Thank you for this!

How does Tab Wrangler recover or prevent going over the storage quota? It would be great to use as much of the storage space as possible, but I want to avoid users ending up with a dead Tab Wrangler because their average AVERAGE_TAB_BYTES_SIZE exceeded 600 and put them in a state where Tab Wrangler can no longer write data.

I intentionally capped max tabs at something much smaller than the max storage space because according to the chrome.storage docs, "Updates that would cause this limit to be exceeded fail immediately and set runtime.lastError when using a callback, or a rejected Promise if using async/await."

How does Tab Wrangler recover or prevent going over the storage quota?

Good question, I haven't tried to check tabwrangler behaviour after exceeding the quota (I just aimed to stay just below the quota)

Updates that would cause this limit to be exceeded fail immediately and set runtime.lastError when using a callback, or a rejected Promise if using async/await.

This makes me think that upon this condition, Tabwrangler would still be able to persist the list of saved tabs after trimming, and so a simple end-user fix would be to lower the amount of saved tabs (thus triggering a trim). But of course, this relies on the storage error being bubbled up to the UI

(I'll try to verify this assumption, but cannot guarantee when I'll be able to do this)

This makes me think that upon this condition, Tabwrangler would still be able to persist the list of saved tabs after trimming, and so a simple end-user fix would be to lower the amount of saved tabs (thus triggering a trim)

Ok so, I confirmed that this is actually how it works.

I modified tabwrangler to not enforce the limit (which on my system is estimated to be about 17000 tabs). I generated an export/import file with a list of saved tabs close to the limit, and then I opened a bunch of 30/50 groups of bookmarks, to let tabwrangler wrangle them.

I checked that the storage used increased gradually, with getBytesInUse() until it finally got quota exceeded

After that, the used bytes was still barely below the quota limit (becaused it failed to write).

This of course means that the wrangled tabs would be lost, but IMO that's ok, because:

1. I intentionally provided a conservative estimate for the space used by a tab
2. my PR wouldn't change the default value, it would just allow people to increase it past 1000
3. when people get to the tab limit nowadays, they also silently lose tabs from tabwrangler (though the difference in that case is that those tabs would be the oldest, instead of the newly wrangled ones)

I also reproduced it while I didn't have the Tabwranglers options page open, and I've still been able to observe the quota exceeded error being logged after opening the console after the fact.

I then decreased the limit to 16001 (not 16000 unlike my comment in the console)

and confirmed that, upon the next wrangle attempt, it trimmed away all the extra 2000 tabs, and the newly wrangled tabs are successfully saved.

@berdario Thank you for the investigation. I will review again and get this merged.

Thanks again for all this work. The code is straightforward, you added the test, and the comments are useful. Merging!