Chrome 110 will automatically discard background tabs
arstechnica.com
arstechnica.com
> The consumer controls for Memory Saver live at 'chrome://settings/performance,' or, for the UI-inclined, hit the menu, then "More tools," then "Performance." You can turn off the feature entirely or add sites to a list called "always keep these sites active."
so it seems like turning it off is in the cards for this feature
Have you used iOS Safari? I use it because of private relay and extensions (I’ve written a few for myself that “clean up” websites I frequent). However, the dropping is atrocious. Because my phone is 4 years old it can barely keep 1 tab in memory. Sometimes the tab is even dropped even if I simply lock my phone and come back to it a few hours later without even opening any other apps. All state is lost. Worse, as I exclusively use private mode, when the tab is dropped the cookies are dropped too. So I can’t even stay logged in to anything. It’s a terrible experience.
It’s a huge battery drain, and might allow some apps to eat away your memory during the update/refresh
Seems that we are going backwards
Seriously, they've been grinding this axe and writing sensationalist misleading articles for years now when it comes to Google (and outright downplaying Apple issues as well).
- It can be disabled via chrome://settings/performance
- Installing browser updates, like 110, may lose tab state anyway.
The only potential "sin" they made here is enabled by default.
However, users also dislike being prompted with new choices when an update installs. They want updates to be "invisible."
It also seems like the first time it discards a background tab, it does warn you that this can be turned off in Settings, which is maybe a fair middle ground.
Most of the people who don't care about browsers will be accustomed to having their state erased anyway when something like a page-reload or a navigation-action has been made, so I think it's fair to enable it by default. The benefit they get by it outweighs the side-effects, specially if you consider your first point.
Thus begins a new arms race, as people who don't want their tabs discarded add hidden forms which are later modified by JS to look like they have useful new state, or adding silent single-pixel auto-playing videos, and the browser try to detect those tricks, so further tricks are devised.
For invasive things like data collection, sure. For what amounts to garbage collection, does it really matter?
This is an annoyance across all browsers; there's no built-in way to bottle up the current state except by printing to PDF, which then becomes unclickable.
I spent time one summer playing around building a capturing system addon for Firefox which bottles a browser page DOM state, all assets, and necessary CSS styles inline into one single blob of HTML for long-term preservation. There were a lot of subtle gotchas to work out (which made for a stimulating exercise) but after some time it worked just fine, so it's 100% totally achievable.
I'd like to understand why this functionality isn't provided with every tier-1 browser.
The problem is native handles. The DOM is easy to restore, but how do you restore a WebSocket connection? An IndexedDB transaction? A `Blob`, a stream, or a WebAssembly module that JavaScript assumes to be initiated?
There are solutions to each one of these, but there are still a lot of remaining pieces that are stateful and not easy to restore or gracefully close. It's possible, but it requires a lot of work.
I think it's available for Chrome as well.
It is able to embed images in the file as well as data urls.
Edit: I use this to capture social media web pages that I have contributed to so that I have a copy in case they get deleted.
It is. It's called swap. They should just throttle the JS down to zero, and then let the OS deal with the memory. Each tab is a separate process. What they need is the opposite of mlock, a way to tell the OS to pretty please swap out this process's memory because it is unlikely to be used in the foreseeable future.
Maybe someone remembers what it was called?
https://source.chromium.org/chromium/chromium/src/+/main:con...
For the purposes of saving memory, which is what tab discarding is intended for, where should it be stored?
* In RAM? It is already there and that is the problem. Maybe you could compress it when inactive, but that won't be effective for pages with much image or other media data.
* To disk? At that point you might as well leave it in RAM but inactive and let the OS page it out as needed. Maybe marking it as inactive internally (so your own garbage collection sweeps don't make the OS think the data is active) and to the OS (assuming there is a concept of this-is-low-priority-so-page-it-out-before-other-stuff at that level).
For storage, save-entire-page-to-disk which has existed in most browsers since early versions does most of the job well enough, with opening the stored result in the same browser being the rehydration method, that it might not be worth creating and maintaining something more complex for the remaining part of the job to cover all the cases that this doesn't. Though the balance may have changed, as even simple pages are often effectively SPAs these days there may be more edge cases that “save all” doesn't work with then cases where it is fine.
> I'd like to understand why this functionality isn't provided with every tier-1 browser.
Lack of demand. While it would be very useful for a minority of users, there aren't enough of them for the feature to affect market share, so the big browsers aren't going to bother. It is the sort of thing niche browsers/add-ons might have more interest in.
I use the extension also on Firefox [2].
[1] https://chrome.google.com/webstore/detail/auto-tab-discard/j...
[2] https://addons.mozilla.org/pl/firefox/addon/auto-tab-discard...
Nothing like replacing a real problem with another potential problem.
https://blogs.windows.com/msedgedev/2020/12/09/sleeping-tabs...
I honestly don’t get it, but it intrigues me.
I have five virtual desktops, two for me and three for my current three customers. One Firefox window per desktop. The windows in the three customers' desktops have the tabs to the web apps (prod, preprod, dev) I'm working on plus documentation sites, CI, cards, administration sites, etc. That's 10 to 20 tabs per customer. Some of them are pinned. The two windows in mine desktops have some pinned tabs and some stuff that keeps there maybe for a long time until I finally close it. Not a big deal, but 102 tabs in one window would be challenging. I bet somebody has 1,000 of them in one window and will write that it's not a big deal too :-)
Then I also have the problem of finding links relating to my personal computing interests, so those build up also until I read them in my free time or reorganize them.
To answer the original question, tab search is entirely necessary.
And with vertical tab list you're not limited to just the icon, but can scroll& read
The bookmarks lack the convenience and immediacy of this even though you can save groups of tabs there as well, but then access is slower/less convenient
I wonder if something like this will get "extended" to work for desktop browsers now if this needs to become more popular?
I think something simple like looping an empty audio or video file will suffice. Not that I agree with that from a "hijack the user's browser without their consent" perspective. Just spitballing workarounds.
Hence why I put a media player I want to run in the background in its own tab and whitelist it.
Or do you mean Safari for iOS? I'm thinking of macOS Safari here.
Really, no, not every developer has 50+ tabs open. And it is very easy to avoid having them without losing anything if you want.
Life gets much simpler when you can just alternate between every active thing on your current desktop area.
I have so many tabs, I have a second computer just for the tabs, which I keep running with all the tabs. For extra measures, they are bookmarked too, but I want them open so I can refer to them. It's like having a library.
I think the most tab I ever had open at the same times was 10 when cross referencing something through a ton of tickets and a secondary system and that was already a pain.
It wouldn't be for me. I tend to keep hundreds of tabs open for various reasons. And when I reboot and Chrome screws up reloading my session, it's super frustrating. Do I _need_ those tabs open? No. Does having them open save me a 5-10 seconds a hundred times a day? Sure does.
I don't even use that much tabs, around 30 at most, but this is the thing I use them for: to keep the current state.
I think the main problem here is the lack of user studies. I think if you watch a hundred people to see how they are using tabs, you'll find a couple of patterns and that the right solution is to do something else with the middle mouse click, such as creating a new reading list, or opening tabs but either not loading them at all, loading them at low priority, or reducing the priority by 15% for each successive tab you've opened from the current page.
If I open 6 tabs from google, I'm most likely going to click on the one that ends up next to my current tab. That one is the priority. Everything else is best effort.
However, an optional feature you can disable, and one that even prompts you to review the setting the first time you access a discarded tab, hardly seems like a good argument to do so.
Like, sure, go to Firefox for their stance of privacy and keeping a user in control. Don't go to Firefox because changing a setting is too difficult.