Bug in codesize.sh, & improvment

• Key: DIR-99
• Type: Bug
• Status: Closed
• Resolution: Fixed
• Priority: Trivial
• Assignee: Alex Karasulu
• Reporter: Emmanuel Lecharny
• Component/s: miscellaneous

Description

The codesize script does not give the good count : if two java files have the same name in two different directories, the latest replace the previous one, thus the count is lost.

I suggest to replace the followin lines :

        mkdir temp
        for file in `find $trunk -type f -regex '.*\.java'`; do
                cp $file temp;
        done;
        cd temp;
        lines=`wc -l * | grep total | awk '{print $1}'`;
        cd ../
        rm -rf temp;

by this one :

        lines=$(for file in $(find $trunk -type f -name "*.java"); do wc -l $file; done | awk '{sum += $col} END {print sum}')

Note that no more copy are made, leading to a small improvment in performance, and in disk occupation ;-)

The result is :

lines of code in ./ldap/trunk = 99208 (was 94685)
lines of code in ./sandbox/trunk = 3630
lines of code in ./protocol/trunk = 583
lines of code in ./dns/trunk = 4242
lines of code in ./authx/trunk = 8325 (was 8109)
lines of code in ./naming/trunk = 15950 (was 15867)
lines of code in ./network/trunk = 14538 (was 14360)
lines of code in ./apacheds/trunk = 44670
lines of code in ./ntp/trunk = 1618
lines of code in ./asn1/trunk = 26796 (was 26752)
lines of code in ./kerberos/trunk = 13292 (was 13083)
lines of code in ./changepw/trunk = 2291
lines of code in ./dhcp/trunk = 4911
lines of code in ./apseda/trunk = 15177
Total: 255231 (was 249978)