PDF-990 captured a major rework of the the text selection algorithm from a text box inclusion mechanism to mouse based start/end detection.
The text box solution work well but had a few corner case issue that didn't allow text on two lines to be easily selected but it work very well for multiple page selections. The new approach in
PDF-990 worked a lot better for selecting text on a line by line bases but was susceptible to missing lines if a user moves very quickly from one page to the next.
This bug will capture the work needed to bring these two approaches together to get a solution that has no limitation.
A community memory has posted an test file that will be very helpful while reworking our text extraction sorting algorithms. The test file can be found here, http://www.icesoft.org/JForum/posts/list/23297.page .