Syntax checking with PHP

freek1

Here's a brainbuster, for anybody who likes a challenge. Although I am not exactly an experienced PHP coder, I do understand PHP and I learn really fast so now I am working on a small PHP script that verifies the syntax of some plain text files I'm using with another application. The basic syntax is:

if
endif

At this point, the script reads the file, breaks it into lines (so I can get the line number and get a detailed report about which line has an error - if any) and then goes further breaking the lines into words. If I have a file that has this code in it:

if

if
endif

and I run it through the script, it will print out an error saying there's an ENDIF missing on line 3 (I know it should say line 2 but like any other syntax checker, the script only spots an error when it runs into something that isn't right). I achieved this by setting an "endif_required" flag whenever an IF is detected so if an ENDIF is NOT detected before another IF, the error is reported. Simple, easy and works like a charm. But here comes the part that's driving me nuts: SOME of the files I'm supposed to check may contain nested events like this:

if
        if
        endif
endif

Of course, the script will print out an error saying there's an ENDIF missing on line 2 which is WRONG. Now, I could easily get rid of this by simply counting the IFs and ENDIFs and compare the total count, but then I wouldn't have the line number report. So my question is (although probably more related to logic than PHP): how should I rewrite/re-think the script so it won't print out an error in case of nested events AND keep the line number report ?

Any help is more than welcome and I thank you in advance !

laserlight

Could try using a stack.

freek1

laserlight wrote:
Could try using a stack.

I could if I only knew how 🙂 As I said, I'm not really an experienced PHP coder so could you please be more specific ?

Thanks !

laserlight

I could if I only knew how As I said, I'm not really an experienced PHP coder so could you please be more specific ?

This has little to do with PHP, and more to do with computer science.

When an "if" is encountered, push the line number to the stack. When an "endif" is encountered, pop the most recent "if" from the stack. If the stack is empty, then there is an unmatched "endif". If at the end of all the parsing the stack is not empty, then what remains in the stack are the line numbers of the unmatched "if"s.

In PHP, arrays can be used as stacks, especially since there are the [man]array_pop/man and [man]array_push/man functions. [man]empty/man could also come in handy.

freek1

This has little to do with PHP, and more to do with computer science.

Yes, that's why I said that it is more about the logic rather than PHP 🙂

So now after reading your last post, I understand what you're saying. Basically every last IF found is "neutralised" by the first ENDIF found, right ? Very good thinking but I think there's a tiny flaw in it. Let's take this code for example:

0-   if
1-           if
2-                   if
3-                   endif
4-   

5-   endif

I indented the lines so it's easier to follow but the files I'm working with may or may NOT be indented - in case you were thinking about using the indentation (tabs or spaces) to keep track of the nesting levels.

So, as you can see, there is an ENDIF missing on line 4 (that should close the IF opened on line 1). So here's how the script would read this:

IF started on line 2 gets "neutralized" by ENDIF found on line 3
IF started on line 1 gets "neutralized" by ENDIF found on line 5
IF started on line 0 will be reported as an error and that is WRONG.

It was a good try tho, I really appreciate your help !

laserlight

IF started on line 0 will be reported as an error and that is WRONG.

That's correct, unless you are working with a language for which indentation is significant, like Python.

Basically, you have to make up your mind: do you want to ignore indentation, or do you want to take it as strictly part of the syntax?

Consider a half-indented example:

if
	if
	if
	endif
	endif
endif

Is there a syntax error, or not?

What language are you working with, anyway?

freek1

Basically, you have to make up your mind: do you want to ignore indentation, or do you want to take it as strictly part of the syntax?

Sorry for the confusion :o I want to completely ignore the indentation.

Is there a syntax error, or not?

Nope, there is no syntax error in your example.

What language are you working with, anyway?

It's called WAC and it is used to program AI movement in (some) Novalogic games.

laserlight

I want to completely ignore the indentation.

In that case the "tiny flaw" is not a flaw at all.

There is another possibility though. You can record the indent level of the "if", and only pop if it matches the indent level of the "endif". If that is so, then when the "endif" on line 5 is encountered, it would notice that the indent level is different from the indent level of the "if" on line 2.

freek1

laserlight wrote:
In that case the "tiny flaw" is not a flaw at all.

I guess you're right since a pre-programmed script can't really tell the difference if the lines are not indented.

laserlight wrote:
You can record the indent level of the "if" ...

True, I can, but since indentation is not mandatory with this language, a script that relies on indentation to work properly wouldn't do much of a job when it has to check a file that is not indented :queasy: hence my problem.

Thanks for baring with me !