9:10 p.m.
I downloaded Ville's latest vc (0.93) and substituted it for what came with
21.5.28. I tried viewing the differences between a file and the latest
version checked into Subversion. XEmacs complained about the variable
'state' not being defined. The ediff-vc-latest-version function does indeed
seem to use an unbound variable named 'state':
(defun ediff-vc-latest-version (file)
"Return the version level of the latest version of FILE in repository."
(if (fboundp 'vc-latest-version)
(vc-latest-version file)
(or (vc-file-getprop file 'vc-latest-version)
(cond ((vc-backend file)
(vc-call state file)
(vc-file-getprop file 'vc-latest-version))
(t (error "File %s is not under version control" file))))
))
I'm not sure why I never saw this before with CVS. Maybe vc-latest-version
was bound in the old vc but not in the new.
--
Skip Montanaro - skip(a)pobox.com - http://www.webfast.com/~skip/
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
9:53 p.m.
vc-call is a macro defined in vc-hooks. The first argument should be unquoted.
Try (vc-call state "somefile")
--michael
I downloaded Ville's latest vc (0.93) and substituted it for what
came with
21.5.28. I tried viewing the differences between a file and the latest
version checked into Subversion. XEmacs complained about the variable
'state' not being defined. The ediff-vc-latest-version function does indeed
seem to use an unbound variable named 'state':
(defun ediff-vc-latest-version (file)
"Return the version level of the latest version of FILE in repository."
(if (fboundp 'vc-latest-version)
(vc-latest-version file)
(or (vc-file-getprop file 'vc-latest-version)
(cond ((vc-backend file)
(vc-call state file)
(vc-file-getprop file 'vc-latest-version))
(t (error "File %s is not under version control" file))))
))
I'm not sure why I never saw this before with CVS. Maybe vc-latest-version
was bound in the old vc but not in the new.
--
Skip Montanaro - skip(a)pobox.com - http://www.webfast.com/~skip/
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
1:03 a.m.
Michael> vc-call is a macro defined in vc-hooks. The first argument
Michael> should be unquoted.
Michael> Try (vc-call state "somefile")
Sorry, I don't know what you are suggesting I do. (Lisp macros have always
confused me.) The call in ediff-vc-latest-version is already
(vc-call state file)
Are you suggesting I should quote "file":
(vc-call state "file")
Skip
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
12:01 p.m.
Michael> vc-call is a macro defined in vc-hooks. The first
argument
Michael> should be unquoted.
Michael> Try (vc-call state "somefile")
Sorry, I don't know what you are suggesting I do. (Lisp macros have always
confused me.) The call in ediff-vc-latest-version is already
(vc-call state file)
Are you suggesting I should quote "file":
(vc-call state "file")
Skip
Oh, I see. Xemacs' vc seems to be completely different from the gnu version.
It does not have vc-call.
This function (ediff-vc-latest-version) in ediff has not changed for over
4 years.
But why does it bother you? XEmacs has vc-latest-version, and this is what
is used then. Gnu emacs does not have it, and so the other if-branch is used.
In that branch a macro vc-call is used (correctly).
--michael
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
12:17 p.m.
Michael> But why does it bother you? XEmacs has vc-latest-version, and
Michael> this is what is used then. Gnu emacs does not have it, and so
Michael> the other if-branch is used. In that branch a macro vc-call is
Michael> used (correctly).
Ville's new vc doesn't have vc-latest-version.
Skip
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
12:25 p.m.
Michael> But why does it bother you? XEmacs has vc-latest-version, and
Michael> this is what is used then. Gnu emacs does not have it, and so
Michael> the other if-branch is used. In that branch a macro vc-call is
Michael> used (correctly).
Ville's new vc doesn't have vc-latest-version.
Skip
Why? Perhaps he can suggest a workaround?
--michael
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
12:36 p.m.
On Monday 20 August 2007, Michael Kifer wrote:
> Michael> But why does it bother you? XEmacs has
vc-latest-version,
> and Michael> this is what is used then. Gnu emacs does not have it, and
> so Michael> the other if-branch is used. In that branch a macro vc-call
> is Michael> used (correctly).
>
> Ville's new vc doesn't have vc-latest-version.
>
> Skip
Why? Perhaps he can suggest a workaround?
The "new vc" is a straight port of vc*.el in GNU Emacs to XEmacs plus a few
additional modes. If it doesn't have something, it's because the GNU Emacs
version doesn't have it either.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
10:07 p.m.
>>>> "skip" == skip <skip(a)pobox.com>
writes:
skip> I downloaded Ville's latest vc (0.93) and substituted it for
skip> what came with 21.5.28. I tried viewing the differences
skip> between a file and the latest version checked into
skip> Subversion. XEmacs complained about the variable 'state'
skip> not being defined. The ediff-vc-latest-version function
skip> does indeed seem to use an unbound variable named 'state':
I've seen this, Skip (I mentioned it earlier today, but the mail have
not come through).
Here it is:
11:45 p.m.
On Monday 20 August 2007, Adrian Aichner wrote:
>>>>> "skip" == skip <skip(a)pobox.com>
writes:
skip> I downloaded Ville's latest vc (0.93)
Thanks for testing! However, the latest is 0.94 but it doesn't have any
changes that would affect this. 0.94 package tarball is also uploaded now to
the location where I guess you got the 0.93 from. (I'm no longer posting the
URL in order to try encourage people to build it from unsupported/scop/vc and
possibly find bugs I haven't caught.)
I've seen this, Skip (I mentioned it earlier today, but the mail
have
not come through).
Here it is:
From: Adrian Aichner <adrian(a)xemacs.org>
Subject: Debugger entered--Lisp error: (void-variable state) in erevision
with unsupported/scop/vc/vc-hooks.el To: XEmacs Beta List
<xemacs-beta(a)xemacs.org>, kifer(a)cs.stonybrook.edu Date: Sat, 28 Oct 2006
23:07:01 +0200
Organization: The XEmacs Project
Message-ID: <r6wscfpm.fsf(a)mx.qsc.de>
And here's a link to that message in archives:
http://calypso.tux.org/pipermail/xemacs-beta/2006-October/009873.html
Adrian, I haven't tested it, but do you think that the recipe you posted in
that message would still fix this?
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
12:55 a.m.
Ville> Thanks for testing! However, the latest is 0.94 but it doesn't
Ville> have any changes that would affect this. 0.94 package tarball is
Ville> also uploaded now to the location where I guess you got the 0.93
Ville> from. (I'm no longer posting the URL in order to try encourage
Ville> people to build it from unsupported/scop/vc and possibly find
Ville> bugs I haven't caught.)
It's not clear to me what I'm supposed to download and build. Do you have a
full URL (or something) for unsupported/scop/vc?
Skip
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
8:04 a.m.
On Monday 20 August 2007, skip(a)pobox.com wrote:
Ville> Thanks for testing! However, the latest is 0.94 but it
doesn't
Ville> have any changes that would affect this. 0.94 package tarball
is Ville> also uploaded now to the location where I guess you got the 0.93
Ville> from. (I'm no longer posting the URL in order to try encourage
Ville> people to build it from unsupported/scop/vc and possibly find Ville>
bugs I haven't caught.)
It's not clear to me what I'm supposed to download and build. Do you have
a full URL (or something) for unsupported/scop/vc?
"unsupported" is a directory below a cvs up'd CVS checkout of the XEmacs
packages tree.
(checkout packages from cvs)
(set up Local.rules from Local.rules.template)
$ cvs up -d # initial checkout doesn't IIRC fetch the "unsupported" tree
$ make autoloads
$ cd unsupported/scop/vc
$ make install
http://www.xemacs.org/Develop/cvsaccess.html
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
4:16 p.m.
> It's not clear to me what I'm supposed to download and
build. Do you
> have a full URL (or something) for unsupported/scop/vc?
Ville> "unsupported" is a directory below a cvs up'd CVS checkout of
the XEmacs
Ville> packages tree.
Ville> (checkout packages from cvs)
Ville> (set up Local.rules from Local.rules.template)
Ville> $ cvs up -d # initial checkout doesn't IIRC fetch the
"unsupported" tree
Ville> $ make autoloads
Ville> $ cd unsupported/scop/vc
Ville> $ make install
Thanks for the instructions. I get this error at the make autoloads step:
% pwd
/Users/skip/src/packages/unsupported
% make autoloads
[ -d scop ] && make -C scop autoloads
[ -d antlr-mode ] && make -C antlr-mode autoloads
make[2]: Nothing to be done for `autoloads'.
[ -d cogre ] && make -C cogre autoloads
make[2]: Nothing to be done for `autoloads'.
make[1]: *** No rule to make target `elib/autoloads.target', needed by
`vc/autoloads.target'. Stop.
make: *** [scop/autoloads.target] Error 2
Suggestions?
Skip
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
4:59 p.m.
On Tuesday 21 August 2007, skip(a)pobox.com wrote:
>> It's not clear to me what I'm supposed to
download and build. Do
>> you have a full URL (or something) for unsupported/scop/vc?
Ville> "unsupported" is a directory below a cvs up'd CVS checkout
of
the XEmacs Ville> packages tree.
Ville> (checkout packages from cvs)
Ville> (set up Local.rules from Local.rules.template)
Ville> $ cvs up -d # initial checkout doesn't IIRC fetch the
"unsupported" tree Ville> $ make autoloads
Ville> $ cd unsupported/scop/vc
Ville> $ make install
Thanks for the instructions. I get this error at the make autoloads step:
% pwd
/Users/skip/src/packages/unsupported
% make autoloads
[ -d scop ] && make -C scop autoloads
[ -d antlr-mode ] && make -C antlr-mode autoloads
make[2]: Nothing to be done for `autoloads'.
[ -d cogre ] && make -C cogre autoloads
make[2]: Nothing to be done for `autoloads'.
make[1]: *** No rule to make target `elib/autoloads.target', needed by
`vc/autoloads.target'. Stop.
make: *** [scop/autoloads.target] Error 2
Suggestions?
Other packages in packages/unsupported/scop than vc and generic-modes are very
much unsupported and may very well be broken, and the above appears to come
from the cogre package. I added elib to unsupported/scop/Makefile's
ASSUMED_PRESENT, perhaps that fixes it though.
But I don't think you need to run "make autoloads" at the
packages/unsupported
or packages/unsupported/scop levels, just cd directly to
packages/unsupported/scop/vc and run it there (after running it at the top
level, ie. in packages/).
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
12:50 a.m.
> Thanks for the instructions. I get this error at the make
autoloads step:
...
> Suggestions?
Ville> But I don't think you need to run "make autoloads" at the
Ville> packages/unsupported or packages/unsupported/scop levels, just cd
Ville> directly to packages/unsupported/scop/vc and run it there (after
Ville> running it at the top level, ie. in packages/).
Thanks. I got it installed, however I still get the same error when trying
to ediff-revision a file. From view-lossage output:
...
Symbol's value as variable is void: state
Loading vc-svn...done
Loading vc-svn...
Loading cl-macs...done
Loading cl-macs...
Loading func-menu...done
Loading func-menu...
Loading ediff...done
Loading ediff...
...
I tried removing the .elc files but that didn't help. What needs to be
fixed, vc, ediff, the byte code compiler or something else?
Thanks,
Skip
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
5:21 p.m.
On Wednesday 22 August 2007, skip(a)pobox.com wrote:
Ville> But I don't think you need to run "make autoloads" at the
Ville> packages/unsupported or packages/unsupported/scop levels, just
cd Ville> directly to packages/unsupported/scop/vc and run it there (after
Ville> running it at the top level, ie. in packages/).
Thanks. I got it installed, however I still get the same error when trying
to ediff-revision a file. From view-lossage output:
...
Symbol's value as variable is void: state
Loading vc-svn...done
Loading vc-svn...
Loading cl-macs...done
Loading cl-macs...
Loading func-menu...done
Loading func-menu...
Loading ediff...done
Loading ediff...
...
I tried removing the .elc files but that didn't help.
Just to clarify, which *.elc did you remove? Offhand, I would guess that
removing ediff's *.elc could help. Also, be sure to remove any old vc
packages possibly in your elisp load path so that the new vc is the only one
there.
What needs to be
fixed, vc, ediff, the byte code compiler or something else?
I hope it's just a matter of rebuilding ediff cleanly against the new vc. But
this issue is a bit convoluted for the moment as long as the new vc lives in
the unsupported area and some others from the regular packages apparently
need to be rebuilt against it - it should be easier to find out what the real
problem is and get it fixed when the new vc is in the regular packages tree.
I'm waiting for Norbert's ack before moving it there.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
6091
days inactive
6094
days old
14 comments
4 participants
participants (4)
-
Adrian Aichner -
Michael Kifer -
Skip Montanaro -
Ville Skyttä