About the shell script

[juno2]

Sorry to post a new question again there, but I really want a solution .

I have a script which can compile a file ( eg. file1.cdr ) , the compiled file is .cbt extension ( eg. file1.cbt ) , how can I mv the compiled file (file1.cbt not file.cdr.cbt ) to the other path ( from /tmp to /usr) ? thx.

$pwd
/tmp
$my_script file1.cdr
$pwd
/usr
$file1.cbt

[Mark Grant]

#!/bin/sh

FILE=`basename $1 .cdr"
compile $1
mv /tmp/$FILE.cbt /usr

This assumes you start the script with something like

myscript myfile.cdir

Also, the "compile" bit is where you do your compile with whatever command you use.

[juno2]

It is great , thx much for help.

[juno2]

thx reply ,

but I still have a problem , I don't know why even the program can't be compiled ( some compile error gerenated ), it still report the result is "0" , so it is not work to use the error checking function " if [ $? -eq 0 ] then .. " , could suggest is there other method to make sure the file will be copy only when the compilation is successful ? thx.

[Mark Grant]

juno2

You'll have to find out in the script, the same way as you do yourself. If the compile produced an error somewhere you'll have to look for it. As an example, suppose the compile writes lots of stuff to the screen and then "ERROR, it didn't work too well" if there is a failure. You would do something like this

your_compile_command > /tmp/compile.log
grep ERROR /tmp/compile.log > /dev/null || {
your_copy_commands here
exit
}
echo "Couldn't copy because the compile broke again"

[juno2]

thx Mark reply, however,

the system will not generate the word "ERROR" , it will generate different words , not exactly the same , do you have other idea ? thx.

[Mark Grant]

juno2,

You could try it the other way around. If it produces no output if it is successful then you you could test for success (nothing in the log) instead of failure (something in the log). Otherwise, I don't think there is much you are going to be able to do.