10:56 a.m.
Hi,
Because of some build problems I looked into what files there are in
packages in CVS and found that I have at least five different md5.el.
----------------------------------------------------------------------
./xemacs-packages/gnus/gnus/lisp/md5.el
./xemacs-packages/w3/lisp/md5.el
./xemacs-packages/ecrypto/md5.el
./unsupported/simon/epop3/md5.el
./unsupported/simon/flim/md5.el
----------------------------------------------------------------------
This doesn't seem to be a good idea. What's the story?
Yours
--
%% Mats
11:21 a.m.
>>>> "Mats" == Mats Lidell
<matsl(a)contactor.se> writes:
Mats> Because of some build problems I looked into what files
Mats> there are in packages in CVS and found that I have at least
Mats> five different md5.el.
Mats> This doesn't seem to be a good idea. What's the story?
I'm not sure what the story is on ecrypto. The Gnus version seems to
be the "canonical" one. None of the others get installed; they are in
CVS because we try to keep our sources as close to the original as
possible to simplify maintenance.
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
My nostalgia for Icon makes me forget about any of the bad things. I don't
have much nostalgia for Perl, so its faults I remember. Scott Gilbert c.l.py
noon
>>>> Stephen wrote:
Stephen> I'm not sure what the story is on ecrypto. The Gnus version
Stephen> seems to be the "canonical" one. None of the others get
Stephen> installed; they are in CVS because we try to keep our sources
Stephen> as close to the original as possible to simplify maintenance.
For the record. I have this in the install tree.
----------------------------------------------------------------------
$ find -name 'md5.elc'
./xemacs-packages/lisp/gnus/md5.elc
./xemacs-packages/lisp/ecrypto/md5.elc
----------------------------------------------------------------------
But maybe that is what you are saying about not knowing about
ecrypto.
Yours
--
%% Mats
12:16 p.m.
"Stephen J. Turnbull" <stephen(a)xemacs.org> wrote:
>>>>> "Mats" == Mats Lidell
<matsl(a)contactor.se> writes:
Mats> Because of some build problems I looked into what files
Mats> there are in packages in CVS and found that I have at least
Mats> five different md5.el.
Mats> This doesn't seem to be a good idea. What's the story?
I'm not sure what the story is on ecrypto. The Gnus version seems to
be the "canonical" one. None of the others get installed; they are in
CVS because we try to keep our sources as close to the original as
possible to simplify maintenance.
Another weird thing is that `md5' is provided as a builtin and md5.c
provides the md5 feature in both Xemacs and Emacs. So you cannot
(require 'md5) to get the other md5- functions from the Rees package.
If you (load "md5") then it will clobber the builtin :-/
--
John Paul Wallington
1:10 p.m.
Mats Lidell <matsl(a)contactor.se> writes:
Hi,
Because of some build problems I looked into what files there are in
packages in CVS and found that I have at least five different md5.el.
----------------------------------------------------------------------
./xemacs-packages/gnus/gnus/lisp/md5.el
./xemacs-packages/w3/lisp/md5.el
./xemacs-packages/ecrypto/md5.el
IMHO all but the ecrypto one should be removed and other packages
should depend on ecrypto for md5.el. Ecrypty doesn't depend on any
other packages, so requiring it is pretty lightweight. If there is
any problems with the ecrypto md5*.el, let's fix it.
./unsupported/simon/epop3/md5.el
./unsupported/simon/flim/md5.el
This is just junk, ignore them.
John Paul Wallington <jpw(a)shootybangbang.com> writes:
Another weird thing is that `md5' is provided as a builtin and
md5.c
provides the md5 feature in both Xemacs and Emacs.
Not with all XEmacs versions that the packages are supposed to work
with, I think.
So you cannot (require 'md5) to get the other md5- functions
from
the Rees package.
(require 'md5) should be all that you need IMHO. Why would you want
the Rees package?
If you (load "md5") then it will clobber the builtin :-/
Then only people that really want the Rees package should use it, and
in that case they get what they asked for.
7:03 p.m.
On Mon, 2002-09-16 at 15:10, Simon Josefsson wrote:
IMHO all but the ecrypto one should be removed and other packages
should depend on ecrypto for md5.el. Ecrypty doesn't depend on any
other packages, so requiring it is pretty lightweight. If there is
any problems with the ecrypto md5*.el, let's fix it.
Sounds good to me. Simon, ISTR you said something about working with
other packages so that they would use md5 from ecrypto, have you taken
(or are you planning to take) a look at the Gnus case? I think you've
already removed the other duplicates...
--
\/ille Skyttä
ville.skytta at xemacs.org
9:20 a.m.
>>>> Ville wrote:
Ville> Sounds good to me. Simon, ISTR you said something about
Ville> working with other packages so that they would use md5 from
Ville> ecrypto, have you taken (or are you planning to take) a look at
Ville> the Gnus case? I think you've already removed the other
Ville> duplicates...
Resent versions of gnus does not contain md5.el. I don't know the
policy for updating the packages but simply moving to a newer
(possibly beta) version of gnus would take care of this particular
problem.
Yours
--
%% Mats
7913
days inactive
7914
days old
6 comments
5 participants
participants (5)
-
John Paul Wallington -
Mats Lidell -
Simon Josefsson -
Stephen J. Turnbull -
Ville Skyttä