[GiNaC-list] Compiling Ginac with MinGW failed in Ginsh compile eventhough, I dont wish to compile it

Christiand tianize at gmail.com
Tue Sep 4 16:52:24 CEST 2012

Hi all,

Thanks for your helps. First of all, I am actually super duper new in
linux things and cross compiling, so frankly I don't understand about
all things that I did, including also the commands for compiler and
configure. I gathered all those by searching internet and reading some
forums and just "try and error".

My final goal is to compile ginac so that I have *.lib (static) or
*.dll (shared) in to use in MSVC. That what brought me to that website
link. I tried to follow all the steps with correction necessary
according to my situation.

As far as understand, CC=cl CXX=cl CCAS=cl is necessary if we want to
compile library with cl.exe (microsoft compiler and linker), of course
CC = cl shows that intention. I dont understand for the remaining two.
That is the reason I put those options in configure.

As said previously, missing in installing libreadline make error in
ginsh compiling process. I am aware of that issue. However, (as i
asked initially) I just dont get it when it is said that ginsh will
not be compiled but in fact it is being compiled.

I finally managed this problem by manually edit the makefile. I erased
ginsh doc tool option in

SUBDIRS = ginac check ginsh tools doc
DIST_SUBDIRS = ginac check ginsh tools doc


SUBDIRS = ginac check
DIST_SUBDIRS = ginac check

then ginac is compiled succesfully. All tests are passed too.

Now, I get the *.lib, I tested it in visual studio, It works well in
release mode but not in debug. I dont know why.

I am also trying to make *.dll. It is succesfully compiled, but all
tests are failed. Again, I dont know what is the reason.

related to Jan's suggestion about putting in .ll ( is it
ginsh_lexer.ll ??), it is failed. Same error appeared.

#ifdef _MSC_VER
   #define YY_NO_UNISTD_H

about alex

Here is my setting :

environment : msys (MinGW)
compiler : cl
ginac : ginac-1.6.2
cln : cln-1.3.2
gmp : gmp-5.0.5

~ chris ~

More information about the GiNaC-list mailing list