Welcome to the Monastery, babp.
You might get better feedback if you format your question a little bit more readable.
Please see Markup in the Monastery (esp. <code>). (Un-)fortunately, the HTML source of this page is more readable, so let me repeat the basic parts of your question:
Perl program:
my $diff1 = system("sh diff.sh");
my $diff2 = system("sh diff1.sh");
diff.sh:
a=diff aaa ccc
diff1.sh:
b=diff aaa bbb
The problem is, that diff.sh and diff1.sh don't do what you (probably) had in mind:
The variable (a or b) is set to the string diff, then
the shell tries to execute(!) aaa with argument bbb or ccc.
Suggestion (non-Perl): Correct diff.sh (analogue diff1.sh), e.g.
diff aaa bbb
a=$?
# do something else with $a
exit $a
Which can be reduced to...
diff aaa bbb
... in case you only want to execute diff. The shell (usually) returns the
exec state of the last statement executed. BTW: Perl has borrowed this idea here
and there (e.g. return) - although this is not best practice.
But then, wouldn't it be easier to avoid the *.sh wrappers and
put that into the Perl script too?
my $diff1 = system("diff aaa ccc >/dev/null 2>&1");
my $diff2 = system("diff aaa bbb >/dev/null 2>&1");
Note the redirections to get rid of diff's output. If your version of diff supports
a switch to suppress output and just return the exec state, then give
the system LIST variant a try for better security.
Please be aware that $diff1 and $diff2 contain 0 on success but not
necessarily 1 on error since catched signals are encoded into the return value
(details in system).
Finally, CPAN offers a lot of modules when searching for diff. Might be worth a check...
HTH |