Re: [Monotone-devel] Re: Future of monotone

[Alvaro Herrera]

Koen Kooi wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Thomas Keller schreef:
>
> | So where lacks monotone the most?
>
> Speed. Try doing simple things like 'mtn log', 'mtn annotate' or even 'm
> tn sync' with an OE database and you start to see how badly mtn scales.
> We keep mtn around because half the coreteam hates git with a passion...

I've said this before but I'll repeat it: I think Monotone is not using
its tools to their full potential.  There was a discussion not long ago about
how suspended branches caused the list of heads to be slower.  There was
some code posted that was extracting all heads and them removing the
ones that were suspended.  Or something like that.  And that made me
think: "this should really be done using an EXCEPT (or perhaps it can be
done with a plain JOIN) clause in the SQL query to get the heads in the
first place".  What you're doing amount to 1 query to get the list, and
then N queries to potentially remove each of the N results.  (I didn't
review it closely, so perhaps I am mistaken.  But that's the impression
one gets.)

Even SQLite can do that, I would think.

I offered to help, but I don't really grok C++ so it takes me a lot of
time to read through the stuff.  Sorry about that.

-- 
Alvaro Herrera                 http://www.amazon.com/gp/registry/DXLWNGRJD34J
"Digital and video cameras have this adjustment and film cameras don't for the
same reason dogs and cats lick themselves: because they can."   (Ken Rockwell)