On Wed, Nov 14, 2018 at 11:41 AM Joe Orton <[email protected]> wrote:
> On Wed, Nov 14, 2018 at 11:09:07AM -0600, William A Rowe Jr wrote: > > Since the documentation offered uses an example of invoking a .pl script, > > the docs are clearly wrong with respect to this feature. It is likely a > > regression > > which changed when the option to specify the user/group was added. > > Gimme a break, "clearly", really?? > > Working script: > > > sys.stdout.write(newValue + "\n") > > Non-working script: > > > > sys.stdout.write(newValue + "/n") > > Spot the difference! I'll bet $10 the docs are right and the code is > fine. > Ack, examining the error log in the first place might have been more informative, I took the python3 invocation feedback verbatim from the post. Curious, Benedikt, whether the transformation now works, without invoking the python3 interpreter, now that you've adjusted the script for "\n" line endings? In other places such as error logs, we've implemented the users choice between invoking the binary, or invoking a shell (allowing us to use shebang line launching of scripts.) This is a possibility for the prg: mapping command invocation as well.
