I'm working on a project (R) with a number of other people. We
generally don't like to have things committed to the trunk until they
are ready to go, passing tests, etc. This means development of changes
that take several days can be a little tricky.
I think the official svn way to do it is by branching, but the trouble
with that is that merging other people's trunk changes into the branch
can be tricky unless done very frequently, and it's a lot of work if
What I've been doing instead is saving patches of my changes, then
reverting, updating the trunk, and re-applying the patch. This seems
However, there's one problem: if I add a file, the following happens:
1. I add the file to the working copy, but don't commit.
2. I create the patch, which will contain a copy of the new file.
3. I revert the changes on the trunk, which undoes the add, but doesn't
delete the file.
4. I update the trunk.
5. I apply the patch. Since the file is still there in the working
copy, it isn't added by the patch. When I get around to committing the
change to the trunk, there's a good chance I'll forget to add it again,
and I end up committing a bad change.
Are there any small changes to this workflow that would get it to work
reliably, i.e. have step 5 get me back to the state at step 1? I'd be
happy if the reversion on step 3 deleted the new file, or if applying
the patch in step 5 overwrote it, but I don't see these options in the
dialogs. (I can't delete all the unversioned files in step 3, because
some of them are libraries from other sources that need to be kept.
What I would need is to delete all files that were added but not committed.)
To unsubscribe from this discussion, e-mail: [users-unsubscribe_at_tortoisesvn.tigris.org].
Received on 2014-06-24 07:33:18 CEST