dewey@sequoia.UUCP (Dewey Henize) (02/10/90)
We are on a Sequent S81, Dynix 3.0.12 running Elm at PL16 (current I believe), and have occasionally run into the following maddening problem: A user goes into elm and prepares to send a message. After entering address, etc, they go into entry mode with vi and compose the message, taking anywhere from less than a minute to quite long periods. After the message is composed, the user exits vi with either :wq or ZZ and gets the message "can't invoke editor". Checking in the elm source, it appears we are getting back a non-zero code from the attempt to invoke vi, yet obviously it WAS invoked... We don't have source to vi, of course. The errno value is 1 in all cases though while the return value is sometimes 1 and sometimes 3. Ideas? Suggestions? Thanks in advance. Dewey -- | Execucom and I often have different ideas. THESE are mine, ok? Ok. | | dewey@execu.com or uunet!sequoia!dewey | |Don't reword the question into such generality that it appears absurd, that's| |a puerile trick. -Barry Shein | But this IS Usenet!! - me |
aeh@j.cc.purdue.edu (Dale Talcott) (02/10/90)
In article <8154@sequoia.UUCP> dewey@sequoia.UUCP (Dewey Henize) writes: >We are on a Sequent S81, Dynix 3.0.12 running Elm at PL16 (current I believe), >... they go into entry mode with vi ... Checking in the elm source, it >appears we are getting back a non-zero code from the attempt to invoke vi, >yet obviously it WAS invoked... Some versions of vi return a non-zero exit status if you ever search for a pattern which isn't found. In fact, the exit status is the number of such failed searches (plus any real errors). Which makes the exit status from such vi's nearly useless. It should be easy to check if this is your problem. Fixing it without sources, however... Dale Talcott, Purdue University Computing Center aeh@j.cc.purdue.edu, {purdue,pur-ee}!j.cc.purdue.edu!aeh, aeh@purccvm.bitnet