On 2012-06-19 11:53:37 +0200, Stefan Sperling wrote: > On Tue, Jun 19, 2012 at 09:46:20AM +0200, Vincent Lefevre wrote: > > On 2012-06-19 01:29:18 -0400, Greg Stein wrote: > > > In 1.6, we erroneously used the containing directory's revision for the > > > file in certain cases. 1.7 is correct: the file is not changed until r4. > > > Maybe the directory has, but that is independent of the file. > > > > But in this case, is it normal that r3 is still in the log of the > > file? For instance, even though r3 is not a changed rev of the file, > > "svn log -v file" gives: > > > > ------------------------------------------------------------------------ > > r4 | vinc17 | 2012-06-19 09:36:50 +0200 (Tue, 19 Jun 2012) | 1 line > > Changed paths: > > D /dir2/file > > A /file (from /dir2/file:3) > > > > mv dir2/file . > > ------------------------------------------------------------------------ > > r3 | vinc17 | 2012-06-19 09:36:47 +0200 (Tue, 19 Jun 2012) | 1 line > > Changed paths: > > D /dir1 > > A /dir2 (from /dir1:2) > > > > mv dir1 dir2 > > ------------------------------------------------------------------------ > > r2 | vinc17 | 2012-06-19 09:36:44 +0200 (Tue, 19 Jun 2012) | 1 line > > Changed paths: > > A /dir1/file > > > > add dir1/file > > ------------------------------------------------------------------------ > > > > IMHO, this can be useful information in the log history, but the > > difference between what is regarded as a "changed rev" of a file > > and the file history that appears in the log doesn't seem to be > > documented. > > 'svn log' shows both simple copy events and content changes and there > is no way to turn off display of pure copy events.
Well, not all copy events should be affected. For instance, if a file is moved from a directory to another one, this should be shown in the log. Here what is copied is not the file itself, but the directory above it. > I would argue this is, if at all, a cosmetic problem in 'svn log' output > that can be worked around with a filtering wrapper. I'm not complaining much on the output except that it is not consistent with the last changed rev when the copy was done. The question is: Is this regarded as a bug? If it is not, this should be documented. > Say there was an option to 'svn log' that caused only commits which changed > file content to be shown. What should this option if 'svn log' is invoked > on a directory, instead of a file? Ignore the option, or complain, or > something else? A directory also has a notion of last changed revision. The rule could be: If when some revision N was committed, it was the last changed rev of the object, then it should be shown in the log. -- Vincent Lefèvre <vinc...@vinc17.net> - Web: <http://www.vinc17.net/> 100% accessible validated (X)HTML - Blog: <http://www.vinc17.net/blog/> Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)
