[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: "svn diff -c" behavior on file copy from an old revision

From: Johan Corveleyn <jcorvel_at_gmail.com>
Date: Wed, 20 Nov 2019 15:21:22 +0100

On Tue, Nov 19, 2019 at 4:53 PM Nathan Hartman <hartman.nathan_at_gmail.com> wrote:
> On Mon, Nov 18, 2019 at 4:44 AM Vincent Lefevre <vincent-svn_at_vinc17.net> wrote:
>> I have the following issue with svn 1.10.6:
>> Assume that I committed "file1" at revision 1, did some unrelated
>> change at revision 2, and for revision 3, copied "file1_at_1" to "file2"
>> with "svn copy"[*] and did some changes in file2 before the commit.
>> [*] The revision older than the latest one is what happens when one
>> does a "svn copy" directly from the working copy without an update
>> first.
>> If I do "svn diff -r 1:3 file2", then I get the changes that have been
>> introduced between file1 and file2. But if I do "svn diff -c 3 file2",
>> which is equivalent to "svn diff -r 2:3 file2", then I get the whole
>> file2 content, as if file2 were an entirely new file.
>> I'm wondering whether this is the expected behavior. In any case,
>> this behavior is rather unintuitive and rather useless. I think there
>> should be an easy way to get the changes introduced by a commit.

[ snipped note ... not sure this is the same problem, see below. ]

> Hello Vincent,
> Thank you for reporting this.
> I think it is the expected behavior because file2 does not exist in revision 2 and by using '-c 3' (equivalent to -r 2:3) you're restricting the revisions that are included in the differencing; as it does not follow the copy source farther back than revision 2, the file is considered a new file.
> In contrast, with -r 1:3, since the revisions are not restricted (or at least not as restricted; revision 1 is now included in the differencing), it follows the copy backwards to revision 1 and shows only the differences.
> I agree it would be nice if it could show only the changes in the copy, but then a command like 'diff -r 2:3' would be ambiguous... How would it know if you want it to trace history beyond the range specified, or if you want it to respect the -r 2:3 limit?
> Thoughts?

This issue is _ancient_ :-).

FWIW: I agree with you, Vincent. I expect the same behavior as you do.
But this has been discussed at length before, and so far no one has
managed to (find consensus to) change this, or to even invent a new
option to get the behavior you and I find intuitive.

For some background, see this thread from dev@ in 2013 [1]:
    1.6 vs. 1.8: strange behavior of 'svn diff -cN WC-FILE' if the
file was created in rev N by copying

In my last post in that thread I tried the "but this is inconsistent /
unpredictable" card, to no avail [2]:
    "Maybe I gave the wrong impression, but I'm not trying to completely
     redefine the behavior of 'diff -c'. However the current behavior, when
     looking at a moved file, is unpredictable: it depends on whether or
     not the move source's revision was exactly 1 revision before the
     move-commit. That unpredictability is not good."

Somewhere in that thread I also dropped a pointer to an earlier thread
in which I observed (in passing) that this was a difference between
'svn' and 'svnlook' [3].
    "So it seems that detecting copies with their sources is broken in
'svn diff'"

[1] https://svn.haxx.se/dev/archive-2013-06/0588.shtml
[2] https://svn.haxx.se/dev/archive-2013-06/0705.shtml
[3] https://svn.haxx.se/dev/archive-2012-11/0480.shtml

Vincent Lefevre also wrote:
>> Note: "svn cat -r... file2" or "svn cat -r... file2_at_3" also shows a
>> similar behavior:
>> -r1: one gets file1_at_1
>> -r2: "Unable to find repository location for..." error
>> -r3: one gets file2_at_3

Hm, it might be related, but I don't see this as exactly the same
problem. IMHO this is normal and correct behavior.

Indeed, in r2 there was no file2. In r1 there was a predecessor of
file2_at_3, namely file1_at_1. You could argue that in r2 it should show the
contents of file1_at_1 (which were incidentally unchanged in r2). But
what if file1 would have been changed in r2 (yet file2_at_3 was a copy of
file1_at_1), what would you expect svn to show? Or what if file1 was
deleted in r2?

I guess the same questions can be asked for 'svn diff -c 3' (what if
file1 had a different content in r2, or was deleted in r2?). Yet in
that case I intuitively expect "diff -c3" to take the copy-source into
account, no matter if it has made a jump through history. As I argued
in [3] above, it seems 'svnlook' can do this with its --diff-copy-from
option ... (which I happen to like for generating diffs in post-commit

Received on 2019-11-20 15:21:48 CET

This is an archived mail posted to the Subversion Users mailing list.