t
wdhandler ''
5
wdhandler ''
)
This has been true forever. I used to do this on purpose to spy on
what J was doing "behind the scenes". Usually to figure out the code
behind a particular IDE wd event (so I could duplicate it).
Of course, that was in J4; since J5, all IDE wd events have been
handled by the single sentence wdhandler'' which makes the strategy
a lot less useful.
I don't remember if m :0 also captures other asynchronous events,
like socket_handler or DLL callbacks (obviously it captures
sys_timer'' because that's a wd event).
Which reminds me: the documentation for -jijx says:
1!:11 ]2216 239 ;~ jpath '~help/user/cmdline.htm'
Jwd tests at the end of the execution of every sentence. If there
is an error and no ijx window, then a message is displayed and the
session is terminated after the message is closed. If there are no
forms, then the sesssion is terminated.
But shouldn't the session stay active if there's ANY possibility of
further processing (c.f. the examples I cite in the earlier
paragraph)? Theoretically, even break handling can cause an event
which would trigger a return to immex which would trip the phrase in
9!:26 '' ...
Perhaps best practice in these cases is to run under jconsole? But,
then, what if one wants both (temporary) forms and asynchronous
events? For example, to configure a web server, then run it in the
background?
For myself in practice, the question is academic. I run all my code
under the full IDE.
-Dan
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
