Johan Corveleyn wrote:
> On Wed, Apr 28, 2010 at 11:58 PM, Neels J Hofmeyr <neels_at_elego.de> wrote:
> [...]
>> We use that in a hook script that forbids
>> commits that replace nodes -- because some users often do that by accident,
>> fragmenting history --
>
> Probably because of this bug?
> http://subversion.tigris.org/issues/show_bug.cgi?id=3429 - "svn mv A
> B; svn mv B A" generates replace without history
>
> Our users also run into that multiple times a month. It's just so easy
> for this to happen in an agile/Java/refactor-often environment...
> You're working on something; move things around a bit; change your
> mind; move some more; ready, commit. If this issue would be fixed,
> accidental replaces would be far less common ...
>
> Anyway, OT for this thread, but this is also one of those issues that
> I really hope to be resolved by wc-ng. Maybe it should be marked
> milestone 1.7?
Hi Johan,
wc-ng is not the solution of all problems. Basically, wc-ng is spring
cleaning and maturing the structure in which we store working copy metadata.
What is true though is that a substantial number of things will get
infinitely easier to fix once we are at 1.7. And it may also be true that
someone takes the time to fix #3429 before the 1.7 release cycle kicks off.
But these events are not necessarily directly correlated. Try to spread your
high hopes evenly about a given time span... :)
The fact that you're mentioning this might already be the cue point for
someone to start fixing this. But if you need it urgently, you can always
allocate man-power directly to this issue (try fixing yourself, or pay
someone to do it, maybe other power users are willing to chip in).
<obligatory-remark>That's what Open Source is all about.</obligatory-remark>
(And, I, too, hope it will get fixed sooner rather than later. Subversion
tracks history explicitly for each node, and that issue is obviously a Bad
break with this paradigm.)
~Neels
Received on 2010-04-29 13:19:54 CEST