Rob van der Heij wrote: >When you issue the READ you get control back when the user hits an AID >key. At that point there's one line stacked for each change on the >screen. You can decide whether XEDIT should update the file being >edited or not. When you have processed the stacked lines, you issue >another READ until you have seen the signal to terminate (eg PFK 3).
>From your earlier posts, Richard, I think you've been doing READ not READ ALL. >As Rob's example below shows, READ ALL is different. >Try something like this to see what happens: >/* */ >'READ ALL NUMBER TAG' >address command 'PIPE stack | cons' >Warning: READ used to be a NOP when there is a line stacked when it is >invoked. I don't see it mentioned anymore in the notes, so it may have >been fixed. It was a very popular cause for such applications to get >into a loop. Still true. Gotta be careful to drain the stack before a READ. Also note that READ NOCHANGE TAG gets you not only everything changed (despite the lack of ALL), but doesn't allow actual screen changes, *AND* interprets CTLCHARs. You can use this for a Q&D (very D!) display manager. HELP uses it -- it's undocumented but not going away. A very useful idiom. It might well be easier for you to format the entire screen and read your "prefix" areas thus, rather than using the actual prefix area. ...phsiii