regex-codeblock calling subroutines?

Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question:

Is it possible to use a subroutine call from inside a regex-codeblock?

This is about separating regexes from application code by storing the regexes in separate files. I have this working nicely and usefully, so far no problem.

But now I am looking to move some code out of the application, more specifically I want to call (dbi-)subroutines from within the codeblock, and pass it a regex-captured value. Here I am running into trouble...

To do database lookups straight after matching seems such a logical step, that I expect others must have progressed further in this direction. Of course I will post complete code if anyone turns out to be interested, but because it needs several files I thought better start with a small, not-running example:

What follows is such a (above-mentioned) separate regex-file.


 (?{
    local $genus   = '';
    local $dbfound = 0;
 })

 ^

    (?:  ([A-Z][A-Z]+)  (?{
                            $genus   = ucfirst(lc($^N));
                            $dbfound = dbilookup($genus);
                        })
    )

 $    

 (?{
    #   
    #  further processing of values 
    #  (give back values to container program) 
 })
[download]

I'd be thankful for enlightenment

TIA, Eric

Comment on regex-codeblock calling subroutines? Download Code

Replies are listed 'Best First'.
Re: regex-codeblock calling subroutines? by dave_the_m (Monsignor) on Oct 15, 2004 at 15:01 UTC
Beware of calling functions that use regular expressions. The regex engine isn't fully re-entrant and strange things may happen. Dave.	[reply]
Re: regex-codeblock calling subroutines? by dragonchild (Archbishop) on Oct 15, 2004 at 15:14 UTC
Why do you want to do it as part of the regex? Why don't you split it out and do it within a while loop? Being right, does not endow the right to be rude; politeness costs nothing. Being unknowing, is not the same as being stupid. Expressing a contrary opinion, whether to the individual or the group, is more often a sign of deeper thought than of cantankerous belligerence. Do not mistake your goals as the only goals; your opinion as the only opinion; your confidence as correctness. Saying you know better is not the same as explaining you know better.	[reply]
Re^2: regex-codeblock calling subroutines? by Anonymous Monk on Oct 15, 2004 at 15:25 UTC
I suppose I could split it out, and will do that if all else fails. But then the functionality is in that one application. I think that if the dbi call can be made to work, then flexibility of the system would greater. I could then, with a certain text that a regexfile is to matched against, control the program flow depending on the outcome of certain found/not found queries. Also, many programs could use that lookup functionaliy.	[reply]
Re: regex-codeblock calling subroutines? by perlcapt (Pilgrim) on Oct 15, 2004 at 18:26 UTC
My understanding of match-time code evaluation is that the code-evaluation is not done in an order that is easy to predict. Regular expressions require that a regex-"engine" be built prior to the evaluation of the string. Secondly, the evaluation is not strictly and directly from left to right. This is more an event-driven code evaluation than a proceedural code evaluation. I.e., even if you can get this to run, it will be very difficult to maintain. We aren't byte/cycle miserly in Perl. There are other languages for that personality of programming. I like what you are trying to do, but don't think this is the design style that will yield the best design. Rebuttal is welcome.	[reply]
Re: regex-codeblock calling subroutines? by erix (Prior) on Oct 15, 2004 at 18:44 UTC
Hmm. Point(s) taken (to both david and capt). Thanks. I must admit that I am not sure if it will be a good idea, or even work at all. I was hoping to find monks who stumbled along the same path, and came out victorious or beaten, but wiser. It seems an attractive idea to have some db-backed 'directive' when crafting a regex. I'll just stumble on for a bit. (erix is 'Anonymous Monk' - lost cookie when I posted that initial question)	[reply]
Re^2: regex-codeblock calling subroutines? by diotalevi (Canon) on Oct 15, 2004 at 19:23 UTC
No really, don't. Any code you call from inside a (?{ ... }) absolutely, cannot use regular expressions at all. Unless you control every part of the code being run inside the expression, the code isn't safe. I expect that you're using DBI and a driver module to work with your database. That's all code that you don't control and that probably uses regular expressions. None of that can be called from within a (?{ ... }) block. Examine the pre-forking used in Regexp::Approx - Use fuzzy regular expressions to see how I fork during INIT and then run the unsafe code in a separate process. This is the only general method which allows you to do this sort of thing.	[reply]
Re^3: regex-codeblock calling subroutines? by erix (Prior) on Oct 15, 2004 at 19:31 UTC
Right, that settles it then. I did not realise the regex-inside-block aspect. I'll scurry back then, instead of stumbling on. And I will have a look at that module. Thanks, Eric	[reply]