Error 400 on Google sites (YouTube, Maps, Search etc)

[H0R1Z0N]

I am a big youtube content consumer, when i wakeup, i put all my interesting new video from subscription into a new tabs, recently, it have a certain limit before error 400 block my access to youtube because of this plugin, so i am sure there might be some chance your code was altered.

Temporary fix i used was flushing last hour history and stuff (work a little more time) and when i cannot do it anymore, i flush all the cookie from youtube.

It's becoming annoying a lot and i am trying to clear my tab in case you don't fix it soon so i could disable the extension for a while.

The chrome store support page of this extension seem to indicate i am not alone with this problem.

[mike-athene]

Looks like this is related to how TGS stores scroll position in the cookie(i.e. it shares the same one as the current domain), if you have too many open but suspended tabs, eventually the size of cookie will reach a limit and users start to get 400.

Any other way to store the scroll position? Maybe local storage?

[H0R1Z0N]

It started happening after the breach or the update after that one, i have around 100-500 tabs open, mostly all suspended with 80% being youtube tabs.

[H0R1Z0N]

What's the dev doing, i might has well start to compile a previous version and upload it instead....

[bxkx]

I wonder if this will ever see a fix...

and I don't have 100+ tabs open, just ~6 YT tabs and yet it's happening. I guess I'll need to search for another suspend extension soon. :/

[deanoemcke]

Hi. Sorry for the delay on this. I can confirm that it is indeed an issue caused by how the scroll position is recorded in the cookie (exactly as @mike-athene described).
The cookie will get cleared on a browser restart, unless you have the 'continue where i left off' setting checked in chrome.
The workaround for now (in the case of youtube) is to go to chrome://settings/content/cookies and type www.youtube.com in the 'search cookies' box, and then click on the trashcan to delete.
It would be nice to get some idea of how badly this is affecting average users. Are their any other sites that have similar issues?
I will work on a fix asap. I'm not sure just yet when that will find it's way into the chrome webstore. That depends on a few other changes that will be bundled into the same release.

[jaug1337]

Oh and here I was thinking I was getting crazy.

[m4gnify]

I'm having this problem too.

[brennebeck]

@deanoemcke

The workaround for now (in the case of youtube) is to go to chrome://settings/content/cookies and type www.youtube.com in the 'search cookies' box, and then click on the trashcan to delete.

Can confirm this works.

It would be nice to get some idea of how badly this is affecting average users. Are their any other sites that have similar issues?

I've only had the issue with youtube. No other sites have given me errors like this and I frequently have ~250 sites open, most of which are suspended (then I use vimium tab search to switch to them and return). However, youtube is a very common site so I would speculate it could have a pretty big affect on the userbase.

[u1735067]

For reference, to clean those cookies, I use Awesome Cookie Manager, then go to chrome-extension://hcpidejphgpcgfnpiehkcckkkemgneif/popup.html and search for Name=gsScrollPos-, then Select all -> Delete.
This way it'll clean for all websites.

@deanoemcke : maybe I missed the information, but without auto-update, how to know if there's an update ? By checking the release date ?

[fzimmermann89]

Can you release a bugfix to the chrome store? This issue is kind of a show stopper right now

[DougWeller]

It also happens with Google Books, which is pretty bad as then I don't know what I was looking at.
Doug

[H0R1Z0N]

It's still fixed, why after more than a week the update has not been push to webstore already ? Do google manually check the source code or the dev's planning to self destroy his own work.

…

On Fri, Aug 4, 2017 at 12:09 PM, DougWeller ***@***.***> wrote: It also happens with Google Books, which is pretty bad as then I don't know what I was looking at. Doug — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#537 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AFphh8RmA2wA3rHE66e4NRU2_7PHYmd6ks5sU0I_gaJpZM4N7oaM> .

[soybomber]

been dealing with this issues since the whole hacker debacle more or less. i generally have a couple hundred or more tabs open, and a lot of them are youtubes. i am constantly in the habit of making multiple youtube playlists, and having to delete cookies and resign in to every damn website i use after words is more than a little frustrating :(
hoping for a fix soon!

[deanoemcke]

The fix is tied up in a dev branch along with a lot of other new code. I'm tying it all up now and want to give the whole release a thorough test before putting in on the chrome webstore. I'm really sorry for the delay everyone.

[PMH232]

The fix is tied up in a dev branch along with a lot of other new code.

Best I remember, that's a big no-no. Fix the bug, *then* write new code. (unless, of course, the fix necessitates new code) (but don't cheat!)

[afedulov]

I really like your plugin, it is a life saver for such an obsessive multi-tasking tab opener as me. For that - many thanks! BUT I also really like when I do not need to delete all google and youtube cookies every half a day. Is there any chance you could pull a feature branch and just hotfix it on the master? This issue is VERY annoying.

[Inverness]

I'm glad you're fixing this @deanoemcke. Please practice better version management in the future though.

This scenario is precisely the reason that patch/bug fix releases have separate versioning in the typical major.minor.patch version scheme. If the released version is 1.0.0 then you should always be able to release a 1.0.1 version to fix bugs in 1.0.0 while simultaneously continuing new feature development for 1.1.0.

[deanoemcke]

😞

that is not actually the whole story however.
you may or may not be aware of the current extraordinarily painful process receiving an automatic update of this extension can be: #526

well the good news is that there was an obvious way to solve this the whole time, but it is going to take 2 new releases before it takes effect (ie. it only kicks in once you have the next new version already installed, and are about to receive the subsequent new version).

so i have been very reluctant to push ANY new version that did not have this new update handling feature. unfortunately, and there is no apologising for this, a lot of other new features ended up creeping into this release at the same time.

@afedulov a temporary hotfix on master IS an option, but i would not want to force the update on anyone automatically. so it would require manual installation. AFAIAA, the only way to update to said version would be to uninstall the extension and reinstall it. this means losing your settings, whitelist, and session history. and the only people that would be doing this are the few of you on this thread.

it's probably not a crazy amount of work to do, so if you really insist i'll do my best.

and once again, big apologies for not handling this in the best manner. semantic version control wasn't even part of my vocabulary when i started this extension - which explains why it's v6.30 :(

[DAOWAce]

So it is caused by this extension, huh..

2nd time it's happened to me; hoping that update's out soon.. with a way to preserve our data!

Edit: Clearing cookies is alright, but we lose things stored in them, which for youtube I think contains our watched/resume history. For someone like me, that's really important, so if the cookies do contain the data, we're forced to manually delete the problematic sections of the cookies rather than wipe them entirely.

[ziobbe]

This awesome addon is my way to arginate the weight of internet content on my browser, I have to use an average number of tabs and id' like to return to use it in yt too.

This is in response to:

It would be nice to get some idea of how badly this is affecting average users. Are their any other sites that have similar issues?

[H0R1Z0N]

Yeah, it also affect store.playstation.com (PSN store from Sony) after 70 cookie, it would take while before it happen. Youtube's affected at 40 cookie and use at less 10 cookie from account login.

[brennebeck]

@bro3234 yup, it does. And I expect it should for most others, but YMMV, of course. 👍

[kalyanac]

@bro3234 yup, it does. And I expect it should for most others, but YMMV, of course. 👍

works great for me. every single time.

[deanoemcke]

@bro3234 @kalyanac just a heads up with session buddy. for me it fails to backup automatically. it seems to only backup when chrome is restarted, which is often quite out of date.
see this issue thread for more information on alternative session backup extensions: #742

[bro3234]

@deanoemcke Thanks. I'll check that out.

[brennebeck]

@deanoemcke thanks for that. I was unaware, actually. (I've mostly switched to FF Dev for daily use now.)

[Kiina]

For those with many chrome windows, make sure to check beforehand if your backup utility not only keeps tabs but also order and windows, else you will end up with 800 tabs in one windows like this:

[TiredFalcon]

For those with many chrome windows, make sure to check beforehand if your backup utility not only keeps tabs but also order and windows, else you will end up with 800 tabs in one windows like this:

I suggest using Session Buddy to avoid this problem. It allows to save windows (or also not) and to export save files.

[Kiina]

I suggest using Session Buddy to avoid this problem. It allows to save windows (or also not) and to export save files.

Well I tried the build in session manager because

These exports should be compatible for import into other session managers as it exports simply as one url per line with windows separated by an extra newline.

but turns out it just exported one long file without any newlines for new windows.

[deanoemcke]

@Kiina what i meant by that was that the data in the export is very basic and easily interpretable by other session managers. i do not expect them however, to automatically interpret an extra newline as a new browser window.

you could manually fix this by editing the session backup and putting each group of tabs into its own file, and then importing them one file at a time.

either that, or reimport the file with the great suspender. it will interpret the file correctly.

[Kiina]

either that, or reimport the file with the great suspender. it will interpret the file correctly.

I did exactly that, didn't work. Guess the export didn't correctly add the newline

[ibash]

The gscrollpos bug was a constant problem for me, I thought it was google that had the bug...

WRT backing up sessions, this works fairly well:

1. Move all your tabs to a single window (I use Tab Organizer to do this quickly)
2. Suspend all tabs (right click, suspend other tabs)
3. Bookmark all tabs to a new folder (Bookmarks -> Bookmark All Tabs...)
4. To restore tabs, open bookmarks manager and hit "Open all tabs in new window"

[jrz]

Why are cookies used on all these sites to restore the scroll position? It makes no sense to me, and it's super annoying. It could easily be set in the url, where it belongs.

I don't even care about the scrollposition for suspended tabs.

[edsilmon]

@jrz search this thread for posts by the addon author. afaik it's 10+ years old and the author was a noob when they made it the first time 🙃 you're welcome

[noahbaxter]

so has this issue been fixed in the latest store version yet? or will it continue to nuke youtube? I've seen reviews claiming it is fixed

[liamjohnston]

It's fixed in the current dev version, AND the latest web store version. However, most people aren't on the latest web store version yet because the rollout isn't complete. The dev wants to iron out the last of the major issues (and test) before rolling out to all users.

If you don't want to wait: you can uninstall and reinstall the program extension from the web store to get the latest version with this fix. However be sure to unsuspend or backup all suspended tabs first, or they'll disappear when you uninstall.

[krisdahl]

@liamjohnston Thanks for the tip. I was running version 6.3 and by removing and installing I'm up to version 7. It seems really odd workaround that it is actually fixed but doesn't get rolled out automatically.

[edsilmon]

@krisdahl did you read liam's comment?!

[krisdahl]

@krisdahl did you read liam's comment?!

Yeah. I understand the issue.

I just think it is lame that multi-year old bugs are actually fixed, but can't be shipped. I think it is more important to roll out bug fixes, and communicate the issue of losing the tab state. Or bite the bullet to fix the tab state on rollouts (sounds like easier said than done).

But between choices of having known bugs or losing my tabs? I'd rather have the bug fixes.

[liamjohnston]

@krisdahl Unfortunately there isn't a way to communicate with people. There's some 'update warning' functionality like this built into the new version, but there's still no guarantee people will see/read it before the update happens. And people on the old (6.x) verison won't see this on their next update. The vast majority of TGS users are normal webstore folks, not Github lurkers.

Only a small percentage of people experience tab loss during an update. The last webstore update was only updated for 5% of webstore users as a trial. But a small percentage of 5% of 1.5m people is still a lot of people, and so it still turns into a shitstorm.

I understand it's frustrating but, but there's a lot of responsibility on the dev's shoulders and it this isn't his full time job. Please bear with us.

[pointyointment]

I just got a 431 error (headers too big) trying to post an annotation on the Halfbakery due to dozens of gsScrollPos cookies. After deleting them it worked fine. So, in case anybody is seeing an HTTP 431 response instead of 400 like OP did, this could be the cause of that too.

[ibash]

Is tab loss the only problem with rolling out an upgrade? Are there other problems?
IS there an issue for the tab loss bug?

[edsilmon]

afaik there's were some other bug fixes so technically yes there were/are other problems but the biggest dealbreaker is fixed and there's no tab loss issue at all if you use the tab saver thing also from these guys. it's all in the text above anyway if you just search for comments by deano

[deanoemcke]

@ibash @edsilmon There have been some other breaking changes since this issue first appears, due to changes in chrome. Most notably, around the favicons not updating and every tab having the default extension icon as it's favicon.
The codebase has been quite significantly rewritten since then, which has introduced it's own host of new bugs. These have mostly been resolved now and I'm currently trying to get this new release out once and for all.

[G-Rath]

@deanoemcke that's great to hear. Might I suggest you create a github milestone?

That should make it far easier for everyone to track the progress towards the final release, and even make it easier for people wanting to help get the release out there figure out what they can work on, since the milestone would contain what you're requiring to be finished for the new release.

Following that, you could make a final comment on this issue linking to said milestone (ideally along with a summary of the whys behind the big delay, since currently the explanations are fragmented and mixed in among the 250+ comments), and then lock the conversation to prevent anymore "wtf is it not released yet?!" style messages :)

For bonus points, you could attach a deadline to the milestone, as I'd imagine that would please the community greatly.

[edsilmon]

i learned that ctrl+tab switches between tabs because of the favicon bug and adding shift reverses lol. i've just been holding it down for a few seconds to scroll through all my tabs...

[deanoemcke]

There is now a milestone for the upcoming release that contains a fix for this youtube issue (and many others): https://github.com/deanoemcke/thegreatsuspender/milestone/1

@G-Rath thanks :)