:47:00 1.13
+++ machine.h 2001/12/02 13:34:57 1.15
@@ -130,14 +130,25 @@
/* gcc 2.95 has a better register allocater */
#define SPREG asm("%esi")
#define RPREG asm("%edi")
-#define TOSREG asm("%ebx")
+#if (__GNUC__>2)
+#define IPREG asm("%ebx")
+/* ebp leads to broken code; eax, ecx, edx produce compile errors */
+#define TOSREG asm("%ecx")
+/* ecx works only for TOS, and eax, edx don't work for anything */
#else
+/* gcc-2.95 manages to fit ip into ebp by itself */
+#define TOSREG asm("%ebx")
+#endif
+#else /* gcc-2.95 or later */
#define IPREG asm("%ebx")
#endif
#else
#if ((__GNUC__==2 && defined(__GNUC_MINOR__) && __GNUC_MINOR__>=95) || (__GNUC__>2))
#define SPREG asm("%esi")
#define RPREG asm("%edi")
+#if (__GNUC__>2)
+#define IPREG asm("%ebx")
+#endif
#else
#define SPREG asm("%ebx")
#endif