Garbage in "bzr commit" output

[John Arbash Meinel]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Aaron Bentley wrote:
> John Arbash Meinel wrote:
>>> So the attached bundle fixes both cases.
>>> 1) Don't clear the terminal if there isn't anything to clear
> 
> This is fine, but it doesn't fix Moy's problem, because he doesn't want
> progress bars displayed at all.

His specific problem is that commit is still throwing garbage, even
though there is no progress bar showing up.

It isn't a 100% fix, because we still need a way to completely disable
progress bars. But it fixes the short-term.

> 
>>> 2) Don't let self.last_updates() grow without bound. It defaults to
>>> being capped at something higher than the default get_eta() cap. We can
>>> make it whatever we want, I just want it capped at *some* value.
> 
> We could actually just store the start time and the number of updates,
> couldn't we?
> 
> Aaron

It depends what kind of ETA you want to produce. There are several ways
of computing it, and the current method is probably slightly more accurate.

The problem with start_time => Now, is that it assumes all steps take
the same amount of time to finish.
Especially with our old weave code, the scaling was O(n_lines), which
means that the last few operations are *much* slower than the first few.

The current implementation can adjust for this fact.
It may not be worth the complexity, though.

John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFErq9kJdeBCYSNAAMRAsdOAJ0Sim4e2Gx3VA2mSHmQqhn/1ZUseACgx+7K
tsg1WrRIwi7KltlAZL353Ao=
=IFVs
-----END PGP SIGNATURE-----