yet another dvcs shootout (this time python-devel)

[Andrew Bennetts]

David Cournapeau wrote:
> Andrew Bennetts wrote:
> >
> > Both.
> >
> > We think we know about the major causes of round trips, but data from actual
> > users has a tendency to surprise developers every now and then :)
> >   
> 
> Hm, does the -Dhpps works only for pull ? I remember I tried it at some
> point (long ago), but I can't see anything when using branch in my
> .bzr.log, now.

-Dhpss, not -Dhpps.  It does work for pull.  In fact it works equally well
for all operations; if there's any smart protocol (i.e. bzr://, bzr+ssh://
or bzr+http://) activity, -Dhpss will log it.

> Now that I think about it, I think git way of outputting progress during
> network information is much better and more informative than bzr. In
> particular, I don't find the bzr progress bar really informative (I
> don't know what the 5 different operations are), and it takes a long
> time before outputting anything at all (~ 15 s for me), which is
> annoying, because you don't know if that's because of a network problem,
> a wrong url, or just because bzr takes time to get information about a
> valid branch). Git just count the number of objects, display it, with
> the the number of bytes downloaded / remaining; this feels more
> informative to me. Maybe it is just me, though :)

Yes, our progress indications are pretty poor at the moment :(

-Andrew.