Almost every day I learn about more little devils that might slip into my code details. For example, I'm slowly getting used to not doing:
but instead doing:
open($fh, '<', $filename) or die "can't open $filename: $!"
Either approach is semantically correct, and there's more than one way to do things, but the second one is better practice because:
- The file handle $fh is lexically scoped, preventing weird action-at-a-distance.
- The mode is explicitly set, preventing unintended operations, i.e. overwriting $filename.
- open is tested, and errors returned by the file system - i.e. $! - are returned.
I am sure there's an endless number of other things that I'm still doing "wrong" because I don't know about them - although my programs execute just fine.
Wouldn't it be a good idea to erect a place on perlmonks where those issues could be discussed in the form of mini code reviews? For example, someone has written a few blocks of code < $maxlength, where $maxlength is something manageable, like 100 lines. The code executes just fine, but the coder is inspired by TheDamian
's book on Perl best practices
and would like further nitpicking for the edification of all and sundry. Other monks move in, and crush submitter's spirit with maddeningly anal comments.
(I am reminded here of UWashington's ecology department's lunch seminar series Eco-Lunch
, where willing grad students present research and field questions. The series is internally known as the Ego-Crunch
The idea is that the code already works - i.e. not SoPW material -, is very much NOT obfuscated and might even be very UNCOOL, i.e. accountancy software.
Your comments please!