Adrian Aichner wrote:
Vin Shelton <acs(a)xemacs.org> writes:
>On 1/16/06, Rodney Sparapani <rsparapa(a)mcw.edu> wrote:
>
>
>>Dear Bug Team!
>>
>>I have the elisp variable inhibit-site-lisp set to the default,
>>nil, yet site-lisp is not added to load-path. I unpacked
>>
>>
Rodney, Vin, I don't think XEmacs has a notion of site-lisp.
It expects to find site-specific lisp code in
site-packages/lisp.
This is also what you both show in your load paths.
Rodney, try
mv c:\XEmacs\Packages\site-lisp\ess\lisp c:\XEmacs\Packages\site-package\lisp\ess
Adrian
>>
>>
>Hi Rodney,
>
>That is strange. I don't have an ess package to test, but I took an
>fsf-compat tarball and unpacked it to e:\XEmacs\site-packages and here
>is the value of load-path when I restarted XEmacs:
>
>`load-path' is a simple built-in variable.
>
>Value: ("e:/acs/lib/lisp/xemacs" "e:/acs/lib/lisp"
>"e:\\XEmacs\\site-packages\\lisp\\"
>"e:\\XEmacs\\site-packages\\lisp\\fsf-compat\\"
>"e:\\XEmacs\\xemacs-packages\\lisp\\"
>"e:\\XEmacs\\xemacs-packages\\lisp\\c-support\\".....
>
>1. I'm wondering if your ess package conforms to the standard package
>format; if you cd to c:/XEmacs/site-packages, what does 'ls -R' print?
>2. Does the behavior change at all if you start XEmacs with the -vanilla option?
>
>HTH,
> Vin
>
>
>
>
>
Hi Guys:
Well, I unpacked ESS into site-lisp with 21.5 on OS X and it does find
it. I don't see why 21.4
on Windows should be different. And, if it was different, then what is
the purpose of the
variable inhibit-site-lisp? I think this is a bug. Oh, and I tried the
site-packages trick
and it didn't find it; only site-packages/lisp was there. But, now I
see why that wouldn't work.
You have the directory site-packages/lisp/fsf-compat. No non-xemacs
package has a directory
structure like that. A non-xemacs packages would have
site-packages/ess/lisp or some-such.
So we need to either fix site-lisp or make the search in site-packages
also find site-packages/*/lisp
Thanks,
Rodney