On 17 July 2013 12:59, Mladen Turk <mt...@apache.org> wrote: > On 07/16/2013 11:42 PM, Christopher Schultz wrote: >> >> All, >> >> While doing the trivial fix for >> https://issues.apache.org/bugzilla/show_bug.cgi?id=55268, I noticed a >> few idioms being used in bin/daemon.sh that struck me as odd. For example: >> >> while [ ".$1" != . ] >> do >> case "$1" in >> --java-home ) >> JAVA_HOME="$2" >> shift; shift; >> continue >> ;; >> >> >> This example actually illustrates the two main questions I had: >> >> 1. Why use [ ".$FOO" != . ] instead of simply [ -n "$FOO" ] (Corollary: >> why use [ ".$FOO" = . ] instead of [ -z "$FOO" ])? >> > > Because some shell scripts does dot handle -z or -n well. > > >> 2. Why have a "continue" at the end of every case option, since the >> whole body of the while loop is nothing but the case construct? >> > > That might be an extra directive, true. > Probably a leftover from multiple case directives in while loop. >
But safer than forgetting to add the continue if another case is ever added. > > >> I may be spoiled by using Linux and bash for most of my career, but I >> believe these are fairly standard POSIX-compliant things that should >> work on all *NIX systems. >> > > Sadly that's not the case IMHO. May I suggest a short comment is added to the script to document why -z and -n are not used? Someone else reading the script in the future is going to wonder the same. > > Regards > -- > ^TM > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org > For additional commands, e-mail: dev-h...@tomcat.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
