Select-next-paragraph doesn't work across scrivenings

I like to select paragraphs of text using the keyboard shortcut option-shift-down arrow, which selects text to the end of the paragraph, and then repeated down-arrows extend the selection by paragraph.

Except, if I’m in scrivening mode with multiple documents selected, it hits a wall at the end of a document and won’t select further (pressing down-arrow does nothing). However, if I switch to the mouse, I can shift-click to extend the selection further (crossing the document boundary), and then the keyboard shortcut will start functioning again. It seem like I should be able to do this without having to switch between keyboard and mouse.

What’s really strange is if you have a selection which crosses a boundary, you can use option-shift-up arrow to shorten the selection, and this will work over the boundary-- but then you can’t get back!

I’ve just tried this.

For me also (Mountain Lion 10.8.1, Scriv 2.3.1) opt-shift-down doesn’t go beyond the boundary the first time, but if you follow the first opt-shift down with an opt-shift-right arrow, the select does cross the boundary, and then you can opt-shift down across any subsequent document boundary. It’s only the first one that doesn’t work…

Odd, isn’t it.

This is the result of checks Scrivener does to try to avoid a different unwanted behaviour. When triple-clicking a paragraph, the separator after it normally gets selected. In scrivenings mode, however, because the separators are uneditable, if you select the last paragraph in a section in this way and try to cut it or edit it, then it won’t work because the selection of the separator will block many actions. Thus Scrivener tries to avoid situations whereby the separator gets selected if the rest of the selection is contained within a section. This is what is causing the problem here (because Scrivener has no way of knowing how the selection is made - whether it is by triple-clicking or using opt-shift arrow or whatever). I have now made improvements to the logic for 2.4 that fixes this particular case, though.

Awesome! Thanks!