gcl-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gcl-devel] Michael Koehne's ANSI pathname patches


From: Camm Maguire
Subject: [Gcl-devel] Michael Koehne's ANSI pathname patches
Date: 24 Jul 2004 11:30:18 -0400
User-agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.2

Greetings!  Just a short note -- I committed your wonderful work into
CVS head as is, as I ran out of time to go over it thoroughly and
didn't want to hold back the project before my week's departure.
Thank you again!

Most things still work.  Maxima cvs and axiom builds are broken
temporarily due to apparently minor pathname issues.  Perhaps some of
you might look into this while I'm gone.  Otherwise we can clean up a
bit when I return in a week's time.

Am uploading a Debian gclcvs package too with preliminary support for
common-lisp-controller.  Haven't tested it yet, and issues doubtlessly
remain, but its a start.

Please accept my apologies for the delay.  I hope your ssh key
registration process is unimpeded.

Take care,

Michael Koehne <address@hidden> writes:

> Moin Camm, Paul & Thomas,
> 
> > Please let me assure you that your valuable work will not be wasted.
> 
>   take a look at the <b>bad news</b> section at www.copyleft.de/lisp/
> 
>   I fear that I have to declare the project "application level of ANSI
>   compatibilty for GCL" as failed to deliver a prototype after 3 month of
>   work, if I need to increase the size of the reverse patch.  This would
>   generate an offspring with a total user base of one person only.
> 
>   The patches are designed to go into the base of GCL, and seldom to
>   become stacked on top. Future plans would be, complete on formatter
>   and pprint (also inside the base of GCL and not on top) and integrate
>   condition into the base to get rid of ./clcs/ in a big error cleanup.
>   A final dream would be to provide some clos that better fits to GCL
>   and to drop ./pcl/, so we would have an ANSI GCL that as small and
>   fast as the current CLtL1 GCL.
> 
>   You are still free to take parts of my patch, as they are under LGPL.
>   But I wont be able to maintain the reverse patches for long, so I
>   might have to drop further progress of this project.
> 
>   Alternate:
> 
>   Reverse your patch - apply my patch - build - do an ansi test.
> 
>       vi ansi-tests/tests.out
>         :$
>         ?failed
>         ma
>         }
>         ?\.$
>         :'a,.w ansi-tests/tests.sum.1
>         :q
>       vi ansi-tests/tests.sum.1
>         :%s/<control-v><control-m>/ /g
>         :g/^$/d
>         :%!sort
>         :wq
> 
>   Next `sh install-clc-gcl.sh` (you'll need usr-lib-common-lisp-bin-gcl.sh
>   in same directory) either as root or with sudo rights. This script might
>   suggest additional Debian/Sarge packages. Those suggestions should be
>   copyied into the Debian control file of the next gclcvs.
> 
>   Do a diff, and install that diff into a fresh checkout, next move things
>   that are crying to become moved, e.g. with-compilation-unit, and perhaps
>   even reinstall the dangerous string-match speedup, to rebuild and check
>   ansi-tests (i hope the cryptic vi commands come handy - i have them on
>   function keys so i just need to type the file name ;) 
> 
>   Commit first my patch, and next your modification, if both versions work
>   in terms that your version give same or less number of ansi-tests failures,
>   and is able to run the same Debian applications with same or less problems.
> 
>   But do not break my patch into pieces, to install broken patches, I would
>   need to revert, as they break my patch. Your mod of is placed on top of my
>   with-standard-io-syntax, while commenting out half of the let. With the
>   result that the C level part of with-standard-io-sytax wont be able to
>   work, if I dont revert your patch first.
> 
>   Next create a way, that I'm able to maintain my patches on a daily base.
> 
>   The complete discussion and idle wait for commit prevented effective
>   work on pprint for nearly a week (the only thing that improved here was
>   a pprint-dispatch bug discovered by test.8 ;) so f2cl does'nt work yet.
> 
> Bye Michael
> -- 
>   mailto:address@hidden             UNA:+.? 'CED+2+:::Linux:2.4.22'UNZ+1'
>   http://www.xml-edifact.org/           CETERUM CENSEO WINDOWS ESSE DELENDAM
> 
> 
> 

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




reply via email to

[Prev in Thread] Current Thread [Next in Thread]