2:30 p.m.
Stephen J. Turnbull wrote:
FKtPp writes:
> Now I want to make the code work correct under XEmacs. But after some
> edebug work, I find out that nxml-mode code using a kind of never
> documented char_class looks like:
>
> (name-start-char "[_[:alpha:]]")
> (name-continue-not-start-char "[-.[:digit:]]")
> (name-continue-char "[-._[:alnum:]]")
>
> And after chat with <kehoea> in #xemacs I know that, there's a very
> easy way to support this kind of thing.
In search.c, line 179 (or so), just change "RE_SYNTAX_EMACS" to
"RE_SYNTAX_EMACS | RE_CHAR_CLASSES". Rebuild. Please let us know how
it works out.
Good news~~ The recompiled XEmacs binary works! And if I make change in
regex.h It works also!
AFAIK this only will work as expected for ASCII files and search
expressions. In a Mule build there are likely to be crashes. For
those reasons, I would not expect this to appear in the mainline
without a lot of work. Definitely we will need testing.
Well, I am looking forward for the this to happen~~
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
4:24 a.m.
New subject: [Bug: 21.5-b28] Could you please turn on `RE_CHAR_CLASSES' for XEmacs?
* It's me FKtPp ;) <m_pupil(a)yahoo.com.cn> writes:
Stephen J. Turnbull wrote:
> In search.c, line 179 (or so), just change "RE_SYNTAX_EMACS" to
> "RE_SYNTAX_EMACS | RE_CHAR_CLASSES". Rebuild. Please let us know how
> it works out.
Good news~~ The recompiled XEmacs binary works! And if I make change
in
regex.h It works also!
> AFAIK this only will work as expected for ASCII files and search
> expressions. In a Mule build there are likely to be crashes. For
> those reasons, I would not expect this to appear in the mainline
> without a lot of work. Definitely we will need testing.
We enabled character classes in SXEmacs 22.1.8 (Mule is enable by
default in SXEmacs), so far no related crashes. Steve, if you can give
me some hairy test cases, I'll be happy to let you know how SXEmacs
stands up to it.
--
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
| SXEmacs - The only _______ you'll ever need. |
| Fill in the blank, yes, it's THAT good! |
|------------------------------------<steve(a)sxemacs.org>---|
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
7:24 a.m.
New subject: [Bug: 21.5-b28] Could you please turn on `RE_CHAR_CLASSES' for XEmacs?
Steve Youngs wrote:
* It's me FKtPp ;) <m_pupil(a)yahoo.com.cn> writes:
> Stephen J. Turnbull wrote:
>> In search.c, line 179 (or so), just change "RE_SYNTAX_EMACS" to
>> "RE_SYNTAX_EMACS | RE_CHAR_CLASSES". Rebuild. Please let us know
how
>> it works out.
> Good news~~ The recompiled XEmacs binary works! And if I make change in
> regex.h It works also!
>> AFAIK this only will work as expected for ASCII files and search
>> expressions. In a Mule build there are likely to be crashes. For
>> those reasons, I would not expect this to appear in the mainline
>> without a lot of work. Definitely we will need testing.
We enabled character classes in SXEmacs 22.1.8 (Mule is enable by
default in SXEmacs), so far no related crashes. Steve, if you can give
me some hairy test cases, I'll be happy to let you know how SXEmacs
stands up to it.
Got some trouble now. :(
I hear that SXEmacs sharing the same packages repository with XEmacs, so
SXEmacs must have a special character class [:word:] defined. Or how
can it handle this, while viewing articles in gnus:
# bind (standard-output stack-trace-on-signal debug-on-signal
stack-trace-on-error debug-on-error)
# (unwind-protect ...)
re-search-forward("\\b\\(\\(www\\.\\|\\(s?https?\\|ftp\\|file\\|gopher\\|nntp\\|news\\|telnet\\|wais\\|mailto\\|info\\):\\)\\(//[-a-z0-9_.]+:[0-9]*\\)?[-a-z0-9_=!?#$＠~%&*+\\/:;.,[:word:]]+[-a-z0-9_=#$＠~%&*+\\/[:word:]]\\)"
nil t)
# bind (inhibit-read-only inhibit-point-motion-hooks case-fold-search
alist beg entry regexp)
# (unwind-protect ...)
# bind (force)
gnus-article-add-buttons()
# (unwind-protect ...)
# bind (--dolist-temp--98428 elem length alist
article-goto-body-goes-to-point-min-p treated-type highlightp val elem
type total-parts part-number condition)
gnus-treat-article(nil 1 1)
# (unwind-protect ...)
# bind (handles inhibit-read-only handle)
# (unwind-protect ...)
# bind (ssw98424)
# (unwind-protect ...)
# bind (ihandles)
gnus-display-mime()
# bind (gnus-article-buffer buffer-read-only inhibit-read-only)
gnus-article-prepare-display()
# (unwind-protect ...)
# bind (result group gnus-tmp-internal-hook summary-buffer gnus-article)
# (unwind-protect ...)
# bind (header all-headers article)
gnus-article-prepare(89684 nil)
# bind (all-header article)
gnus-summary-display-article(89684 nil)
# (unwind-protect ...)
# bind (article all-headers gnus-summary-display-article-function
article pseudo force all-headers)
gnus-summary-select-article(nil nil pseudo)
# bind (lines)
gnus-summary-scroll-up(1)
# bind (command-debug-status)
call-interactively(gnus-summary-scroll-up)
(dispatch-event "[internal]")
# (condition-case ... . error)
# (catch top-level ...)
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
4:32 a.m.
New subject: [Bug: 21.5-b28] Could you please turn on `RE_CHAR_CLASSES' for XEmacs?
* It's me FKtPp ;) <m_pupil(a)yahoo.com.cn> writes:
I hear that SXEmacs sharing the same packages repository with
XEmacs,
so SXEmacs must have a special character class [:word:] defined. Or
how can it handle this, while viewing articles in gnus:
Yes, SXEmacs has character classes, but because (featurep 'xemacs) => t
in SXEmacs, reading articles in Gnus is the same as if you were using
XEmacs.
--
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
| SXEmacs - The only _______ you'll ever need. |
| Fill in the blank, yes, it's THAT good! |
|------------------------------------<steve(a)sxemacs.org>---|
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
5:03 p.m.
New subject: [Bug: 21.5-b28] Could you please turn on `RE_CHAR_CLASSES' for XEmacs?
On Mon, Feb 25 2008, Steve Youngs wrote:
* It's me FKtPp ;) <m_pupil(a)yahoo.com.cn> writes:
> I hear that SXEmacs sharing the same packages repository with XEmacs,
> so SXEmacs must have a special character class [:word:] defined. Or
> how can it handle this, while viewing articles in gnus:
Yes, SXEmacs has character classes, but because (featurep 'xemacs) => t
in SXEmacs, reading articles in Gnus is the same as if you were using
XEmacs.
The relevant code in Gnus doesn't test (featurep 'xemacs) but...
(string-match "[[:digit:]]" "1") ;; support POSIX?
So if in SXEmacs this returns non-nil, `gnus-button-url-regexp's value
will be like in Emacs.
Bye, Reiner.
--
,,,
(o o)
---ooO-(_)-Ooo--- | PGP key available | http://rsteib.home.pages.de/
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
10:06 p.m.
New subject: [Bug: 21.5-b28] Could you please turn on `RE_CHAR_CLASSES' for XEmacs?
* Reiner Steib <reinersteib+gmane(a)imap.cc> writes:
On Mon, Feb 25 2008, Steve Youngs wrote:
> * It's me FKtPp ;) <m_pupil(a)yahoo.com.cn> writes:
>
> > I hear that SXEmacs sharing the same packages repository with XEmacs,
> > so SXEmacs must have a special character class [:word:] defined. Or
> > how can it handle this, while viewing articles in gnus:
>
> Yes, SXEmacs has character classes, but because (featurep 'xemacs) => t
> in SXEmacs, reading articles in Gnus is the same as if you were using
> XEmacs.
The relevant code in Gnus doesn't test (featurep 'xemacs)
but...
(string-match "[[:digit:]]" "1") ;; support
POSIX?
So if in SXEmacs this returns non-nil, `gnus-button-url-regexp's
value
will be like in Emacs.
Indeed...
gnus-button-url-regexp
=>
"\\b\\(\\(www\\.\\|\\(s?https?\\|ftp\\|file\\|gopher\\|nntp\\|news\\|telnet\\|wais\\|mailto\\|info\\):\\)\\(//[-a-z0-9_.]+:[0-9]*\\)?\\(?:[-a-z0-9_=#$＠~%&*+\\/[:word:]!?:;.,]+([-a-z0-9_=#$＠~%&*+\\/[:word:]!?:;.,]+[-a-z0-9_=#$＠~%&*+\\/[:word:]]*)[-a-z0-9_=#$＠~%&*+\\/[:word:]]\\|[-a-z0-9_=#$＠~%&*+\\/[:word:]!?:;.,]+[-a-z0-9_=#$＠~%&*+\\/[:word:]]\\)\\)"
Looking back through the changelog, we had to add `:word:' to get Gnus
working again.
--
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
| SXEmacs - The only _______ you'll ever need. |
| Fill in the blank, yes, it's THAT good! |
|------------------------------------<steve(a)sxemacs.org>---|
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
9:40 a.m.
New subject: [Bug: 21.5-b28] Could you please turn on `RE_CHAR_CLASSES' for XEmacs?
Steve Youngs writes:
We enabled character classes in SXEmacs 22.1.8 (Mule is enable by
default in SXEmacs), so far no related crashes. Steve, if you can give
me some hairy test cases, I'll be happy to let you know how SXEmacs
stands up to it.
Well, on second thought I take the worries about crashes back. These
are just character classes, same as any other.
The main question is how the character classes are constructed in Mule.
Eg, you would hope that "^[[:word:]]+ [[:word:]]+$" would match
"Hrvoje Nikšić" and
"^半田 健一$", as well as "Steve Youngs".
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
4:12 a.m.
New subject: [Bug: 21.5-b28] Could you please turn on `RE_CHAR_CLASSES' for XEmacs?
* Stephen J Turnbull <stephen(a)xemacs.org> writes:
Steve Youngs writes:
> We enabled character classes in SXEmacs 22.1.8 (Mule is enable by
> default in SXEmacs), so far no related crashes. Steve, if you can give
> me some hairy test cases, I'll be happy to let you know how SXEmacs
> stands up to it.
Well, on second thought I take the worries about crashes back.
These
are just character classes, same as any other.
The main question is how the character classes are constructed in
Mule.
Eg, you would hope that "^[[:word:]]+ [[:word:]]+$" would
match
"Hrvoje Nikšić" and "^半田 健一$", as well as "Steve Youngs".
Yeah, you'd hope, but unfortunately it fails miserably and only matches
"Steve Youngs". I guess we need to work on that.
It would be interesting to learn how GNU/Emacs copes with that
situation. Can somebody with a recent Emacs let me know?
--
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
| SXEmacs - The only _______ you'll ever need. |
| Fill in the blank, yes, it's THAT good! |
|------------------------------------<steve(a)sxemacs.org>---|
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
5:38 a.m.
New subject: [Bug: 21.5-b28] Could you please turn on `RE_CHAR_CLASSES' for XEmacs?
Steve Youngs writes:
It would be interesting to learn how GNU/Emacs copes with [Mule and
POSIX character classes]. Can somebody with a recent Emacs let me
know?
I'll bet it works correctly, they completely redid the regex engine
for 21 or 22. It probably still uses the same API, though, so it
might be pretty port-able.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
4:48 a.m.
New subject: [Bug: 21.5-b28] Could you please turn on `RE_CHAR_CLASSES' for XEmacs?
* Stephen J Turnbull <stephen(a)xemacs.org> writes:
Steve Youngs writes:
> It would be interesting to learn how GNU/Emacs copes with [Mule and
> POSIX character classes]. Can somebody with a recent Emacs let me
> know?
I'll bet it works correctly, they completely redid the regex
engine
for 21 or 22. It probably still uses the same API, though, so it
might be pretty port-able.
How much was that bet? It gave the same results as SXEmacs... nil on
Hrvoje's name and the Japanese, 0 on "Steve Youngs" (Emacs CVS HEAD
23.0.something).
--
|---<Steve Youngs>---------------<GnuPG KeyID: A94B3003>---|
| SXEmacs - The only _______ you'll ever need. |
| Fill in the blank, yes, it's THAT good! |
|------------------------------------<steve(a)sxemacs.org>---|
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
3:54 p.m.
New subject: [Bug: 21.5-b28] Could you please turn on `RE_CHAR_CLASSES' for XEmacs?
Steve Youngs writes:
How much was that bet?
I owe you a beer next time we meet, your place or mine. :-( Not that
I object to to buying you a beer, but I'm shocked, SHOCKED, I say....
It gave the same results as SXEmacs... nil on Hrvoje's name and
the
Japanese, 0 on "Steve Youngs" (Emacs CVS HEAD 23.0.something).
How sad. Can I hope I misunderstood the standard ....
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
6113
days inactive
6118
days old
10 comments
4 participants
participants (4)
-
FKtPp -
Reiner Steib -
Stephen J. Turnbull -
Steve Youngs