Arch Linux User Repository

@yan12125: that's a good question, I never tried! From your experience it seems plausible. I'll play with it - if I can ditch a dependency that's never a bad thing.

Is python2 required to run insync? Apparently insync comes with its own python executable (/usr/lib/insync/py), and it works if I remove the python2 package forcibly.

Alright folks! (@auroq and @techlive in particular). I've added the dependencies suggested after some research. They are not listed in the .deb file I'm pulling from insync, but the insync binaries do expect them. Hopefully that saves some trouble for others in the future.

For anyone updating a working insync who finds themselves surprised by new dependencies that they don't have installed, please comment so I can be aware of anything extra.

@thenaterhood, Thanks for maintaining this package! I ran into the following error which was fixed by installing "libxcomposite". You might want to add that to dependencies.

Traceback (most recent call last): File "<string>", line 6, in <module> File "main.py", line 128, in <module> File "maininsync.py", line 125, in <module> File "isyncd/linux/headless/cli.py", line 49, in run File "maininsync.py", line 46, in main File "isyncd/linux/platform_impl.py", line 2, in <module> File "isyncd/linux/appui_impl.py", line 2, in <module> File "isyncd/linux/htmlui_base.py", line 7, in <module> File "PySide2/QtWebEngineWidgets.py", line 26, in <module> File "PySide2/QtWebEngineWidgets.py", line 17, in _bbfreeze_import_dynamic_module ImportError: libXcomposite.so.1: cannot open shared object file: No such file or directory

techlive: I'm happy to add those as dependencies if they're needed, but I need to look more into that. I'm concerned with the fact that this is the first I'm hearing of a major dependency problem since I adopted the package - which may mean something else is going on, or that nobody else has complained - who knows.

On a fresh installed Arch Linux, no GUI, start Insync in headless mode, get "Insync doesn't seem to be running. Start it first." on every command.

.config/Insync/out.txt says a lot about missing libraries, pkgfile the missing library one by one, trial and error, finally get it working. shouldn't these be added as dependencies?

$ pkgfile libGL.so.1
extra/libglvnd

$ pkgfile libsmime3.so
core/nss

$ pkgfile libXcursor.so.1
extra/libxcursor

$ pkgfile libXi.so.6
extra/libxi

$ pkgfile libXtst.so.6
extra/libxtst

$ pkgfile libXrandr.so.2
extra/libxrandr

$ pkgfile libasound.so.2
extra/alsa-lib

Also, refer to this article https://medium.com/@insync/run-google-drive-via-command-line-c3c504425fac for headless setup.

RE: headless mode - I'm in agreement for that. RE: systemd userspace - That sounds reasonable to me.

I'll look into both of those.

This is right. I personally think that systemd service should start insync always in headless mode.

I was having trouble getting the service to work on a headless box running insync.

The solution was to edit the service file to point to insync-headless instead of insync.

The proposed insync.service for systemd userspace has a After=local-fs.target network.target definition. IMHO this should be removed because systemd in userspace does not have these targets (only systemd system has both of them).