Re: [Axiom-developer] Axiom/GCL on windows

[Vanuxem Gregory]

Le lundi 11 décembre 2006 à 15:04 -0600, Gabriel Dos Reis a écrit :
> On Mon, 11 Dec 2006, Vanuxem Gregory wrote:
> 
> | Le lundi 11 décembre 2006 à 19:42 +0100, Vanuxem Gregory a écrit :
> |
> | [...]
> |
> | > I don't know exactly where is the problem but you're rerooting (function
> | > reroot) Axiom when you compile some interpreter's files. Reroot accepts
> | > an argument which has to be the root of Axiom. You give him (to reroot)
> | > a unix path and that's a problem (consequently $spadroot is wrong).
> | > $spadroot is used everywhere to create path relative to it so may be
> | > this is the problem ($spadroot is used to define |
> | > $defaultMsgDatabaseName|).
> |
> | After further investigation, $spadroot is more likely set, when building
> | depsys, via makedep.lisp.
> 
> Hmm, depsys builds fine and translates/compiles files with no problem.

Yes, I was wrong. What I said is applicable to depsys (via build-depsys)
only. Building interpsys (via build-interpsys) is similar, the Lisp
'spad' variable holds the root of Axiom. build-depsys and
build-interpsys call initroot (with spad as argument) which calls reroot
(and set $spadroot). So I hardcoded the SPAD variable in AX_FLAGS
(src/Makfile and src/interp/Makefile) to a usable path for GCL
(C:/staging/build/directory) for testing purpose and tried to rebuild
Axiom. The build process reached "layer 1" of the algebra and I stopped
there.

I wonder how it'll be possible to have a "correct" SPAD variable for
Windows with the autotools or any other tools. For Cygwin there is
cygpath to do that.


Greg