Discussion:
GM2 problems
(too old to reply)
Jürgen Lerch
2009-04-05 17:11:44 UTC
Permalink
Saluton!

In my continuing quest to write about a gazillion vapourware
roguelikes, I wrestled again with gm2 (version 0.68/gcc 4.1.2
(Debian etch)).

a) I had a procedure which took a (_non_-VAR) CARDINAL parameter.
Calling it with a SHORTCARD produced a parameter mismatch
compiler error. I since changed all SHORTCARDs to CARDINALs,
which made the problem disappear, but I consider this a
suboptimal solution in light of at least my idea of the
Modula-2 type system.

b) Now when I try to compile the program, the linker complains
about an undefined reference to a _M2_JuLANSI_finish in a
function ,,finish'' in some weird .cpp file in /tmp. What's
the problem?

Ad Astra!
JuL
--
***@gmx.de / Never anger a dragon, for you will be
Jürgen ,,JuL'' Lerch / crunchy and taste good with ketchup
Lee
2009-04-06 05:43:12 UTC
Permalink
maybe you wanna post to gm2 mailing list

http://www.nongnu.org/gm2/index-15.html#Mailing%20list
Jürgen Lerch
2009-04-07 18:15:20 UTC
Permalink
Saluton!
Post by Lee
maybe you wanna post to gm2 mailing list
http://www.nongnu.org/gm2/index-15.html#Mailing%20list
Well, I had rather hoped to get an answer here instead
of subscribing to an additional mailing list.

Ad Astra!
JuL
--
***@gmx.de / Never anger a dragon, for you will be
Jürgen ,,JuL'' Lerch / crunchy and taste good with ketchup
Christoph Schlegel
2009-04-08 16:49:03 UTC
Permalink
Post by Jürgen Lerch
Saluton!
Post by Lee
maybe you wanna post to gm2 mailing list
http://www.nongnu.org/gm2/index-15.html#Mailing%20list
Well, I had rather hoped to get an answer here instead
of subscribing to an additional mailing list.
Ad Astra!
JuL
According to

http://lists.nongnu.org/mailman/listinfo/gm2

it is not required to subscribe. Just send your question to

***@nongnu.org
Gaius Mulley
2009-04-09 14:49:00 UTC
Permalink
Post by Jürgen Lerch
Saluton!
In my continuing quest to write about a gazillion vapourware
roguelikes, I wrestled again with gm2 (version 0.68/gcc 4.1.2
(Debian etch)).
a) I had a procedure which took a (_non_-VAR) CARDINAL parameter.
Calling it with a SHORTCARD produced a parameter mismatch
compiler error. I since changed all SHORTCARDs to CARDINALs,
which made the problem disappear, but I consider this a
suboptimal solution in light of at least my idea of the
Modula-2 type system.
Hi,

very true, thanks for this error report. Now fixed in the current
cvs..
Post by Jürgen Lerch
b) Now when I try to compile the program, the linker complains
about an undefined reference to a _M2_JuLANSI_finish in a
function ,,finish'' in some weird .cpp file in /tmp. What's
the problem?
not sure although there have been many fixes applied to the cvs since
release 0.68. Is there any chance you could try this with the latest
cvs gm2? Or alternatively you could wait until I've finished the
current round of bug fixing and I'll release a new debian etch gm2.
Lastly - if you are willing - I'd happily download your code and check
that the latest gm2 can compile/link it (aka black box test)

regards,
Gaius
Jürgen Lerch
2009-04-24 22:39:57 UTC
Permalink
Saluton!

a) Thanks for answering.
b) Sorry for answering so late myself.
Post by Gaius Mulley
Post by Jürgen Lerch
a) I had a procedure which took a (_non_-VAR) CARDINAL parameter.
Calling it with a SHORTCARD produced a parameter mismatch
compiler error. I since changed all SHORTCARDs to CARDINALs,
which made the problem disappear, but I consider this a
suboptimal solution in light of at least my idea of the
Modula-2 type system.
very true, thanks for this error report. Now fixed in the current
cvs..
Good to hear!
Post by Gaius Mulley
Post by Jürgen Lerch
b) Now when I try to compile the program, the linker complains
about an undefined reference to a _M2_JuLANSI_finish in a
function ,,finish'' in some weird .cpp file in /tmp. What's
the problem?
not sure although there have been many fixes applied to the cvs since
release 0.68. Is there any chance you could try this with the latest
cvs gm2? Or alternatively you could wait until I've finished the
current round of bug fixing and I'll release a new debian etch gm2.
I downloaded the cvs-latest-tarball from the gm2 page but
didn't get to try it yet, so we'll see whether I get to it
first or you release the new .deb ...
Post by Gaius Mulley
Lastly - if you are willing - I'd happily download your code and check
that the latest gm2 can compile/link it (aka black box test)
I don't mind in princple, it'll be open soure anyway (so
people have yet another example of strange & bad programming
practices to avoid ;-)), but I don't have a web site yet (and
I don't like those usual free services).

Ad Astra!
JuL
--
***@gmx.de / Reality is a crutch for those who can't
Jürgen ,,JuL'' Lerch / cope with fantasy
Loading...