On Tue, 2004-04-27 at 22:44, Daniel F Garcia wrote:
>
> * I have a developer (Developer 1) who made extensive changes to a source
> file
> * The last known time he had the file was last Tuesday 20th April
> * Since then there have been two commits from another developer (Developer
> 2) (Thursday 22nd April and Tuesday 27th April)
> * On Thursday 22nd of April Developer 1 did an update. The Commit from
> Developer 2 was still running, but the file in question had already been
> committed. According to the web logs the update did not touch the file in
> question.
This is to be expected. When a checkout starts, it latches onto a
single revision tree, which is immutable by definiton. Similarly, a
commmit-in-progress is invisible to readers, until the commit finishes
completely and becomes a new revision.
> * This morning the Developer 1 did an update then a commit
> * The update did not raise a conflict, but we are unsure whether it did a
> merge.
If the file was locally modified, and the server's version was newer,
then yes, 'svn update' always does a merge. There's no question about
it.
> * After that the first developer went to work on his file and his changes
> are gone
There must be more to this story. The server logs don't say anything,
other than "somebody ran an update". And if the developer was using
TortoiseSVN, then there's even less the main svn developers can do;
maybe the Tortoise folks have some insight.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Thu Apr 29 17:49:20 2004