Ebook reader should suspend on idleness

[jg]

We'd like a good way to suspend explicitly: we don't want to force the laptop to be closed to suspend (though it should do that too), because the antennae may have been carefully positioned for best use (closing the laptop will put the antennae possibly down on the ground, which is a really bad place for them to be).

So by moving shutdown someplace else (#1395), we free the button for suspend use. When suspended this way (or when the laptop is closed) the screen should be turned off.

This shouldn't really be done until the UI for shutdown is designed by Eben.

[jg]

Bug #1290 is an example bug report for another reason to make the button suspend rather than shutdown.

[Eben]

#1395 Now has a UI design for placement of the shutdown button in software.

[kimquirk]

#1311 is a dup of this one.

[jg]

CJB has a patch to do the first of these (replacing shutdown with suspend), though it seems to have some issues to be sorted out in testing I did on 6/15.

I'll assign it to him for the moment, to clean up the first piece (making the power button suspend).

[jg]

Shutdown is now implemented; Chris, please get your sugar patch for the suspend button to the sugar folks ASAP.

[cjb]

Replying to jg:

Shutdown is now implemented; Chris, please get your sugar patch for the suspend button to the sugar folks ASAP.

I can (and the patches have already been on the sugar list), but I would rather we fix the SCI bug (#1729) first, else after the first suspend we will continuously drop to suspend, as you saw when testing out the patch.

[jg]

Chris has the patch in hand, please apply to the ebook reader as soon as possible for Trial-2.

[dcbw]

'touch /home/olpc/ebook-enable-sleep' to turn it on, then relaunch ebook

[cjb]

Let's leave this open, since we might want the file test to go away (or to track getting the touched file in the Trial-2 build) before this is finished. There's still also #1729, but that's not related to the sugar work.

[cjb]

This isn't in the build yet, and is blocking on #2401.

[jg]

This is in the build, but is latent; a file must be touched to make it start working. We have to fix #2401 to get full ebook.

[cjb]

Will retest.

[kimquirk]

Once all the suspend/resume problems are cleared out, this can be re-evaluated. Moving it to FRS.

[cjb]

This isn't working anymore; Read is being disallowed from sending dbus messages to olpc-hardware-manager.

[marco]

Patches make sense to me but I don't see a reason to turn it on/off depending on the presence of a file anymore. Unless someone has good reason for it, I'd just to turn it always on.

[cjb]

Isn't jhbuild one such reason, or qemu, or using an older XO without suspend support?

We are going to provide the file in the builds by default, of course, so it will be default-on.

[marco]

Replying to cjb:

Isn't jhbuild one such reason, or qemu, or using an older XO without suspend support?

With jhbuild, olpc-hardware-manager will not be running, which is handled by the try/except. I tend to think the other two cases should be handled by olpc-hardware-manager (if they are not already) and not by the activity).

[marco]

Built an rpm with Reinier patch: olpc-hardware-manager-0.4.2-1.olpc2

[tomeu]

Read-41.xo will check for /etc/inhibit-ebook-sleep at startup.

[rwh]

Thanks; makes more sense to me!

[rwh]

Verified in joyride 1546, please tag olpc-hardware-manager-0.4.2-1.olpc2 and Read-41.xo for update-1.