|
xamphear posted:If I search for something on Amazon, and then ctrl-click to open up 3 product description pages in the background, then click in the address bar to type in a fresh URL, the browser chugs so hard that I can type out half the address before it catches up and even shows what I've been typing. I once opened 10 Amazon pages at once. Firefox has never forgiven me for that one mistake. Amazon just does something horrible to Firefox.
|
#
?
May 21, 2015 01:21
|
|
|
|
| # ? Apr 23, 2024 20:47 |
|
|
I just tried opening 3 product pages in the background and WORKS FINE FOR ME  .
|
|
#
?
May 21, 2015 01:26
|
|
|
xamphear posted:So then what is going to fix Firefox's serious ongoing jank problem? If process separation isn't going to do it, what is? Also, for the record, I don't see jank nearly as bad as you're describing, I wonder if you're seeing issues related primarily to slow disk responsiveness from trying to read the places.sqlite history database for autocomplete, while also saving the content from the pages you just opened into the cache. I have a fast SSD (840 Evo, latest firmware, RAPID) and while there was jank it was only about two characters behind.
|
|
#
?
May 21, 2015 01:50
|
|
|
Nalin posted:I once opened 10 Amazon pages at once. Firefox has never forgiven me for that one mistake. Amazon just does something horrible to Firefox. Amazon has been really, really bad in Firefox for me for the past month or two. The search results not as much, but product pages take a good 30 seconds to a minute to get stable and quit making everything else unresponsive.
|
|
#
?
May 21, 2015 02:24
|
|
|
I don't care about RAM usage, in fact, I'd gladly let Firefox double or quadruple its memory footprint if it meant it would be faster and more responsive. I bought all this RAM, something should use it! And maybe Amazon is A/B testing again. The product detail pages chug for me, even with addons disabled, on a 4.2Ghz processor, GTX 970, and SSD. Also, "only two characters of jank" is still two characters too many, in my opinion. Why should Firefox users have to settle for any jank, when IE and Chrome are, generally speaking, fully capable of no jank at all. xamphear fucked around with this message at 03:02 on May 21, 2015 |
|
#
?
May 21, 2015 03:00
|
|
|
xamphear posted:If memory serves me right, their position was that "the problems people have with Firefox's responsiveness can be fixed more easily in other ways" which rang as patently untrue to my ears at the time. Multi-process is only one of several different ways to improve responsiveness, and compared to other solutions, the main benefit of a multi-process architecture (on desktop*) is for security, not speed. In fact, multi-process can easily be less efficient (especially for memory usage) than other solutions. An alternative to multi-process is multi-thread; the main difference here is that it's quite a bit simpler to share memory in a single-process, multi-thread application. In fact, the Firefox developers have spent the past several years moving more and more things out of the UI thread. This provides many of the same benefits as using multiple processes, but without the additional overhead of creating a new process; this is especially important for the Windows platform, since it's considerably more difficult to efficiently create a new process on Windows. Unlike POSIX systems (Linux, OS X, Android), you can't really "clone" a process on Windows**, which is pretty important for one-process-per-tab. There's an awful lot of shared code that each tab needs (the JS environment, the DOM, etc), and Windows' architecture makes it pretty hard to share this code efficiently. You could probably use shared libraries for some of it, but I think that's still more expensive than the POSIX way: create a template process with everything you need and clone it every time you open a new tab. I'm more of a POSIX developer than a Windows developer though, so I'm not sure what the most-efficient way to do things here is. Still, Microsoft explicitly recommends using threads instead of processes when porting UNIX programs to Windows... * On memory-constrained devices, like a phone, multi-process Firefox makes it easier to handle out-of-memory cases. For instance, Firefox OS uses a multi-process Gecko, with one process per app, so that the OOM killer can kill apps just by killing their associated process. ** For those who know about POSIX development, I'm of course talking about fork(2) vs CreateProcess. If you really want to get deep into how Firefox does multi-process, I recommend searching for info about "Nuwa processes".
|
|
#
?
May 21, 2015 03:05
|
|
|
dud root posted:Go for it. That's usually how I discover new and useful extensions I didn't know existed  I'm not a  I swear. Anything addressing the efficacy of these privacy extensions is my immediate source of curiosity. Or anything I might like but am missing. For instance, should I use Lightbeam over Ghostery?The search stuff is so that I can highlight stuff (say, a model number) and then just click an icon to search craigslist, ebay, pcpartspicker, et al for priceshopping. Easy wikipedia/definition is a bonus. I like the way The Fox, Only Better hides my address bar, etc. More screen real estate. FireGestures gives me what I missed from old Opera; right click + mousewheel to change tabs.
|
|
#
?
May 21, 2015 03:43
|
|
|
m2pt5 posted:Amazon has been really, really bad in Firefox for me for the past month or two. The search results not as much, but product pages take a good 30 seconds to a minute to get stable and quit making everything else unresponsive. Amazon does not like whatever blocks I'm using with uBlock. Scripts go haywire, CSS fails, all kinds of fun stuff. With it turned off, Amazon is just noticeably (a few seconds) laggy and scrolling is intermittently sluggish while something flails in the background.
|
|
#
?
May 21, 2015 05:10
|
|
|
Extension news: As flashblock has been broken for some time, I started using a new extension that fixes the built-in "click to play" plugin functionality to the point where its useable. Edit: Click to play per element * I picked the wrong addon and URL while on my phone. Normally all plugins on a page get allowed at once, and this apples to all pages within a site for your session. I also get popup nag bubbles on lots of sites asking if I want to enable flash. Nope. This fixes all that and makes things work like Chrome. I'd normally happily uninstall the flash plugin and use HTML5 video, but YouTube videos have been broken on FF stable for me on my work PC (Win7). Good news for YouTube html 5 and Firefox : In the latest nightly YouTube videos work, even on my work PC with outdated HD4000 GPU drivers. They don't in stable last I checked. Not clear what they've changed, bit this is wonderful. Chrome is more responsive than Firefox when you have a few tabs open, but it quickly slows to a crawl, pisses away tons of ram and CPU and does not unload or background any unused tabs, ever. Once it exhausts the ram on your system (easy with 4GB) it becomes unusable. Firefox can handle ~100 tabs with no sweat on the same system. wooger fucked around with this message at 12:03 on May 21, 2015 |
|
#
?
May 21, 2015 08:55
|
|
|
wooger posted:Extension news: As flashblock has been broken for some time, I started using a new extension that fixes the built-in "click to play" plugin functionality to the point where its useable. Says there it's useless after Firefox 24? I've been using this and it works pretty well when you switch the about :config preference. Jon Do posted:
Update your Web of Trust extension: https://www.mywot.com/en/download - it hasn't been updated in the Mozilla Add-ons site in about 18 months. Also, Context Search might be of interest, though I think it be a more limited version if your Add to Search Bar extension. Edit: Is Firefox usable for anyone in Windows 10 at the moment? I'm getting repeated crashes on any version I use (could be the GPU drivers - known issues with AMD stuff at the moment). WattsvilleBlues fucked around with this message at 12:16 on May 21, 2015 |
|
#
?
May 21, 2015 12:08
|
|
|
xamphear posted:So then what is going to fix Firefox's serious ongoing jank problem? If process separation isn't going to do it, what is? This is something with amazon: https://www.amazon.com/gp/help/customer/forums/ref=cs_hc_g_tv?ie=UTF8&forumID=Fx1SKFFP8U1B6N5&cdThread=Tx2Z1E6IEDSGMTJ Basically, amazon reloads each page several times. If you do that with normal pages, it's fine, but with the heavy weight javascript and all the other pictures and poo poo they have... Happens to a lot of people it seems, (definitely happens to me), doesn't seem like amazon is interested in fixing it, and it's not limited to firefox. Truga fucked around with this message at 14:56 on May 21, 2015 |
|
#
?
May 21, 2015 14:54
|
|
|
Amazon is just one example, albeit an extreme one lately. Any set of large/complex pages can do it. Avenging Dentist had a lot of awesome detail in his post which on the one hand interests me on a technical level, but on the other hand, I just kinda don't care. I just want Firefox not to chug when I ask it to do several things at once, EVEN IF I have umpteen extensions adding to the workload for each page render. I know it's not too much to ask for, because I've seen other browsers do it, and I know that a quad-core system running at over 4Ghz should be more than capable of rendering 5 pages at the same time while also letting me load another page in the foreground. These problems with Firefox have been around for too long.
|
|
#
?
May 21, 2015 16:15
|
|
|
I'm actually not seeing the same jank problems you have either. Open three Amazon pages, have the url bar momentarily lag one or two characters behind, but nothing too jarring. And that's on a recent Atom SOC with 2GB RAM. But basically I'm here to say the problems with the non-showing pictures seem to be fixed after all in 38.0.1. The weirdness I saw earlier seems to have been a one time occurrence. Apparently. Have stress tested in the mean time.
|
|
#
?
May 21, 2015 16:37
|
|
|
xamphear posted:Amazon is just one example, albeit an extreme one lately. Any set of large/complex pages can do it. Multi-process isn't helping with the jank so far in Firefox 40. But you're right about the performance, there' no excuse for a browser in 2015 to break a sweat on a quad core CPU with tons of RAM on an SSD (my current setup).
|
|
#
?
May 21, 2015 16:43
|
|
|
WattsvilleBlues posted:Says there it's useless after Firefox 24?
|
|
#
?
May 21, 2015 16:49
|
|
|
xamphear posted:I just want Firefox not to chug when I ask it to do several things at once, EVEN IF I have umpteen extensions adding to the workload for each page render. I know it's not too much to ask for, because I've seen other browsers do it, and I know that a quad-core system running at over 4Ghz should be more than capable of rendering 5 pages at the same time while also letting me load another page in the foreground. It's not really fair to compare Firefox-with-extensions to other-browsers-with-extensions, since Firefox lets you do a lot more with extensions (read: basically anything). There have been a number of auto-installed extensions that did horrible things to Firefox's performance. A notable example was the Skype extension, which was installed by default with Skype. It searched every page for anything that looked like a phone number so it could wrap it in a link that would open up Skype. Just that one extension caused massive performance problems, so you can imagine what would happen with several such extensions. Unfortunately, this means that Firefox users have to be somewhat-aware of the performance costs of each extension they install. Firefox can sometimes help detect slow extensions, but if there's one thing I've learned about software, it's that there are infinitely-many ways to write bad software. I'm not sure Firefox could detect them all. I also have a theory that there are some obscure bugs that cause Firefox profiles to get hosed up and slow things down, but my only evidence for that is that "Reset Firefox" seems to fix a surprising number of issues.
|
|
#
?
May 21, 2015 20:12
|
|
|
Flipperwaldt posted:I'm actually not seeing the same jank problems you have either. Flipperwaldt posted:Open three Amazon pages, have the url bar momentarily lag one or two characters behind, but nothing too jarring. Avenging Dentist posted:It's not really fair to compare Firefox-with-extensions to other-browsers-with-extensions, since Firefox lets you do a lot more with extensions (read: basically anything). Maybe another profile reset will help some, but it's not going to completely eliminate the jank. Plus I just don't feel like spending the chunk of time it'll require me to get everything set back up again.
|
|
#
?
May 21, 2015 22:04
|
|
|
A profile reset should never even be necessary unless you had bad plugins loving with settings that they should've left alone. However, a profile reset is the panacea for every problem that crops up because Firefox can apparently not even be trusted to not gently caress up its own settings. Heck, one time it got so hosed that Places Maintenance couldn't even loving parse my profile anymore to fix it. It's like formatting and reinstalling your computer. Sure, it works, but it takes forever to get all the esoterically hidden settings to act the way you want them again. It's ridiculous that it's considered to be a valid solution to problems that shouldn't exist in the first place. Anyway, my current gripe for the last few months is that they broke the jumplist icons again. I posted about this way back when. poo poo got fixed in version 36 and then subsequently broke again in 37, still is broken in 38. However, now it's not putting 0 byte .ico files in the folder it stores them in, it just straight up doesn't store icons at all anymore. At least now the icons look like the default page with a Firefox logo, but that's not the favicon they're supposed to be. My notebook was still on 36 and the moment I updated it, the icons broke on it as well. So I doubt it's a problem with my computer. It's pretty minor, but it's also kinda bullshit that it's broken. Fake edit: Also, I'm not seeing any of the jank you guys describe with Amazon, at least not on the .co.uk version. Sure, it does its spazzy reload the page a dozen times thing, but I'm not getting any lag when entering an URL in a new tab while Amazon is loading a bunch of product descriptions.
|
|
#
?
May 21, 2015 22:38
|
|
|
xamphear posted:Hmm, okay, I guess not everyone sees it with the Amazon example. I feel like you're a loving idiot, I also feel like your avatar is actually you. If it's so inexcusable, then by all means go use another browser since you've seen them do it better. Hurr Durr I feel a Quad Core.
|
|
#
?
May 21, 2015 22:40
|
|
|
Im_Special posted:I feel like you're a loving idiot, I also feel like your avatar is actually you. If it's so inexcusable, then by all means go use another browser since you've seen them do it better. Hurr Durr I feel a Quad Core. Cool, thanks!
|
|
#
?
May 21, 2015 23:05
|
|
|
xamphear posted:Except... You are? Jank is jank. Two characters or ten characters. Firefox shouldn't jank AT ALL. This isn't an unreasonable request! Computers are pretty powerful these days. No one's arguing it should happen at all. Obviously. I don't know what's happening on your end and Firefox should get their poo poo together in many ways, but maybe what you're seeing isn't typical. Flipperwaldt fucked around with this message at 23:53 on May 21, 2015 |
|
#
?
May 21, 2015 23:12
|
|
|
Flipperwaldt posted:It's relevant in the sense that it isn't as simple as "Firefox is a huge resource hog" or it wouldn't run on this tablet at all. But it does and does it better than your beefy computer, since what I get is a barely noticeable blip versus your aggravating torment. The difference in the quantity of jank probably comes down to extensions. I've got uBlock, Disconnect, and Greasemonkey, and each have a hit to page loads. I fully understand that each of those extensions is adding to the equation. I'm saying that regardless of the load they add into rendering content, the UI should not hitch up.
|
|
#
?
May 21, 2015 23:23
|
|
|
xamphear posted:The difference in the quantity of jank probably comes down to extensions. I've got uBlock, Disconnect, and Greasemonkey, and each have a hit to page loads. I fully understand that each of those extensions is adding to the equation. I'm saying that regardless of the load they add into rendering content, the UI should not hitch up. What happens when you use an adblocking host file (which typically also blocks most tracker stuff) and disable Disconnect?
|
|
#
?
May 21, 2015 23:39
|
|
|
Wow. The defensive nature of this thread to real life problems amazes me. I too am a huge Firefox fan and have been using since Phoenix. The browser is downright terrible lately. Amazon is unusable. Scrolling is usually laggy on OSX and Windows. I have to restart Firefox regularly due to what seems to be memory leaking to the point of massive jank to open a new tab. New profile doesn't help for longer than a day or two. Extensions or not, the browser starts to crawl. Something is wrong with the codebase. Devs are working hard on the situation but I think the problem is they aren't testing in ways that users do; long term runs with tabs open for long periods of time. The code rot is real. As a developer myself, I understand the situation to try to improve performance how the Mozilla team has. I just don't think the performance improvements are enough if the perceived performance is downright terrible. I am praying that e10s alleviates some of these issues with the UI locking up. I could then get over the terrible page rendering performance. Will I switch browsers? No. I haven't since the start. Will anyone I know switch to Firefox? Hell no. Isn't that what matters in the current browser war?
|
|
#
?
May 21, 2015 23:42
|
|
|
I have been having a lot of issues with slowdown, though I have a lot of extensions. I thought it was one of the extensions. Amazon gets really, really choppy to the point where I do my shopping/research in Chrome a lot. I don't think Chrome would run great with versions of all those extensions made for Chrome installed on it, either, however.
|
|
#
?
May 21, 2015 23:47
|
|
|
jink posted:Wow. The defensive nature of this thread to real life problems amazes me. A significant reason for this is that many people in this thread aren't experiencing problems. I for one have had very few issues with Firefox (at least in the past couple versions). jink posted:Something is wrong with the codebase. Devs are working hard on the situation but I think the problem is they aren't testing in ways that users do; long term runs with tabs open for long periods of time. The code rot is real. Have you looked at the Gecko codebase? If not, I'd recommend against making proclamations about the quality of the code. (For reference, I've worked with Gecko as part of some of the Thunderbird patches I've contributed, and while there are some definite problems, most of the ones I've encountered are related to bleeding-edge JS features.)
|
|
#
?
May 22, 2015 00:39
|
|
|
Nintendo Kid posted:What happens when you use an adblocking host file (which typically also blocks most tracker stuff) and disable Disconnect? Host files can't hide elements and disabling Disconnect makes my browsing a lot more trackable. I don't want to give up either of those things. I also don't want to give up Tree Style Tabs. I also like Firefox, in large part because it makes all of those things possible. I'm not in here talking about the shortcomings I see in Firefox just because I like making GBS threads on crap I don't care about. I'm mentioning it because I like Firefox and I want it to be better. If you're running Firefox and are perfectly satisfied with it, then that's awesome, and I'm envious.
|
|
#
?
May 22, 2015 02:29
|
|
|
Avenging Dentist posted:A significant reason for this is that many people in this thread aren't experiencing problems. I for one have had very few issues with Firefox (at least in the past couple versions). I understand what you are saying. I wasn't trying to say that everyone has issues. To pretend Firefox doesn't have any lag issues isn't fair either. My experience in real life is that I have suggested friends try Firefox, many quit using it because of it's laggy nature. They use Chrome instead because it feels fast. That perceived speed is what matters. The Chrome UI is responsive and feels like it's ready to go. Firefox tends to get extremely bogged down and lock up entirely. e10s might be the silver bullet. I haven't worked in the codebase. I don't have the low level experience to help fix memory leaks and UI thread work. The devs have been making massive strides with the age of the codebase (and their mobile/Firefox OS work, wow!). What I hope to see is more of the UI speedup work (I miss memshrink and snappy projects). I stand by Mozilla and their efforts. I am hoping they will get it right but the current state isn't ideal.
|
|
#
?
May 22, 2015 02:35
|
|
|
Alereon posted:While WebM may be lovely, its a lot less lovely than GIF for files of non-trivial size. But it's generally lovely when compared to APNG which, unlike WEBM, is supposed to be used as an image format.
|
|
#
?
May 22, 2015 16:51
|
|
|
But unlike APNG, webms are actually used by people
|
|
#
?
May 22, 2015 17:22
|
|
|
xamphear posted:Host files can't hide elements and disabling Disconnect makes my browsing a lot more trackable. I don't want to give up either of those things. Dog uBlock can cover all of that in conjunction with a hosts file, christ. You're effectively running redundant extensions.
|
|
#
?
May 22, 2015 18:53
|
|
|
Nintendo Kid posted:Dog uBlock can cover all of that in conjunction with a hosts file, christ. You're effectively running redundant extensions. Not sure why you'd need the hosts file at all, just uBlock, with the various subscriptions. Easy privacy is the one you want to cover disconnect's functionality, and it's in the default list options. One check box to tick.
|
|
#
?
May 22, 2015 21:27
|
|
|
wooger posted:Not sure why you'd need the hosts file at all, just uBlock, with the various subscriptions. Hosts file is just for the paranoiac tendency he's displaying, and because it requires 0 browser resources.
|
|
#
?
May 22, 2015 21:29
|
|
|
Superb Owls posted:But it's generally lovely when compared to APNG which, unlike WEBM, is supposed to be used as an image format. No one ever advocates for .MNG. 
|
|
#
?
May 22, 2015 22:17
|
|
|
The Dark One posted:No one ever advocates for .MNG. APNG actually got implemented by something popular ever (firefox). What ever implemented MNG?
|
|
#
?
May 22, 2015 22:22
|
|
|
Nintendo Kid posted:Hosts file is just for the paranoiac tendency he's displaying, and because it requires 0 browser resources. Thanks for being helpful. If I posted the other extensions I use, could you give me some additional recommendations? I really like YouTube Center because it makes the videos default to the highest resolution rather than letting YT decide to autoplay them at 360p. What undiagnosed mental health problems might that signify?
|
|
#
?
May 22, 2015 22:32
|
|
|
You're the one scared of people tracking you so much that you're using at least 2 extensions against it man.
|
|
#
?
May 22, 2015 22:34
|
|
|
 Jesus Christ, chill out kids.Superb Owls posted:But it's generally lovely when compared to APNG which, unlike WEBM, is supposed to be used as an image format. As horribly non-standard and patent-encumbered as it is, BPG (an H.265 I-frame) is pretty cool for lossy image compression because JPEG is archaic and WebP is garbage. H.265 is the current state of the art and it's loving time to get something better.
|
|
#
?
May 22, 2015 22:36
|
|
|
Nintendo Kid posted:APNG actually got implemented by something popular ever (firefox). What ever implemented MNG? Ironically, Firefox did, for a time. Then they removed it because it added a couple hundred KB to the filesize of Firefox or something like that.
|
|
#
?
May 23, 2015 03:14
|
|
|
|
| # ? Apr 23, 2024 20:47 |
|
|
xamphear posted:I really like YouTube Center because it makes the videos default to the highest resolution rather than letting YT decide to autoplay them at 360p. What undiagnosed mental health problems might that signify?
|
|
#
?
May 23, 2015 16:50
|
|
