[HTTPS-Everywhere] Startup time measurements in Chrome and Firefox

[Peter Eckersley]

Ooops, sorry, I wasn't following the list closely enough.  We haven't
replicated any of your work, yet.

Mike has been worrying about the oracle attack where someone measures
performance differences to tell whether someone has visited site X
before, based on whether their HTTPSE instance needs to go to disk to
fetch a ruleset from sqlite.

On the other hand, the SQLite solution should be let us scale not just
to ~10K rulesets but if we're lucky to more like ~100K.

On Fri, Jan 17, 2014 at 05:18:10PM -0800, Jacob Hoffman-Andrews wrote:
> > It's suspicious to me that the parser is almost exactly twice as fast in
> > Chrome as it is in Firefox.  But either way, Mike and I are hacking
> > today on a way to fix these problematic slowdowns at browser load time.
> 
> You might be interested in the SQLite branch I emailed about last week. I
> think it's in
> pretty good shape, so I made a pull request:
> 
> https://github.com/EFForg/https-everywhere/pull/90/files
> 
> This speeds up Firefox load time dramatically. The main issue is that we
> hit the DB
> synchronously on some resource loads. But we only do this once per target
> per
> browser lifetime, and only when we know there is a rule for that target.
> 
> Also you might be interested in this issue: "HTTPSRules.init gets called
> three times during Firefox startup."
> 
> https://github.com/EFForg/https-everywhere/issues/84

We clobbered that a couple of hours ago :)

And closed the equivalent bug in trac.torproject.org.

I think we need to close down one of those bug trackers ASAP...

-- 
Peter Eckersley                            pde at eff.org
Technology Projects Director      Tel  +1 415 436 9333 x131
Electronic Frontier Foundation    Fax  +1 415 436 9993