[Ecls-list] Problems with GMP on Win32

[Juan Jose Garcia-Ripoll]

On Fri, Aug 1, 2008 at 9:01 PM, Robin Lee Powell
<rlpowell at digitalkingdom.org> wrote:
> I have no idea how to do that; the make structure of ECL is
> extremely complicated, and I've never built anything significant on
> Windows before.  I've done a lot of Linux building (I'm a sysadmin
> for a living), but I wasn't looking to have an adventure here, just
> get a working CL.  Anyways, like I said, the GMP bits seem to be
> working after I hacked mpn/x86/x86-defs.m4

Your statements are a bit contradictory. On the one hand you do not
know how to install libraries with cygwin's setup (execute the same
setup.exe, select gmp from the list, install), but on the other hand
you are perfectly capable of hacking GMP's autoconfiguration process.

ECL ships a particular version of GMP that may not work on all
platforms of all versions, but works most of the time. Now, in the
case in which it does not work, one typically installs the GMP library
by other means (on any Linux platforms that is just 2 clicks away) and
then configures ECL using that library with the flag
--with-system-gmp. So, no need to hack GMP no need to know ECL's
"complicated make structure".

Now, if you just want a common lisp that runs in Windows, the
Microsoft Visual C++ compiler may be a better option, at least with
ECL, since it builds ECL always (there are less variations), it is
free, produces executables with better performance and in our case it
is the only version that is automatically tested on our compiler farm
(http://ecls.sourceforge.net/logs.html)

Juanjo

-- 
Facultad de Fisicas, Universidad Complutense,
Ciudad Universitaria s/n Madrid 28040 (Spain)
http://juanjose.garciaripoll.googlepages.com