Firefox - reading out urls of opened tabs from the command-line

Question

I sometimes have quite a big range of tabs open in Firefox and I prefer it to save them to a file, rather then using the build-in bookmarks.

Therefore I (manually) copy the urls from the about:preferences page, save them to a file and process the file with: tr '|' '\n' in a little bash script.

Later when I want to reopen the tabs from the textfile I run this little loop:

#!/bin/bash

# usage: $bash Open-tabs.sh file-with-bookmarks.txt

 while read -r line; do
     firefox -new-tab "$line" 2>/dev/null &
     sleep 2
 done < "$1"

and it opens all tabs with a delay of 2 seconds.

I would like to know if there is a way, I can read-out the urls of the opened tabs from the command line, so I could include it to my script?

score 12 · Answer 1 · edited Jul 07 '20 at 15:06

12

this works for Firefox 57+. You'll need lz4 (via pip). The file header is gathered from the length of b'mozLz40\0'. Use an environment variable for the filepath if you want to use it in a oneliner, replace with \n and \t accordingly and merge lines.

export opentabs=$(find ~/.mozilla/firefox*/*.default/sessionstore-backups/recovery.jsonlz4);

python3 <<< $'import os, json, lz4.block
f = open(os.environ["opentabs"], "rb")
magic = f.read(8)
jdata = json.loads(lz4.block.decompress(f.read()).decode("utf-8"))
f.close()
for win in jdata["windows"]:
    for tab in win["tabs"]:
        i = int(tab["index"]) - 1
        urls = tab["entries"][i]["url"]
        print(urls)'

edited Jul 07 '20 at 15:06

Caio S. Souza

3
2

answered Aug 30 '17 at 21:04

wbob

298
1
5

To get the latest needed file from firefox : `export opentabs=$(ls -t ~/.mozilla/firefox*/*/sessionstore-backups/recovery.jsonlz4 | sed q)` – Gilles Quénot Jan 04 '18 at 00:26
Can you explain `magic = f.read(8)` ? You don't use `magic` var later. – Gilles Quénot Jan 04 '18 at 12:19
1

the first 8 bytes in recovery.jsonlz4 are part of mozillas custom file format - passing the length of the magic to read(8) causes the next read() within json.loads to continue **after** the header, making it a valid decompressable lz4 stream. For the reasoning to choose a custom file header, you can find more details in this github [gist](https://gist.github.com/mnordhoff/25e42a0d29e5c12785d0) linking to the relevant firefox source. – wbob Jan 06 '18 at 11:11
@wbob Now I get `UnicodeDecodeError: 'utf-8' codec can't decode byte 0xf2 in position 12: invalid continuation byte`, can you help ? – SebMa Jun 14 '18 at 17:22
@Sebma I updated to lz4 2.0.0 and got the error too, your edit fixed it, thank you. I went ahead and merged the module imports – wbob Jun 20 '18 at 19:25
@GillesQuenot using: `ls -t ~/.mozilla/firefox*/*/sessionstore-backups/recovery.jsonlz4 | sed q` I get: `ls: cannot access '/home/user/.mozilla/firefox*/*/sessionstore-backups/recovery.jsonlz4': No such file or directory` – nath Jun 23 '18 at 12:09
@wbob using: `find ~/.mozilla/firefox*/*.default/sessionstore-backups/recovery.jsonlz4` I get: `find: /home/user/.mozilla/firefox*/*.default/sessionstore-backups/recovery.jsonlz4': No such file or directory` – nath Jun 23 '18 at 12:11
got it, on my system it is called `.../recovery.js` still I get the error: `Traceback (most recent call last): File "", line 1, in ModuleNotFoundError: No module named 'js' ` – nath Jun 23 '18 at 12:17
@nath absence of a jsonlz4 file hints at a firefox version prior to 57 - then the answer by raymii/huntersthompson is to be prefered and this one will not work. Nonetheless: your traceback reads like the import line somehow got cut up in the middle of "json" - as this is what python will throw for an "import js" statement. – wbob Jun 23 '18 at 23:57
@wbob ah thanks you are right I'm running `Firefox 52.8.1 (32-Bit)` – nath Jun 24 '18 at 19:57
@wbob On firefox > 57, I get `UnicodeDecodeError: 'ascii' codec can't decode byte 0x9c in position 8: ordinal not in range(128)` What is the encoding of the `recovery.jsonlz4` file ? – SebMa Aug 28 '18 at 13:24
@wbob I've uncompressed the `recovery.jsonlz4` file with the [mozlz4a.py script](https://gist.github.com/Tblue/62ff47bef7f894e92ed5) and it turns out the `recovery.json` is utf-8 encoded, I don't understand why it does not work. – SebMa Aug 28 '18 at 14:17
@wbob My mistake, I had forgotten the `magic = f.read(8)` instruction by the time I used your script. – SebMa Aug 28 '18 at 14:55

score 7 · Accepted Answer · edited May 05 '21 at 09:56

7

Source(Changed file path) : Get all the open tabs

This snippet gets the current firefox tab url's. It uses the recovery.js[onlz4] file in your profile folder. That file is updated almost instantly, however it will not always be the correct url.

Get all the open tabs:

python -c '
import io, json, pathlib as p
fpath = next(iter(p.Path("~/.mozilla/firefox").expanduser().glob("*.default/sessionstore-backups/recovery.js*")))
with io.open(fpath, "rb") as fd:
    if fpath.suffix == ".jsonlz4":
        import lz4.block as lz4
        fd.read(8)  # b"mozLz40\0"
        jdata = json.loads(lz4.decompress(fd.read()).decode("utf-8"))
    else:
        jdata = json.load(fd)
    for win in jdata.get("windows"):
        for tab in win.get("tabs"):
            i = tab["index"] - 1
            print(tab["entries"][i]["url"])
'

edited May 05 '21 at 09:56

ankostis

491
5
11

answered Aug 09 '17 at 18:04

Hunter.S.Thompson

8,839
7
26
41

is it possible, the path on debian is `~/.mozilla/firefox/*.default/sessionstore-backups/recovery.js`? – nath Aug 09 '17 at 18:48
Yes sorry I edited the question with that but I forgot the sessionrestore-backups before recovery.js. Thank you for pointing it out. Edited Answer – Hunter.S.Thompson Aug 09 '17 at 18:49
I always get `Traceback (most recent call last):` `File "", line 2, in ` `IOError: [Errno 2] No such file or directory: '~/.mozilla/firefox/*.default/sessionstore-backups/recovery.js'` with `cat` I can read the file... – nath Aug 09 '17 at 19:15
I edited the answer. copy and paste it, but instead of add your username. – Hunter.S.Thompson Aug 09 '17 at 19:17
Python probably does not understand the `~/` syntax being `/home/`. – Hunter.S.Thompson Aug 09 '17 at 19:19
This snippet is outdated, check wbob answer – Gilles Quénot Jan 04 '18 at 00:27
@GillesQuenot still works for me :-) – nath Jun 23 '18 at 12:28
jq can be used instead of python: `jq '.windows | .[].tabs | .[].entries | .[].url' $(find ~/.mozilla/ -name recovery.js)` but the tab list seems very off/outdated – Will Aug 09 '18 at 15:03
Didn't work since _firefox_ uses its own `.jsonlz4` formatted files nowdays - so i ported the code to Python-3 and the locating of the recovery-file is now automated, so now the code work as-is. – ankostis Apr 24 '21 at 23:05

score 2 · Answer 3 · answered Mar 24 '20 at 17:18

Some of these answers reference the "[random chars].default" directory. Starting with version 67, users can have profiles for different update channels (e.g., release, beta, nightly, etc.).

On my Ubuntu 18 system, this directory was "[random chars].default-release". I still had a "[...].default" directory but it was mostly empty. Keep that in mind if you get an error that "sessionstore-backups" can't be found.

score 1 · Answer 4 · answered Dec 04 '19 at 22:08

I recommend using https://github.com/balta2ar/brotab for this purpose:

pip install brotab
brotab install

Install the web extension as well: https://addons.mozilla.org/en-US/firefox/addon/brotab/

Restart Firefox, and you can use brotab list and parse it as so:

bt list | awk -F'\t' '{
    print "Downloading "$2
    system("curl --silent --output \""$2"\" \""$3"\"")
}'

Firefox - reading out urls of opened tabs from the command-line

4 Answers4