Adblock Plus is probably the reason Firefox and Chrome are such memory hogs
351–360 of 419 posts
Re: Adblock Plus is probably the reason Firefox and Chrome are such memory hogs
#352Earlier quoted context omitted.
Firefox differs from Chrome in that it will enable Flash for a whole page, rather than per-element. This means you still need an extension for blocking Flash ads on Flash video sites.
Firefox's model is better, I think. Plenty of sites use hidden flash videos for various utility functionality. How are you supposed to click on these hidden videos so that you can enable them?
Re: Adblock Plus is probably the reason Firefox and Chrome are such memory hogs
#353People are using µBlock now for precisely this reason. https://github.com/gorhill/uBlock/wiki/%C2%B5Block-vs.-ABP:-... https://chrome.google.com/webstore/detail/%C2%B5block/cjpalh... https://github.com/gorhill/uBlock
Re: Adblock Plus is probably the reason Firefox and Chrome are such memory hogs
#354because ad block won't work on mobiles, and devices owned more by Google/apple than yourself.
if you want to block malicious scripts and plugins you should use noscript.
both will actually improve memory use and performance, not to mention getting closer to solving the problem in the first place
Re: Adblock Plus is probably the reason Firefox and Chrome are such memory hogs
#355Earlier quoted context omitted.
> Even though chromium is open it won't be good for the ecosystem. You didn't mention how or why. Personally, I hope Mozilla goes away and some other group comes along to fill the void because I don't think Mozilla consists of the right group of people to make anything worth using. I'm not using any Mozilla browser until they learn how to respect the native operating system UI guidelines. Instead they're trying to bu…
> learn how to respect the native operating system UI guidelines Your comment is almost worthwhile just for the irony in this. Tabbed browsing basically hijacks window management away from the OS. Tabs are always implemented customly in-app (UI break) - and break the UX of window finding, switching, etc. (doubly annoying in a tiling WM, btw). That's all browsers. > Personally, I hope Mozilla goes away and some other…
Re: Adblock Plus is probably the reason Firefox and Chrome are such memory hogs
#356Earlier quoted context omitted.
This is what I've never understood about the people claiming that "they have no choice" or that they feel morally justified in denying revenue to authors and other content creators. I don't use ABP, but it's really not that much effort to have to see bad ads on each "bad" domain _once_, and then blacklist and not have to deal with that entire domain and their bad ads ever again. I do something quite similar with a JS…
> "feel morally justified in denying revenue to authors and other content creators" I do think that only a very small portion of adblocking users wants to "justify" anything in this regard. Most just don't want to see ads and skip the faulty and economically biased moral debate around it. edit: While I agree that the blacklisting solution shouldn't be too much work, the charming "setup and forget" experience most adb…
That said, thankfully the new design has fewer ads, which should help a lot. It really just depends, it's easy to go down a rabbit whole with ads... but we get 90% if our ad revenue from 40% of our ads. So cutting out a lot changes the layout of income without nearly so much impact on revenue.
Re: Adblock Plus is probably the reason Firefox and Chrome are such memory hogs
#357There's a couple inaccurate things in that article that stood out at me. It should be pretty easy to measure the CPU hit when running ABP. It looks like the author is running windows so he should be able to use perfmon. He also mentions that "As with all proxy servers, though, the one caveat is that it doesn’t work with HTTPS connections" which is very much wrong.
Also, the article's claim that most pages on the web make use of several iframes is laughable conjecture. Using their own site's poorly put-together home page as an example does not prove their point. And claiming that all free content on the web only exists because of advertising is disingenuous, to say the least.
Re: Adblock Plus is probably the reason Firefox and Chrome are such memory hogs
#358Sounds like if Mozilla wants to improve Firefox memory usage, they should work on blocking ads natively so people don't need Adblock Plus. <:D
You're going to love Tracking Protection: http://monica-at-mozilla.blogspot.com.au/2014/11/tracking-pr... It's an in-development feature to block parts of websites that track you. In practice, it mostly blocks ads, and it blocks most ads. And it doesn't have the performance/memory hit of AdBlock Plus for two reasons: * The domain matching is simpler (no regexps) and thus much faster. * It doesn't use CSS to fix up th…
Wondering about catching on-error events for iframes.. or load errors, etc... that may help in detection, and reflow. I'm more concerned about things continuing to look nice with adblockers than circumventing them.
Re: Adblock Plus is probably the reason Firefox and Chrome are such memory hogs
#359Earlier quoted context omitted.
this is mainly due to having many images on your website (memory can store uncompressed images to save CPU cycles when rendering) Also, last I checked, most phones have 1 to 2 gigs of RAM, and they're not keeping many tabs in memory. If the RAM is there, using it is not really an issue (until you start getting into the whole virtual memory stuff I guess)
> this is mainly due to having many images on your website (memory can store uncompressed images to save CPU cycles when rendering) If your webpage has that many images, you need to redesign your webpage. An uncompressed image at my screen resolution is 21MB - your webpage needs ten times that in memory? For maybe 500 words of prose and some meme/cat picture? Modern web design is just so bad that they can't help them…
Not that all or most do.. but there are a lot of really nice sites that use a lot of large images for full content display. Other techniques can be used to reserve space and load when a section comes into view, but that's not always prudent.
Images are crucial, and optimizing your rendering is also becoming more crucial.. just the same, not everything on the web is text.. and I'd say people spend more time on the web looking at pictures than text, not to mention video.
Re: Adblock Plus is probably the reason Firefox and Chrome are such memory hogs
#360He's definitely on to something here, FF with a hundred or so tabs open: With ABP: 12961 XXXXXX 20 0 2615960 1.483g 58692 S 45.2 9.5 2:13.00 firefox Without: 13098 XXXXXX 20 0 2157212 1.109g 57476 R 142.6 7.1 1:26.25 firefox Both after a complete stop and start of the browser, restoring every tab by activating it.
I guess the problem here is there is no way to cache this css injection. Maybe FF devs should do something about it instead of just whining about how an add-on is unable to handle the browser's shortcomings. We could say the same about jQuery or bootstrap. You could have pages with many iFrames to other pages using the same js or css files. Right now these pages waste memory on the browsers too...
It's not clear that we want to make those tradeoffs.
We _should_ of course make the changes that don't involve such a tradeoff. They're not trivial, and have significant opportunity costs, so haven't been the top priority for Gecko, unfortunately. You can tell how long https://bugzilla.mozilla.org/show_bug.cgi?id=77999 has been around, for example.
Just to be clear, what we're talking about here is a stylesheet which has about 13000 rules. The data structures used inside Gecko to optimize dynamic changes to documents without having to recompute all style for the document are about 200 bytes per style rule. So you get about 3MB per document, which can add up once you have a large number of documents. Ignoring ABP, having that many rules is not that common, and having that many rules in random dummy iframes sprinkled in your page is something that basically never happens.
Disclaimer: I'm a style system module peer for Gecko, and I've looked in quite some detail at the ABP sitation here; see https://bugzilla.mozilla.org/show_bug.cgi?id=988266#c20 and following.