You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by "C. Michael Pilato" <cm...@collab.net> on 2007/07/17 01:11:32 UTC
Re: svn commit: r25763 - trunk/subversion/libsvn_client
dlr@tigris.org wrote:
> Author: dlr
> Date: Mon Jul 16 17:45:21 2007
> New Revision: 25763
>
> Log:
> Change "revert" terminology to "rollback" throughout to differentiate
> between Subversion's 'revert' subcommand and a subtractive 'merge'.
Actually, I'm concerned about the use of the term rollback to mean "any ol'
reversal of a changeset". If I'm not mistaken, there are VC system in which
rollback means something different, namely the pruning of last N changes to
a line of history. In other words, given a file added in r1, changed in r3,
r6, and r9, it means nothing to "rollback r6". You can rollback *to* r6 (by
reversing r9). You can rollback *to* r1 (by reversing r3, r6, and r9). But
rollbacks don't have gaps in them.
I can't readily point to documentation of other VC systems to back my claim
(which is based entirely on now-fading memories). Am I the only person who
thinks of the term "rollback" in this way?
--
C. Michael Pilato <cm...@collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
Re: svn commit: r25763 - trunk/subversion/libsvn_client
Posted by Daniel Rall <dl...@collab.net>.
On Mon, 16 Jul 2007, C. Michael Pilato wrote:
> dlr@tigris.org wrote:
> > Author: dlr
> > Date: Mon Jul 16 17:45:21 2007
> > New Revision: 25763
> >
> > Log:
> > Change "revert" terminology to "rollback" throughout to differentiate
> > between Subversion's 'revert' subcommand and a subtractive 'merge'.
>
> Actually, I'm concerned about the use of the term rollback to mean "any ol'
> reversal of a changeset". If I'm not mistaken, there are VC system in which
> rollback means something different, namely the pruning of last N changes to
> a line of history. In other words, given a file added in r1, changed in r3,
> r6, and r9, it means nothing to "rollback r6". You can rollback *to* r6 (by
> reversing r9). You can rollback *to* r1 (by reversing r3, r6, and r9). But
> rollbacks don't have gaps in them.
>
> I can't readily point to documentation of other VC systems to back my claim
> (which is based entirely on now-fading memories). Am I the only person who
> thinks of the term "rollback" in this way?
FWIW, svnmerge.py uses the term "rollback" in exactly the way used in
this commit.
Re: svn commit: r25763 - trunk/subversion/libsvn_client
Posted by Daniel Rall <dl...@collab.net>.
On Mon, 16 Jul 2007, Blair Zajac wrote:
> C. Michael Pilato wrote:
> >dlr@tigris.org wrote:
> >>Author: dlr
> >>Date: Mon Jul 16 17:45:21 2007
> >>New Revision: 25763
> >>
> >>Log:
> >>Change "revert" terminology to "rollback" throughout to differentiate
> >>between Subversion's 'revert' subcommand and a subtractive 'merge'.
> >
> >Actually, I'm concerned about the use of the term rollback to mean "any ol'
> >reversal of a changeset". If I'm not mistaken, there are VC system in
> >which
> >rollback means something different, namely the pruning of last N changes to
> >a line of history. In other words, given a file added in r1, changed in
> >r3,
> >r6, and r9, it means nothing to "rollback r6". You can rollback *to* r6
> >(by
> >reversing r9). You can rollback *to* r1 (by reversing r3, r6, and r9).
> >But
> >rollbacks don't have gaps in them.
> >
> >I can't readily point to documentation of other VC systems to back my claim
> >(which is based entirely on now-fading memories). Am I the only person who
> >thinks of the term "rollback" in this way?
>
> Well, I would guess that people coming from a database background will
> think of rollbacks the way you describe them.
>
> Why don't we call them reverse merges? There's no ambiguity there.
Yup. Just seemed a little long for a variable name...
Re: svn commit: r25763 - trunk/subversion/libsvn_client
Posted by Blair Zajac <bl...@orcaware.com>.
C. Michael Pilato wrote:
> dlr@tigris.org wrote:
>> Author: dlr
>> Date: Mon Jul 16 17:45:21 2007
>> New Revision: 25763
>>
>> Log:
>> Change "revert" terminology to "rollback" throughout to differentiate
>> between Subversion's 'revert' subcommand and a subtractive 'merge'.
>
> Actually, I'm concerned about the use of the term rollback to mean "any ol'
> reversal of a changeset". If I'm not mistaken, there are VC system in which
> rollback means something different, namely the pruning of last N changes to
> a line of history. In other words, given a file added in r1, changed in r3,
> r6, and r9, it means nothing to "rollback r6". You can rollback *to* r6 (by
> reversing r9). You can rollback *to* r1 (by reversing r3, r6, and r9). But
> rollbacks don't have gaps in them.
>
> I can't readily point to documentation of other VC systems to back my claim
> (which is based entirely on now-fading memories). Am I the only person who
> thinks of the term "rollback" in this way?
Well, I would guess that people coming from a database background will
think of rollbacks the way you describe them.
Why don't we call them reverse merges? There's no ambiguity there.
Regards,
Blair
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: svn commit: r25763 - trunk/subversion/libsvn_client
Posted by David Glasser <gl...@mit.edu>.
On 7/16/07, Eric S. Raymond <es...@thyrsus.com> wrote:
> C. Michael Pilato <cm...@collab.net>:
> > Actually, I'm concerned about the use of the term rollback to mean "any ol'
> > reversal of a changeset". If I'm not mistaken, there are VC system in which
> > rollback means something different, namely the pruning of last N changes to
> > a line of history. In other words, given a file added in r1, changed in r3,
> > r6, and r9, it means nothing to "rollback r6". You can rollback *to* r6 (by
> > reversing r9). You can rollback *to* r1 (by reversing r3, r6, and r9). But
> > rollbacks don't have gaps in them.
> >
> > I can't readily point to documentation of other VC systems to back my claim
> > (which is based entirely on now-fading memories). Am I the only person who
> > thinks of the term "rollback" in this way?
>
> No. New VC mode, following some early VCSes (SCCS, RCS) and some very
> recent ones (notably Mercurial), uses "rollback" in exactly that way.
(To be pedantic, Mercurial doesn't refer to it as "rollback r6" *or*
"rollback to r6" because it supports exactly one revision worth of
rollback: you just "rollback the last transaction". But yeah, it's
more like what cmpilato said than like svnmerge.py rollback.)
--dave
--
David Glasser | glasser@mit.edu | http://www.davidglasser.net/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: svn commit: r25763 - trunk/subversion/libsvn_client
Posted by "Eric S. Raymond" <es...@thyrsus.com>.
C. Michael Pilato <cm...@collab.net>:
> Actually, I'm concerned about the use of the term rollback to mean "any ol'
> reversal of a changeset". If I'm not mistaken, there are VC system in which
> rollback means something different, namely the pruning of last N changes to
> a line of history. In other words, given a file added in r1, changed in r3,
> r6, and r9, it means nothing to "rollback r6". You can rollback *to* r6 (by
> reversing r9). You can rollback *to* r1 (by reversing r3, r6, and r9). But
> rollbacks don't have gaps in them.
>
> I can't readily point to documentation of other VC systems to back my claim
> (which is based entirely on now-fading memories). Am I the only person who
> thinks of the term "rollback" in this way?
No. New VC mode, following some early VCSes (SCCS, RCS) and some very
recent ones (notably Mercurial), uses "rollback" in exactly that way.
It is *not* a synonym for reversion.
--
<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: svn commit: r25763 - trunk/subversion/libsvn_client
Posted by Nathan Kidd <na...@spicycrypto.ca>.
C. Michael Pilato wrote:
> dlr@tigris.org wrote:
>>
>> Change "revert" terminology to "rollback" throughout to differentiate
>> between Subversion's 'revert' subcommand and a subtractive 'merge'.
>
> Actually, I'm concerned about the use of the term rollback to mean "any ol'
> reversal of a changeset". If I'm not mistaken, there are VC system in which
> rollback means something different, namely the pruning of last N changes to
> a line of history. In other words, given a file added in r1, changed in r3,
> r6, and r9, it means nothing to "rollback r6". You can rollback *to* r6 (by
> reversing r9). You can rollback *to* r1 (by reversing r3, r6, and r9). But
> rollbacks don't have gaps in them.
>
> I can't readily point to documentation of other VC systems to back my claim
> (which is based entirely on now-fading memories). Am I the only person who
> thinks of the term "rollback" in this way?
Microsoft Visual Sourcesafe (ha!) has a "Rollback" button in its history
view, which will permanently erase all history after the selected file
version, making that version of the file the latest.
On the Windows side I think a very large amount of users (if not the
majority) are upgrading to SVN from VSS. Avoiding any confusion from
similar terminology will surely help with the support burden down the road.
-Nathan
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org