Nook STRg doesn't eject completely?

[JimmXinu]

I recently switched from a Win7 laptop to a new Win8 one.

On the new computer, when I eject my Nook STRg, it doesn't eject completely, if you will.

The Nook still shows the 'USB Mode' screen, and calibre sorta still thinks it's connected. See the attached image--after I ejected using calibre, there's no Device icon, but if I try to delete a book, I get the delete confirmation for when there is a device attached. I have to physically unplug the Nook to prevent that.

Since it still ejects completely as I expect on the Win7 laptop, I suspect either it's a Win8 issue or a USB driver issue of some kind.

Does anyone have any ideas or advice?

[kovidgoyal]

Run in debug mode and see if you get any helpful output

[JimmXinu]

On connect I get:

Spoiler:

Worker Launch took: 0.142999887466
Job: 9 Get device information started
DeviceJob: 9 Get device information done, calling callback
DeviceJob: 9 Get device information callback returned
Job: 10 Get list of books on device started
DEBUG: 7792.3 USBMS: Fetching list of books from device. Device= NOOK_COLOR onca
rd= None
DEBUG: 7792.3 USBMS: dirs are: G:\ My Files
DEBUG: 7792.3 USBMS: scan from root True G:\
DEBUG: 7792.3 USBMS: count found in cache: 0, count of files in metadata: 0, nee
d_sync: False
DEBUG: 7792.3 USBMS: Finished fetching list of books from device. oncard= None
DEBUG: 7792.3 USBMS: Fetching list of books from device. Device= NOOK_COLOR onca
rd= carda
DEBUG: 7792.3 USBMS: dirs are: I:\
Job: 9 Get device information finished
No details available.
DEBUG: 7800.9 USBMS: scan from root True I:\
DEBUG: 7804.2 USBMS: count found in cache: 2773, count of files in metadata: 277
3, need_sync: False
DEBUG: 7804.2 USBMS: Finished fetching list of books from device. oncard= carda
DEBUG: 7804.2 USBMS: Fetching list of books from device. Device= NOOK_COLOR onca
rd= cardb
DeviceJob: 10 Get list of books on device done, calling callback
READING LIST: Metadata available: {'A': {u'device_store_uuid': u'de84e019-fe07-4
e40-8bd6-020f308c9ec3', u'location_code': 'A', u'device_name': u'The Nook', u'ca
libre_version': u'1.3.0', u'prefix': 'I:/', u'last_library_uuid': u'e835a254-143
e-4293-b466-d075973e44a4', u'date_last_connected': u'2013-09-20T03:23:31.836000+
00:00'}, 'main': {u'device_store_uuid': u'1164341b-3489-49e5-8ef2-386a547850c4',
 u'location_code': 'main', u'device_name': u'The Nook', u'calibre_version': u'1.
3.0', u'prefix': 'G:/', u'last_library_uuid': u'e835a254-143e-4293-b466-d075973e
44a4', u'date_last_connected': u'2013-09-20T03:23:31.817000+00:00'}}
READING LIST: Sync Now - force_sync= False
READING LIST: Device found to sync to: The Nook de84e019-fe07-4e40-8bd6-020f308c
9ec3
READING LIST: No books on 'Add to Nook' list need to be synced
DeviceJob: 10 Get list of books on device callback returned
Job: 10 Get list of books on device finished
No details available.

I would wonder about it reporting as a NOOK_COLOR while the User-defined device info says NOOK_SIMPLETOUCH, but it reports as NOOK_COLOR on the Win7, too.

On eject, I get:

Eject failed: [Error 2] The system cannot find the file specified

But only when the SD card is in. Without, I don't get that message, but the same behavior still. And I also get that message on Win7, too.

I know some of the output above is from plugins, so I ran with --ignore-plugins, too. No change.

If I eject each drive (the SD card and the Nook) as drives first, it disconnects as expected, but the drives remain defined with media removed. (This is how B&N says to eject.) When I eject from calibre, the drives themselves are removed from the system.

This all seems so familiar that feel like it's a problem I solved on the last computer, too. All I remember about it is that I had to plug it into the SS-USB port and not the USB on the other side. I'm using ports labeled SS-USB on the new one, too...

[kovidgoyal]

Not much I can tell you, calibre uses the CM_Request_Device_Eject api call to eject devices. That is apparently failing on your system, not being a windows user, I haven't a clue as to why that might be.

[kovidgoyal]

Furthermore if you are getting no error message when the sd card is not present that means the API call is reporting success.

[FizzyWater]

I've never bothered to debug (to be honest, don't even know how). But since I switched to Win 7 I've always had to eject first from the taskbar, then from Calibre. If I don't, I have the issue with excess battery draining.

For what it's worth...

[kovidgoyal]

Looking over the source code, the reason for the error message with the SD card inserted is that the code attempts to eject the device using the drive letters in order. So when the eject is run on the main memory, all drives are removed, as the entire device is ejected. This causes the subsequent eject on the SD card to fail, harmlessly.

[JimmXinu]

Well, I've continued looking, but either I'm the only one having this issue or I'm the only one who cares.

The answer from B&N is to eject each drive (nook & SD card) separately from the file explorer.

The answer from Microsoft is 'Safely Remove and Eject' from systray.

The programmatic answer from Microsoft is CM_Request_Device_Eject()--which calibre already does and is equivalent to 'Safely Remove and Eject'.

All I've come up with so far is that using PowerShell I can do this:

$o = new-object -com Shell.Application
$o.Namespace("G:\").Self.InvokeVerb("eject")
$o.Namespace("H:\").Self.InvokeVerb("eject")

I'm trying--with little success so far-- to lift code from Open With to run a PowerShell script from SmartEject.

[kovidgoyal]

I'll see if it is possible to create an eject utility that first ejects individual drives and then uses CM_Request_Device_Eject

That might do the trick, or it might not...microsofts device subsytem is truly terrible.

[kovidgoyal]

Try the attached eject utility, to use it run

calibre-eject.exe E F

replace E and F with the drives of your nook.

This will first eject the individual drives using the technique described here: http://support.microsoft.com/kb/165721

and then call CM_Request_Eject_Device

[kovidgoyal]

Oh and the code is here, in case you want to play with it.

https://github.com/kovidgoyal/calibr...506f6472277b4a

[JimmXinu]

Originally Posted by kovidgoyal

Try the attached eject utility, to use it run

calibre-eject.exe E F

replace E and F with the drives of your nook.

This will first eject the individual drives using the technique described here: http://support.microsoft.com/kb/165721

and then call CM_Request_Eject_Device

Perfect! Drives are both unmounted and removed, Nook returns to 'On' while still connected. Works with Kindle, too.

[PeterT]

For a long time I've made use of one of the Sysinternal's tools; SYNC.

UNIX provides a standard utility called Sync, which can be used to direct the operating system to flush all file system data to disk in order to insure that it is stable and won't be lost in case of a system failure. Otherwise, any modified data present in the cache would be lost. Here is an equivalent that I wrote, called Sync, that works on all versions of Windows. Use it whenever you want to know that modified file data is safely stored on your hard drives. Unfortunately, Sync requires administrative privileges to run. This version also lets you flush removable drives such as ZIP drives.

My experience is that it does not require admin access.

I often do a "sync -e p:" where P is my Kobo.

[JimmXinu]

For those following along at home, Kovid has committed changes for the next calibre release (presumably) that will add a new command line program for the windows install, 'calibre-eject', described as a 'Helper program for ejecting connected reader devices'.

Since I'm the only one asking for it, I completely agree with his decision not to add it to the eject code used by everyone.

Thanks, Kovid, I appreciate it.

[kovidgoyal]

@JimmXinu: The new calibre-eject helper program will be used for everyone. See the changes to usbms/device.py that I linked to earlier.