Scrounger can not open any of my 20 possible session restore files
My computer crashed and when I openend firefox, all my tabs were gone, about 140, it was impossible to open recently closed tabs or previous session. I found about 20 files that might contain all the tabs but none of them are opened properly in the scrounger at https://www.jeffersonscher.com/ffu/scrounger.html When I try to export it uncompressed, i get the message that the context of the box does not look like json, that's true because the box is empty after opening all 20 files. Even though one file even is 1.16mb. Does anyone has other advice on at least getting the url's from the file? You would expect at least one file to work right?
Bewurke troch yorick_123 op
Alle antwurden (11)
Sorry to hear about that. The Scrounger should also work in Chrome and the new Edge, if you want to try it there.
If the Scrounger can't decompress the file and show you its raw JSON-format contents, you can do it manually through the Browser Console tool (Ctrl+Shift+J). The script and a necessary setting for that are here:
Thank you for the quick reply. The console gave this error with all the files I could open: Error: uncaught exception: Invalid header (no magic number) - Data: I:\Firefox\1\recovery.jsonlz4
Hmm, that's shouldn't happen with intact files. I'm not sure why all of them have corrupted headers. Is it possible the files were modified by another program?
I don't think so
If you make a copy of one of the smaller files and change its file extension to .txt and try viewing it in Notepad, what does the beginning of the file look like? This is what I would expect (I've redacted some of the data in the middle, as you can see). mozLz40 is critical to proper decoding, and you should be able to see a bit of the JSON data despite the compression. (The unprintable control characters are not shown by Notepad.)
In Notepad++ it has an endless list of nul symbols. In Visual Studio Code it has an endless list of question marks in a diamond shape. In Notepad It does not show anything. It really starts to feel like it's all lost. I got some important pages back by scrolling through all history but missing about a hundred tabs now.
Null symbols are a bad sign. Sometimes you get that when reconstructing deleted files using Recuva. I haven't heard of that in a normal crash scenario. Even if a crash wipes out recovery.jsonlz4, usually the other files are fine. Very strange.
So depending on how long you had most of those tabs open, another option would be to mine your last Windows restore point for session history files. To avoid further data loss, don't use System Restore! You can use a utility program to look in your restore points and export possibly useful files to a convenient location (i.e., not directly back into the profile folder). See:
Thanks, with Shadow Explorer I found 9 more files, the first 3 url's were able to be extracted but at line 56, i have a variety of symbols for a long time, but then they suddenly all turn to nul symbols until the end of the file. 99% of that line is a nul symbol on the previous.jsonlz4. One of the few that were able to open in the scrounger. On exporting from the scrounger I get this error message now: Not saving due to failure parsing JSON: SyntaxError: JSON.parse: bad control character in string literal at line 1 column 23892 of the JSON data
Wow, I'm surprised that most of those files also are unusable. Perhaps part of your hard drive is going bad. You can use the "Unstructured URL List" button to extract readable links from the file as a reference, but it doesn't sound as though the file could be swapped in for the current file to restore that session.
Interesting, the C: drive seems to be in good health, 94% life left, no warnings in CrystalDiskInfo
I guess that's comforting for the future at least.