Feature Request: Device Caching...

[Marauder]

Whenever I connect my Kindle DX to Calibre, I have to wait 10-20+ minutes for it to run through the "Get List of books on device" task before I can do anything with it. And of course Calibre is rightfully hesitant to terminate any process that's actively communicating with the device (silly Flash memory...)

This can be a royal pain in the tail, especially when I want to quickly update my news feeds before heading off to work.

How difficult would it be for Calibre to cache the metadata information on devices such that the next time it sees the same device Calibre could quickly ready the FS' directory tree to see what, if any, changes have been made (new/removed files, updated mtime, etc.), and just update those instead of opening and parsing the metadata for every single title on the device?

Prime example. I'm sitting here, waiting for a ride to the airport that could arrive any minute. All of my news feeds for this morning have successfully downloaded, but 17 minutes on it's still "Getting list of books on device..." and won't upload anything until that task finishes...

Thanks.

(I know, I know, pull up the source and make the modifications yourself, the joys of FLOSS software... guess I may have to sit down and learn Python... I hope that Dr. Goyal who is infinitely more familiar with the code, could at least give some insight into the feasibility of such a feature.)

:-)

(...several minutes pass while composing the message, this morning it ended up taking 20:28 and the last of the news feeds just finished uploading...) :-)

Anyway, thoughts, ideas, insights, patches?

[Marauder]

...Of course NOW, I see a nifty "read metadata from files on device." option within the Kindle plugins, disabling that drops connect time down to 19s... huge improvement, though a "best of both worlds" would be nice... ;-)

[kovidgoyal]

on device metadata caching is on the TODO list. But there are only so many hours in the day, so unless you're willing to help, you will have to be patient. And I simply dont understand sarcastic comments about the joys of open source software. How is having the option to do something yourself bad?

[Marauder]

Originally Posted by kovidgoyal

And I simply dont understand sarcastic comments about the joys of open source software. How is having the option to do something yourself bad?

I'm sorry, I was not beong sarcastic. I am a huge proponent of the Open Source development model for this very reason. Because I DO have the option to dig into the code and do it myself. If anything, the comments are more about my personal lack of Python skill.

I wanted to be sure it was on the radar as an issue that impacts user experience. I understand that time is always a resource with hard limits. The "do not read metadata" switch is acceptable for my needs. If I did have requisite Python skills I would be happy to give it a shot. As I said, I may have to take a look at some Python docs.