3:16 a.m.
Are there any plans to create the "back to the future" package, with
stuff that older XEmacsen need for forward compatibility? I've synced
up xemacs-base/easy-mmode.el with Emacs 21.3, and it works great with
current 21.5 CVS. However, older XEmacsen won't be able to use it, due
to 2 missing functions, and an older version of a 3rd.
In particular, I need to make the xemacs-base package depend on the
future package, and the future package needs to contain something like
this:
------------------------------------------------------------------------
;; This function is defined in subr.el for XEmacs >= 21.5.7
(if (not (fboundp 'propertize))
(defun propertize (string &rest properties)
"Return a copy of STRING with text properties added.
First argument is the string to copy.
Remaining arguments form a sequence of PROPERTY VALUE pairs for text
properties to add to the result."
(let ((str (copy-sequence string)))
(add-text-properties 0 (length str)
properties
str)
str)))
;; This function is defined in subr.el for XEmacs >= 21.5.12
(if (not (fboundp 'replace-regexp-in-string))
(defun replace-regexp-in-string (regexp rep string &optional
fixedcase literal subexp start)
"Replace all matches for REGEXP with REP in STRING.
Return a new string containing the replacements.
Optional arguments FIXEDCASE, LITERAL and SUBEXP are like the
arguments with the same names of function `replace-match'. If START
is non-nil, start replacements at that index in STRING.
REP is either a string used as the NEWTEXT arg of `replace-match' or a
function. If it is a function it is applied to each match to generate
the replacement passed to `replace-match'; the match-data at this
point are such that match 0 is the function's argument.
To replace only the first match (if any), make REGEXP match up to \\'
and replace a sub-expression, e.g.
(replace-regexp-in-string \"\\(foo\\).*\\'\" \"bar\" \" foo
foo\" nil nil 1)
=> \" bar foo\"
"
(let ((l (length string))
(start (or start 0))
matches str mb me)
(save-match-data
(while (and (< start l) (string-match regexp string start))
(setq mb (match-beginning 0)
me (match-end 0))
;; If we matched the empty string, make sure we advance by one char
(when (= me mb) (setq me (min l (1+ mb))))
;; Generate a replacement for the matched substring.
;; Operate only on the substring to minimize string consing.
;; Set up match data for the substring for replacement;
;; presumably this is likely to be faster than munging the
;; match data directly in Lisp.
(string-match regexp (setq str (substring string mb me)))
(setq matches
(cons (replace-match (if (stringp rep)
rep
(funcall rep (match-string 0 str)))
fixedcase literal str subexp)
(cons (substring string start mb) ; unmatched prefix
matches)))
(setq start me))
;; Reconstruct a string from the pieces.
(setq matches (cons (substring string start l) matches)) ; leftover
(apply #'concat (nreverse matches))))))
;; This function is defined in core.
;; It has the 3rd argument in XEmacs >= 21.5.16.
(when (eq (function-max-args #'pos-visible-in-window-p) 2)
(fset 'pos-visible-in-window-p
`(lambda (&optional pos window partially)
"Returns t if position POS is currently on the frame in WINDOW.
Returns nil if that position is scrolled vertically out of view.
If a character is only partially visible, nil is returned, unless the
optional argument PARTIALLY is non-nil.
POS defaults to point in WINDOW's buffer; WINDOW, to the selected window."
(funcall ,(symbol-function 'pos-visible-in-window-p) pos window))))
------------------------------------------------------------------------
Once that is in place, we can put the easy-mmode.el patch into place.
I've seen several Lisp packages now that use the new functionality, so
we need this patch before they can be ported to XEmacs.
--
Jerry James
http://www.ittc.ku.edu/~james/
4:58 p.m.
>>>> "Jerry" == Jerry James
<james(a)xemacs.org> writes:
Jerry> Are there any plans to create the "back to the future"
Jerry> package, with stuff that older XEmacsen need for forward
Jerry> compatibility? I've synced up xemacs-base/easy-mmode.el
Jerry> with Emacs 21.3, and it works great with current 21.5 CVS.
Jerry> However, older XEmacsen won't be able to use it, due to 2
Jerry> missing functions, and an older version of a 3rd.
What about moving the functionality out of core to the xemacs-base package?
Is that functionality used in core itself?
Jerry> In particular, I need to make the xemacs-base package
Jerry> depend on the future package, and the future package needs
Jerry> to contain something like this:
Jerry> ------------------------------------------------------------------------
Jerry> ;; This function is defined in subr.el for XEmacs >= 21.5.7
Jerry> (if (not (fboundp 'propertize))
Jerry> (defun propertize (string &rest properties)
Jerry> "Return a copy of STRING with text properties added.
Jerry> First argument is the string to copy.
Jerry> Remaining arguments form a sequence of PROPERTY VALUE pairs for text
Jerry> properties to add to the result."
Jerry> (let ((str (copy-sequence string)))
Jerry> (add-text-properties 0 (length str)
Jerry> properties
Jerry> str)
Jerry> str)))
Jerry> ;; This function is defined in subr.el for XEmacs >= 21.5.12
Jerry> (if (not (fboundp 'replace-regexp-in-string))
Jerry> (defun replace-regexp-in-string (regexp rep string &optional
Jerry> fixedcase literal subexp start)
Jerry> "Replace all matches for REGEXP with REP in STRING.
Jerry> Return a new string containing the replacements.
Jerry> Optional arguments FIXEDCASE, LITERAL and SUBEXP are like the
Jerry> arguments with the same names of function `replace-match'. If START
Jerry> is non-nil, start replacements at that index in STRING.
Jerry> REP is either a string used as the NEWTEXT arg of `replace-match' or a
Jerry> function. If it is a function it is applied to each match to generate
Jerry> the replacement passed to `replace-match'; the match-data at this
Jerry> point are such that match 0 is the function's argument.
Jerry> To replace only the first match (if any), make REGEXP match up to \\'
Jerry> and replace a sub-expression, e.g.
Jerry> (replace-regexp-in-string \"\\(foo\\).*\\'\"
\"bar\" \" foo foo\" nil nil 1)
Jerry> => \" bar foo\"
Jerry> "
Jerry> (let ((l (length string))
Jerry> (start (or start 0))
Jerry> matches str mb me)
Jerry> (save-match-data
Jerry> (while (and (< start l) (string-match regexp string start))
Jerry> (setq mb (match-beginning 0)
Jerry> me (match-end 0))
Jerry> ;; If we matched the empty string, make sure we advance by one char
Jerry> (when (= me mb) (setq me (min l (1+ mb))))
Jerry> ;; Generate a replacement for the matched substring.
Jerry> ;; Operate only on the substring to minimize string consing.
Jerry> ;; Set up match data for the substring for replacement;
Jerry> ;; presumably this is likely to be faster than munging the
Jerry> ;; match data directly in Lisp.
Jerry> (string-match regexp (setq str (substring string mb me)))
Jerry> (setq matches
Jerry> (cons (replace-match (if (stringp rep)
Jerry> rep
Jerry> (funcall rep (match-string 0 str)))
Jerry> fixedcase literal str subexp)
Jerry> (cons (substring string start mb) ; unmatched prefix
Jerry> matches)))
Jerry> (setq start me))
Jerry> ;; Reconstruct a string from the pieces.
Jerry> (setq matches (cons (substring string start l) matches)) ; leftover
Jerry> (apply #'concat (nreverse matches))))))
Jerry> ;; This function is defined in core.
Jerry> ;; It has the 3rd argument in XEmacs >= 21.5.16.
Jerry> (when (eq (function-max-args #'pos-visible-in-window-p) 2)
Jerry> (fset 'pos-visible-in-window-p
Jerry> `(lambda (&optional pos window partially)
Jerry> "Returns t if position POS is currently on the frame in WINDOW.
Jerry> Returns nil if that position is scrolled vertically out of view.
Jerry> If a character is only partially visible, nil is returned, unless the
Jerry> optional argument PARTIALLY is non-nil.
Jerry> POS defaults to point in WINDOW's buffer; WINDOW, to the selected
window."
Jerry> (funcall ,(symbol-function 'pos-visible-in-window-p) pos window))))
Jerry> ------------------------------------------------------------------------
Jerry> Once that is in place, we can put the easy-mmode.el patch into place.
Jerry> I've seen several Lisp packages now that use the new functionality, so
Jerry> we need this patch before they can be ported to XEmacs.
Jerry> --
Jerry> Jerry James
Jerry> http://www.ittc.ku.edu/~james/
--
Adrian Aichner
mailto:adrianï¼ xemacs.org
http://www.xemacs.org/
6:21 a.m.
>>>> "APA" == Adrian Aichner
<adrian(a)xemacs.org> writes:
APA> What about moving the functionality out of core to the
APA> xemacs-base package?
This will surely cause hard-to-explain bugs in older installations as
the xemacs-base version diverges from the old core version. (I think
in this case the syntax error will be easy to interpret, but suppose
later changes involve an incorrect result, not a syntax error?)
The point of the "back-to-the-future" package is to ensure warnings or
errors when the packaged version is unavailable.
--
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.
7:46 a.m.
"Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
This will surely cause hard-to-explain bugs in older installations
as
the xemacs-base version diverges from the old core version. (I think
in this case the syntax error will be easy to interpret, but suppose
later changes involve an incorrect result, not a syntax error?)
The point of the "back-to-the-future" package is to ensure warnings or
errors when the packaged version is unavailable.
Do we need a separate package for this or could we stick these things
into xemacs-base which has versions too? If I'm not mistaken
xemacs-base would depend on the b-t-t-f package, wouldn't it? This
implies, sort of, that the user would have to download three packages
in order to get the whole thing rolling.
But my mind might not grasp the full picture of it.
norbert.
9:41 a.m.
>>>> "viteno" == Norbert Koch
<viteno(a)xemacs.org> writes:
viteno> Do we need a separate package for this or could we stick
viteno> these things into xemacs-base which has versions too?
We currently don't have a way to depend on package versions, and
21.4 will never get one, except via (require 'back-to-the-future). :-)
--
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.
12:15 p.m.
"Stephen J. Turnbull" <stephen(a)xemacs.org> writes:
We currently don't have a way to depend on package versions, and
21.4 will never get one, except via (require 'back-to-the-future). :-)
:-) Okay, so that's settled. Jerry, is there anything I can/should do
to pave the path?
norbert.
3:37 a.m.
Norbert Koch <viteno(a)xemacs.org> wrote:
:-) Okay, so that's settled. Jerry, is there anything I
can/should do
to pave the path?
We have to decide what kind of content the back-to-the-future package
will contain, and how it is structured. Stephen suggested that it
contain forms that cause "you aren't recent enough" errors if older
XEmacsen try to use newer functionality. That's great, but I think we
should provide that newer functionality when possible. That's the
upshot of the request that started this thread.
I propose that the future package contain one future-x.el file for each
x.el file containing a future function. For two of the functions I am
interested in, that would mean a future-subr.el file. We should also
have a core.el or future-core.el file for changes in core. That would
take care of the third function I'm concerned with.
What do you think?
--
Jerry James
http://www.ittc.ku.edu/~james/
5:51 a.m.
Jerry James <james(a)xemacs.org> writes:
I propose that the future package contain one future-x.el file for
each
x.el file containing a future function. For two of the functions I am
interested in, that would mean a future-subr.el file. We should also
have a core.el or future-core.el file for changes in core. That would
take care of the third function I'm concerned with.
What do you think?
From my point of view this sounds like a good approach. If nobody
else objects to the idea, send me the file(s) you already have for
this package and I try to get them packed up and distributed.
norbert.
10:53 a.m.
Moving to xemacs-design. Reply-To set.
>>>> "viteno" == Norbert Koch
<viteno(a)xemacs.org> writes:
viteno> From my point of view this sounds like a good approach.
viteno> If nobody else objects to the idea, send me the file(s)
viteno> you already have for this package and I try to get them
viteno> packed up and distributed.
Please don't. It's really important that we get this either right the
first time, or wrong in a way that doesn't screw the users no matter
what we do to fix it.
We really should think about what problems we're trying to solve, what
APIs we want to provide, etc. Is it really enough to have a future-x
file? What do we do for 21.4 when 22.0 has x-foo-new and 22.2 has a
changed version of x-foo-new and a new function x-bar-new? Do you
realize that you (I mean "Norbert") may have to build separate
versions of these things for different old versions of XEmacs (21.4
v. 22.0)? This has been tried before (the APEL package). It's no
accident that "APEL" is a 4-letter word!
--
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.
7456
days inactive
7462
days old
8 comments
4 participants
participants (4)
-
Adrian Aichner -
Jerry James -
Norbert Koch -
Stephen J. Turnbull