2:20 p.m.
"Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
>>>>> "Simon" == Simon Josefsson
<jas(a)extundo.com> writes:
Simon> This makes blowfish generate the same values under XEmacs
Simon> as the previous (XEmacs incompatible) version does under
Simon> Emacs, for me. Does this help, Karl?
I prefer the separate versions, but there is a hack in mule-ucs which
uses four character strings of hex digits, and converts them from hex
strings to integers using a home-grown hex-to-int function.
If Emacs doesn't add the CL hex constant stuff, that is probably the
best solution, but let's hope Emacs support CL hex constants. (Maybe
by starting to use it, someone will feel the pressure to implement it.)
9:48 p.m.
New subject: [PATCH] ecrypto/blowfish.el integer constant fix
If Emacs doesn't add the CL hex constant stuff, that is probably
the
best solution, but let's hope Emacs support CL hex constants. (Maybe
by starting to use it, someone will feel the pressure to implement it.)
Try Emacs-21 one day. You might like it ;-)
Stefan
11:19 p.m.
New subject: [PATCH] ecrypto/blowfish.el integer constant fix
"Stefan Monnier"
<monnier+gmane.emacs.xemacs.beta/news/(a)cs-www.cs.yale.edu> writes:
> If Emacs doesn't add the CL hex constant stuff, that is
probably the
> best solution, but let's hope Emacs support CL hex constants. (Maybe
> by starting to use it, someone will feel the pressure to implement it.)
Try Emacs-21 one day. You might like it ;-)
It works! :-) OK, then the patch that went in really was the right
thing all things considered.
I wonder if the blowfish code generate the same encryptions as other
blowfish implementations though...
1:37 a.m.
New subject: [PATCH] ecrypto/blowfish.el integer constant fix
>>>> "Simon" == Simon Josefsson
<jas(a)extundo.com> writes:
Simon> I wonder if the blowfish code generate the same encryptions
Simon> as other blowfish implementations though...
Regression test! Regression test!
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
2:11 p.m.
New subject: [PATCH] ecrypto/blowfish.el integer constant fix
"Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
>>>>> "Simon" == Simon Josefsson
<jas(a)extundo.com> writes:
Simon> I wonder if the blowfish code generate the same encryptions
Simon> as other blowfish implementations though...
Regression test! Regression test!
Hm, is there a regression test infrastructure for elisp packages?
That would be neat.
In any case, I verified manually that key=0 and plaintext=0 generated
the correct ciphertext, so it is probably close to be in a good shape.
2:42 p.m.
New subject: [PATCH] ecrypto/blowfish.el integer constant fix
>>>> "Simon" == Simon Josefsson
<jas(a)extundo.com> writes:
Simon> Hm, is there a regression test infrastructure for elisp
Simon> packages? That would be neat.
tests/automated/test-harness.el
I plan to package it; Martin thinks it's a bad idea, but I don't
really understand what he's talking about.
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
5:50 p.m.
New subject: [PATCH] ecrypto/blowfish.el integer constant fix
"Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
>>>>> "Simon" == Simon Josefsson
<jas(a)extundo.com> writes:
Simon> Hm, is there a regression test infrastructure for elisp
Simon> packages? That would be neat.
tests/automated/test-harness.el
It seem to be for core only, but I guess it could be used for the
packages too.
I plan to package it; Martin thinks it's a bad idea, but I
don't
really understand what he's talking about.
The only bad about it I can see is that it takes away time from more
important things, but that appears to hold for everything we do here
anyway, so I guess I don't see the problem either.
FWIW, when I suggested a test suite for Emacs to RMS several years ago
he didn't like it either, because he felt that emacs went through a
self-test during bootstrap. I guess I didn't explain carefully enough
that bootstrapping doesn't test for regression in elisp package APIs.
Having a self test for, e.g., the mail standard stuff in Gnus would be
neat to have when people propose large modifications... OTOH, this
sounds like work.
4:39 a.m.
New subject: [PATCH] ecrypto/blowfish.el integer constant fix
>>>> "Simon" == Simon Josefsson
<jas(a)extundo.com> writes:
Simon> The only bad about it I can see is that it takes away time
Simon> from more important things,
There is nothing more important than testing. Among other things,
there are a number of areas where the best documentation we have for
certain counter-intuitive (to me, anyway) functionality is a test case
explicitly written to make sure XEmacs works that way.
It's also a good way to get bugs fixed. I put some tests in, the
Voice of Dog (that's God in Australian, which, if not backwards, is
upside-down) said "THERE WILL BE NO RELEASE UNTIL THOU SHALT REPENT OF
ALL TEST FAILURES," and by golly, the tests failures miraculously
disappeared (and not because the tests were disabled, either!)
Simon> FWIW, when I suggested a test suite for Emacs to RMS
Simon> several years ago he didn't like it either, because he felt
Simon> that emacs went through a self-test during bootstrap.
That's not surprising. rms doesn't believe in engineering. He's a
hacker to the bone.
But he's just plain wrong, at least if you have Ben Wing and Mike
Sperber keeping things stirred up. Having the test suite to keep us
honest has helped us return to stability more quickly after several
big changes, in my somewhat informed opinion.
Now, if I can come up with some tests for specifier behavior and
Crushtomize....
--
Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp
University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Ask not how you can "do" free software business;
ask what your business can "do for" free software.
7:51 a.m.
New subject: [PATCH] ecrypto/blowfish.el integer constant fix
>>>> Stephen wrote:
Stephen> That's not surprising. rms doesn't believe in engineering.
Stephen> He's a hacker to the bone.
Hackers seems to like testing if you disguise it as part of doing
XP. ;-)
Does this mean that we are now ready, and maybe even encouraged ;-),
to set up an automatic smoketest including test for XEmacs-21.5. You
know the cycle: new things in cvs, checkout, build, test, report?
Yours
--
%% Mats
7824
days inactive
7826
days old
8 comments
4 participants
participants (4)
-
Mats Lidell -
Simon Josefsson -
Stefan Monnier -
Stephen J. Turnbull