You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Jim Blandy <ji...@red-bean.com> on 2002/07/26 17:39:18 UTC
When did a file last change?
Is there any reliable way to find the revision number at which a file
was last changed?
One command that Emacs's VC package provides is vc-diff (C-x v =).
Normally, vc-diff shows you your uncommitted changes. But you can
also invoke in a way that lets you get a diff between arbitrary
revisions of the file. It prompts you for the "from" and "to"
revisions. These are supposed to default to immediately before and
immediately after the last time the file was changes, so it's trivial
to see the last thing that happened to a given file.
Under CVS, if you've got a file at 1.38, say, then you know its
immediate predecessor is 1.37. But since Subversion uses global
revision numbers, it's not so easy. A file may be at revision '1093',
but you have no idea what revision it last changed in.
Does the working copy have this info? Is there some straightforward
way to get it?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: building apache for svn under OS X
Posted by Hamilton Link <he...@sandia.gov>.
Thanks for your response, it confirms my suspicion. I don't have the
hard drive space for fink, and I think that's been the source of some of
my problems -- because you can build apache and subversion and a bunch
of things from the sources, but autoconf and libtool both have to be
darwinized to work. The fink versions have been tailored, the versions I
had weren't.
I finally found a README.platforms in httpd-2.0 that led me to
darwinized versions of just the two things I needed. If it all works I'm
preparing a "HOWTO build svn in OS X without fink" email that might save
the next guy some README-crawling.
hamilton
David Mankin wrote:
>
> This won't be a very helpful reply, but I'll send it just in case it is...
>
> It's been a couple months since I tried, but I'm pretty sure I had
> not too much trouble getting this all to work. (I did have to tweak
> one apache script so that it could find apxs.) I got most of the
> build-tool dependencies out of fink. Are you using fink? I have the
> fink libtool14 and libtool14-shlib packages installed.
>
> HTH,
> -David
>
> >I know this isn't an apache forum, but the reason I'm hung up trying to
> >get apache building smoothly is because neon is having trouble finding
> >expat when I build svn later on...
> >
> >I think I'm using the CVS head of apache, along with the latest apr and
> >apr-util from CVS, the 9am Jul 25th tarballs didn't work for me, and the
> >9am Jul 26th tarballs didn't work either (I'm getting them from
> >http://cvs.apache.org/snapshots/httpd-2.0/, apr/, and apr-util/, isn't
> >this correct?
> >
> >The bottom line is, should I report this to devapache and call it a day
> >or do I still have some chance of getting subversion working? Is the
> >approach of replacing apr and apr-util's copies of libtool with a newer
> >version before I run make advisable or not? Has anyone else seen this
> >problem or got a clue to what I'm doing wrong?
> >
> >Following the directions in svn/INSTALL for building apache with apr and
> >apr-util (and it's not like I've never built apache before) and not
> >swapping apr and apr-util's self-generated libtool 1.3.5 for anything
> >else, buildconf gives me warnings about deprecated stuff but everything
> >configures and make depends perfectly (looks like) and I eventually end
> >up with the following error during make:
> >
> >/bin/sh /Users/helink/archives/subversion/httpd-2.0/srclib/apr/libtool
> >etc
> >ld: -undefined error must be used when -twolevel-namespace is in effect
> >/usr/bin/libtool: internal link edit command failed
> >make[3]: *** [libapr.la] Error 1
> >make[2]: *** [all-recursive] Error 1
> >make[1]: *** [all-recursive] Error 1
> >make: *** [all-recursive] Error 1
> >
> >Now, in OS X /usr/bin/libtool isn't actually libtool, it's (apparantly)
> >something totally unrelated -- /usr/bin/glibtool is actually libtool.
> >Dunno if this matters because it appears to be calling it's own private
> >libtool when it gets the error, and why that's not the right version of
> >libtool (the httpd docs say it requires 1.4.2, buildconf detects it, and
> >they build themselves 1.3.5) I don't know.
> >
> >hamilton
> >
> >
> >---------------------------------------------------------------------
> >To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
> >For additional commands, e-mail: dev-help@subversion.tigris.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: building apache for svn under OS X
Posted by Hamilton Link <he...@sandia.gov>.
I know this isn't an apache forum, but the reason I'm hung up trying to
get apache building smoothly is because neon is having trouble finding
expat when I build svn later on...
I think I'm using the CVS head of apache, along with the latest apr and
apr-util from CVS, the 9am Jul 25th tarballs didn't work for me, and the
9am Jul 26th tarballs didn't work either (I'm getting them from
http://cvs.apache.org/snapshots/httpd-2.0/, apr/, and apr-util/, isn't
this correct?
The bottom line is, should I report this to devapache and call it a day
or do I still have some chance of getting subversion working? Is the
approach of replacing apr and apr-util's copies of libtool with a newer
version before I run make advisable or not? Has anyone else seen this
problem or got a clue to what I'm doing wrong?
Following the directions in svn/INSTALL for building apache with apr and
apr-util (and it's not like I've never built apache before) and not
swapping apr and apr-util's self-generated libtool 1.3.5 for anything
else, buildconf gives me warnings about deprecated stuff but everything
configures and make depends perfectly (looks like) and I eventually end
up with the following error during make:
/bin/sh /Users/helink/archives/subversion/httpd-2.0/srclib/apr/libtool
etc
ld: -undefined error must be used when -twolevel-namespace is in effect
/usr/bin/libtool: internal link edit command failed
make[3]: *** [libapr.la] Error 1
make[2]: *** [all-recursive] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all-recursive] Error 1
Now, in OS X /usr/bin/libtool isn't actually libtool, it's (apparantly)
something totally unrelated -- /usr/bin/glibtool is actually libtool.
Dunno if this matters because it appears to be calling it's own private
libtool when it gets the error, and why that's not the right version of
libtool (the httpd docs say it requires 1.4.2, buildconf detects it, and
they build themselves 1.3.5) I don't know.
hamilton
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: error finding expat under OS X
Posted by Hamilton Link <he...@sandia.gov>.
I thought I had.
The subversion tarball worked to get me a client and I was able to check
out the latest version of the source from the repository with it
yesterday. I rm -rfed it and did it again this morning, so I'm using rev
2737 at this point.
I can't get cvs to go through the firewall here at work, so I got the
"every six hours snapshot of CVS" tarballs of apache, apr, and apr-util
(last night at like 5pm MST). I untarred httpd-2.0 and unpacked apr and
apr-util into them, ran configure, replaced libtool with the newer
version, and did the remaining make steps.
I downloaded release versions of Python 2.2.1, Berkeley DB 4.0.14, and
Neon 0.21.3. I built Python and BDB on their own, they worked fine.
My new autoconf (2.53) builds fine, but if I run "make check" it reports
a bunch of failures.
Libtool 1.4.2 I got out of Berkeley DB after I built it.
Neon generates the expat problem when it's in a neon directory in the
latest svn directory.
I'm using OS X 10.1.5.
hamilton
Ben Collins-Sussman wrote:
>
> "Hamilton Link" <he...@sandia.gov> writes:
>
> > Hamilton Link wrote:
> > > I just did XYZ and it appears to make things work properly.
> >
> > ... that is to say, until it dives into neon (untarred into the
> > checked-out svn directory and renamed to neon, like in the instructions)
> > and complains that it can't find expat.h or any other XML stuff. When I
> > look for it, /usr/local/apache/include/expat.h exists. I *thought* the
> > apache build process worked fine, perhaps I was wrong. Any advice?
> >
> > Freshly untarred, httpd-2.0 (with apr and apr-util unpacked into srclib)
>
> Are you using the CVS head of apache, like the INSTALL instructions
> say? Along with the latest apr and apr-util from CVS?
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: dev-help@subversion.tigris.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: error finding expat under OS X
Posted by Ben Collins-Sussman <su...@collab.net>.
"Hamilton Link" <he...@sandia.gov> writes:
> Hamilton Link wrote:
> > I just did XYZ and it appears to make things work properly.
>
> ... that is to say, until it dives into neon (untarred into the
> checked-out svn directory and renamed to neon, like in the instructions)
> and complains that it can't find expat.h or any other XML stuff. When I
> look for it, /usr/local/apache/include/expat.h exists. I *thought* the
> apache build process worked fine, perhaps I was wrong. Any advice?
>
> Freshly untarred, httpd-2.0 (with apr and apr-util unpacked into srclib)
Are you using the CVS head of apache, like the INSTALL instructions
say? Along with the latest apr and apr-util from CVS?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
error finding expat under OS X
Posted by Hamilton Link <he...@sandia.gov>.
Hamilton Link wrote:
> I just did XYZ and it appears to make things work properly.
... that is to say, until it dives into neon (untarred into the
checked-out svn directory and renamed to neon, like in the instructions)
and complains that it can't find expat.h or any other XML stuff. When I
look for it, /usr/local/apache/include/expat.h exists. I *thought* the
apache build process worked fine, perhaps I was wrong. Any advice?
Freshly untarred, httpd-2.0 (with apr and apr-util unpacked into srclib)
dies in make, I think because apr and apr-util both prepare themselves
libtool 1.3.5, replacing these with libtool 1.4.2 before making the
system makes it complete, but it generates some non-fatal errors and
warnings along the way. I should probably write to apache's mailing
list, but has anyone here had more success building this under OS X?
hamilton
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ltconfig error under OS X
Posted by Hamilton Link <he...@sandia.gov>.
Hamilton Link wrote:
>
> I'm in the home stretch of building the svn server under OS X, and I'm
> getting the following error when I run
>
> % ./configure --with-apr=/usr/local/apache2
> --with-apr-util=/usr/local/apache2 --enable-maintainer-mode
>
> ...
> checking build system type... powerpc-apple-darwin5.5
> checking host system type... powerpc-apple-darwin5.5
> ...
> ltconfig: you must specify a host type if you use `--no-verify'
> Try `ltconfig --help' for more information.
> configure: error: libtool configure failed
>
> Now, I look in configure and it does call ltconfig with --no-verify, can
> I just remove that and move on?
I just did this and it appears to make things work properly.
hamilton
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
ltconfig error under OS X
Posted by Hamilton Link <he...@sandia.gov>.
I'm in the home stretch of building the svn server under OS X, and I'm
getting the following error when I run
% ./configure --with-apr=/usr/local/apache2
--with-apr-util=/usr/local/apache2 --enable-maintainer-mode
...
checking build system type... powerpc-apple-darwin5.5
checking host system type... powerpc-apple-darwin5.5
...
ltconfig: you must specify a host type if you use `--no-verify'
Try `ltconfig --help' for more information.
configure: error: libtool configure failed
Now, I look in configure and it does call ltconfig with --no-verify, can
I just remove that and move on, or does someone have a more
sophisticated multiplatform solution? I'll probably try it my way before
someone responds, but I'm wondering if I'm doing something horribly
wrong despite reading through all the various INSTALL files about six
times.
I also find that I have to replace apr and apr-util's versions of
libtool with libtool 1.4.2 (they build 1.3.5 for themselves, for some
reason... perhaps they're making a copy of the one on my system, and
finding the wrong one in the path?). This is very hinky but not the
problem of the svn team.
Anyway if someone could help me resolve this ltconfig issue that'd be
great.
thanks
hamilton
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: When did a file last change?
Posted by Ben Collins-Sussman <su...@collab.net>.
Jim Blandy <ji...@red-bean.com> writes:
> Ben Collins-Sussman <su...@collab.net> writes:
> > Yes, it's in the status output, already cached in the entries file.
> >
> > Try 'svn st -v', and 'svn help st'.
>
> Okay, this works because when I see something like this:
>
> zenia:subversion$ svn status -v CHANGES
> _ 2711 2667 kfogel CHANGES
>
> the second number is guaranteed to be no greater than the first
> number. Thus the recent fix to the docs.
Right.
> However, in your conversation with Alexis Huxley, you pondered the
> possibility of changing this so that the second number might be
> greater than the first, if you passed `svn status' the `-u' flag, to
> allow it to consult the repository and see the latest revision.
>
> I understand that was just idle discussion, but if it does happen, I'd
> like to request that there remain *some* way to find the most recent
> change that produced the text the user has in hand. Running `svn
> status -u' really shouldn't drop that older number in favor of the
> newer number.
OK, then maybe put a 'future' rev number next to the asterisk, or
something, should we decide to implement this feature.
>
> > Alternately, 'svn diff -rPREV' will show you the last change to the
> > file. :-)
>
> The `-rPREV' part should appear in the `diff' command's help,
> shouldn't it?
Yeah. I guess so. It's fully documented in the Handbook (nudge,
nudge...), but it sort of begs the general question: how much
information is in the handbook, and how much can be built into 'svn
help'?
I mean, it already seems like pretty extreme to me that 'svn help
propset' now documents *every* special svn:property. What if we add
new svn: props? Are we going to remember to keep the Handbook and
help text in sync?
I'm of the opinion that the built-in help should document syntax, and
be changed when syntax changes. But the existence of special props,
or special revision names... that kind of stuff can change all the
time. That's Handbook material.
The Handbook covers a wide array of topics: Modules, Vendor branches,
special props and revision names... things we could never fit into the
help system. I'd rather the help system give a pointer to the
Handbook.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: When did a file last change?
Posted by Ben Collins-Sussman <su...@collab.net>.
Branko Čibej <br...@xbc.nu> writes:
> Well, then let's make the * part of the first column, which then becomes:
>
> __L+*
>
Fine with me. :-)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: When did a file last change?
Posted by Branko Čibej <br...@xbc.nu>.
Ben Collins-Sussman wrote:
>Branko Čibej <br...@xbc.nu> writes:
>
>
>
>>Now, if we change "svn st -u" to print the info that's current in the
>>repository, the second result might change to:
>>
>> _ * 2711 2767 jimb CHANGES
>>
>>It's not enough to print the "current" last-changed rev number, you
>>have to print the author, too. Now that you've gone that far, it
>>strikes me that the asterisk in the second column is extraneous -- you
>>_already_ know your working copy is not up-to-date, because the
>>last-changed revision is greater than the wc revision.
>>
>>
>
>Oh c'mon, Brane. You want to look down a list of 20 items, and
>*manually* notice exactly which ones have CR's greater than the
>working rev? That's absurd. There's no way I want to think that
>hard!
>
>The whole point of the asterisk is to quickly scan down the column and
>just instantly *know* what the next update will do, without doing
>calculations. In particular, you can instantly predict conflicts by
>looking for "M *" on the same line.
>
Well, then let's make the * part of the first column, which then becomes:
__L+*
Or, in regex-speak, /^[_!?ACDM][_!ACDM ][L ][+ ][* ]/.
--
Brane Čibej <br...@xbc.nu> http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: When did a file last change?
Posted by Ben Collins-Sussman <su...@collab.net>.
Branko Čibej <br...@xbc.nu> writes:
> Now, if we change "svn st -u" to print the info that's current in the
> repository, the second result might change to:
>
> _ * 2711 2767 jimb CHANGES
>
> It's not enough to print the "current" last-changed rev number, you
> have to print the author, too. Now that you've gone that far, it
> strikes me that the asterisk in the second column is extraneous -- you
> _already_ know your working copy is not up-to-date, because the
> last-changed revision is greater than the wc revision.
Oh c'mon, Brane. You want to look down a list of 20 items, and
*manually* notice exactly which ones have CR's greater than the
working rev? That's absurd. There's no way I want to think that
hard!
The whole point of the asterisk is to quickly scan down the column and
just instantly *know* what the next update will do, without doing
calculations. In particular, you can instantly predict conflicts by
looking for "M *" on the same line.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: When did a file last change?
Posted by Branko Čibej <br...@xbc.nu>.
Karl Fogel wrote:
>But the main point is, since the information is local anyway, I think
>it's safe to say that there will *always* be a way to get at it using
>the command-line client. Probably we'd add a new column to the output
>of "svn status -v" (rather than changing the meaning of the existing
>column), but either way, there'll be plenty of warning, and it'll be
>before 1.0.
>
>
No, please, I promise I'll be good, not another column ...
It's not necessary. "svn status" meanse "Tell me what you know about
this file". So, if you do a "svn st -v" to get
_ 2711 2667 kfogel CHANGES
the correct interpretation is: "You have revision 2711, and the contents
were created in 2667 by kfogel." This information is always correct and
unambiguous. Now, if you do a "svn st -uv", you may well get:
_ * 2711 2667 kfogel CHANGES
The above interpretation is still correct, you just add "but the file
has changed in the repository since then."
Now, if we change "svn st -u" to print the info that's current in the
repository, the second result might change to:
_ * 2711 2767 jimb CHANGES
It's not enough to print the "current" last-changed rev number, you have
to print the author, too. Now that you've gone that far, it strikes me
that the asterisk in the second column is extraneous -- you _already_
know your working copy is not up-to-date, because the last-changed
revision is greater than the wc revision.
So, instead of adding another column, I vote for _removing_ one. And
if/when we do that, we can widen the "author" column from 8 chars to at
least something that'll hold the username of every Subversion developer. :-)
--
Brane Čibej <br...@xbc.nu> http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: When did a file last change?
Posted by Karl Fogel <kf...@newton.ch.collab.net>.
Jim Blandy <ji...@red-bean.com> writes:
> the second number is guaranteed to be no greater than the first
> number. Thus the recent fix to the docs.
>
> However, in your conversation with Alexis Huxley, you pondered the
> possibility of changing this so that the second number might be
> greater than the first, if you passed `svn status' the `-u' flag, to
> allow it to consult the repository and see the latest revision.
This does touch on something important, though, I think:
The last-changed-rev is available *without* talking to the repository,
because it's recorded in the .svn/entries file:
<entry
committed-rev="2566"
name="Makefile.in"
text-time="2002-07-17T16:21:58.000000Z"
committed-date="2002-07-17T15:41:55.685214Z"
checksum="WMQuKeJsE1SjuP1yJtNBJA=="
last-author="sussman"
kind="file"
prop-time="2002-07-17T16:21:58.000000Z"/>
And "svn status -v" isn't the only thing that prints it; there's also
"svn info"...
But the main point is, since the information is local anyway, I think
it's safe to say that there will *always* be a way to get at it using
the command-line client. Probably we'd add a new column to the output
of "svn status -v" (rather than changing the meaning of the existing
column), but either way, there'll be plenty of warning, and it'll be
before 1.0.
-Karl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: When did a file last change?
Posted by Jim Blandy <ji...@red-bean.com>.
Ben Collins-Sussman <su...@collab.net> writes:
> Yes, it's in the status output, already cached in the entries file.
>
> Try 'svn st -v', and 'svn help st'.
Okay, this works because when I see something like this:
zenia:subversion$ svn status -v CHANGES
_ 2711 2667 kfogel CHANGES
the second number is guaranteed to be no greater than the first
number. Thus the recent fix to the docs.
However, in your conversation with Alexis Huxley, you pondered the
possibility of changing this so that the second number might be
greater than the first, if you passed `svn status' the `-u' flag, to
allow it to consult the repository and see the latest revision.
I understand that was just idle discussion, but if it does happen, I'd
like to request that there remain *some* way to find the most recent
change that produced the text the user has in hand. Running `svn
status -u' really shouldn't drop that older number in favor of the
newer number.
> Alternately, 'svn diff -rPREV' will show you the last change to the
> file. :-)
The `-rPREV' part should appear in the `diff' command's help,
shouldn't it?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: When did a file last change?
Posted by Ben Collins-Sussman <su...@collab.net>.
Yes, it's in the status output, already cached in the entries file.
Try 'svn st -v', and 'svn help st'.
Alternately, 'svn diff -rPREV' will show you the last change to the
file. :-)
Jim Blandy <ji...@red-bean.com> writes:
> Is there any reliable way to find the revision number at which a file
> was last changed?
>
> One command that Emacs's VC package provides is vc-diff (C-x v =).
> Normally, vc-diff shows you your uncommitted changes. But you can
> also invoke in a way that lets you get a diff between arbitrary
> revisions of the file. It prompts you for the "from" and "to"
> revisions. These are supposed to default to immediately before and
> immediately after the last time the file was changes, so it's trivial
> to see the last thing that happened to a given file.
>
> Under CVS, if you've got a file at 1.38, say, then you know its
> immediate predecessor is 1.37. But since Subversion uses global
> revision numbers, it's not so easy. A file may be at revision '1093',
> but you have no idea what revision it last changed in.
>
> Does the working copy have this info? Is there some straightforward
> way to get it?
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: dev-help@subversion.tigris.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: When did a file last change?
Posted by Ben Collins-Sussman <su...@collab.net>.
"Bill Tutt" <ra...@lyra.org> writes:
> We'd talked about supporting a -r -N syntax that would make things like
> this easier. I don't know if your question can be easily answered atm.
That '-r -N' syntax he's describing aren't two switches; he's talking
about implementing negative revisions:
svn diff -r-3 foo.c
...might show you how foo looked three "changes" ago, however far back
that may be.
For now, 'svn log foo.c' will list all revs in which foo.c changed,
and the most recent rev of this set is printed by 'svn st -v'.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
RE: When did a file last change?
Posted by Bill Tutt <ra...@lyra.org>.
We'd talked about supporting a -r -N syntax that would make things like
this easier. I don't know if your question can be easily answered atm.
I think somebody might have logged an issue about it as well.
FYI,
Bill
> From: Jim Blandy [mailto:jimb@red-bean.com]
>
>
> Is there any reliable way to find the revision number at which a file
> was last changed?
>
> One command that Emacs's VC package provides is vc-diff (C-x v =).
> Normally, vc-diff shows you your uncommitted changes. But you can
> also invoke in a way that lets you get a diff between arbitrary
> revisions of the file. It prompts you for the "from" and "to"
> revisions. These are supposed to default to immediately before and
> immediately after the last time the file was changes, so it's trivial
> to see the last thing that happened to a given file.
>
> Under CVS, if you've got a file at 1.38, say, then you know its
> immediate predecessor is 1.37. But since Subversion uses global
> revision numbers, it's not so easy. A file may be at revision '1093',
> but you have no idea what revision it last changed in.
>
> Does the working copy have this info? Is there some straightforward
> way to get it?
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: dev-help@subversion.tigris.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org