-
-
Notifications
You must be signed in to change notification settings - Fork 29
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Closing tabs but not all disappear from tabfern #129
Comments
Thanks for reporting! Was this when you had a lot of tabs open, like you mentioned in the other thread? It does take some nonzero time to save, so if you hibernated while the save was in progress, strange things might have happened. I will add a spinner to 0.1.17 so you can at least see if that's the case. Thanks for checking the error logs! Not all messages are on by default. Maybe I should add a "developer options" setting or magic hotkey :) . To enable, in the console, go to the middle iframe (tree.html), and say |
No it was not at the same time. But I had 4 windows with 2-5 tabs each and one more with 10 tabs. |
It just happened again now. Edit: In fact it was 12 tabs, and 2 were closed "correctly". |
Thanks for the detailed report! I tried reproducing with the same configuration and set of tabs but was not able to trigger the behaviour. I also looked through the source briefly and don't see the code path that gave you that result. Here's a picture of the four steps in turning on the full debug output in the developer console, if you're willing to do so---
The tabs showing under "recovered" after closing and reopening the TF window means that TF thought those tabs were open in an unsaved window the last time the tree was saved. Saves happen automatically when tabs are added/deleted/moved. You can also refresh the TabFern window (F5) to fix this sort of situation, rather than closing and re-opening TF.
That makes sense. TF assumes that the list in the TF window matches the order of tabs in the Chrome window, so any difference between those two will cause the wrong tabs to be activated. I am working on relaxing that assumption as part of #35. |
I have set the debug mode as asked. We will see if something happens. I'm running an old extension, and its behavior (when used in incognito mode) has change since my "return". I was previously running 67.0.3396.99
That's what I thought, but I wanted to describe the whole behavior. |
It happened again. It's only cut and paste, sorry if it's not readable.
tab 599 is the ghost tab. It was the last tab of the window. This is not the first ghost, it's difficult to know if all ghosts had the same "error" or if ghosts >1 are a consequence of TF failing on ghost one. Do you know if there is a way to timestamp console lines ? Edit: the current tab is in bold in TF. But it was broken when it had ghosts. Even on windows without ghosts. Sometimes bold, most of the time not. And going back to a previously bold, not bold. |
I don't know about timestamps for console lines. I did add a DEBUG-level log to each step of |
v0.1.17-pre.1 loaded and in debug |
No ghost but I have noticed that the bold-style of the current tab is missing after a while. After a refresh F5, it goes back to normal. (5 min after writing the lines above) : Tabs count is low: 6 in total in 4 windows. |
You are seeing a number of things I've not seen in the whole time TF has existed :D and I'm curious what the difference might be.
|
What I have done: I'll try in a new profile and let you know. [edit: see below] Win 7 32bits, chrome 68.0.3440.106
|
Could this be chrome's fault ? |
It could most certainly be Chrome's fault :D :D . However, the Awesome cookie manager change may have been a change to that extension's incognito mode - see, e.g., here. |
Last version of Awesome cookie manager is: 1.0.0.0 - September 22, 2016 Have you already played with multiple chrome versions on the same computer, or with installing an older version? All pages related to this issue are a bit outdated. Even binaries archives don't work properly or are more than 2 years old. I have finally found a copy of 67.0.3396.79, but don't know how to handle it properly with my current installation. I can't correctly extract its content without installing it. |
TF v0.1.17-pre.1 Still happening in this version. It seems less frequent (than 0.1.16.1337). Usually I don't notice it immediately so I can't paste a relevant log here.
See "tree.js:2456 {Bailing - no tab_node_id: 270, removeinfo: {…}}" ? |
By the way, I have discovered why "Awesome cookie manager" doesn't display incognito cookies. Either google has changed the default value of |
Thank you for the additional detail! Updated based on the logs --- please grab the latest. When you catch a
Expand all and send me that log (obfuscate URLs if you want ;) ). I know this is tedious, but it's the best I can think of at the moment. My guess is that the tab is being removed prematurely. Explanation: somehow, TF's internal record of the tab lost its connection to the node ID. I will add some more error checking but am not able to do so right at the moment. |
Here it is. Nothing related to 552 after that.
|
Here is a new one. I wasn't browsing much. Watching a video on another site (not youtube), and closed this page (459).
|
Thanks! I think I may have it. Right at the top of the log there is a "replaced". Those are not yet handled! Let me see what I can figure out. Do you happen to remember how you opened 459? I've never seen a Replaced event in my own use. |
I have to say I don't know what it is. This is not the first instance of "replace" I see. I think I have seen it each time I had a ghost tab. Edit: can't reproduce it. |
I also can't repro this morning, but I just pushed a first version of (Separately, I am adding i18n in #135. Would you by any chance be willing to contribute a French translation of the TabFern UI sometime, or any other languages you happen to know? Not urgent by any means; I am still converting code to access the translation files.) |
about translation: Do you plan on adding language support with the standard i18n (the user has no choice), or with an alternative like in tampermonkey. I have made my own in alias bookmarks. |
Translation: thank you! Would you please join me over at #135 for that discussion? |
@Procyon-b Have you had a chance to try the updates in e236de5? If so, are you still seeing ghost tabs? I'm now using the v0.1.17 release candidate (in |
No, sorry. I have not paid attention to this update. The newer version is now in place. |
Looks good. Edit: |
Last comment to confirm that I haven't seen any problem since I upgraded to 0.1.17-pre.3 |
@Procyon-b Thank you! I am planning to deploy 0.1.17 early next week. |
chrome 68.0.3440.84
tabfern 0.1.16.1337
I have noticed that some tabs that I have closed moments ago are still in tabfern's list. The ghost tabs are in different windows. I have looked at the error log for the tabfern window as well as the background script. No errors.
On the windows with ghost tabs, when I click in tabfern past the ghost, the wrong tab is selected (depending on how many ghosts there are).
Note that windows was hybernating. Has it somehow screwed chrome ? ...
Edit:
I have moved this tab from a single-tab window to another one, and coincidently placed it on place of a ghost tab. Now tabfern identifies this tab (github) with the identity of the ghost tab
The text was updated successfully, but these errors were encountered: