[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Silly unedit behavior w.r.t. timestamp.
From: |
Kaz Kylheku |
Subject: |
Re: Silly unedit behavior w.r.t. timestamp. |
Date: |
Sun, 13 Oct 2002 16:38:10 -0700 (PDT) |
I've noticed that if you execute a cvs unedit, and choose to revert
the changes, it reverts the timestamp as well! The behavior should
be like the new -C option of update.
The problem with reverting the timestamp is that the file might
look older than some dependent file, and so the dependent file won't
be rebuilt; it will continue to be based on the edited version
that was thrown away.
Another problem I noticed is that when there are no changes, unedit
re-creates the file from scratch, and preserves its timestamp. It
should just leave the darn file alone if it has no changes.
In Meta-CVS, this causes annoying behavior, because a hard link is
broken for no good reason, and then there are two identical objects
that have the same timestamp:
$ mcvs edit NEWS
$ mcvs unedit NEWS
The following error has occured:
MCVS/F-30C08E3165D36B410E702C27410DCCCE and NEWS have the same modification
time.
You have these alternatives:
?) Re-print this menu.
C) (Continue) Continue synchronizing files.
A) Auto-continue all continuable errors.
T) Terminate.
These special alternatives are also available:
1) take MCVS/F-30C08E3165D36B410E702C27410DCCCE; clobber NEWS.
2) take NEWS; clobber MCVS/F-30C08E3165D36B410E702C27410DCCCE.