1374731 - Remove built-in Flash test tables

Status: RESOLVED FIXED

(Toolkit :: Safe Browsing, defect, P1)

Description

[:Felipe Gomes (needinfo for replies!)]

On bug 1307604, the test entries from itisatrap.org to test the flash blocking lists were added to the hardcoded list in SafeBrowsing.addMozEntries(), following the pattern of the other existing ones (from the malware/tracking-protection features).

However, all these Flash entries are present in the actual lists served by shavar, and it's been agreed that it's not necessary to have them hardcoded by the browser.

It's likely that this will represent a perf win, or at least open the way for one. When bug 1307604 landed, it caused bug 1337043. That regression was recovered by some other means, but removing this is strictly less work to be done on startup.

Also, I wish to remove more stuff from this in the future, but that involves more conversations, and this is the easy first step to get those conversations started.

Comment 1

[:Felipe Gomes (needinfo for replies!)]

Attachment: Bug 1374731 - Don't use hardcoded test entries for the Flash classification tables.

Review commit: https://reviewboard.mozilla.org/r/151060/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/151060/

Comment 2

[:Felipe Gomes (needinfo for replies!)]

I wasn't sure if it was necessary to add the test tables to disallow completions during tests, so I added them just to be sure.

Comment 3

[Robin Steuber (she/her) [:bytesized]]

Comment on attachment 8879719 [details]
Bug 1374731 - Don't use hardcoded test entries for the Flash classification tables.

https://reviewboard.mozilla.org/r/151060/#review155908

Looks good to me.

Comment 4

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/100a1fdb6daa
Don't use hardcoded test entries for the Flash classification tables. r=bytesized

Comment 5

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/100a1fdb6daa

Comment 6

[François Marier [:francois]]

BTW, next time we should ensure that one of the URL Classifier peers (Dimi, Henry, GCP or me) takes a look a the patch too before landing.

(I just had a look through it and it looks all good to me.)