Bernd Schmidt <[EMAIL PROTECTED]> wrote:

> One more question (I'm experimenting... excuse the stupid questions
> but  I'd rather not accidentally mess up the branch).  What I did now was,
> first,
>
>    ~/svnmerge.py merge  -r 96659-96679
>
> just to get started, then
>
>    ~/svnmerge.py merge -F -r 96681,<big list of numbers>,99999
>
> with the list of numbers cut&pasted from the output of "~/svnmerge.py
> avail".  (That second command took a rather long time).

There is no need to cut & paste the exact revision numbers. "svnmerge avail"
shows you the exact revisions that are not merged yet, but "svnmerge merge" is
smart and won't merge twice a revision already merged, nor a revision which
does not refer to a commit to another branch. If you want to merge everything
up to revision 100000, you can just do: "~/svnmerge.py merge -r96681-100000",
or even "~/svnmerge.py merge -r1-100000".

> Property changes on: .
> ___________________________________________________________________
> Name: svnmerge-integrated
>     - /trunk:1-96656
>     + /trunk:1-100000,<snip>,103808-103809,1
>
> i.e. it seems to contain exactly the revision numbers I didn't want to
> merge yet.  Is this expected?

Yes, even if it might look confusing. If you try "svn log -r103809
http://gcc.gnu.org/svn/gcc/trunk";, you'll see that the log is empty. If you
then try "svn log -v -r103809 http://gcc.gnu.org/svn/gcc";, you'll see that it's
a commit on the 4.0 branch. Basically, svnmerge is reporting as "integrated"
revisions that won't ever need to be merged because they're not on trunk. It's
doing it so that, when you are finished merging from trunk, your
svnmerge-integrated property will be a solid range, like "1-108904" or
whatever. In short: trust it :)

As for the slowness, there are a couple of low-hanging fruits which I haven't
address yet. One glaring issue is that generating the log message for a big
merge operation can take longer than merging itself (ewwww). If you feel
familiar with Python, feel free to comment out the log generation code in
action_merge.

Giovanni Bajo


Reply via email to