headers
Gary King | 7 Mar 2007 14:22
Favicon
Gravatar

ASDF copyright year+

Shouldn't the copyright year in the texinfo source be updated? It reads:

asdf Copyright  <at> copyright{} 2001-2004 Daniel Barlow and contributors

This manual Copyright  <at> copyright{} 2001-2004 Daniel Barlow and
contributors

Is there any reason not to bump 2004 to 2007?
--
Gary Warren King, metabang.com
Cell: (413) 885 9127
Fax: (206) 338-4052
gwkkwg on Skype * garethsan on AIM

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 0 comments |
headers
Gary King | 7 Mar 2007 16:45
Favicon
Gravatar

An ASDF patch

The following is a patch to asdf that does the following:

* Adds, exports and documents system-relative-pathname
* ensures that a name or type in *default-pathname-defaults* doesn't  
confuse the asdf preference file loading mechanism
* When not resolving symlinks, uses *load-pathname* rather than *load- 
truename* when running operations. The latter may have already  
resolved the links and that can leave one up a creek.

Comments welcome (and desired); if none are forthcoming, this will be  
checked in later this week.

thanks,

Index: asdf.lisp
===================================================================
RCS file: /cvsroot/cclan/asdf/asdf.lisp,v
retrieving revision 1.106
diff -u -w -r1.106 asdf.lisp
--- asdf.lisp	14 Feb 2007 15:16:24 -0000	1.106
+++ asdf.lisp	7 Mar 2007 15:40:34 -0000
 <at>  <at>  -13,7 +13,7  <at>  <at> 
;;; is the latest development version, whereas the revision tagged
;;; RELEASE may be slightly older but is considered `stable'
-;;; Copyright (c) 2001-2003 Daniel Barlow and contributors
+;;; Copyright (c) 2001-2007 Daniel Barlow and contributors
;;;
;;; Permission is hereby granted, free of charge, to any person  
obtaining
;;; a copy of this software and associated documentation files (the
(Continue reading)

Permalink | Reply | 2 comments |
headers
Christophe Rhodes | 7 Mar 2007 16:58
Favicon

Re: An ASDF patch

Gary King <gwking <at> metabang.com> writes:

> Comments welcome (and desired); if none are forthcoming, this will be  
> checked in later this week.

What are the expected results when a defsystem form is evaluated
outside of a file?

> +(defvar *resolve-symlinks* t)

What is this new variable for?  It doesn't seem to be either exported
or bound.

> +  (let ((*default-pathname-defaults*
> +	 (make-pathname :name nil :type nil
> +			:defaults *default-pathname-defaults*)))

I'd be surprised if this actually did what you might expect it to in
all implementations, knowing what I do about some of them... only some
of them treat :name nil as explicit provision of a name; others will
merrily merge over the top of that NIL from the defaults.

> 				   (or ,pathname
> -				       (when *load-truename*
> +					;; If not resolving symlinks, then
> +					;; use *load-pathname* instead of
> +					;; *load-truename* since the latter,
> +					;; in some implementations, has
> +					;; already resolved a symlink.
> +					(if *resolve-symlinks*
(Continue reading)

Permalink | Reply | 1 comment |
headers
Gary King | 7 Mar 2007 20:48
Favicon
Gravatar

Re: An ASDF patch

Hi Christophe,

I'm going to skip over some of your questions completely and get back  
to them when I make more time to think.

>> +  (let ((*default-pathname-defaults*
>> +	 (make-pathname :name nil :type nil
>> +			:defaults *default-pathname-defaults*)))
>
> I'd be surprised if this actually did what you might expect it to in
> all implementations, knowing what I do about some of them... only some
> of them treat :name nil as explicit provision of a name; others will
> merrily merge over the top of that NIL from the defaults.

Hm, OK. Is there a portable solution? Otherwise, I think I'd like to  
keep it since it will at least work some of the time. I was actually  
bit by this not too long ago which is why I thought adding it would  
be a good thing.

Secondly, I take it that the system-relative-pathname hunk is not  
objectionable. Unless I hear otherwise, I think I'll include that  
patch separately and then we can keep discussing the other ones.

thanks,
--
Gary Warren King, metabang.com
Cell: (413) 885 9127
Fax: (206) 338-4052
gwkkwg on Skype * garethsan on AIM
(Continue reading)

Permalink | Reply | 0 comments |
headers
Gary King | 21 Mar 2007 23:08
Favicon
Gravatar

ASDF changes

I've updated ASDF with the system-relative-pathname changes.

--
Gary Warren King, metabang.com
Cell: (413) 885 9127
Fax: (206) 338-4052
gwkkwg on Skype * garethsan on AIM

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 0 comments |
headers
Attila Lendvai | 29 Mar 2007 17:35
Picon
Gravatar

with-compilation-unit

dear list,

asdf's OPERATE is wrapping basically everything in a big
WITH-COMPILATION-UNIT, even TEST-OP's.

on sbcl it locks the *big-compiler-lock* which basically stops all
other threads and because of that i can't start the UCW tests from the
'test-op... removing the w-c-u solves this problem.

i suspect that it accumulates the compiler warnings/errors, but what
else does it do?

any chance of removing it?

or is it sbcl's fault that it locks that big lock?

any ideas/suggestions?

--

-- 
- attila

"- The truth is that I've been too considerate, and so became
unintentionally cruel...
 - I understand.
 - No, you don't understand! We don't speak the same language!"
(Ingmar Bergman - Smultronstället)
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
(Continue reading)

Permalink | Reply | 2 comments |
headers
Gary King | 29 Mar 2007 17:43
Favicon
Gravatar

Re: with-compilation-unit

Hi Attila,

I don't know SBCL but from where I sit (in Amherst, MA) it seems like  
wcu and the *bcl* should be separate issues.

Another thought is that this reminds me that traverse is a big ball  
of mud. IIRC, the reason traverse doesn't use method-combination is  
because CLISP didn't used to lack support for it. I wonder if this  
decision should be revisited and then perhaps wcu could be applied  
more intelligently. Of course, altering traverse is a big job (and so  
a bit scary!).

Them's my two cents!

--
Gary Warren King, metabang.com
Cell: (413) 885 9127
Fax: (206) 338-4052
gwkkwg on Skype * garethsan on AIM

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 0 comments |
headers
Robert Goldman | 29 Mar 2007 17:49
Favicon

Re: with-compilation-unit

Attila Lendvai wrote:
> dear list,
> asdf's OPERATE is wrapping basically everything in a bigWITH-COMPILATION-UNIT, even TEST-OP's.
> on sbcl it locks the *big-compiler-lock* which basically stops allother threads and because of that i
can't start the UCW tests from the'test-op... removing the w-c-u solves this problem.
> i suspect that it accumulates the compiler warnings/errors, but whatelse does it do?
> any chance of removing it?
> or is it sbcl's fault that it locks that big lock?
> any ideas/suggestions?
> -- - attila

I think that we really need it; without it we will get floods of 
spurious undefined function warnings.  Now, it might be possible to be 
more refined about the scope of the with-compilation-unit so that, for 
example, it doesn't wrap a test-op.

Do you need multi-threading in anything other than the test-op?  It 
seems kinda weird to want multi-threading in the compilation process..

> 

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
Permalink | Reply | 0 comments |

Gmane