If i put 1 in the paste buffer (from an html source, eg ijx or ijs windows) and i paste into ijx and return, ijx evaluates 1. Good.
If i do 5 pastes so i get 11111 on the command line and i still get 1 as the evaluation result! Interspersing the pastes of 1 with numbers so i have 012131415 i get 15. Doing the same setup but moving the cursor(^): 0121^31415 i get 13 Doing the same setup but moving the cursor(^): 012^131415 i get 13 Doing the same setup but moving the cursor(^): 01^2131415 i get 12 Doing the same setup but moving the cursor(^): 0^12131415 i get 12 Doing the same setup but moving the cursor(^): 012131^415 i get 14 Doing the same setup but doubling the typed numbers and moving the cursor(^): 0012213^3144155 i get 133 Doing the same setup and moving the cursor(^): 0012213314^4155 i get 144 my guess is from the cursor position the IJX is taking the text of last paste and the tokens until the next paste as the input line. i seem to get the same results using , or ' for delimeters and either ff or chrome (on W7 and XP) in the JUM If i do a paste of a number cut from some window where that digit has a style, that style is preserved in the paste line - so my guess is the current (where the cursor is) 'text' of a markup plus the succeeding typed entries only, are used. This was verified by taking copies from a prepared page, and found to go up only one markup level. Previous and succeeding markups are cut out. They are not in the scroll buffer nor executed - though they were pasted. So a solution is to eliminate markup, before inserting from the paste buffer. After that its elimination becomes problematic as the user can 'jump' to any text spot at any time and hierarchy (which correct XML and its tools requires) is not necessarily preserved. greg ~krsnadas.org ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
