Add support for moving items up/down 10 rows and to top or bottom of …

[snarkophilus]

…a reading list.

[JimmXinu]

Found a few bugs.

If I select several books and click up/down 10 or beginning/end, I end up with blank rows appearing in the list. Here's a simple way to see it (Win10, Calibre 64 bit):

1. Add four items to an empty list
2. Edit List
3. Select top 3 items
4. Click 'Move to bottom'
5. Watch empty rows appear
6. Click 'Move to bottom' or 'Move to top' more times to see more weirdness.

Looping on move_rows_up/down() is inelegant and inefficient for large lists, especially with large selections. I had it hang for several seconds with less than a 1000 entry list. Many calibre users have libraries of 10s of thousands.

There's a comment in move_rows_up(self) that says:

        # Workaround for strange selection bug in Qt which "alters" the selection
        # in certain circumstances which meant move down only worked properly "once"

I wonder if that's part of it?

[snarkophilus]

For the first bug, blank rows appearing, I can reproduce that - probably just need to add a bounds check? This was obviously a case I didn't check. Ta, will look into this.

For the second bug/comment, yes, looping is inefficient. I see a small delay with my current 300ish list. This was the least invasive way to implement this. I'd also wager than people with a 1000 entry list wouldn't move items or selections of items through lists, at least not without breaking their mouse button :).

Last - I think I should have created a branch then applied the patch in my local git repo? Still new git.

[snarkophilus]

Jim- for the first bug, I can reproduce the problem with the existing (unmodified) plugin with the same steps, except clicking on the down arrow button repeatedly. With four books in a list , the first click moves the top three items down to the bottom, the following clicks then insert a blank row per click.

[JimmXinu]

So it does. Joy.

I think I figured it out and pushed a fix: f9ab494

You want to try that with your code?

[snarkophilus]

So it does. Joy.

I think I figured it out and pushed a fix: f9ab494

You want to try that with your code?

Yep, that works for me! Thanks.

[JimmXinu]

I'm not sure why the commit has a different number, but I did incorporate this: 6debdf5