Hi all, I finally suceeded porting cln and Ginac to win32, but just without gmp and ginsh. What disadvantages do I have to exspect from that (especially from the missing of gmp) regards Tilman Utz - To UNSUBSCRIBE, email to ginac-list@ginac.de with a subject of "unsubscribe".
On Thu, 7 Dec 2000, Tilman Utz wrote:
I finally suceeded porting cln and Ginac to win32, but just without gmp and ginsh. What disadvantages do I have to exspect from that (especially from the missing of gmp)
Missing GMP: For handling large numbers it will be slower. It will still behave asymptotically ideal, but the factor in front is just larger. This matters only if your interest lies in establishing world records in number theory and such. Missing Ginsh: No twisty little maze of passages to explore. :-) Actually I don't see what the problem is with getting things to work on win32. I was under the assumption that it had already been ported for Cygwin by David Billinghurst. Regards -richy. -- Richard Kreckel <Richard.Kreckel@Uni-Mainz.DE> <http://wwwthep.physik.uni-mainz.de/~kreckel/> - To UNSUBSCRIBE, email to ginac-list@ginac.de with a subject of "unsubscribe".
Hi, thanks for the information "Richard B. Kreckel" schrieb:
Actually I don't see what the problem is with getting things to work on win32. I was under the assumption that it had already been ported for Cygwin by David Billinghurst.
Yes, but there was still a reliance on the cygwin-dll which provides the posix background. With the lib, I built now I'm able to produce executables _totally_ independent of any third party dll's (Someone did that already with cln 1.0 as I discovered a couple of hours before, but I didn't check out wheher his patch works on 1.0.3, too) regards Tilman
Regards -richy. -- Richard Kreckel <Richard.Kreckel@Uni-Mainz.DE> <http://wwwthep.physik.uni-mainz.de/~kreckel/>
- To UNSUBSCRIBE, email to ginac-list@ginac.de with a subject of "unsubscribe".
On Thu, 7 Dec 2000, Tilman Utz wrote:
Actually I don't see what the problem is with getting things to work on win32. I was under the assumption that it had already been ported for Cygwin by David Billinghurst.
Yes, but there was still a reliance on the cygwin-dll which provides the posix background. With the lib, I built now I'm able to produce executables _totally_ independent of any third party dll's (Someone did that already with cln 1.0 as I discovered a couple of hours before, but I didn't check out wheher his patch works on 1.0.3, too)
I was refering to patches by Mumit Khan that went into CLN 1.0.1, see CLN's NEWS file enty for that release. I do not know about other aspects of the port and I am pretty sure the other developers don't know either since none of them uses Windows. But if you know what you are doing and think it is good work you are welcome to follow the last two lines of GiNaC's README. ;-) (But please do check out CLN 1.1 from CVS and GiNaC from there and work with these sources since a release is around the corner -- there are no show-stoppers any more right now.) Regards -richy. -- Richard Kreckel <Richard.Kreckel@Uni-Mainz.DE> <http://wwwthep.physik.uni-mainz.de/~kreckel/> - To UNSUBSCRIBE, email to ginac-list@ginac.de with a subject of "unsubscribe".
Hi Tilman, I just found out that there seems to be a possibility to use cln/ginac under win32. Please, can you (or anyone else listening to the list) tell me + what to do (and how long it will take) + whether there are any files etc. downloadable anywhere in order to get ginac running under win (preferably without the cygwin-dll, i.e. in the way you described in your mail.) Hopefully, that's another step to convince the "Skip C++, use Mathematica/Maple/whatever instead" people around here ;-) - Roland Tilman Utz wrote:
"Richard B. Kreckel" schrieb:
Actually I don't see what the problem is with getting things to work on win32. I was under the assumption that it had already been ported for Cygwin by David Billinghurst.
Yes, but there was still a reliance on the cygwin-dll which provides the posix background. With the lib, I built now I'm able to produce executables _totally_ independent of any third party dll's (Someone did that already with cln 1.0 as I discovered a couple of hours before, but I didn't check out wheher his patch works on 1.0.3, too)
- To UNSUBSCRIBE, email to ginac-list@ginac.de with a subject of "unsubscribe".
Roland Richter schrieb:
Please, can you (or anyone else listening to the list) tell me
+ what to do (and how long it will take) + whether there are any files etc. downloadable anywhere
in order to get ginac running under win (preferably without the cygwin-dll, i.e. in the way you described in your mail.)
Hi, assuming, you have cygwin (dll-version 1.1.7, older ones didn't seem to work) installed, cln 1.1, GiNaC 0.7.0 of course. Beyond that, you need the mingw-versions of some libraries. The url, where to download them and some explanation, why you need them can be found at http://www.xraylith.wisc.edu/~khan/software/gnu-win32/mno-cygwin-howto.txt just install them as told in the how-to. Next, I changed 2 files: replace, in your cygwin direcory, /usr/include/mingw/unistd.h by /usr/include/unistd.h , then, insert /usr/include/mingw/process.h into /usr/include/mingw/sys/unistd.h Now I set the following compiler flags: export CPPFLAGS="-I/usr/local/mingw32/include -mno-cygwin -I/usr/local/include" export CXXFLAGS="-I/usr/local/mingw32/include -mno-cygwin -O2 -I/usr/local/include" export LDFLAGS="-L/usr/local/mingw32/lib -mno-cygwin -L/usr/local/lib" ./configure --disable-shared --without-gmp I omitted gmp, didn't check out, if it works with the new cygwin dll... make .... for GiNaC: ./configure --disable-shared make ... Preparation shouldn't take no longer than half an hour, Building cln takes about 2 hours minimum, testing and installing again about half an hour. Building GiNaC goes faster, but I still end up with an error building Ginsh, which didn't seem to be too severe: Testing worked. Well, that's it, hope I didn't miss anything, it's half past two, and I have to go to bed, today, I'm writing an exam. Let me know, if you have any problems. regards Tilman - To UNSUBSCRIBE, email to ginac-list@ginac.de with a subject of "unsubscribe".
Tilman Utz schrieb:
Roland Richter schrieb:
Please, can you (or anyone else listening to the list) tell me
+ what to do (and how long it will take) + whether there are any files etc. downloadable anywhere
in order to get ginac running under win (preferably without the cygwin-dll, i.e. in the way you described in your mail.)
Hi, assuming, you have cygwin (dll-version 1.1.7, older ones didn't seem to work) installed, cln 1.1, GiNaC 0.7.0 of course. Beyond that, you need the mingw-versions of some libraries. The url, where to download them and some explanation, why you need them can be found at http://www.xraylith.wisc.edu/~khan/software/gnu-win32/mno-cygwin-howto.txt just install them as told in the how-to.
Next, I changed 2 files: replace, in your cygwin direcory, /usr/include/mingw/unistd.h by /usr/include/unistd.h , then, insert /usr/include/mingw/process.h into /usr/include/mingw/sys/unistd.h
Now I set the following compiler flags: export CPPFLAGS="-I/usr/local/mingw32/include -mno-cygwin -I/usr/local/include" export CXXFLAGS="-I/usr/local/mingw32/include -mno-cygwin -O2 -I/usr/local/include" export LDFLAGS="-L/usr/local/mingw32/lib -mno-cygwin -L/usr/local/lib"
./configure --disable-shared --without-gmp
I omitted gmp, didn't check out, if it works with the new cygwin dll...
make ....
for GiNaC: ./configure --disable-shared make ...
Preparation shouldn't take no longer than half an hour, Building cln takes about 2 hours minimum, testing and installing again about half an hour. Building GiNaC goes faster, but I still end up with an error building Ginsh, which didn't seem to be too severe: Testing worked.
I just made all these steps again, to see, if it's really working for me, and I now recall why I get this error with Ginsh: The configure script didn't find the headers for libreadline. They are on my system, but when I set their path to the compiler flags, I got some strange compilation error I didn't understand so far, so I preferred a working Ginac without Ginsh.
Well, that's it, hope I didn't miss anything, it's half past two, and I have to go to bed, today, I'm writing an exam. Let me know, if you have any problems.
regards Tilman
- To UNSUBSCRIBE, email to ginac-list@ginac.de with a subject of "unsubscribe".
- To UNSUBSCRIBE, email to ginac-list@ginac.de with a subject of "unsubscribe".
participants (3)
-
Richard B. Kreckel
-
Roland Richter
-
Tilman Utz