OT: broken/useless ansi - console driver??
Kenneth Brody
kenbrody at bestweb.net
Tue Oct 25 11:44:52 PDT 2005
Quoting Bob Stockler (Tue, 25 Oct 2005 13:13:24 -0400):
[...]
> But I don't see this as filePro's problem, but the responsibility
> of the shell programmer to know how to write shell scripts for
> a specific platform, or to write generalized scripts that will run
> on any platform under any shell.
True, but you know that it's going to be a nightmare for fpsupport
with people calling up saying "it used to work" and "it works if I
run the script from a shell prompt", and "it only fails when run it
from a filePro menu".
[...]
> And anyone familiar with filePro and shell scripts knows to trap
> the INT and QUIT signals in the shell scripts to protect them
> from those signals when issued from within filePro.
There's a world of difference between SIGINT causing the shell to
terminate any further execution of the script (as is the case with
/bin/sh) and the shell itself terminating (as /u95/bin/sh does).
There is no excuse for the shell itself to exit while dclerk is
still running.
While the "echo -n" is relatively minor, and the change in "["s
behavior can be explained as "your shell script is broken and always
has been", the SIGINT change is a major issue.
How do you explain to the 99% of the people out there who will refuse
to see "it worked in filePro 5.0 and it fails in filePro 5.6" as
being anything but a filePro bug?
While we have worked around this problem by replacing the SCO-supplied
system() function with one that correctly calls "/bin/sh", it's still
an issue related to the curent thread.
--
KenBrody at BestWeb dot net spamtrap: <g8ymh8uf001 at sneakemail.com>
http://www.hvcomputer.com
http://www.fileProPlus.com
More information about the Filepro-list
mailing list