> On Mon, Aug 08, 2011 at 08:56:36AM -0500, Dennis Williamson wrote:
>On Mon, Aug 8, 2011 at 3:47 AM, Roger <rogerx....@gmail.com> wrote:
>> I'm doing some research for one of my scripts and always liked the C style
>> ifdef inline debug statements.
...
>Another way to write the _debug() function:
>
>#!/bin/bash
>debug=true
>
># Function to optionally handle executing included debug statements
>_debug()
>{
> debug && "$@"
>}
>
>Using this method, you don't really need a function at all. You can
>just use the debug variable directly (you could use an underscore
>prefix):
>
> $debug && printf "Checking depedencies...\n"
>
> $debug && {
> # it's even more like ifdef since you can conditionally
>execute blocks of code
> foo
> bar
> }
>
>You have to make sure that $debug contains only "true" or "false" or
>you'll get errors. There are exceptions to this, but the complexity
>isn't worth the effort.
This is interesting because I'm guessing it might save one or two CPU
cycles, although it still does use a test at "&&", and is still
readable, if not more readable then before.
>I prefer using lowercase or mixed case variable names by habit to
>reduce the chances of name collisions with predefined variables
>(although that's not an issue with this specific script).
I know the system capitalized defined variables, so I usually just use
all capitalized variable for global script variables within the
header/top of the bash script. I've heard of the risk of collision,
and have tried mixed-case variable names, but they were not as readable
and the risk seemed more minimal ... unless a Linux distribution makes
up a system variable that conflicts with one of my all capitalized
variable names.
I then use lower case for locally used variables within the Bash script or
within the functions of the script. Similar to C. It just makes good
sense, and I know where to look for predefined capitalized variable names
for their definition.
>Since you're writing a script for Bash, you might as well use Bash
>features. Here is the main line of your function a couple of more
>different ways (using the original capitalization and value):
>
> [[ $DEBUG != 0 ]] && "$@" # string comparison
> (( DEBUG != 0 )) && "$@" # integer comparison (note,
>the $ is unnecessary)
Yup. I was thinking of sticking with 0's & 1's, as text comparison
requires more CPU cycles ... although negligible these days, it's just
good programming practice.
--
Roger
http://rogerx.freeshell.org/
