robtu@itx.isc.com (Rob Tulloh) (03/19/91)
I think I have determined what is happening when my machine crashes
after calling ASyncRun(). (This is in reference to my last post
concerning problems with popen/pclose found in popen.lzh on
abcfd20.larc.nasa.gov - apologies for the multiple posts!). The problem
definitely traces to something in ASyncRun().
Apparently, if you do not specify pcb_Output to be a valid file handle,
the thing just simply doesn't work. Now, the ARP docs claim that if you
pass 0, the program will inherit your CLI window stdout. Something is
amiss here. Even if I tell it the result of Output(), it will still
crash. I tried digging out the pr_COS handle from the Task structure
and it crashed. I tried pulling the output file handle from the cli
structure (don't remember which field), and it still crashed. I tried
to open the window again (i.e. Open("*",mumble mode)) and it still crashed!
The only thing that seems to work is if I open up some other file handle
like NIL: or ram:foo or sys:t/foo and hand that to ASyncRun(). Now, this
is fine if you don't care to see the output right away (or at all), but...
I tried playing around with some of the controls which can be set
via pcb_Control in the PCB struct. One other result that didn't crash
the machine is specifying pcb.pcb_Control = PRF_STDIO. Doing this means
that a new window is opened and ARP will supply a default window and
name (if you don't give it one). The new window is spawned but the whole
process just hangs with no output produced in either window.
I am wondering if anyone else has had this type of problem with ARP?
I am using V39 which I think is the latest rev. of the library. I don't
understand why it crashes the machine when it has the valid file handles.
--
INTERACTIVE Systems Corp. Tel: (512) 343 0376 Ext. 116
9442 Capital of Texas Hwy. North Fax: (512) 343 0376 Ext. 161 (not a typo!)
Arboretum Plaza One, Suite 700 Net: robtu@itx.isc.com (polled daily)
Austin, Texas 78759 GEnie: R.TULLOH (polled monthly)