top | item 10274245

Mozilla fixed a 14-year-old bug in Firefox, Adblock Plus uses less memory

249 points| devNoise | 10 years ago |venturebeat.com | reply

117 comments

order
[+] AdmiralAsshat|10 years ago|reply
Great for everyone still using Adblock Plus. The rest of us switched to uBlock Origin and never looked back.
[+] ethana|10 years ago|reply
What people doesn't know is that uBlock makes use of EasyList as one of its default block lists. EasyList is maintained by ABP's parent company. That list itself has built-in whitelist that could not be turn off without revamping the list itself. That list contains upward of 60k of rules and no one has time to evaluate that for free.

uBlock can make all kind of independence claims it wants, but at the end of the day, uBlock is still relying on ABP for its blocking rules--which is no guarantee at all.

edit: lol why are people downvoting me for pointing out that uBlock is relying on ABP? uBlock never acctually claimed to not having acceptable ads. By using ABP's EasyList, they could already be running an "acceptable ads" whitelist.

edit 2: Here's the link to easylist for anyone that care to check out the included whitelist: https://easylist-downloads.adblockplus.org/easylist.txt

Warning, the list is over 1mb.

[+] nnethercote|10 years ago|reply
[Here's a comment that I made on this topic on my blog back when this was discussed in July, which is still relevant. The numbers have probably changed a bit, but I doubt enough to affect my point. (The link is https://blog.mozilla.org/nnethercote/2015/07/01/firefox-41-w...).]

I’ve seen variations on this comment many times in discussions of this post. It’s totally wrong-headed.

Here are the usage stats for these add-ons:

- AdBlock Plus: 20.3 million daily users (https://addons.mozilla.org/en-US/firefox/addon/adblock-plus/...)

- uBlock: 0.22 million daily users (https://addons.mozilla.org/en-US/firefox/addon/ublock/statis...)

- uBlock Origin: 0.08 million daily users (https://addons.mozilla.org/en-US/firefox/addon/ublock-origin...)

There are ~70x as many AdBlock Plus users as there are uBlock users. Unless you personally convince all 20 million AdBlock Plus users to convert to uBlock, they won’t see any benefit. (And good luck explaining to them the uBlock vs. uBlock Origin split! What a mess that is.) Meanwhile, when Firefox 41 comes out in September, all 20 million of those AdBlock Plus users will immediately benefit, without having to lift a finger, thanks to Cameron’s patch.

Sometimes it’s worth thinking outside the tech elite bubble (the one where “everybody knows uBlock is better”) and thinking about ordinary users in the real world (where most people haven’t even heard of uBlock).

[+] unKlever|10 years ago|reply
Do you have any tips for performance. I use chrome and ublock, umatrix and disconnect. Disconnect is a few second penalty but worth it for a free proxy. However, I am not sure whether ublock runs before uMatrix, or how much that matters. I was hoping to find a few great regex heuristics to speed things up and procide better blocking. I personally don't use 3rd party block by default because honestly why bother, sites display so shitty you might as well just noscript and whitelist.

tldr; parent or anyone interested do you have suggestions surrounding: > good filter configurations in ublock > whitelist & regex config > use uMatrix? > how to optimize for performance?

Ideally in chrome.

[+] WatchDog|10 years ago|reply
I tried out ublock, didn’t seem to work on twitch.tv and a couple of other random things. So I moved back to ABP.
[+] ssalazar|10 years ago|reply
I switched back to ABP when uBlock started universally blocking Sourceforge. I understand Sourceforge has engaged in a number of very shady practices lately, but I need Numpy for my job and I don't have time to work around nanny software.
[+] 9fb29947|10 years ago|reply
I like how 10 years passes between comment 4 and 6 in Bugzilla: https://bugzilla.mozilla.org/show_bug.cgi?id=77999#c4
[+] logn|10 years ago|reply
And this comment should be a classic:

  Peter 2001-12-01 13:07:33 PST

  Will this really land by Dec 11?
[+] vxNsr|10 years ago|reply
Also how casual comment 6, never mind that we've waited 10 years to look at this thing again, I've got a question.
[+] disillusioned|10 years ago|reply
ABP memory usage in Chrome made me switch to router-level ad blocking. Added bonus: it works on mobile too. I have weird page renderings sometimes, and whitelisting a legit site is a pain, but it's overall a huge value to off-load the blocking to the router.
[+] jtriangle|10 years ago|reply
If you're not on your network and you're on android and you're rooted you can sideload adaway to get ad blocking everywhere, including app-based ads.
[+] btgeekboy|10 years ago|reply
Do you have any documentation on this? It's something I've been interested in setting up.
[+] eponeponepon|10 years ago|reply
This amused me - about twenty minutes ago I'd noticed Firefox using way less memory that it usually does, but thought nothing more of it.

Jung would have something to say about synchronicity, I'm sure.

[+] bitwarrior|10 years ago|reply
Not really a bug, more of an enhancement.
[+] padenot|10 years ago|reply
This is Mozilla jargon. Anything that is tracked in Bugzilla is a "bug". I've opened bug in Bugzilla to change the chair at my desk, for example (I work at Mozilla).
[+] elinchrome|10 years ago|reply
Cem Kaner said that all the pedantry about whether or not something is a bug is irrelevant. He says a bug is anything a user perceives as wrong.

I used to worry about calling things not-really-bugs, but when I adopted his viewpoint it gave me a better perspective on what's important in software.

[+] kzrdude|10 years ago|reply
Does it have a benefit on something else than ABP?
[+] heycam|10 years ago|reply
There are minor benefits without ABP. For each page/iframe that is created, we can avoid running the CSS cascade (and share those data structures) for the UA style sheets. It's saves something like 100 KiB per document.

If other add-ons insert a common style sheet into all documents (it's plausible Stylish does something like this? though I've never checked) then we'll again be able to avoid running the cascade and having duplicate data structures for the cascade across different documents.

[+] dragon88|10 years ago|reply
I honestly use use Ghostery which blocks all trackers. It prevents a lot of ads from loading/working because they often have to redirect to serve me the right one (which Ghostry prevents). Waaay less of a resource suck.
[+] zurn|10 years ago|reply
I don't see any percentage-wise memory savings in the article or here in comments. How big a difference does a couple of MB per document make?
[+] bzbarsky|10 years ago|reply
Depends on the documents and how many there are.

On your typical 50-100MB web-appy page (e.g. gmail) a couple MB is a few percent.

On about:blank, which is normally a few hundred KB, a couple MB is several thousand percent.

So if a page has lots of subframes and those subframes don't have much in them, you get large wins. The canonical things that have lots of subframes are techcrunch and the like, which have "like" buttons for several different social networks and whatnot all over them. Each of those "like" buttons is a separate iframe. https://bugzilla.mozilla.org/show_bug.cgi?id=988266#c7 has some relevant numbers for techcrunch; the difference there was 300MB vs 520MB or so (which means that the page has about 100 iframes, of course).

[+] creshal|10 years ago|reply
Depends on how many tabs you have, and how many iframes etc. each tab contains.
[+] addicted|10 years ago|reply
I really prefer Ghostery. Gives a much better analysis of what's being blocked, with fine grain controls, and much nicer UI.

I haven't measured the memory consumption, but I havent had any noticeable issues so far.

[+] PhasmaFelis|10 years ago|reply
My one huge problem with Ghostery is that it doesn't appear to provide any way to pinpoint the source of an ad. I only want to block intrusive ad providers, but the only way to go from "see an annoying ad" to "block its provider" in Ghostery is to play trial-and-error with potentially dozens of different trackers, blocking and reloading one at a time until the offending ad goes away. In ABP, you could do this with a couple of clicks.

Am I missing something?

[+] dexterdog|10 years ago|reply
I moved to Ghostery a few years ago and all I notice is that whenever I disable it most sites are far slower.
[+] frandroid|10 years ago|reply
Someone needs to get in there and do better bug prioritization...
[+] nnethercote|10 years ago|reply
It's been known for 14 years that sharing this data might help some pages. But it was only last year that it became clear that it would provide big wins in a compelling use case (i.e. for AdBlock Plus users). See https://bugzilla.mozilla.org/show_bug.cgi?id=988266#c8

And then it took another 15 months to get fixed because it was a significant and complicated change.

[+] reitanqild|10 years ago|reply
Hopefully they could also get around to fix the usability bug where the back button on the right click menu disappear if anything is selected...
[+] kazinator|10 years ago|reply
Wow, that's quite weird. Not just the back button but refresh, bookmark (star icon), Save Page As ...

As I type this comment, I'm noticing that you get almost this same reduced context menu when you right click on the edit box (with additional items like Paste and Check Spelling).

Maybe the intent is that you don't unintentionally reload the page or back out when you're selecting.

But then Backspace will navigate back regardless of whether you have a selection, and the main navigation bar's back button also works.

[+] thephyber|10 years ago|reply
Mozilla is a non-profit organization who supports an open source browser for free.

It's not entirely out of the question for people on HackerNews complaining about Firefox features to think about submitting patches.

[+] reitanqild|10 years ago|reply
Hopefully they could also het arpund to fix the usability bug where the back button on the right click menu disappear if anything is selected...