[HTTPS-Everywhere] more rules mean more time

Peter Eckersley pde at eff.org
Fri Mar 8 15:32:40 PST 2013


Adding more rulesets does not slow the browser down.  The algorithm for
querying them is O(1).

There are some other real performance issues though.  I just ran a test on
4.0dev6, and it seems that on my fairly modern laptop, parsing the rulesets
and building the ruleset data structures there takes 1.8 seconds.  For 3.1.4,
it takes one second.

Those numbers might be slight exaggerations (they're measures of gross elapsed time,
not time actually spent on the JS thread), but they're large enough that we
should start to think about doing things differently at some point.

One obvious move would moving default.rulesets to Binary XML.  But it might be
that the real perf hit is happening in the JavaScript object generation and
memory allocation process, in which case we'd need to work out a better way to
store everything.

On Thu, Mar 07, 2013 at 11:07:18PM -0600, John A. Wallace wrote:
> I think this is a great project, and I use the extension in Firefox without
> hesitation. Even so, at the risk of incurring great wrath, I would like to
> ask a practical question or two. Does the browser get slowed down by more
> and more sets of rules getting added to this extension? Is there not a point
> of diminishing returns? Why would it not be preferable to have an ad hoc
> approach such as one that "on the fly" determines whether another SSL
> website exists and then simply have the browser switch to it and right that
> rule into "this" browser since it would, in all likelihood, be one to which
> "this user" returns at some time, instead of pushing these rules off into
> every browser be default? No doubt I am overlooking some important
> consideration, but it seems curious to me nevertheless.
> 
> 
> John A. Wallace
> 
> The pen is mightier than the sword, but only if you get in the first stroke.
> 
> 

> _______________________________________________
> HTTPS-everywhere mailing list
> HTTPS-everywhere at mail1.eff.org
> https://mail1.eff.org/mailman/listinfo/https-everywhere


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




More information about the HTTPS-everywhere mailing list