View Single Post
Old 09-29-2010, 02:56 PM   #144
chaley
Grand Sorcerer
chaley ought to be getting tired of karma fortunes by now.chaley ought to be getting tired of karma fortunes by now.chaley ought to be getting tired of karma fortunes by now.chaley ought to be getting tired of karma fortunes by now.chaley ought to be getting tired of karma fortunes by now.chaley ought to be getting tired of karma fortunes by now.chaley ought to be getting tired of karma fortunes by now.chaley ought to be getting tired of karma fortunes by now.chaley ought to be getting tired of karma fortunes by now.chaley ought to be getting tired of karma fortunes by now.chaley ought to be getting tired of karma fortunes by now.
 
Posts: 12,476
Karma: 8025702
Join Date: Jan 2010
Location: Notts, England
Device: Kobo Libra 2
Quote:
Originally Posted by kacir View Post
I came home from work. Calibre was still greyed out. I closed the program and started it again.
It looks like Calibre recovered from the situation very gracefully. I did not have to kill it, I did not have to restart computer, the database does not seem to be damaged, the change I have tried to commit is done.
We found a couple of things.

1) There was definitely some fighting going on between the backup and the bulk edit. Fixed by disabling the backup during bulk edit, and also changing the way bulk edit reports progress on the screen. On my test library of 1000 books, I don't have any problems with performance of bulk edit. Changing tags on all books takes less than 2 seconds. Changing authors and titles is slower (around 1/4 second per book), which is caused by the file system operations.

2) However... the code we added to reselect the books seems not to be optimal. It takes over a minute on my machine to reselect the 1000 books after the bulk edit finishes. This isn't 10 (or however many) hours, but it is annoying. Kovid has fixed it.

3) I had one case (before fixes) where the replace finished, and then everything hung. I am unable to repeat it. It is possible that the former progress-reporting code (multiple threads) could deadlock (although it isn't obvious how), and perhaps that is what was happening. With the thread gone, this problem should be gone.

Kovid is preparing a release that contains all these fixes.

Quote:
Thank you dear developers for your hard work.
You are very welcome.
chaley is offline   Reply With Quote