Re: [Monotone-devel] slow update/status on NFS

[Richard Levitte]

In message <address@hidden> on Wed, 23 May 2007 15:36:14 +0200, Markus 
Schiltknecht <address@hidden> said:

markus> Richard Levitte wrote:
markus> > Have a time_t that gets initialised when stat() is used,
markus> > then compare it with the result of time() at any use,
markus> 
markus> To really reduce memory consumption, you'd have to compare
markus> *all* cached stat results and throw away expired ones.

????????????

We're holding entire files in memory, and you worry about the memory
consumption of struct stat?????  Let's get things in perspective!  I
was suggesting doing the checks as needed (lazy check, so to say),
nothing else.  The cache will die with the path object anyway.

markus> > I can't really say if this is really needed, since most
markus> > paths should be fairly short-lived objects...
markus> 
markus> That, and for the sake of simplicity, I'd rather like to bind
markus> the cached status to the lifetime of the any_path object.

Good, I like that we agree on this.

Cheers,
Richard

-----
Please consider sponsoring my work on free software.
See http://www.free.lp.se/sponsoring.html for details.

-- 
Richard Levitte                         address@hidden
                                        http://richard.levitte.org/

"When I became a man I put away childish things, including
 the fear of childishness and the desire to be very grown up."
                                                -- C.S. Lewis