(In reply to rosenp from comment #0)
> With the previous version, everything worked as intended. With the latest
> version, only the Documents folder gets mounted automatically. If I use
> gvfs-mount manually, it works but then Rhythmbox is unable to see the iPod.
> 
> Version is 1.28.1
> 
> iPod is an iPod Touch 2G with iOS 4.2.1

We've done that because there's nothing interesting in the "normal" mount anymore for more recent versions of iOS. Obviously, your device is older, and doesn't have that same problem.

If we mounted it automatically, we would do that based on the version of iOS used, so that only older versions of iOS get mounted automatically for Rhythmbox, gtkpod, etc. to use, and not newer ones.

http://www.libimobiledevice.org/ mentions that 4.3.5 might be the cut-off point.

What's the output of:
ideviceinfo -q com.apple.mobile.iTunes -k DBVersion
on your device?

4

interetingly enough i have a 3GS running on iOS 6.1.6 with the same DB version even though the iPhone 4 has version 5 on the same iOS version if i remember correctly.

I do wonder if sync-ing to a DBVersion 5 device is possible. It requires a proprietary .so file that i mirrored on my github page.

What DBVersion does the latest stuff use?

Any progress on this? I played with this a few days ago and on devices newer than DBVersion 4, you can still listen to whatever music is sync'ed to the device, just not sync new music. Not through MTP of course(or PTP). Would it make sense to revert based on that reason alone?

(In reply to rosenp from comment #3)
> Any progress on this?

No.

> I played with this a few days ago and on devices newer
> than DBVersion 4, you can still listen to whatever music is sync'ed to the
> device, just not sync new music. Not through MTP of course(or PTP). Would it
> make sense to revert based on that reason alone?

I'm not reverting changes that make things better for current devices, to help with obsolete devices that aren't supported by their makers and full of security holes.

If somebody presents a patch that implements this, I'll certainly test and apply, but I don't have time to spend on this right now.

Update: I recently tried backing up a friend's Pictures using MTP but it did not work. Neither in Windows nor in Fedora did the pictures show up under MTP but they did show up under AFC.

Can we revert this change? AFC is both more reliable and faster. Extra data as well.

I do also remember another use case regarding copying music already sync'ed to the iDevice.

(In reply to rosenp from comment #5)
> Update: I recently tried backing up a friend's Pictures using MTP but it did
> not work. Neither in Windows nor in Fedora did the pictures show up under
> MTP but they did show up under AFC.

The fact that it doesn't work under Windows is irrelevant. It works with recent versions of libmtp as well, please check with your distribution.

> Can we revert this change? AFC is both more reliable and faster. Extra data
> as well.
> 
> I do also remember another use case regarding copying music already sync'ed
> to the iDevice.

Still a no, sorry. This is a corner case, and you can already accessing this data by entering the device's URI manually in nautilus, no need to show it to all users all the time.

If anybody needs a version with the "old" behaviour: I have created a ppa for Ubuntu 16.04 with the commit reverted. https://launchpad.net/~ingo/+archive/ubuntu/gvfs-afc-ipod

(In reply to Ingo Haschler from comment #7)
> If anybody needs a version with the "old" behaviour: I have created a ppa
> for Ubuntu 16.04 with the commit reverted.
> https://launchpad.net/~ingo/+archive/ubuntu/gvfs-afc-ipod

If you want to add special-casing for older iOS versions where the iTunes database version would be supported out-of-the-box with libgpod's code (and without additional binary-only blobs), please file a new bug about that problem.

I would probably advise to start looking into adding support for mounting access to those files in the applications themselves though, but that's probably too much work for obsolete hardware.

Well, I already added support for the mount to rhythmbox which makes the gvfs patch redundant. I also think handling the new gvfs behaviour in the apps is the right way to go. 
A remark to "obsolete" hardware: I for example do deliberately buy such "old" iPods (with DB version < 5) to be able to use them with libgpod/rhythmbox. Perhaps I am not the only one ;-)
Anyway, thanks for your reply.

(In reply to Ingo Haschler from comment #9)
> Well, I already added support for the mount to rhythmbox which makes the
> gvfs patch redundant. I also think handling the new gvfs behaviour in the
> apps is the right way to go. 
> A remark to "obsolete" hardware: I for example do deliberately buy such
> "old" iPods (with DB version < 5) to be able to use them with
> libgpod/rhythmbox. Perhaps I am not the only one ;-)
> Anyway, thanks for your reply.

I wouldn't trust a device that can go to the Internet but hasn't seen security updates in many years (full disclosure, I still have an iPad 1 I bought on release day).

*gg* Absolutely agree on your point - this is not for the unexperienced user . Internet access is turned off on mine for sure; it is just used as a luxury music player...