Regular expression to find accidental assignments in a conditional statement

Question

I've gotten bitten more than once by a bug where you accidentally assign a variable inside a conditional statement, like if.

I'm looking to grep my sources for such occurances. I created an expression that works in PHP for the simple case, but was wondering if anyone could optimize it or handle the more interesting cases.

Samples:

if ($var = 3)       //MATCH
if($var = 3)        //MATCH
if($var=3)          //MATCH
if ( $var = 3)      //MATCH
if ($var == $var2)  
if ($var = $var3)   //MATCH
if ( $var === 7)
if( $var == 8 || $var = 9) //MATCH
if (($var == 7 ) && ($var ==10))

The simple cases are handled well by if\s*$[\$a-zA-Z\d\s]*=[\$a-zA-Z\d\s]*$ but it would be nice to come up with something that works for the extended versions at the bottom of the sample.

Any ideas on better expressions?

@sputnick regex has an if: `(?(ifregex)thenregex|elseregex)` — Bohemian, Dec 09 '12 at 22:30
you'll also want to catch any other conditionals, which are all potential assignments. There's probably no way to catch everything unless you were parsing tokenized. — chris, Dec 09 '12 at 22:31
You could just check if the whole line contains both `if` and `==` - assuming the blocks are line-wrapped — Bergi, Dec 09 '12 at 23:14

score 4 · Answer 1 · edited May 23 '17 at 11:45

4

Regular expressions are not the best tool to parse code.

For most languages there are linting tools that check your code. I have no experience with PHP but you'll find plenty of pointers here: Is there a static code analyzer [like Lint] for PHP files?

edited May 23 '17 at 11:45

Community

1
1

answered Dec 09 '12 at 22:36

akuhn

27,477
2
76
91

Hey thanks for that link, i'll look at some of those tools. I was looking for something quick and dirty, but once one of those tools is up and running, it should give more accurate results. – akatakritos Dec 09 '12 at 23:22

Shahbaz · Accepted Answer · 2012-12-09T23:31:08.387

3

First, let's assume you don't have cases like this (C example):

if ((a = b) == c)

because understanding these cases needs an actual parser.

Now, simply put, you want to match = but not ==\+. Therefore:

if\([^=]\|==\+\)*[^=]=[^=]\([^=]\|==\+\)*$

What it says:

if: match the first if.
[^=]: any character but =
==\+: == and ===
$[^=]\|==\+$*: anything that is not = or is == or ===. This includes all the whitespace, the beginning (, ending ) etc. This comes both before and after the isolated =.
[^=]=[^=]: isolated =.

edited Dec 09 '12 at 23:31

answered Dec 09 '12 at 22:53

Shahbaz

46,337
19
116
182

Awesome. I tried it on my project and it didnt find anything. So I guess that means I wrote good code this time. Ha! Cheers! – akatakritos Dec 09 '12 at 23:23
@akatakritos, I'm not a php developer, but I would have expected whatever compiler/interpreter reads that code should give you warnings, doesn't it? – Shahbaz Dec 09 '12 at 23:32
Accidental assignment, AFAIK is not any sort of warning in PHP. I miss C# where its a compile error. :( – akatakritos Dec 10 '12 at 00:20

Regular expression to find accidental assignments in a conditional statement

2 Answers2