versioncache considered harmful?

[Joel Peshkin]

Maxwell Kanat-Alexander wrote:

>	Hey there. Do we have any evidence that versioncache actually helps
>performance more than it hurts our code quality? (It's really a
>nightmare -- I just spent about six hours yesterday trying to refactor
>it.)
>
>	See https://bugzilla.mozilla.org/show_bug.cgi?id=110503#c8 and let me
>know what you think (on the list, not in Bugzilla).
>
>	-Max
>
>  
>
You are right.

Compared to the effort to generate the versions, we do a lot of 
subsequent work to figure out which of them we care about at the 
moment.  Let's move toward generating it on the fly unless we see an 
actual problem with that path.