Hi all,

The Firefox version of HTTPSE reads its rulesets from a sqlite file and
caches them in memory. The current version does this read synchornously
the first time a given ruleset is encountered, which has the potential
to lock up the UI thread when disk is slow.

I've got a branch going that switches to reading asynchronously from
SQLite. To make it work I had to borrow a subtle hack from AdBlock Plus:
If we get a request and we don't yet have the information about what to
do with it, we redirect the request to its own URL, then suspend it.
Once we get back data from SQLite, we result the request. The redirect
handler fires a second time, but now we have the data cached and can
rewrite immediately. It's a pretty tricksy change, so I'd like some help
testing it out. Branch is here:

https://github.com/EFForg/https-everywhere/compare/async?expand=1

Package for testing is here, along with a signature:

https://jacob.hoffman-andrews.com/https-everywhere-5.1.3asyncbeta-eff.xpi.html

Thanks,
Jacob
_______________________________________________
HTTPS-Everywhere mailing list
[email protected]
https://lists.eff.org/mailman/listinfo/https-everywhere

Reply via email to