Reported by Kevin Reid at
http://www.eros-os.org/pipermail/e-lang/2007-February/011896.html
? interp.waitAtTop(stdout <- print("hello sometime!"))
# stdout: hello sometime!
? print("padding")
# stdout: padding
? print("padding")
# stdout: padding
Under E-on-CL's updoc implementation, this passes. Under E-on-Java's,
"hello sometime!" is collected as part of the first "padding" step's
stdout.
It seems to me that output happening before a waitAtTop completes
would be most usefully associated with the step during which the wait
was initiated.
This is a problem for me at the moment as I have an updoc script I
want to use on both E-on-CL and E-on-Java, and it will fail on one or
the other depending on how I arrange the output.