Re: [Gcl-devel] delayed pathname.d patch

[Camm Maguire]

Greetings!

Michael Koehne <address@hidden> writes:

> Moin Mike Thomas,
> 
> > Whether this patch is really the cause or it has just triggered the strange
> > path behaviour seen in the Maxima build I don't know, but the build now
> > fails trying to compile pcl/pcl_pkg.lsp because the C compiler is fed a path
> > with no device (C:) as follows:
> 
>   I wonder how this patch sneaked out - impossile, as I delayed make-pathname
>   to get asdf up patch, and told Camm, that pathname.d is a big can of worms.
> 

Its actually not your patch, which I've not yet received, and which I
agree with you should be kept out of stable.  I did decide to fix that
bug you found -- the one you thought was critical, and I agree -- this
did not break the windows build, but rather the *features* change, but
only in a very simple and easily fixable way.  All should be well now,
or at least back to the same old mysteries.  Would appreciate
confirmation.

Take care,

>   My current pathname.d has a complete rewrite of parse-namesting and
>   merge-namesting and need an other overhaul before I attack make-pathname
>   again. I'm currently even thinking about implementing translate-pathname
>   at a low level - valid pathname strings are now :
> 
>   Lisp like
>   [host   ':'] [';'] [directory ';']* [name ['.' type ['.' version]]]
>   RCP like
>   [host   ':'] ['/'] [directory '/']* [name ['.' type ['.' version]]]
>   DOS like
>   [device ':'] ['/'] [directory '/']* [name ['.' type ['.' version]]]
>   URL like
>   [device ':'] ['//' host '/'] [directory '/']* [name ['.' type ['.' 
> version]]]
> 
>   I only know two systems who need device names: VM and some other.
>   I dont have any Lisp under VM and I avoid to touch the other system.
> 
>   Both have single characters as devices, e.g: #P"C:FOO.LISP" or
>   #P"FOO LISP C". Both systems provide own facilities to ACCESS a
>   local or network drive or directory as a device. (ACCESS 191 A
>   on VM/370) I therefore deceided that single character string in
>   RCP/DOS notations are devices, while longer are hosts.
>   
>   The ANSI standard does'nt tell how to map devices to a
>   system that does not have such a concept. So #P"C:FOO.LISP"
>   is untested, currently. But I plan to use translate-pathname
>   for LOGICAL-DEVICENAME-TRANSLATIONS similar to logical-pathname.
> 
>   I could mail you my next version, to check if they are able to
>   build a gcl and your favourite software under the 'other os',
>   that I avoid to touch.
> 
> Bye Michael
> -- 
>   mailto:address@hidden             UNA:+.? 'CED+2+:::Linux:2.4.22'UNZ+1'
>   http://www.xml-edifact.org/           CETERUM CENSEO WINDOWS ESSE DELENDAM
> 
> 
> _______________________________________________
> Gcl-devel mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/gcl-devel
> 
> 
> 

-- 
Camm Maguire                                            address@hidden
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah