Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
endlessmonotony
Nov 4, 2009

by Fritz the Horse

Computer viking posted:

I haven't noticed the youtube thing myself yet, but my partner happens to be a browser QA guy (for Vivaldi, though), so I quickly talked to him. The obvious question is which video codec youtube is pushing when it fails - they keep changing what they send, and may have started pushing AV1 to more people. Video decoders are exactly the sort of thing that can eat performance and memory, and crash both the browser and drivers if you suddenly start exercising not-quite-solid hardware decoding, so it certainly sounds plausible.

On desktop, you can look at the "stats for nerds" to see the codec used - so if you have some videos that work and some that fail, you could check if they use different codecs?

Not relevant to the problem.

Adbot
ADBOT LOVES YOU

Computer viking
May 30, 2011
Now with less breakage.

endlessmonotony posted:

Not relevant to the problem.

How so? If youtube has e.g. switched some users from VP9 to AV1, and the HW accelerated AV1 decoding they use has issues, that directly causes firefox to eat memory/crash/etc on some videos - but not all of them, since not every video is available in AV1.

I mean, ideally there should be no way for youtube to crash a browser no matter what they send down the wire. That it still happens suggests that it's probably something outside the HTML/JS/CSS path, since those have been thoroughly battle tested. Given that it's youtube, a video decoding problem would make sense?

endlessmonotony
Nov 4, 2009

by Fritz the Horse

Computer viking posted:

How so? If youtube has e.g. switched some users from VP9 to AV1, and the HW accelerated AV1 decoding they use has issues, that directly causes firefox to eat memory/crash/etc on some videos - but not all of them, since not every video is available in AV1.

I mean, ideally there should be no way for youtube to crash a browser no matter what they send down the wire. That it still happens suggests that it's probably something outside the HTML/JS/CSS path, since those have been thoroughly battle tested. Given that it's youtube, a video decoding problem would make sense?

JS is well known to cause issues like this, and it being video decoding would make very little sense given it happens when no video is playing.

Computer viking
May 30, 2011
Now with less breakage.

JS causing weird performance issues, sure - but it feels like outright crashes have been unusual for years?

Good point about it happening without playing the video, though. I hope and assume they're smart enough to just show a jpeg thumbnail until you actually press play, but they do probably preload all the JS.


E: Knowing Google it's probably WebP, not jpeg. But anyway.

endlessmonotony
Nov 4, 2009

by Fritz the Horse

Computer viking posted:

but it feels like outright crashes have been unusual for years?

Buggy JS has been causing problems like this constantly for longer than I can accurately recall.

Computer viking
May 30, 2011
Now with less breakage.

endlessmonotony posted:

Buggy JS has been causing problems like this constantly for longer than I can accurately recall.

I genuinely don't think I've seen it crash desktop FF for a decade, but I accept that I may just be lucky or unobservant.

jeeves
May 27, 2001

Deranged Psychopathic
Butler Extraordinaire

jeeves posted:

Sort of Firefox-related, since it is Thunderbird:

For the past two updates, I've noticed a "Subscription Summary" folder that is being auto-generated on my Thunderbird accounts. I don't know if it is a Gmail or Thunderbird thing, but I can't delete the folder without it coming back after a Thunderbird restart.

Anyone else notice it? I haven't found any information on it, since "subscription summary" are apparently way too generic of terms to search for or something?

I found an answer to my own problem, it was my iOS email app Edison causing this and through IMAP it was making me think it was Thunderbird's fault.

Klyith
Aug 3, 2007

GBS Pledge Week
Whatever youtube was doing, they seem to have stopped doing it, so I guess we'll never know.


I doubt it was AV1 video though. AV1 is new-ish, but it's been around a while and it's not crashing peoples PCs. The software decoder for it has been around for years and works fine -- it uses a fair amount of CPU power to decode, which is why YT isn't switching to it, but it doesn't take multiple GB of memory. And very few people have hardware decoders for it. Only the most recent generation of video cards & intel's most recent CPU have hardware. I doubt all the people ITT have AV1 hardware.


Also, the time that I got the memory hit from youtube yesterday, it was more or less instant. My firefox main thread went from 1GB to 3GB as soon as I opened the youtube vid. That says javascript fuckup to me.

nielsm
Jun 1, 2009



It's more likely that someone on YouTube hosed up and pushed some debugging code to production.

Szmitten
Apr 26, 2008
Ironically, Youtube's bigger problem for me has been the playlists or in-profile video list tab where there are no videos playing at all. That or multiple hour long videos.

BlankSystemDaemon
Mar 13, 2009



I'm just gonna stick to using the open with add-on to open all videos in mpv using youtube-dl(p).

Truga
May 4, 2014
Lipstick Apathy

BlankSystemDaemon posted:

I'm just gonna stick to using the open with add-on to open all videos in mpv using youtube-dl(p).

oh poo poo, this was one of the add-ons that broken when firefox "upgraded" to new addon engine, nice to see they got it to work again

BlankSystemDaemon
Mar 13, 2009



Truga posted:

oh poo poo, this was one of the add-ons that broken when firefox "upgraded" to new addon engine, nice to see they got it to work again
They now cheat by using a python script, but since python exists basically everywhere it's fairly consistent.
I think I had to modify it to use a /usr/bin/env hashbang path for python instead of calling python directly.

And let's not kid ourselves, getting rid of XUL wasn't the terrible idea it's been made out to be - even if it took Mozilla much longer to deliver on the promises of API equivalence, XUL had plenty of big issues that necessitated its removal.

WattsvilleBlues
Jan 25, 2005

Every demon wants his pound of flesh
XUL absolutely had to go, the performance degradation forced me off Firefox for a year until version 57 came out. I used Vivaldi, first time I'd used anything but Firefox for my primary browser since 2004.

Szmitten
Apr 26, 2008
I'm migrating my Firefox profile from one computer to another, but it was 5gb. It was all my Storage folder and was apparently retaining every website/server I hit every day since 01-01-2021 and was taking forever because it was a billion low kb files. Just to be sure, I'm not losing anything vital if I skip this entire folder am I?

BlankSystemDaemon
Mar 13, 2009



Szmitten posted:

I'm migrating my Firefox profile from one computer to another, but it was 5gb. It was all my Storage folder and was apparently retaining every website/server I hit every day since 01-01-2021 and was taking forever because it was a billion low kb files. Just to be sure, I'm not losing anything vital if I skip this entire folder am I?
I believe that's local storage and DOM storage, but I'm surprised you've managed to hit 5GB of it in less than a few years - especially considering it's supposed to have a quota: dom.storage.default_quota=5120 and dom.storage.default_site_quota=25600, respectively.

Szmitten
Apr 26, 2008

BlankSystemDaemon posted:

I believe that's local storage and DOM storage, but I'm surprised you've managed to hit 5GB of it in less than a few years - especially considering it's supposed to have a quota: dom.storage.default_quota=5120 and dom.storage.default_site_quota=25600, respectively.

Those both have those numbers. The fodlers list that as the earliest date, but this profile technically began in 2013 and is again technically migrated from something older than that, but there's nothing dating beyond that.

As a layman, will I be okay starting from scratch with this? I know the cache and relevant history are different, so is this just an issue of the first load of every site taking a fraction longer because it'll be "the first time"?

Klyith
Aug 3, 2007

GBS Pledge Week

Szmitten posted:

I'm migrating my Firefox profile from one computer to another, but it was 5gb. It was all my Storage folder and was apparently retaining every website/server I hit every day since 01-01-2021 and was taking forever because it was a billion low kb files. Just to be sure, I'm not losing anything vital if I skip this entire folder am I?

Storage is like cookies. There may be stuff that you'll be slightly sad to see go like website settings & customization. (Also, lots of tracking data!) However, extensions also have storage in this folder. So that can be a bunch of your extension settings that you probably want to save.


Two ways you could handle this:

1. Copy the extension storage only by looking in \storage\default\ for everything that starts with "moz-extension"

2. Clear out old crap by going into Settings -> Privacy & Security -> Cookies & Site Data -> Manage Data... and sort by last used. Then shift-select everything from more than a year or 2 old and remove. (It will take a while.) This will also get rid of old cookie poo poo as well as a bonus. And stuff you actively use won't be lost.



edit: also if this is a profile that started in 2013 and migrated from something older, it might be a good opportunity to actually make a new profile? It's not something you need to do frequently or anything, but honestly I think profiles that date from before v57 are likely to have a *lot* of old crap in them.

Klyith fucked around with this message at 16:19 on Dec 3, 2022

BlankSystemDaemon
Mar 13, 2009



Szmitten posted:

Those both have those numbers. The fodlers list that as the earliest date, but this profile technically began in 2013 and is again technically migrated from something older than that, but there's nothing dating beyond that.

As a layman, will I be okay starting from scratch with this? I know the cache and relevant history are different, so is this just an issue of the first load of every site taking a fraction longer because it'll be "the first time"?
sessionStorage goes away when the browser window does (unless you're in private mode, then it goes away when the tab does), while localStorage stays around.

You should be fine to remove both, but if in doubt you can always put the entire folder in a compressed archive of some sort and then delete it to see what happens.

Szmitten
Apr 26, 2008

Klyith posted:

Storage is like cookies. There may be stuff that you'll be slightly sad to see go like website settings & customization. (Also, lots of tracking data!) However, extensions also have storage in this folder. So that can be a bunch of your extension settings that you probably want to save.


Two ways you could handle this:

1. Copy the extension storage only by looking in \storage\default\ for everything that starts with "moz-extension"

2. Clear out old crap by going into Settings -> Privacy & Security -> Cookies & Site Data -> Manage Data... and sort by last used. Then shift-select everything from more than a year or 2 old and remove. (It will take a while.) This will also get rid of old cookie poo poo as well as a bonus. And stuff you actively use won't be lost.



edit: also if this is a profile that started in 2013 and migrated from something older, it might be a good opportunity to actually make a new profile? It's not something you need to do frequently or anything, but honestly I think profiles that date from before v57 are likely to have a *lot* of old crap in them.

I wanna thank you both because #2 here revealed that sorting by storage size that the single biggest stored file is one cookie for one site that is 1.9gb, fandom wiki being 300mb, and my food delivery site being 190mb, and many others sub-100mb. So that's interesting.

If I make a new profile won't I have to manually add bookmarks etc though? It's kinda unappealling.

CatHorse
Jan 5, 2008

Szmitten posted:

I'm migrating my Firefox profile from one computer to another, but it was 5gb. It was all my Storage folder and was apparently retaining every website/server I hit every day since 01-01-2021 and was taking forever because it was a billion low kb files. Just to be sure, I'm not losing anything vital if I skip this entire folder am I?

Just zip it on one pc then unzip on the other.

Klyith
Aug 3, 2007

GBS Pledge Week

Szmitten posted:

If I make a new profile won't I have to manually add bookmarks etc though? It's kinda unappealling.

Assuming you're not using firefox sync, you'd export bookmarks to a backup file and then import them. Easy-peasy button for that in manage bookmarks. Saved passwords is more annoying -- there's an export to plaintext but no import. And then you have all the extensions. Lots of extensions have pretty good backup export/import features though.


I dunno, if FF is working well and you aren't having any notable problems you don't need to bother. I guess it also depends how vanilla you keep your firefox. I'm a tweaker and probably one of the reasons I noticed a big benefit was clearing out all the accumulated tweaks that dated back to the early days of firefox.

Szmitten
Apr 26, 2008
Firefox is working very well to a shocking degree considering the age pf the profile, I only have Ublock Origin and View Image Context Menu Item, there may have been various tiny tweaks as time has gone on, I was just very surprised the profile was so large - and it seems it's because nearly half of it is because of literally one 2gb cookie (which I could possibly flush and retry anyway). Also my workflow is one part logins, one part bookmarks, and one part typing two letters into the url bar and getting served the thing I want and it'll be impractical to rebuild that. I'll dick around later and see.

Klyith
Aug 3, 2007

GBS Pledge Week

Szmitten posted:

Firefox is working very well to a shocking degree considering the age pf the profile ... Also my workflow is one part logins, one part bookmarks, and one part typing two letters into the url bar and getting served the thing I want and it'll be impractical to rebuild that. I'll dick around later and see.

Eh forget about it then, if nothing is weird or busted it's not necessary.


Szmitten posted:

and it seems it's because nearly half of it is because of literally one 2gb cookie

Local storage isn't just cookies, it can also be dynamic parts of a website. Like, webapps will put stuff in local storage to make themselves run better and reduce bandwidth even more than relying on browser cache. But the cookie-like functions (and extension storage) are what you'd miss if you ditched the whole folder.


Was the 2gb from megaupload or some other download site? IIRC mega abuses local storage for in-progress downloads, or at least did at one point.

Szmitten
Apr 26, 2008

Klyith posted:

Eh forget about it then, if nothing is weird or busted it's not necessary.

Local storage isn't just cookies, it can also be dynamic parts of a website. Like, webapps will put stuff in local storage to make themselves run better and reduce bandwidth even more than relying on browser cache. But the cookie-like functions (and extension storage) are what you'd miss if you ditched the whole folder.


Was the 2gb from megaupload or some other download site? IIRC mega abuses local storage for in-progress downloads, or at least did at one point.

Mega is there but only 20-something mb. The offending site I just removed, went back to it, it remembered my login and I only had to re-set dark mode, and now it's only 2mb. So it might just be a matter of an occasional manual flush if I need to do this again. I've got more of a handle than I did so at the very least I'll grab the moz-extensions to be sure. Thanks.

mobby_6kl
Aug 9, 2009

by Fluffdaddy
Does Firefox actually resume (as opposed to restart) downloads? It seems that whenever I've had an issue and a download stopped, it would always start from the beginning. Not a problem at home at 100mbs but I'm now traveling and the connection maxes out at like 70KB/s it seems. So I'll probably have to run the download overnight and can't afford to restart it from 0.

Malloc Voidstar
May 7, 2007

Fuck the cowboys. Unf. Fuck em hard.
does the server you're downloading from support resuming?

mobby_6kl
Aug 9, 2009

by Fluffdaddy

Malloc Voidstar posted:

does the server you're downloading from support resuming?
Is that still even a problem? I thought we figured this poo poo out in the late 90s.

I don't actually know for sure, I've been trying to download this tool from DJI for the past two hours, it always fails eventually, and restarts from 0: https://dl.djicdn.com/downloads/dji_assistant/20190327/DJI+Assistant+2+1.2.5.exe

Malloc Voidstar
May 7, 2007

Fuck the cowboys. Unf. Fuck em hard.
It's surprisingly common not to support it, but that server does support resuming downloads, and I'm pretty sure Firefox does too, so :shrug:
I'd try getting a copy of wget and running it with --continue, or any other download manager type thing, or a torrent

Knormal
Nov 11, 2001

In my experience Firefox can support resuming downloads, but not as robustly as wget or curl or similar. I don't know what the difference is in how it's requesting the resume, but Firefox resumes seem to work maybe 50% of the time, while wget works nearly every time.

Truga
May 4, 2014
Lipstick Apathy
a lot of browsers, firefox included, will just delete a .part file when download fails unless it fails under what I assume are very specific conditions, so normally there's nothing to resume from

e: also yeah, i've had resume in firefox work maybe once or twice so far, and i've been using it for like 20 years lmao

Computer viking
May 30, 2011
Now with less breakage.

The one place I've had it work reliably is from sharepoint - thankfully, since I've recently had to download about a TB of 20GB files by hand from a really unstable share point instance in the US. It actually resumed correctly every time ... except for when it claimed the download was complete but also way too small, but I blame the server for that.

CatHorse
Jan 5, 2008
A note on curl on Windows. Windows 10 and 11 bundle curl, but you have to use the full curl.exe name to use it. Just curl is an alias to powershell command.

Malloc Voidstar
May 7, 2007

Fuck the cowboys. Unf. Fuck em hard.

MikusR posted:

A note on curl on Windows. Windows 10 and 11 bundle curl, but you have to use the full curl.exe name to use it. Just curl is an alias to powershell command.
Windows bundles a real curl since ~2018
code:
> curl --version
curl 7.83.1 (Windows) libcurl/7.83.1 Schannel
Release-Date: 2022-05-13
Protocols: dict file ftp ftps http https imap imaps pop3 pop3s smtp smtps telnet tftp
Features: AsynchDNS HSTS IPv6 Kerberos Largefile NTLM SPNEGO SSL SSPI UnixSockets
> (Get-Command curl).path
c:\windows\system32\curl.exe
It was pretty out of date for a while, though seems they updated it
I think Windows provided "wget" as a non-compatible powershell command though it appears gone on my system

CatHorse
Jan 5, 2008

Malloc Voidstar posted:

Windows bundles a real curl since ~2018
code:
> curl --version
curl 7.83.1 (Windows) libcurl/7.83.1 Schannel
Release-Date: 2022-05-13
Protocols: dict file ftp ftps http https imap imaps pop3 pop3s smtp smtps telnet tftp
Features: AsynchDNS HSTS IPv6 Kerberos Largefile NTLM SPNEGO SSL SSPI UnixSockets
> (Get-Command curl).path
c:\windows\system32\curl.exe
It was pretty out of date for a while, though seems they updated it
I think Windows provided "wget" as a non-compatible powershell command though it appears gone on my system

On my Windows 11 both curl and wget are aliases to Invoke-WebRequest. curl.exe runs the real one. But that is by using the bundled powershell 5
PowerShell 7 does not have those aliases.

Geemer
Nov 4, 2010



Since version 108 this occasionally pops up.


I'm a) a split bar supremacist and b) never ever going to intentionally want to search google using the address bar. What do I turn off to make this never appear again?

Armauk
Jun 23, 2021


Geemer posted:

I'm a) a split bar supremacist
What’s the thinking behind that?

Geemer
Nov 4, 2010



Armauk posted:

What’s the thinking behind that?

:ssh: I don't use the search part for searching either. I use it as a shortcut to get to the search engine.

If it can be helped, I'd prefer to never use in-browser based search. I don't have much of a specific reason for it other than not wanting random-rear end searches popping up in my url history suggestions.

isndl
May 2, 2012
I WON A CONTEST IN TG AND ALL I GOT WAS THIS CUSTOM TITLE

Armauk posted:

What’s the thinking behind that?

For me, the URL bar is for URLs. If I'm typing up there it's because I know where I want to go, and I don't want to end up on a search page with random garbage in my search history because I made a typo.

Adbot
ADBOT LOVES YOU

Geemer
Nov 4, 2010



isndl posted:

For me, the URL bar is for URLs. If I'm typing up there it's because I know where I want to go, and I don't want to end up on a search page with random garbage in my search history because I made a typo.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply