Am 06.09.2013 14:53, schrieb M.L.:
> All Linux here (apart from odd machines running Solaris etc., which
> aren't relevant to the test).
> it appears to want explicit start commands, i.e. one or both of
> (prefixed for Hercules) /start 00c and /start rdr
> I can now collect the file in my cmsuser session so the first card
> (ID...) is probably OK, I need to put more time into the read control
> card but will report back later.
> Netcat is not closing the connection reliably itself, it has to have a
> manual kill using ^C. Provided that 00c and/or rdr have been started,
> when I ^C nc the Hercules console correctly reports the disconnection.
> I didn't want to say this earlier when I was attempting to follow
> instructions, but this isn't the first time that I've had termination
> problems with Netcat on Linux. Where possible, I prefer the Netpipes
> programs (faucet and hose), hence
> $ hose localhost 3505 --out cat TEST.RJE
> which transfers and terminates without manual intervention.
If you code the ID card (the second one) the same way as I did,
USERID BERND CLASS A NAME TESTP3 PLI
:READ TESTP3 PLI A1
you can simply issue READ * in your CMS machine, and the file will be read and
stored as a CMS file (replaced, if necessary).
The first card tells VM to which virtual machine the file has to be sent.
I, too, did non want to do the editing using the VM editors, because there is no
XEDIT (in this old version of VM) and I didn't want to spend the time to test
the other available editors. The solution with the transfer from the PC works
very well. For the other way round I punch the files and extract them from the
resulting PC file, see my definition
000D 3525 /herc370/vm/pun/pun00d.txt ascii
I did the whole work on the Stanford compiler using this approach. The sources
reside on the PC. Hercules and VM is used for compile and test only. (Much the
same way as I do it in my "real" working place with z/OS and the "real"