From: ············@gmail.com
Subject: should load translate-logical-pathname
Date: 
Message-ID: <1114250624.149985.80920@f14g2000cwb.googlegroups.com>
Hi,

my reading of the standard is that load should
translate-logical-pathname; LW, CMUCL, OpenMCL do, CLISP does not. Is
it implementation dependent or a bug in CLISP.

David

From: Wade Humeniuk
Subject: Re: should load translate-logical-pathname
Date: 
Message-ID: <e6sae.497$tg1.159@edtnps84>
············@gmail.com wrote:
> Hi,
> 
> my reading of the standard is that load should
> translate-logical-pathname; LW, CMUCL, OpenMCL do, CLISP does not. Is
> it implementation dependent or a bug in CLISP.
> 
> David
> 

LWW works fine for me, always has: (from an init file)

(setf (logical-pathname-translations "RUNNING")
       `(("**;*"
          ,(make-pathname :host (pathname-host *load-truename*)
                          :device (pathname-device *load-truename*)
                          :directory (append (pathname-directory *load-truename*) 
'(:wild-inferiors))
                          :name :wild
                          :type :wild
                          :version :wild))))

(load "RUNNING:lww-gdi;defsys.lisp")

Sure your are specififying your logical-pathname correcly?

Wade
From: ············@gmail.com
Subject: Re: should load translate-logical-pathname
Date: 
Message-ID: <1114265636.346321.183450@z14g2000cwz.googlegroups.com>
> Sure your are specififying your logical-pathname correcly?


Of course I do. LW does, CLISP does not.
From: ··············@hotmail.com
Subject: Re: should load translate-logical-pathname
Date: 
Message-ID: <1114273047.035224.56670@l41g2000cwc.googlegroups.com>
············@gmail.com wrote:
> > Sure your are specififying your logical-pathname correcly?
>
>
> Of course I do. LW does, CLISP does not.

Perhaps it is obvious to others what the problem is, but to me this
statement is not conclusive.

Can you provide more detail?

 In particular, does CLISP's translate-logical-pathname do what you
want, but the load does not? Or is CLISP's translate-logical-pathname
somehow differing from the other implementations?
From: ············@gmail.com
Subject: Re: should load translate-logical-pathname
Date: 
Message-ID: <1114279909.843345.80820@g14g2000cwa.googlegroups.com>
··············@hotmail.com wrote:
> ············@gmail.com wrote:
> > > Sure your are specififying your logical-pathname correcly?
> >
> >
> > Of course I do. LW does, CLISP does not.
>
> Perhaps it is obvious to others what the problem is, but to me this
> statement is not conclusive.
>
> Can you provide more detail?
>
>  In particular, does CLISP's translate-logical-pathname do what you
> want, but the load does not? Or is CLISP's translate-logical-pathname
> somehow differing from the other implementations?

[15]> (setf (logical-pathname-translations "home") `(
  ("**;*" "/Users/dvd/**/*")))
((#S(LOGICAL-PATHNAME :HOST "HOME" :DEVICE :UNSPECIFIC
     :DIRECTORY (:ABSOLUTE :WILD-INFERIORS) :NAME :WILD :TYPE NIL
:VERSION NIL)
  "/Users/dvd/**/*"))
[16]> (load "home:Library;Lisp;defsystem.lisp")

*** - LOAD: A file with name home:Library;Lisp;defsystem.lisp does not
exist
Break 1 [17]> :a

[18]> (load (translate-logical-pathname
"home:Library;Lisp;defsystem.lisp"))
;; Loading file /Users/dvd/library/lisp/defsystem.lisp ...
;; Loaded file /Users/dvd/library/lisp/defsystem.lisp
T
[19]>
From: Kalle Olavi Niemitalo
Subject: Re: should load translate-logical-pathname
Date: 
Message-ID: <87sm1g2kpm.fsf@Astalo.kon.iki.fi>
·············@gmail.com" <············@gmail.com> writes:

> [16]> (load "home:Library;Lisp;defsystem.lisp")
>
> *** - LOAD: A file with name home:Library;Lisp;defsystem.lisp does not
> exist

By default, CLISP 2.33.2 does not recognize a namestring as
a logical pathname namestring just because it begins with
alphanumeric characters and a colon; you must instead call
LOGICAL-PATHNAME or TRANSLATE-LOGICAL-PATHNAME explicitly.
To change this, do (setq custom:*parse-namestring-ansi* t)
or start CLISP with the -ansi option.
From: ············@gmail.com
Subject: Re: should load translate-logical-pathname
Date: 
Message-ID: <1114380292.375718.282780@g14g2000cwa.googlegroups.com>
> To change this, do (setq custom:*parse-namestring-ansi* t)
> or start CLISP with the -ansi option.

That was it, thanks a lot. I've added (setq custom:*ansi* t) to the
initialization script.

David Tolpin
From: Adam Warner
Subject: Re: should load translate-logical-pathname
Date: 
Message-ID: <pan.2005.04.23.13.14.59.870066@consulting.net.nz>
On Sat, 23 Apr 2005 03:03:44 -0700, ············@gmail.com wrote:
> my reading of the standard is that load should
> translate-logical-pathname; LW, CMUCL, OpenMCL do, CLISP does not. Is
> it implementation dependent or a bug in CLISP.

David, while I'm yet to use logical pathnames there seems to be no
ambiguity in the specification for LOAD:

   Function LOAD

   Syntax:

   load filespec &key verbose print if-does-not-exist external-format

   ... If filespec is a logical pathname, it is translated into a physical
   pathname as if by calling translate-logical-pathname.

That is, it is a bug if LOAD doesn't load the same physical pathname that
would have resulted if you had first called TRANSLATE-LOGICAL-PATHNAME
yourself.

"as if by calling" just means the result must be the same even if the
implementation uses a different method to obtain it (e.g. for efficiency
reasons).

Regards,
Adam