[dev] [commits] Horde-Hatchery branch master updated. a617531b1810f57ef80e11dbcbf504af2528d79b

Michael M Slusarz slusarz at horde.org
Sat Jul 25 18:30:53 UTC 2009


Quoting Chuck Hagenbuch <chuck at horde.org>:

> commit 4fe032e49d6042d5cf69b714369034b08859af37
> Author: Chuck Hagenbuch <chuck at bluestatedigital.com>
> Date:   Fri Jul 24 17:31:47 2009 -0400
>
>     for now, if a repo file doesn't exist, default to 60 seconds of  
> caching. In the future we should cache git SHA1s infinitely, since  
> they're immutable
>
>  framework/Vcs/lib/Horde/Vcs.php |   12 ++++++++++--
>  1 files changed, 10 insertions(+), 2 deletions(-)

Do we really want unexpiring caches?  If I look up a random file  
version for a file that has 1000's of commits, what are the odds that  
we are going to hit that file again (especially if it is a huge  
commit)?  There's going to be a lot of cache garbage that is really  
not useful - it is probably less expensive to do the lookup on these  
rarely accessed objects again rather than keep the cache object around.

I agree with caching in this instance, but only for a certain time  
period (a day? a week?).

michael

-- 
___________________________________
Michael Slusarz [slusarz at horde.org]



More information about the dev mailing list