[MERGE][bug #111702] bzr init <remote_branch> connects multiple times

[Vincent Ladeuil]

>>>>> "john" == John Arbash Meinel <john at arbash-meinel.com> writes:

    john> Vincent Ladeuil wrote:
    john> ....

    >> Fire, fire ! I'm refactoring :-)

    john> By the way, before you get too far down the track, I
    john> would like to get your changes merged.

Nice to hear ;-)

    john> So if you can come up with a semi-clean result, we can
    john> merge it, and then we will continue to merge your work
    john> as you finish up.

I'm pretty much finished and I have long realized that the
merging may cause problems because I address several related
different issues.

So I'll publish a branch on lp, send a merge request with a
summary of all the modifications and and we can discuss on if and
how I can redefine the patches.

BUT, I think merging the transport refactoring in 0.17 is far too
risky. I had relied on the test suite only, so the risks are
related to the test suite coverage of the transport abstraction.

I trust the test suite, but I prefer to merge when bzr.dev is
reopened ;-)

On the other hand, I will *greatly* appreciate if the review can
occur before that :-P

    john> That way you don't feel like your work misses the
    john> boat. As long as you feel sufficiently motivated to
    john> finish :)

The day I will lack motivation (if ever) I'll make you know about
it in advance so that the good bits can be saved ;-)

<snip/>

    john> Also, the way you have written this, the transport will
    john> only be re-used if it is identically the same. I
    john> thought we were trying to also handle cases when it is
    john> a parent/child. Or is that something you are still
    john> working on?

The funny thing is that, so far, this is enough to catch the
reuses... but I'm still debugging merge and that assumption may
fail.


<snip/>

    >> This work is done now, I will now begin to address the multiple
    >> connections bugs. So far, using ftp for the tests and disabling
    >> the cache reveals ('b' is the actual number of connections):

Quick update:

- pull is fixed,

- merge is under work.

- all other were false positive caused by make_branch_and_tree in
  the tests.

    john> Well, you may also want to check what happens when you open:

    john> http://bzr.arbash-meinel.com/branches/bzr/ünicóde-test

Quick check with bzr.dev:

bzr branch http://bzr.arbash-meinel.com/branches/bzr/ünicóde-test
bzr: ERROR: Unsupported protocol for url "http://bzr.arbash-meinel.com/branches/bzr/ünicóde-test"
vila:~/tmp :( $ 

I'm pretty sure it's due to the *directory* using unicode chars,
and the error is misleading.

I came across that during the refactoring, we focus on unicode
*filenames* support but we ignore the *directories* (well, those
above /.bzr).

    john> However, support for that seems broken at the moment
    john> (and maybe never worked).  Shame. I wanted that to work
    john> rather than requiring people to type:
    john> http://bzr.arbash-meinel.com/branches/bzr/%c3%bcnic%c3%b3de-test/

This one works.

    john> So I guess don't worry about it (though it might be
    john> nice to have that work eventually)

Whatever we decide, I too, prefer to not worry about it now.

         Vincent