Beta test of major new features

[chaley]

Originally Posted by kovidgoyal

Unfortunately, you need the storage id to have non null values for VEN and PROD, otherwise you cannot have more than one such driver in the calibre code base.

Why? From what I see in the code, the hex VENDOR_ID and PRODUCT_ID are checked first on all systems before card matching happens. Can't I have multiple drivers as long as the hex IDs are different?

[ccayer]

OK, I just got my Jetbook back, and was just about to test the "Book on Device" feature (using 0.6.92), but it won't open at all. The Task Manager show calibre.exe running with 0 CPU usage, and a constant 58Mb Mem usage, but no UI appears. I'm not sure how to gather the appropriate info to help diagnose the issue; is there a message I missed that explains what you're looking for? I keep seeing mention of "open a ticket", but still haven't found where to do that either. Thanks!

[chaley]

Originally Posted by ccayer

OK, I just got my Jetbook back, and was just about to test the "Book on Device" feature (using 0.6.92), but it won't open at all.

Does calibre start if the device is not plugged in? What OS are you running?

I'm not sure how to gather the appropriate info to help diagnose the issue; is there a message I missed that explains what you're looking for?

The first thing to do is start calibre in debug mode. You do that by running 'calibre-debug -g' (no quotes) from a command line. How you get a command line box depends on the system you are running.

There will likely be some output in the command box (terminal window) that will help debug the problem.

I keep seeing mention of "open a ticket", but still haven't found where to do that either. Thanks!

Go to http://bugs.calibre-ebook.com/ and register. Once done, you can create tickets. You can look at existing tickets without registering.

Thank you for your help. I very much appreciate it.

[ccayer]

Does calibre start if the device is not plugged in? What OS are you running?

The first thing to do is start calibre in debug mode. You do that by running 'calibre-debug -g' (no quotes) from a command line.

Sigh. OK, this one's my fault. I moved Calibre & the library from my flash drive to my network at home. Everything worked great at first UNTIL I rebooted my machine. The drive letter mapped to the Library share was "Disconnected" and Calibre wouldn't start. Simply clicking on the drive in Explorer reconnected it, which allowed Calibre to run.

Any chance there could be a check to make sure the library is available when starting Calibre?

[chaley]

Originally Posted by ccayer

Any chance there could be a check to make sure the library is available when starting Calibre?

I just tried your scenario and found the same behavior you described. I would have expected calibre to complain and die.

I have opened a ticket for the problem.

[chaley]

Originally Posted by user_none

Just curious, are you not using the metadata caching code that I was using with the on device column? It included id's associated with the library entry so the file names didn't have to be modified?

Well, I think I was wrong. Having a Sony that caches (in a form) metadata, I didn't realize how *slow* opening a device can be. Using my phone pseudo-driver, and given ccayer's description, I now see.

I will put the cache into the beta, ignoring the sony (for the moment).

[DoctorOhh]

Originally Posted by chaley

Well, I think I was wrong. Having a Sony that caches (in a form) metadata, I didn't realize how *slow* opening a device can be. Using my phone pseudo-driver, and given ccayer's description, I now see.

I will put the cache into the beta, ignoring the sony (for the moment).

Just a side note: There have been a few instances on this board where people with quite a few books on a SD card had to turn off metadata reading from the device so they wouldn't have to wait 20 minutes after hooking up their device to send other books to it.

[theducks]

Originally Posted by dwanthny

Just a side note: There have been a few instances on this board where people with quite a few books on a SD card had to turn off metadata reading from the device so they wouldn't have to wait 20 minutes after hooking up their device to send other books to it.

I send books to my device while the library is building.
Even with Meta-data-off, the process takes about a minute per 100 books
and
Also blocks changing of preferences (i use 2 different save templates. Series and No series., that are quite different).

Would it be possible to have Calibre acknowledge a detected device ONLY (not build a library)

and (regardless of the above)

Add a Manual Refresh library list, per Storage detected?
Current Calibre requires ejecting the device, then reattaching and allowing the entire process to run. again.

Lots of minutes used up

[chaley]

Originally Posted by chaley

I will put the cache into the beta, ignoring the sony (for the moment).

I have added user_none's metadata caching code. It is 50% faster on my HTC, but I don't know whether that is indicative of performance on other devices. Certainly USB1 devices will speed up much more. I also found that when I put PDFs on my phone, the old code slowed down while the new code didn't. Bottom line: it helps, but we will need to see how much.

Also, I removed the ID from the filenames for devices other than Sonys, because it serves no purpose.

Finally, note that the first time you use this code to open a device, it will not be any faster, and could be slower. It must build the metadata cache.

Kovid, the code is in the my lp repository for your review. Also, my HTC driver is there. I would appreciate it if we could keep that driver at least until the release, because it is incredibly useful for my testing.

[chaley]

Originally Posted by theducks

Would it be possible to have Calibre acknowledge a detected device ONLY (not build a library)

and (regardless of the above)

Add a Manual Refresh library list, per Storage detected?

FWIW: I think we should see how much the metadata cache helps before going down this road. And in this best of all possible worlds, it won't be needed. (Ever the optimist, aren't I? )

[kovidgoyal]

Originally Posted by chaley

I have added user_none's metadata caching code. It is 50% faster on my HTC, but I don't know whether that is indicative of performance on other devices. Certainly USB1 devices will speed up much more. I also found that when I put PDFs on my phone, the old code slowed down while the new code didn't. Bottom line: it helps, but we will need to see how much.

It depends on the ebook format being used on the devices. PDF is the slowest, EPUB and MOBI are relatively quick.

Also, I removed the ID from the filenames for devices other than Sonys, because it serves no purpose.

Isn't it used by the On device column?

Finally, note that the first time you use this code to open a device, it will not be any faster, and could be slower. It must build the metadata cache.

We can always add a progress dialog for the initial cache build.

Kovid, the code is in the my lp repository for your review. Also, my HTC driver is there. I would appreciate it if we could keep that driver at least until the release, because it is incredibly useful for my testing.

Unfortunately, I'm travelling the rest of this week (leaving tonight) so I don;t know if I will have the time to review it and build an updated beta, let's see.

[chaley]

Originally Posted by kovidgoyal

Isn't it used by the On device column?

To determine the OnDevice status, devices use cached metadata. It checks, in order of preference the book's UUID then the book's 'application_id, then the db_id from the filename. If none of them are there (why would that happen?), then author matching is tried. I changed the OnDevice cache construction as appropriate.

The db_id is still included in Sony pathnames, and because the other values aren't there, it is used for OnDevice.

Unfortunately, I'm travelling the rest of this week (leaving tonight) so I don;t know if I will have the time to review it and build an updated beta, let's see.

Having fun, I hope.

[DoctorOhh]

Originally Posted by chaley

I have added user_none's metadata caching code. It is 50% faster on my HTC, but I don't know whether that is indicative of performance on other devices. Certainly USB1 devices will speed up much more.

Looking forward to it.

This may be obvious, I delete a book from my device as I finish reading it, how well does it handle books deleted from the device upon reconnecting?

[chaley]

Originally Posted by dwanthny

This may be obvious, I delete a book from my device as I finish reading it, how well does it handle books deleted from the device upon reconnecting?

When the device is connected, the cached metadata is cleaned up to reflect the books on the library at that time.

Books that appear are added to the cache, as long as they are in a directory that calibre knows about, matching books imported into calibre since the device was last connected. However, the books are not automatically imported into calibre -- this release does not include synchronizing.

[speakingtohe]

I have been using the Beta for a few days and finally added some custom fields.
I thought wouldn't it be wonderful if you could bulk edit the customs fields, looked, and there it was. How cool is that.
Thanks