11:50 a.m.
Could somebody please take a look at this. Adrian, you were the last
person to sync VC, perhaps you could give it a look.
Thanks.
--
|---<Steve Youngs>---------------<GnuPG KeyID: 9E7E2820>---|
| XEmacs - It's not just an editor. |
| It's a way of life. |
|------------------------------------<youngs(a)xemacs.org>---|
12:42 a.m.
>>>> "SY" == Steve Youngs
<youngs(a)xemacs.org> writes:
SY> Could somebody please take a look at this. Adrian, you were
SY> the last person to sync VC, perhaps you could give it a look.
Last time I touched it was to fix a pcl-cvs interaction issue.
I have more pressing issues on my list for a while, sorry.
Also, I have not been using RCS since my times on UNIX.
CVS rules!
Perhaps I can help Nathan to debug the problem, Nathan?
Best regards,
Adrian
SY> Thanks.
SY> From: froydnj(a)rose-hulman.edu (Nathan Froyd)
SY> Subject: vc problems in XEmacs 21.1.14
SY> Newsgroups: comp.emacs.xemacs,comp.emacs
SY> Date: 13 Mar 2001 00:56:09 GMT
SY> Using XEmacs 21.1.14, I attempt to do the following:
SY> Open an existing file
SY> C-x v v to place it under version control (RCS 5.7)
SY> C-x v v to check it out for editing
SY> However, this doesn't work. The second `C-x v v' gives me an error,
SY> informing me that running `ci' failed with status code 1. Another `C-x
SY> v v' is necessary to get everything to work correctly.
SY> Furthermore, GNU Emacs exhibits differing behavior. Not only does it
SY> act as expected above, but after the first `C-x v v', it correctly
SY> changes the modeline to reflect that this file is now under RCS control.
SY> This is the second instance where I have discovered that GNU Emacs does
SY> something correctly where XEmacs does not. I hope that this will be
SY> rectified in the next release of XEmacs.
SY> --
SY> </nathan> froydnj(a)rose-hulman.edu |
http://www.rose-hulman.edu/~froydnj/
SY> Yes, God had a deadline. So He wrote it all in Lisp.
SY> ----------
SY> --
SY> |---<Steve Youngs>---------------<GnuPG KeyID: 9E7E2820>---|
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/
8:19 a.m.
Adrian, the vc problem also exists with CVS. And I presume the other vc
problem I mentioned a while back still exists, where vc-cancel-version
could potentially move files to other directories.
Adrian Aichner writes:
From: Adrian.Aichner(a)t-online.de (Adrian Aichner)
To: Steve Youngs <youngs(a)xemacs.org>, froydnj(a)rose-hulman.edu (Nathan Froyd)
Cc: XEmacs Beta <xemacs-beta(a)xemacs.org>, Adrian Aichner <adrian(a)xemacs.org>
Subject: Re: [comp.emacs.xemacs,comp.emacs] vc problems in XEmacs 21.1.14
Date: 15 Mar 2001 01:42:41 +0100
>>>>> "SY" == Steve Youngs <youngs(a)xemacs.org> writes:
SY> Could somebody please take a look at this. Adrian, you were
SY> the last person to sync VC, perhaps you could give it a look.
Last time I touched it was to fix a pcl-cvs interaction issue.
I have more pressing issues on my list for a while, sorry.
Also, I have not been using RCS since my times on UNIX.
CVS rules!
Perhaps I can help Nathan to debug the problem, Nathan?
Best regards,
Adrian
SY> Thanks.
SY> From: froydnj(a)rose-hulman.edu (Nathan Froyd)
SY> Subject: vc problems in XEmacs 21.1.14
SY> Newsgroups: comp.emacs.xemacs,comp.emacs
SY> Date: 13 Mar 2001 00:56:09 GMT
SY> Using XEmacs 21.1.14, I attempt to do the following:
SY> Open an existing file
SY> C-x v v to place it under version control (RCS 5.7)
SY> C-x v v to check it out for editing
SY> However, this doesn't work. The second `C-x v v' gives me an error,
SY> informing me that running `ci' failed with status code 1. Another `C-x
SY> v v' is necessary to get everything to work correctly.
SY> Furthermore, GNU Emacs exhibits differing behavior. Not only does it
SY> act as expected above, but after the first `C-x v v', it correctly
SY> changes the modeline to reflect that this file is now under RCS control.
SY> This is the second instance where I have discovered that GNU Emacs does
SY> something correctly where XEmacs does not. I hope that this will be
SY> rectified in the next release of XEmacs.
SY> --
SY> </nathan> froydnj(a)rose-hulman.edu |
http://www.rose-hulman.edu/~froydnj/
SY> Yes, God had a deadline. So He wrote it all in Lisp.
SY> ----------
SY> --
SY> |---<Steve Youngs>---------------<GnuPG KeyID: 9E7E2820>---|
SY> | XEmacs - It's not just an editor. |
SY> | It's a way of life. |
SY> |------------------------------------<youngs(a)xemacs.org>---|
12:54 a.m.
>>>> "Albert" == Albert L Ting
<alt(a)artisan.com> writes:
Albert> Adrian, the vc problem also exists with CVS. And I
Albert> presume the other vc problem I mentioned a while back
Albert> still exists, where vc-cancel-version could potentially
Albert> move files to other directories.
Hi Albert,
do you have a simple test-case for this for cvs?
I'll give it a try.
Later,
Adrian
Albert> Adrian Aichner writes:
> From: Adrian.Aichner(a)t-online.de (Adrian Aichner)
> To: Steve Youngs <youngs(a)xemacs.org>, froydnj(a)rose-hulman.edu (Nathan Froyd)
> Cc: XEmacs Beta <xemacs-beta(a)xemacs.org>, Adrian Aichner
<adrian(a)xemacs.org>
> Subject: Re: [comp.emacs.xemacs,comp.emacs] vc problems in XEmacs 21.1.14
> Date: 15 Mar 2001 01:42:41 +0100
>
> >>>>> "SY" == Steve Youngs <youngs(a)xemacs.org>
writes:
>
SY> Could somebody please take a look at this. Adrian, you were
SY> the last person to sync VC, perhaps you could give it a look.
>
> Last time I touched it was to fix a pcl-cvs interaction issue.
>
> I have more pressing issues on my list for a while, sorry.
>
> Also, I have not been using RCS since my times on UNIX.
>
> CVS rules!
>
> Perhaps I can help Nathan to debug the problem, Nathan?
>
> Best regards,
>
> Adrian
>
SY> Thanks.
>
SY> From: froydnj(a)rose-hulman.edu (Nathan Froyd)
SY> Subject: vc problems in XEmacs 21.1.14
SY> Newsgroups: comp.emacs.xemacs,comp.emacs
SY> Date: 13 Mar 2001 00:56:09 GMT
>
SY> Using XEmacs 21.1.14, I attempt to do the
following:
>
SY> Open an existing file
SY> C-x v v to place it under version control (RCS 5.7)
SY> C-x v v to check it out for editing
>
SY> However, this doesn't work. The second
`C-x v v' gives me an error,
SY> informing me that running `ci' failed with status code 1. Another `C-x
SY> v v' is necessary to get everything to work correctly.
>
SY> Furthermore, GNU Emacs exhibits differing
behavior. Not only does it
SY> act as expected above, but after the first `C-x v v', it correctly
SY> changes the modeline to reflect that this file is now under RCS control.
>
SY> This is the second instance where I have
discovered that GNU Emacs does
SY> something correctly where XEmacs does not. I hope that this will be
SY> rectified in the next release of XEmacs.
SY> --
SY> </nathan> froydnj(a)rose-hulman.edu |
http://www.rose-hulman.edu/~froydnj/
>
SY> Yes, God had a deadline. So He wrote it all in
Lisp.
SY> ----------
>
>
>
SY> --
SY> |---<Steve Youngs>---------------<GnuPG KeyID: 9E7E2820>---|
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/
3:35 a.m.
Adrian Aichner writes:
From: Adrian.Aichner(a)t-online.de (Adrian Aichner)
To: "Albert L. Ting" <alt(a)artisan.com>
Cc: Adrian.Aichner(a)t-online.de, youngs(a)xemacs.org, froydnj(a)rose-hulman.edu,
xemacs-beta(a)xemacs.org, adrian(a)xemacs.org
Subject: Re: [comp.emacs.xemacs,comp.emacs] vc problems in XEmacs 21.1.14
Date: 17 Mar 2001 01:54:18 +0100
>>>>> "Albert" == Albert L Ting <alt(a)artisan.com> writes:
Albert> Adrian, the vc problem also exists with CVS. And I
Albert> presume the other vc problem I mentioned a while back
Albert> still exists, where vc-cancel-version could potentially
Albert> move files to other directories.
Hi Albert,
do you have a simple test-case for this for cvs?
Create a cvs tree, then create a new file. Now use "c-x v i", vc-register,
to register the file. It's been added to the cvs tree, yet the emacs
buffer doesn't reflect this.
So in order to do the commit, I now have to kill the buffer, revisit the
file, then do a "c-x c-q", vc-toggle-read-only.
10:17 p.m.
>>>> "Albert" == Albert L Ting
<alt(a)artisan.com> writes:
Albert> Adrian Aichner writes:
>> From: Adrian.Aichner(a)t-online.de (Adrian Aichner)
>> To: "Albert L. Ting" <alt(a)artisan.com>
>> Cc: Adrian.Aichner(a)t-online.de, youngs(a)xemacs.org, froydnj(a)rose-hulman.edu,
>> xemacs-beta(a)xemacs.org, adrian(a)xemacs.org
>> Subject: Re: [comp.emacs.xemacs,comp.emacs] vc problems in XEmacs 21.1.14
>> Date: 17 Mar 2001 01:54:18 +0100
>>
>> >>>> "Albert" ==
Albert L Ting <alt(a)artisan.com> writes:
>>
Albert> Adrian, the vc problem also exists with CVS. And I
Albert> presume the other vc problem I mentioned a while back
Albert> still exists, where vc-cancel-version could potentially
Albert> move files to other directories.
>
> Hi Albert,
>
> do you have a simple test-case for this for cvs?
Albert> Create a cvs tree, then create a new file. Now use "c-x v
Albert> i", vc-register, to register the file. It's been added to
Albert> the cvs tree, yet the emacs buffer doesn't reflect this.
Thanks for the test-case, Albert.
I will investigate this real soon now.
Best regards,
Adrian
Albert> So in order to do the commit, I now have to kill the
Albert> buffer, revisit the file, then do a "c-x c-q",
Albert> vc-toggle-read-only.
--
Adrian Aichner
mailto:adrian＠xemacs.org
http://www.xemacs.org/
1:59 a.m.
Thanks a bunch Adrian! This is the other vc related problem, where xemacs
will sometimes move cvs'd files to other directories.
This bug used to exist with GNU Emacs but was fixed about 1-2 years ago.
I don't remember how it was fixed.
Create the following directory setup. I'm using RCS as an example, but the
same problem occurs with CVS.
$ ls *
dir1:
RCS/ file1
dir2:
file2
With file1 already checked in. Now run xemacs, and load the vc.el
package. In one frame, open three windows with these buffers:
file2
------------
file2
------------
file1
Go to the third window, checkout file1, edit the text, then save. Now
do a vc-revert-buffer and say "yes" to cancel the changes. dir1/file1
does not change. However, a fresh copy is incorrectly stored in
dir2/file1!
Albert
7:09 p.m.
Well, I found a patch that solves one of the vc problems. The one where vc
would move files to another directory when you attempt to do a
vc-revert-buffer. This is essentially the same code used in the fsf emacs
version.
*** vc.el Wed Apr 11 11:03:21 2001
--- vc.el.orig Mon Apr 9 17:53:23 2001
***************
*** 2012,2029 ****
;; This operation should always ask for confirmation.
(vc-suppress-confirm nil)
(obuf (current-buffer)) (changed (vc-diff nil t)))
!
! (if changed
! (unwind-protect
! (if (not (yes-or-no-p "Discard changes? "))
! (error "Revert cancelled"))
(if (and (window-dedicated-p (selected-window))
(one-window-p t 'selected-frame))
(make-frame-invisible (selected-frame))
! (delete-window))))
! (set-buffer obuf)
(vc-backend-revert file)
! (vc-resynch-window file t t))
)
;;;###autoload
--- 2012,2033 ----
;; This operation should always ask for confirmation.
(vc-suppress-confirm nil)
(obuf (current-buffer)) (changed (vc-diff nil t)))
! (if (and changed (not (yes-or-no-p "Discard changes? ")))
! (progn
(if (and (window-dedicated-p (selected-window))
(one-window-p t 'selected-frame))
(make-frame-invisible (selected-frame))
! (delete-window))
! (error "Revert cancelled"))
! (set-buffer obuf))
! (if changed
! (if (and (window-dedicated-p (selected-window))
! (one-window-p t 'selected-frame))
! (make-frame-invisible (selected-frame))
! (delete-window)))
(vc-backend-revert file)
! (vc-resynch-window file t t)
! )
)
Albert L. Ting writes:
From: "Albert L. Ting" <alt(a)artisan.com>
To: Adrian.Aichner(a)t-online.de
Cc: alt(a)artisan.com, youngs(a)xemacs.org, froydnj(a)rose-hulman.edu,
xemacs-beta(a)xemacs.org, adrian(a)xemacs.org
Subject: Re: [comp.emacs.xemacs,comp.emacs] vc problems in XEmacs 21.1.14
Date: Mon, 19 Mar 2001 17:59:41 -0800
Thanks a bunch Adrian! This is the other vc related problem, where xemacs
will sometimes move cvs'd files to other directories.
This bug used to exist with GNU Emacs but was fixed about 1-2 years ago.
I don't remember how it was fixed.
Create the following directory setup. I'm using RCS as an example, but the
same problem occurs with CVS.
$ ls *
dir1:
RCS/ file1
dir2:
file2
With file1 already checked in. Now run xemacs, and load the vc.el
package. In one frame, open three windows with these buffers:
file2
------------
file2
------------
file1
Go to the third window, checkout file1, edit the text, then save. Now
do a vc-revert-buffer and say "yes" to cancel the changes. dir1/file1
does not change. However, a fresh copy is incorrectly stored in
dir2/file1!
Albert
8:59 p.m.
Took a stab at fixing this problem. It's a subtle bug, below is a patch.
*** vc.el Thu Apr 12 12:58:27 2001
--- vc.el.orig Mon Apr 9 17:53:23 2001
***************
*** 573,580 ****
(message "Running %s on %s..." command file))
(let ((obuf (current-buffer)) (camefrom (current-buffer))
(squeezed nil)
(olddir default-directory)
! vc-file status)
(set-buffer (get-buffer-create buffer))
(set (make-local-variable 'vc-parent-buffer) camefrom)
(set (make-local-variable 'vc-parent-buffer-name)
--- 573,581 ----
(message "Running %s on %s..." command file))
(let ((obuf (current-buffer)) (camefrom (current-buffer))
(squeezed nil)
+ (vc-file (and file (vc-name file)))
(olddir default-directory)
! status)
(set-buffer (get-buffer-create buffer))
(set (make-local-variable 'vc-parent-buffer) camefrom)
(set (make-local-variable 'vc-parent-buffer-name)
***************
*** 586,596 ****
(mapcar
(function (lambda (s) (and s (setq squeezed (append squeezed (list s)))))
)
flags)
! ;; do NOT call vc-name until (eq last 'MASTER). vc-name calls vc-backend,
! ;; which sets the file's 'vc-backend prop. If we preset vc-file in the
! ;; let statement above, the 'vc-backend prop also gets set when (eq last
! ;; 'WORKFILE), which is too soon (occurs when one does a vc-register).
! (if (and (eq last 'MASTER) file (setq vc-file (vc-name file)))
(setq squeezed (append squeezed (list vc-file))))
(if (eq last 'WORKFILE)
(progn
--- 587,593 ----
(mapcar
(function (lambda (s) (and s (setq squeezed (append squeezed (list s)))))
)
flags)
! (if (and vc-file (eq last 'MASTER))
(setq squeezed (append squeezed (list vc-file))))
(if (eq last 'WORKFILE)
(progn
***************
Steve Youngs writes:
From: Steve Youngs <youngs(a)xemacs.org>
To: XEmacs Beta <xemacs-beta(a)xemacs.org>
Cc: Adrian Aichner <adrian(a)xemacs.org>
Subject: [comp.emacs.xemacs,comp.emacs] vc problems in XEmacs 21.1.14
Date: 14 Mar 2001 21:50:12 +1000
Could somebody please take a look at this. Adrian, you were the last
person to sync VC, perhaps you could give it a look.
Thanks.
From: froydnj(a)rose-hulman.edu (Nathan Froyd)
Newsgroups: comp.emacs.xemacs,comp.emacs
Subject: vc problems in XEmacs 21.1.14
Date: 13 Mar 2001 00:56:09 GMT
MIME-Version: 1.0
Using XEmacs 21.1.14, I attempt to do the following:
Open an existing file
C-x v v to place it under version control (RCS 5.7)
C-x v v to check it out for editing
However, this doesn't work. The second `C-x v v' gives me an error,
informing me that running `ci' failed with status code 1. Another `C-x
v v' is necessary to get everything to work correctly.
Furthermore, GNU Emacs exhibits differing behavior. Not only does it
act as expected above, but after the first `C-x v v', it correctly
changes the modeline to reflect that this file is now under RCS control.
This is the second instance where I have discovered that GNU Emacs does
something correctly where XEmacs does not. I hope that this will be
rectified in the next release of XEmacs.
--
</nathan> froydnj(a)rose-hulman.edu | http://www.rose-hulman.edu/~froydnj/
Yes, God had a deadline. So He wrote it all in Lisp.
1:12 a.m.
On Thu, Apr 12, 2001 at 12:59:28PM -0700, Albert L. Ting wrote:
Took a stab at fixing this problem. It's a subtle bug, below is
a patch.
Sorry for taking so long to reply. I was on break last week and then
the rigors of school got me down. Anyway...
*** vc.el Thu Apr 12 12:58:27 2001
--- vc.el.orig Mon Apr 9 17:53:23 2001
These changes *are* present in the version I have -- which, according to
_pkg.el, is vc 1.26. vc 1.26 continues to exhibit the bug for me, just
as vc 1.25 did. If vc has been upgraded in 21.4.1, I will attempt to
tickle the bug with that version.
--
</nathan> froydnj(a)rose-hulman.edu | http://www.rose-hulman.edu/~froydnj/
Yes, God had a deadline. So He wrote it all in Lisp.
8612
days inactive
8649
days old
9 comments
4 participants
participants (4)
-
Adrian Aichner -
Albert Ting -
Nathan Froyd -
Steve Youngs