4:24 a.m.
In trying to use the mail archive of the xemacs-patches list, I have
stumbled again upon the fact that the files from the archive can not
be applied as-is. For example, Olivier's patch:
http://list-archive.xemacs.org/xemacs-patches/200206/msg00175.html
Has the following HTML in it:
<PRE>
It was only a matter of fiddling with the includes.
I can't commit, I didn't take the time to try to find out what is
wrong.
OG.
2002-06-27 Olivier Galibert &lt;galibert(a)pobox.com&gt;
* input-method-motif.c: Fix the #includes.
</PRE>
<PRE>
Index: input-method-motif.c
===================================================================
RCS file: /pack/xemacscvs/XEmacs/xemacs/src/input-method-motif.c,v
retrieving revision 1.13
diff -u -r1.13 input-method-motif.c
--- input-method-motif.c 2002/03/13 08:52:48 1.13
+++ input-method-motif.c 2002/06/27 20:02:11
＠＠ -26,7 +26,8 ＠＠
#include <config.h>
#include <X11/Xlocale.h> /* More portable than
<locale.h> ? */
#include "lisp.h"
-#include "console-x.h"
+#include "frame-impl.h"
+#include "console-x-impl.h"
#include "device.h"
#include "frame.h"
#include "EmacsFrame.h"
</PRE>
All those "<", ">", and """ need to
be untranslated in
order to apply the patch. But since the whole thing is bound between
a <PRE> - </PRE> pair, those htmlizations are not necessary, right?
So I have 2 questions:
1. Can we fix whatever it is that makes the list-archive.xemacs.org
entries so that it doesn't htmlize the actual patches?
2. Is there an easy way to un-htmlize a file?
Thanks,
Vin
6:26 a.m.
>>>> "vin" == Vin Shelton <acs(a)xemacs.org>
writes:
vin> All those "<", ">", and
""" need to be
vin> untranslated in order to apply the patch. But since the
vin> whole thing is bound between a <PRE> - </PRE> pair, those
vin> htmlizations are not necessary, right?
The PRE element is still parsed, it just respects whitespace and
forbids certain other elements.
vin> 1. Can we fix whatever it is that makes the
vin> list-archive.xemacs.org entries so that it doesn't htmlize
vin> the actual patches?
Not easily. Currently the headers are parsed and made mousable,
references are added, and the entire body is simply enclosed in a PRE
element. I'm not sure how attachments are handled, but most patches
are not submitted as attachments.
vin> 2. Is there an easy way to un-htmlize a file?
PSGML probably has an `html-unquote' function or something like that.
You could also suck down the monthly archives, they're about 3MB each
on average.
--
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
7:33 a.m.
On Tue, 2002-07-02 at 08:26, Stephen J. Turnbull wrote:
vin> All those "<", ">", and
""" need to be
vin> untranslated in order to apply the patch. But since the
vin> whole thing is bound between a <PRE> - </PRE> pair, those
vin> htmlizations are not necessary, right?
The PRE element is still parsed, it just respects whitespace and
forbids certain other elements.
...and if the HTML escaping would not be done, it would be a security
hole [1].
vin> 1. Can we fix whatever it is that makes the
vin> list-archive.xemacs.org entries so that it doesn't htmlize
vin> the actual patches?
Not easily. Currently the headers are parsed and made mousable,
references are added, and the entire body is simply enclosed in a PRE
element. I'm not sure how attachments are handled, but most patches
are not submitted as attachments.
ISTR that attachments would work. Wonder if it would be possible to
provide a "show raw message" link, something akin to how it's handled
for example in FreeBSD lists [2]?
vin> 2. Is there an easy way to un-htmlize a file?
PSGML probably has an `html-unquote' function or something like that.
You could also suck down the monthly archives, they're about 3MB each
on average.
[1] <http://www.cert.org/advisories/CA-2000-02.html>
[2] <http://docs.freebsd.org/mail/>
--
Ville Skyttä
ville.skytta at xemacs.org
8:07 a.m.
>>>> "Ville" == Ville Skytt <Ville>
writes:
Ville> ISTR that attachments would work.
Right, I'm pretty sure they end up as .bin links.
Ville> Wonder if it would be
Ville> possible to provide a "show raw message" link, something
Ville> akin to how it's handled for example in FreeBSD lists [2]?
We already have that, but only for the first post of the month, and
only until the second post arrives. ;-) The problem is that we simply
put the whole thing into an mbox file, and wilma (the HTMLizer)
processes that. So we don't have a maildir lying around anywhere.
--
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
6:47 p.m.
On Tue, 2002-07-02 at 10:07, Stephen J. Turnbull wrote:
Ville> Wonder if it would be
Ville> possible to provide a "show raw message" link, something
Ville> akin to how it's handled for example in FreeBSD lists [2]?
We already have that, but only for the first post of the month, and
only until the second post arrives. ;-) The problem is that we simply
put the whole thing into an mbox file, and wilma (the HTMLizer)
processes that. So we don't have a maildir lying around anywhere.
:)
But what's wilma and are we using it for something? I only see
references to "MHonArc 2.4.7" in the HTML sources of various pages at
<http://list-archive.xemacs.org/>;. Oh, and MHonArch 2.4.7 is pretty
old, from 10/2000 and the current one seems to be 2.5.8 from a couple of
days ago.
Surely, it wouldn't be too tough a task to hack up a Perl CGI script
that shows a particular message from the mbox files, based on
Message-Id?
--
Ville Skyttä
ville.skytta at xemacs.org
7:10 p.m.
On Tue, Jul 02, 2002 at 04:07:45PM +0900, Stephen J. Turnbull wrote:
>>>>> "Ville" == Ville Skytt <Ville>
writes:
Ville> ISTR that attachments would work.
Right, I'm pretty sure they end up as .bin links.
Not if they're text/plain. My patches (since one has been used as the
example) are always changelog in the message, patch as an attachment
tagged text/plain.
Somebody apply them please btw if it's not done already? :-)
OG.
7:10 p.m.
>>>> "Stephen" == Stephen J Turnbull
<stephen(a)xemacs.org> writes:
Stephen> PSGML probably has an `html-unquote' function or
Stephen> something like that.
Last I looked I could not find one. So, I wrote:
;; Define the reverse of html-quote-region.
(defun html-unquote-region (begin end)
"Reverses html-quote-region. \"UnQuote\" any characters in the
region that have special HTML meanings. This converts HTML escapes
for <'s, >'s, and &'s back into their literal representation."
(interactive "r")
(save-excursion
(goto-char begin)
(while (search-forward "&" end t)
(forward-char -5)
(delete-char 5)
(insert "&")
(setq end (- end 4)))
(goto-char begin)
(while (search-forward "<" end t)
(forward-char -4)
(delete-char 4)
(insert "<")
(setq end (- end 3)))
(goto-char begin)
(while (search-forward ">" end t)
(forward-char -4)
(delete-char 4)
(insert ">")
(setq end (- end 3)))))
Gleb
5:44 a.m.
Gleb Arshinov <gleb(a)barsook.com> writes:
>>>>> "Stephen" == Stephen J Turnbull
<stephen(a)xemacs.org> writes:
Stephen> PSGML probably has an `html-unquote' function or
Stephen> something like that.
Last I looked I could not find one. So, I wrote:
[...]
This looks like it mistakenly translates "&lt;" to "<". To
avoid
that bug, handle "&" last.
Also, the docstring is incorrect. To do this right, i.e. convert
"any characters in the region that have special HTML meanings", you
should at least handle """, the numeric, and the hexadecimal
entities in the 0-255 range ("&#ddd;" and "&#xhh").
Also note that the ";" delimiter is optional in some cases; if the
entity is followed by non-alphanumerics or if it is at the end of
buffer, the ";" is optional. "1 < 2" is a valid way to write
"1 <
2" in HTML-ese.
12:48 a.m.
>>>> "Hrvoje" == Hrvoje Niksic
<hniksic(a)xemacs.org> writes:
Hrvoje> Gleb Arshinov <gleb(a)barsook.com> writes:
>>>>>> "Stephen" == Stephen J Turnbull
<stephen(a)xemacs.org>
>>>>>> writes:
>
Stephen> PSGML probably has an `html-unquote' function or
Stephen> something like that.
> Last I looked I could not find one. So, I wrote:
Hrvoje> [...]
Hrvoje> This looks like it mistakenly translates "&lt;" to
Hrvoje> "<". To avoid that bug, handle "&" last.
Yes it does. Nice catch. Thanks.
Gleb
12:01 a.m.
|--==> "VS" == Vin Shelton <acs(a)xemacs.org> writes:
VS> In trying to use the mail archive of the xemacs-patches list, I have
VS> stumbled again upon the fact that the files from the archive can not
VS> be applied as-is. For example, Olivier's patch:
VS> http://list-archive.xemacs.org/xemacs-patches/200206/msg00175.html
VS> Has the following HTML in it:
VS> <PRE>
[...]
Hi Vin!
I haven't done this in ages but what I used to do is view the archived
message in W3 and then just do 'M-x ediff-patch-file RET' from the W3
buffer. I'm pretty sure it used to work great. :-)
--
|---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---|
| XEmacs - It's not just an editor. |
| It's a way of life. |
|------------------------------------<youngs(a)xemacs.org>---|
8151
days inactive
8179
days old
11 comments
7 participants
participants (7)
-
Gleb Arshinov -
Hrvoje Niksic -
Olivier Galibert -
Stephen J. Turnbull -
Steve Youngs -
Ville Skyttä -
Vin Shelton