Weave.join(), and pr."ogress indicator fixes

[Robey Pointer]

On 29 Jan 2006, at 17:12, John Arbash Meinel wrote:

> Robey Pointer wrote:
>>
>> On 27 Jan 2006, at 19:43, John A Meinel wrote:
>>
>
> ...
>
>> * I caught the progress bar saying "merge weave merge" at one  
>> point.  It
>> isn't part of your patch, but sometime that should probably be  
>> changed
>> to a meaningful phrase.
>
> What about 'merging weaves'?

Sounds good to me.


>> * There's still a long pause between the last progress bar and "All
>> changes applied successfully."  I don't know why, maybe just my slow
>> computer.
>
> As near as I can tell, this is the time it is actually merging the
> changes onto the disk. How long is a long pause? Minutes, seconds?  
> I see
> maybe <5s on my slower machine.
>
> I would like it to be faster, but I'm not sure where the code  
> should lie.

I clock it at 15 seconds on my laptop, with both branches on the  
local disk (merging bzr.dev.sftp into bzr.jam.weave).  I don't think  
a progress bar is necessary.  I just wince a little whenever bzr is  
spending more than 1 second on something but leaves the display blank  
-- it looks like it might be frozen or broken.  It's just a picky UI  
thing, I guess.

What if it just said "Cleaning up..." ?


>> * You no longer need the [brackets] in this line in python 2.4:
>>     n1 = set([self._names[i] for i in self_parents])
>>   :)
>
> General question, do we want to use the new generator syntax?
> The above creates a list, which is passed to set(), Robey's suggestion
> would create a generator.

Either way, it will get compacted into a set.  I just find the lisp-y  
([ ]) visually jarring.  If you want a less personal reason, I guess  
I could say "it avoids the temporary list creation".  But really it's  
that the [ ] are superfluous and therefore visual noise to me. :)

robey