Regular Expressions

coreys · #1 12-18-2009, 08:25 AM

Every once in awhile I run across a situation where I wish I could use regular expressions in GS2. Since this is no trivial task to implement myself, not to mention it would most likely be slow if not done at a language level, it would be really nice if regular expressions were added into GS2.

Programmer · #2 12-18-2009, 01:19 PM

Quote:

Originally Posted by coreys

Every once in awhile I run across a situation where I wish I could use regular expressions in GS2. Since this is no trivial task to implement myself, not to mention it would most likely be slow if not done at a language level, it would be really nice if regular expressions were added into GS2.

I completely agree. I have also come across several situations where Regex would make my life so much easier. Seeing as how several libraries offer this in C++, it should not be so difficult to implement in GScript at all.

Immolate · #3 12-18-2009, 01:21 PM

Yeah, I support this.

coreys · #4 12-23-2009, 09:02 PM

Bump!
There's a couple things I hope to use regex for that would make my life so much better.
Better living through regular expressions!

fowlplay4 · #5 12-23-2009, 09:36 PM

Could you show some theory usage of how it would work? I haven't touched regex before.

Crow · #6 12-23-2009, 09:44 PM

Lua syntax, it's from one of my WoW addons:

PHP Code:


			
msg = gsub(msg, "^|Hchannel:(%S+)|h%[([%d%.%s]*)([^%]]+)%]|h ", "|Hchannel:%1|h" .. channelFormat .. "|h ");

Basically, this replaces the display format of the channel escape sequence (WoW stuff) at the start of the string msg (if any escape sequence exists) with a predefined new format stored in channelFormat.
WoW channel escape sequences work like this:

PHP Code:


			
|Hchannel:CHANNEL_HERE|hCHANNEL_DISPLAY_NAME_HERE|h

If you want to know some more details, either ask, or check out patterns on the Lua website. Also, as far as I know, Lua implements regular expressions a tad differently compared to other languages, so you might also want to check out the PHP documentation.

WhiteDragon · #7 12-23-2009, 09:44 PM

I've never needed regex for coding anything in Graal. Not to mention they can lead to some seriously inefficient code if you don't know what you are doing (e.g., when checking against a non-regular language, unlimited backreferences are an NP-complete problem).

DustyPorViva · #8 12-23-2009, 09:50 PM

I've ran into a couple situations where I really wish I had it. I think getstringkeys() is probably the most common.

coreys · #9 12-23-2009, 09:54 PM

I'm familiar with regular expressions as they're used in Python. They're a tad bit complicated to learn, but very useful for searching for patterns.

And inefficient code is always an issue, no matter what you do, so it's not really a reason to not support this. I personally could find good use for regex.

WhiteDragon · #10 12-23-2009, 10:12 PM

Quote:

Originally Posted by coreys

And inefficient code is always an issue, no matter what you do, so it's not really a reason to not support this. I personally could find good use for regex.

I'm just saying that regex is particularly complex and therefore more prone to inefficiencies.

coreys · #11 12-23-2009, 10:44 PM

Quote:

Originally Posted by WhiteDragon

I'm just saying that regex is particularly complex and therefore more prone to inefficiencies.

Yeah, but user error is user error, no reason not to have it

Loriel · #12 12-24-2009, 12:12 AM

Pft, who needs regexps when you could instead build some awesome parser combinator system.

WhiteDragon · #13 12-24-2009, 12:28 AM

Quote:

Originally Posted by Loriel

Pft, who needs regexps when you could instead build some awesome parser combinator system.

Whoever implements parsec first wins a cookie!

Loriel · #14 01-16-2010, 12:00 AM

Quote:

Originally Posted by Loriel

Pft, who needs regexps when you could instead build some awesome parser combinator system.

Bumping this too!!

cbk1994 · #15 01-15-2010, 11:07 PM

Bumping this because I just ran into a place where this would be perfect.

coreys · #16 01-16-2010, 11:19 PM

Pl0x pl0x pl0x

coreys · #17 01-22-2010, 10:26 AM

bump

further reasoning:

We can become graal superheros if we get regex.

fowlplay4 · #18 01-22-2010, 03:52 PM

thats so lame.

Imperialistic · #19 01-22-2010, 04:43 PM

ya, im going to pretend that i never saw that.

Liberated · #20 01-22-2010, 04:47 PM

That's so funny.

Loriel · #21 01-22-2010, 07:43 PM

Regular expressions are goddamn retarded, here is the perl regexp for matching an email address

NPC Code:



(?:(?:\r\n)?[ \t])*(?:(?:(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t]

)+|\Z|(?=[\["()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:

\r\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(

?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ 

\t]))*"(?:(?:\r\n)?[ \t])*))*@(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\0

31]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\

](?:(?:\r\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+

(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:

(?:\r\n)?[ \t])*))*|(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z

|(?=[\["()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)

?[ \t])*)*\<(?:(?:\r\n)?[ \t])*(?:@(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\

r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[

 \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)

?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t]

)*))*(?:,@(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[

 \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*

)(?:\.(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t]

)+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*))*)

*:(?:(?:\r\n)?[ \t])*)?(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+

|\Z|(?=[\["()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r

\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:

\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t

]))*"(?:(?:\r\n)?[ \t])*))*@(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031

]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](

?:(?:\r\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?

:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?

:\r\n)?[ \t])*))*\>(?:(?:\r\n)?[ \t])*)|(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?

:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?

[ \t]))*"(?:(?:\r\n)?[ \t])*)*:(?:(?:\r\n)?[ \t])*(?:(?:(?:[^()<>@,;:\\".\[\] 

\000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|

\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?:[^()<>

@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|"

(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)?[ \t])*))*@(?:(?:\r\n)?[ \t]

)*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\

".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?

:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[

\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*))*|(?:[^()<>@,;:\\".\[\] \000-

\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|\\.|(

?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)?[ \t])*)*\<(?:(?:\r\n)?[ \t])*(?:@(?:[^()<>@,;

:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([

^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\"

.\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\

]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*))*(?:,@(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\

[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\

r\\]|\\.)*\](?:(?:\r\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] 

\000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]

|\\.)*\](?:(?:\r\n)?[ \t])*))*)*:(?:(?:\r\n)?[ \t])*)?(?:[^()<>@,;:\\".\[\] \0

00-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|\\

.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?:[^()<>@,

;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|"(?

:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)?[ \t])*))*@(?:(?:\r\n)?[ \t])*

(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".

\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t])*(?:[

^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\]

]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*))*\>(?:(?:\r\n)?[ \t])*)(?:,\s*(

?:(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\

".\[\]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)?[ \t])*)(?:\.(?:(

?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[

\["()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)?[ \t

])*))*@(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t

])+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*)(?

:\.(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|

\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*))*|(?:

[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".\[\

]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)?[ \t])*)*\<(?:(?:\r\n)

?[ \t])*(?:@(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["

()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*)(?:\.(?:(?:\r\n)

?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>

@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*))*(?:,@(?:(?:\r\n)?[

 \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,

;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*)(?:\.(?:(?:\r\n)?[ \t]

)*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\

".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*))*)*:(?:(?:\r\n)?[ \t])*)?

(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\["()<>@,;:\\".

\[\]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)?[ \t])*)(?:\.(?:(?:

\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z|(?=[\[

"()<>@,;:\\".\[\]]))|"(?:[^\"\r\\]|\\.|(?:(?:\r\n)?[ \t]))*"(?:(?:\r\n)?[ \t])

*))*@(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])

+|\Z|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*)(?:\

.(?:(?:\r\n)?[ \t])*(?:[^()<>@,;:\\".\[\] \000-\031]+(?:(?:(?:\r\n)?[ \t])+|\Z

|(?=[\["()<>@,;:\\".\[\]]))|\[([^\[\]\r\\]|\\.)*\](?:(?:\r\n)?[ \t])*))*\>(?:(

?:\r\n)?[ \t])*))*)?;\s*)

i am not really sure this is what we need to deliver a better gameplay experience in graal

DustyPorViva · #22 01-22-2010, 08:35 PM

Quote:

Originally Posted by Loriel

Regular expressions are goddamn retarded, here is the perl regexp for matching an email address

...

i am not really sure this is what we need to deliver a better gameplay experience in graal

That is such an extreme example and you know it :P In fact, I've never seen a regex expression that long before.

coreys · #23 01-22-2010, 09:31 PM

Quote:

Originally Posted by DustyPorViva

That is such an extreme example and you know it :P In fact, I've never seen a regex expression that long before.

Neither have I.

Also, what? I've seen much shorter regular expressions to match an email address, you must be mistaken.

Edit:
I just looked it up, you are way off. The regular expression to find an email address (may vary slightly between implementations) is:

NPC Code:



\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}\b

In this instance using grep

WhiteDragon · #24 01-22-2010, 09:40 PM

Quote:

Originally Posted by coreys

Neither have I.

Also, what? I've seen much shorter regular expressions to match an email address, you must be mistaken.

Edit:
I just looked it up, you are way off. The regular expression to find an email address (may vary slightly between implementations) is:

NPC Code:



\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}\b

In this instance using grep

That technically doesn't match the RFC 2822 specifications for an e-mail address.

Anyways, like I said, 90% of people who use regex don't actually understand what the hell it does or how it works, even if they understand the syntax.

This results in a lot of copying-and-pasting into your codebase from other places, and therefore a lot of potential edge-cases where your script could grind to a halt.

Loriel · #25 01-22-2010, 10:09 PM

Quote:

Originally Posted by coreys

Neither have I.

Also, what? I've seen much shorter regular expressions to match an email address, you must be mistaken.

Edit:
I just looked it up, you are way off. The regular expression to find an email address (may vary slightly between implementations) is:

NPC Code:



\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}\b

In this instance using grep

that does not even match the .museum tld

WhiteDragon · #26 01-22-2010, 07:58 PM

If this gets implemented I better get my enhancements on anonymous functions.

xXziroXx · #27 01-22-2010, 09:44 PM

I don't see how we need regex at all.

coreys · #28 01-22-2010, 09:44 PM

Who cares? Just because some people don't understand how to use it doesn't make it any less applicable for those that do, especially when it can be extremely useful. There's a reason many languages have regular expressions in their standard library (such as Perl, Python, etc), and most that don't have extremely popular libraries that provide the functionality, like C++'s Boost.Regex.

WhiteDragon · #29 01-22-2010, 10:38 PM

Quote:

Originally Posted by coreys

Who cares?

I do.

I don't want any more codebases that I work on to get splattered with crazy stuff that no one understands except the people who need to come in and waste their time fixing it later.

Quote:

"omg why doesn't this code run faster"
"That is because the GS2 implementation of regex uses the backtracking algorithm which, while allowing for more expressive patterns, is not as efficient as other algorithms that construct a NFA to a DFA."
"wat??????? can u fix it"
"!@#$"

I seriously have not ever had a use for regex in my whole time scripting GS2 and I don't want any of these potential situations, so I'd prefer for it to not be implemented.

Skyld · #30 01-22-2010, 10:30 PM

No. Regular expressions are a fantastic way to make code harder to edit and understand (a lot of scripters manage to make things terribly overcomplicated as it is without them!), especially for newer people who will take one look at a regular expression and not know what it is, little less what to do with it. Let's focus on functions that make gscript easier, perhaps.

coreys · #31 01-22-2010, 11:14 PM

Quote:

Originally Posted by Skyld

No. Regular expressions are a fantastic way to make code harder to edit and understand (a lot of scripters manage to make things terribly overcomplicated as it is without them!), especially for newer people who will take one look at a regular expression and not know what it is, little less what to do with it. Let's focus on functions that make gscript easier, perhaps.

God forbid we have a useful feature that not everybody can use. God forbid people who don't know how to use something...oh, I don't know, not use it.

I have at least one useful feature in mind where regular expressions would be invaluable, and I'm not the only one. So basically, we can **** off because there are a lot of idiots who want to script on Graal? Sounds reasonable.

fowlplay4 · #32 01-23-2010, 01:11 AM

Quote:

Originally Posted by coreys

God forbid we have a useful feature that not everybody can use. God forbid people who don't know how to use something...oh, I don't know, not use it.

I have at least one useful feature in mind where regular expressions would be invaluable, and I'm not the only one. So basically, we can **** off because there are a lot of idiots who want to script on Graal? Sounds reasonable.

GScript2 is a game scripting language though, and it's not really needed at all in game server development.

The only way I can really see this getting implemented is if Stefan finds that he needs it one day, until then it'll be low on priorities as far as adding features/functions to GScript2 goes.

cbk1994 · #33 01-22-2010, 11:34 PM

This is such a silly debate. People will have trouble with any feature they don't understand, it's no reason to not add something. From what others have said, it doesn't sound like it would be hard to add, either.

Inverness · #34 01-22-2010, 11:36 PM

Regex is in a whole other class of difficulty, imho. I've considered learning regex before but I'm intimidated and haven't encountered a need to learn it yet.

DustyPorViva · #35 01-23-2010, 12:34 AM

The only time I think I've ever needed regex was with getstringkeys(). Those instances are few and far between though, and I don't even remember them. Would I have loved to be able to use regex in the few times I actually needed it? Hell yes. Do I really think it's needed? Not really. It's just not something I see being used too much to justify the effort of adding it. Now that can be said about a lot of things, but I feel this is one of those additions that everyone gets all giddy over and then they get it and don't know what the **** they actually want to do with it, and forgets about it. However, if there were any regex added, a very simple form of it would suffice. At least, if not regex then a rip-off wildcard system that format uses(%s, %d, and so on). Anything more complicated would definitely not be used in my opinion.

coreys · #36 01-23-2010, 03:25 AM

Quote:

Originally Posted by DustyPorViva

However, if there were any regex added, a very simple form of it would suffice. At least, if not regex then a rip-off wildcard system that format uses(%s, %d, and so on). Anything more complicated would definitely not be used in my opinion.

That alone would make me an extremely happy scripter.