9:31 a.m.
In package-admin-find-top-directory:
(while path-list
(if (equal (substring (car path-list) -16) "xemacs-packages/")
(setq top-dir (car path-list)))
(setq path-list (cdr path-list))))
((eq type 'mule)
(while path-list
(if (equal (substring (car path-list) -14) "mule-packages/")
(setq top-dir (car path-list)))
(setq path-list (cdr path-list)))))))
But on win32, the path-list returned has '\\' as path separator. So that
just can't match. Possible fix attached.
Other related questions:
- what is the status of 'mule' packages wrt to 21.5+Mule? I guess
mule-ucs and mule-base packages are not needed anymore (integrated). But
what about the others (locale, latin-unity ...) and what about their
location? Shouldn't they move into xemacs-packages/ in 21.5 since Mule
is considered to be the default?
--
Fabrice
C:\>diff -u xemacs\xemacs-21.5\lisp\package-admin.el
xemacs\xemacs-21.5\lisp\package-admin.el
--- xemacs\xemacs-21.5\lisp\package-admin.el Mon Mar 24 17:43:55 2003
+++ xemacs\xemacs-21.5\lisp\package-admin.el Wed Apr 09 10:29:54 2003
＠＠ -170,21 +170,21 ＠＠
;; First, check the environment var.
(if env-value
(let ((path-list (paths-decode-directory-path env-value 'drop-empties)))
(cond ((eq type 'std)
(while path-list
- (if (equal (substring (car path-list) -16)
"xemacs-packages/")
+ (if (string-match "xemacs-packages[/\\]$" (car path-list))
(setq top-dir (car path-list)))
(setq path-list (cdr path-list))))
((eq type 'mule)
(while path-list
- (if (equal (substring (car path-list) -14) "mule-packages/")
+ (if (string-match "mule-packages[/\\]$" (car path-list))
(setq top-dir (car path-list)))
(setq path-list (cdr path-list)))))))
;; Wasn't in the environment, try `user-init-directory' if
;; USER-DIR is non-nil.
(if (and user-dir
(not top-dir))
(cond ((eq type 'std)
(setq top-dir (file-name-as-directory
(expand-file-name "xemacs-packages"
user-init-directory))))
((eq type 'mule)
＠＠ -195,14 +195,14 ＠＠
(let ((path-list (nth 1 (packages-find-packages
emacs-data-roots
(packages-compute-package-locations user-init-d
irectory)))))
(cond ((eq type 'std)
(while path-list
- (if (equal (substring (car path-list) -16)
"xemacs-packages/")
+ (if (string-match "xemacs-packages[/\\]$" (car path-list))
(setq top-dir (car path-list)))
(setq path-list (cdr path-list))))
((eq type 'mule)
(while path-list
- (if (equal (substring (car path-list) -14) "mule-packages/")
+ (if (string-match "mule-packages[/\\]$" (car path-list))
(setq top-dir (car path-list)))
(setq path-list (cdr path-list)))))))
;; Now return either the directory or nil.
2:05 p.m.
>>>> "Fabrice" == Fabrice Popineau
<Fabrice.Popineau(a)supelec.fr> writes:
Fabrice> - what is the status of 'mule' packages wrt to 21.5+Mule?
Fabrice> I guess mule-ucs and mule-base packages are not needed
Fabrice> anymore (integrated).
Correct, although mule-base may get resurrected. I'm not very happy
with Ben's rearrangement. I don't think language environments are
sufficiently low-level to be in core.
Fabrice> But what about the others (locale, latin-unity ...) and
Fabrice> what about their location? Shouldn't they move into
Fabrice> xemacs-packages/ in 21.5 since Mule is considered to be
Fabrice> the default?
So what? The question is "Does there exist an XEmacs that barfs when
loading a Mule package?" If so, the mule-packages hierarchy is
necessary.
latin-unity is a special case; it will eventually be integrated with
21.5, when I get the time.
--
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.
9:46 a.m.
|--==> "FP" == Fabrice Popineau <Fabrice.Popineau(a)supelec.fr> writes:
FP> In package-admin-find-top-directory:
FP> (while path-list
FP> (if (equal (substring (car path-list) -16) "xemacs-packages/")
FP> (setq top-dir (car path-list)))
FP> (setq path-list (cdr path-list))))
FP> ((eq type 'mule)
FP> (while path-list
FP> (if (equal (substring (car path-list) -14) "mule-packages/")
FP> (setq top-dir (car path-list)))
FP> (setq path-list (cdr path-list)))))))
FP> But on win32, the path-list returned has '\\' as path separator. So that
FP> just can't match. Possible fix attached.
I have this fixed in an as yet un-checked in workspace. But thank you
very much for reminding me. Probably be in CVS later tonight.
--
|---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---|
| XEmacs - It's not just an editor. |
| It's a way of life. |
|------------------------------------<youngs(a)xemacs.org>---|
11:35 a.m.
Steve Y, this patch may be better than using directory-sep-char.
On windows people mix / and \\ a lot and they both work.
We should also handle the case more gracefully (easy to understand
error message) where the package path does not contain mule-packages
or xemacs-packages at all (think Debian).
Just my 0.02~
Adrian
>>>> "Fabrice" == Fabrice Popineau
<Fabrice.Popineau(a)supelec.fr> writes:
Fabrice> In package-admin-find-top-directory:
Fabrice> (while path-list
Fabrice> (if (equal (substring (car path-list) -16)
"xemacs-packages/")
Fabrice> (setq top-dir (car path-list)))
Fabrice> (setq path-list (cdr path-list))))
Fabrice> ((eq type 'mule)
Fabrice> (while path-list
Fabrice> (if (equal (substring (car path-list) -14)
"mule-packages/")
Fabrice> (setq top-dir (car path-list)))
Fabrice> (setq path-list (cdr path-list)))))))
Fabrice> But on win32, the path-list returned has '\\' as path separator.
So that
Fabrice> just can't match. Possible fix attached.
Fabrice> Other related questions:
Fabrice> - what is the status of 'mule' packages wrt to 21.5+Mule? I guess
Fabrice> mule-ucs and mule-base packages are not needed anymore (integrated). But
Fabrice> what about the others (locale, latin-unity ...) and what about their
Fabrice> location? Shouldn't they move into xemacs-packages/ in 21.5 since
Mule
Fabrice> is considered to be the default?
Fabrice> --
Fabrice> Fabrice
Fabrice> C:\>diff -u xemacs\xemacs-21.5\lisp\package-admin.el
xemacs\xemacs-21.5\lisp\package-admin.el
Fabrice> --- xemacs\xemacs-21.5\lisp\package-admin.el Mon Mar 24 17:43:55 2003
Fabrice> +++ xemacs\xemacs-21.5\lisp\package-admin.el Wed Apr 09 10:29:54 2003
Fabrice> ＠＠ -170,21 +170,21 ＠＠
Fabrice> ;; First, check the environment var.
Fabrice> (if env-value
Fabrice> (let ((path-list (paths-decode-directory-path env-value
'drop-empties)))
Fabrice> (cond ((eq type 'std)
Fabrice> (while path-list
Fabrice> - (if (equal (substring (car path-list) -16)
"xemacs-packages/")
Fabrice> + (if (string-match "xemacs-packages[/\\]$"
(car path-list))
Fabrice> (setq top-dir (car path-list)))
Fabrice> (setq path-list (cdr path-list))))
Fabrice> ((eq type 'mule)
Fabrice> (while path-list
Fabrice> - (if (equal (substring (car path-list) -14)
"mule-packages/")
Fabrice> + (if (string-match "mule-packages[/\\]$" (car
path-list))
Fabrice> (setq top-dir (car path-list)))
Fabrice> (setq path-list (cdr path-list)))))))
Fabrice> ;; Wasn't in the environment, try `user-init-directory' if
Fabrice> ;; USER-DIR is non-nil.
Fabrice> (if (and user-dir
Fabrice> (not top-dir))
Fabrice> (cond ((eq type 'std)
Fabrice> (setq top-dir (file-name-as-directory
Fabrice> (expand-file-name
"xemacs-packages" user-init-directory))))
Fabrice> ((eq type 'mule)
Fabrice> ＠＠ -195,14 +195,14 ＠＠
Fabrice> (let ((path-list (nth 1 (packages-find-packages
Fabrice> emacs-data-roots
Fabrice> (packages-compute-package-locations
user-init-d
Fabrice> irectory)))))
Fabrice> (cond ((eq type 'std)
Fabrice> (while path-list
Fabrice> - (if (equal (substring (car path-list) -16)
"xemacs-packages/")
Fabrice> + (if (string-match "xemacs-packages[/\\]$"
(car path-list))
Fabrice> (setq top-dir (car path-list)))
Fabrice> (setq path-list (cdr path-list))))
Fabrice> ((eq type 'mule)
Fabrice> (while path-list
Fabrice> - (if (equal (substring (car path-list) -14)
"mule-packages/")
Fabrice> + (if (string-match "mule-packages[/\\]$" (car
path-list))
Fabrice> (setq top-dir (car path-list)))
Fabrice> (setq path-list (cdr path-list)))))))
Fabrice> ;; Now return either the directory or nil.
--
Adrian Aichner
mailto:adrian＠xemacs.org
http://www.xemacs.org/
2:56 a.m.
|--==> "APA" == Adrian Aichner <adrian(a)xemacs.org> writes:
>>>>>"Fabrice" == Fabrice Popineau
<Fabrice.Popineau(a)supelec.fr> writes:
Fabrice> (while path-list
Fabrice> - (if (equal (substring (car path-list) -16)
"xemacs-packages/")
Fabrice> + (if (string-match "xemacs-packages[/\\]$" (car
path-list))
Fabrice> (setq top-dir (car path-list)))
APA> Steve Y, this patch may be better than using directory-sep-char.
APA> On windows people mix / and \\ a lot and they both work.
Ew! How obtuse! Does this mean that 'directory-sep-char' won't work
here? That just plain sucks. I'll never understand why we support
that poor excuse for an OS anyway.
Wouldn't the best fix be to fix 'directory-sep-char' so it can handle
this type of bogosity? I don't mind using Fabrice's patch in the
interim, if that's what's needed.
APA> We should also handle the case more gracefully (easy to
APA> understand error message) where the package path does not
APA> contain mule-packages or xemacs-packages at all (think
APA> Debian).
Improvements could probably be made in this regard, yes. Patches
welcome.:-)
APA> Just my 0.02~
Does that include GST?
--
|---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---|
| XEmacs - It's not just an editor. |
| It's a way of life. |
|------------------------------------<youngs(a)xemacs.org>---|
8:41 a.m.
>>>> "SY" == Steve Youngs
<youngs(a)xemacs.org> writes:
SY> |--==> "APA" == Adrian Aichner <adrian(a)xemacs.org> writes:
>>>>>> "Fabrice" == Fabrice Popineau
<Fabrice.Popineau(a)supelec.fr> writes:
Fabrice> (while path-list
Fabrice> - (if (equal (substring (car path-list) -16)
"xemacs-packages/")
Fabrice> + (if (string-match "xemacs-packages[/\\]$"
(car path-list))
Fabrice> (setq top-dir (car path-list)))
APA> Steve Y, this patch may be better than using directory-sep-char.
APA> On windows people mix / and \\ a lot and they both work.
SY> Ew! How obtuse! Does this mean that 'directory-sep-char' won't
work
SY> here? That just plain sucks. I'll never understand why we support
SY> that poor excuse for an OS anyway.
SY> Wouldn't the best fix be to fix 'directory-sep-char' so it can
handle
SY> this type of bogosity? I don't mind using Fabrice's patch in the
SY> interim, if that's what's needed.
Don't think so since 'directory-sep-char' is just a char.
APA> We should also handle the case more gracefully (easy to
APA> understand error message) where the package path does not
APA> contain mule-packages or xemacs-packages at all (think
APA> Debian).
SY> Improvements could probably be made in this regard, yes. Patches
SY> welcome.:-)
Yep, I know.
APA> Just my 0.02~
SY> Does that include GST?
What's GST?
SY> --
SY> |---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---|
SY> | XEmacs - It's not just an editor. |
SY> | It's a way of life. |
SY> |------------------------------------<youngs(a)xemacs.org>---|
--
Adrian Aichner
mailto:adrian＠xemacs.org
http://www.xemacs.org/
2:39 p.m.
>>>> On 15 Apr 2003 09:41:32 +0200, Adrian Aichner
<adrian(a)xemacs.org> said:
>>>> "SY" == Steve Youngs
<youngs(a)xemacs.org> writes:
SY> |--==> "APA" == Adrian
Aichner <adrian(a)xemacs.org> writes:
APA> We should also handle the case more gracefully (easy to
APA> understand error message) where the package path does not
APA> contain mule-packages or xemacs-packages at all (think Debian).
Next debian package will use xemacs-packages rather than packages.
iirc packages was once the default and it was changed to
xemacs-packages. At the time I was building a new version it was
easier to go back to the old default. I'll fix this is the next
upload to debian.
Jim
--
＠James LewisMoss <dres(a)lewismoss.org> | Blessed Be!
＠ http://www.lewismoss.org/~dres | Linux is kewl!
＠"Argue for your limitations and sure enough, they're yours." Bach
3:03 p.m.
>>>> "Dres" == James LewisMoss
<dres(a)lewismoss.org> writes:
Dres> iirc packages was once the default and it was changed to
Dres> xemacs-packages.
It was; Steve Baur's original intention was not to distinguish between
"official" and "unofficial" packages, and he didn't know about
Mule
(in particular, that multibyte characters would cause syntax errors
when loading), so he couldn't know that a distinction between Mule and
non-Mule would be necessary if both were installed.
Dres> I'll fix this is the next upload to debian.
Great!
--
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.
12:11 p.m.
|--==> "APA" == Adrian Aichner <adrian(a)xemacs.org> writes:
>>>>>"SY" == Steve Youngs
<youngs(a)xemacs.org> writes:
SY> Wouldn't the best fix be to fix
'directory-sep-char' so it can handle
SY> this type of bogosity? I don't mind using Fabrice's patch in the
SY> interim, if that's what's needed.
APA> Don't think so since 'directory-sep-char' is just a char.
So who was the idiot who decided to use two characters as a directory
separator anyway? This type of thing is incredibly annoying.
APA> Just my 0.02~
SY> Does that include GST?
APA> What's GST?
Goods and Services Tax. The Australian Government in its infinite
wisdom decided it was something we just couldn't do without. We've
been living with it for a while now and all those /promises/ the Govt
made, saying nothing would increase in price, have all been
broken.:-( Maybe it's time for a move, I wonder what the Internet
connectivity out of the Cayman Islands is like? Their taxes a pretty
good.:-P
--
|---<Steve Youngs>---------------<GnuPG KeyID: 10D5C9C5>---|
| XEmacs - It's not just an editor. |
| It's a way of life. |
|------------------------------------<youngs(a)xemacs.org>---|
2:12 p.m.
So who was the idiot who decided to use two characters as a
directory
separator anyway? This type of thing is incredibly annoying.
Some guy who was responsible for win32 internals. The fact is that '/'
can be used as a directory separator for every single win32 function needing a
filename. But this is not understood by windows apps at the user level
... except for apps ported from unix! Most of the time it was easy and
seemed to be nice to say that both '/' and '\\' are separators than to
change the
separator. I must admit that I agree with you.
Fabrice
7685
days inactive
7692
days old
9 comments
5 participants
participants (5)
-
Adrian Aichner -
Fabrice Popineau -
James LewisMoss -
Stephen J. Turnbull -
Steve Youngs