You think you can write regex?

[James]

This is a regex to match every single word in the English Dictionary.

Crazy.

[Jon]

Aww, it doesn't like it.

#include <Constants.au3>


$result = InetGet("https://gist.github.com/noprompt/6106573/raw/fcb683834bb2e171618ca91bf0b234014b5b957d/word-re.clj", @TempDir & "\regexp.txt")

If Not $result Then
    MsgBox($MB_SYSTEMMODAL, "Error", "Unable to download regexp")
EndIf

$regexp = FileRead(@TempDir & "\regexp.txt")

While 1
    $word = InputBox("Enter word", "Enter word", "neccessary")
    If @error Then Exit


    $result = StringRegExp($word, $regexp)

    If $result = 1 Then
        MsgBox($MB_SYSTEMMODAL, "Result", "Matched" & @CRLF & "@Error: " & @error & @CRLF & "@Extended:" & @extended)
    Else
        MsgBox($MB_SYSTEMMODAL, "Result", "Not Matched" & @CRLF & "@Error: " & @error & @CRLF & "@Extended:" & @extended)
    EndIf

WEnd

[James]

You spelt "necessary" wrong

It's so big that I can't even test it anywhere else.

Edited August 12, 2013 by James

[JohnOne]

#include <Constants.au3>

Local $result = 0

If Not FileExists(@TempDir & "\regexp.txt") Then
    $result = InetGet("https://gist.github.com/noprompt/6106573/raw/fcb683834bb2e171618ca91bf0b234014b5b957d/word-re.clj", @TempDir & "\regexp.txt")
Else
    $result = 1
EndIf

If Not $result Then
    MsgBox($MB_SYSTEMMODAL, "Error", "Unable to download regexp")
EndIf

$regexp = FileRead(@TempDir & "\regexp.txt")

While 1
    $word = InputBox("Enter word", "Enter word", "")
    If @error Then Exit


    $result = StringRegExp($word, $regexp)

    If $result = 1 Then
        MsgBox($MB_SYSTEMMODAL, "Result", "Matched" & @CRLF & "@Error: " & @error & @CRLF & "@Extended:" & @extended)
    Else
        MsgBox($MB_SYSTEMMODAL, "Result", "Not Matched" & @CRLF & "@Error: " & @error & @CRLF & "@Extended:" & @extended)
    EndIf

WEnd

Doesn't work for any word here.

EDIT:

neither does the other

$regexp = InetRead("https://gist.github.com/noprompt/6106573/raw/fcb683834bb2e171618ca91bf0b234014b5b957d/word-re.clj")

While 1
    $word = InputBox("Enter word", "Enter word", "")
    If @error Then Exit


    $result = StringRegExp($word, $regexp)

    If $result = 1 Then
        MsgBox($MB_SYSTEMMODAL, "Result", "Matched" & @CRLF & "@Error: " & @error & @CRLF & "@Extended:" & @extended)
    Else
        MsgBox($MB_SYSTEMMODAL, "Result", "Not Matched" & @CRLF & "@Error: " & @error & @CRLF & "@Extended:" & @extended)
    EndIf

WEnd

Edited August 12, 2013 by JohnOne

[guinness]

It's 1.52MB when saved to a file. I have noticed that with PCRE it doesn't like having large optional groups.

[Jon]

It's not going to work while @error = 2, that means that it thinks there is a problem with the pattern - doesn't even get as far as checking the word.

[Edano]

now come on, boys and girls, this is a hoax. simply not possible.

[guinness]

now come on, boys and girls, this is a hoax. simply not possible.

Care to elaborate?

[Edano]

a stringregexpr for every single english word would assume that all words follow a pattern. this is untrue with a very high probability.

[James]

a stringregexpr for every single english word would assume that all words follow a pattern. this is untrue with a very high probability.

You're an idiot.

It's not going to work while @error = 2, that means that it thinks there is a problem with the pattern - doesn't even get as far as checking the word.

Yeah, I finally managed to test it and it didn't seem to work there. It was generated with something called frak.

Edited August 12, 2013 by James

[Edano]

and you are a prole, James.

[guinness]

It was generated with something called frak.

Now the pieces seems to fit into place. I mean that's some massive regular expression to not only create but debug too.

[FireFox]

It was generated with something called frak.

Nice tool.

[James]

Nice tool.

Yeah, hopefully the example will be fixed. I'll raise an issue, asking for confirmation that it does in fact work.

[guinness]

Yeah, hopefully the example will be fixed. I'll raise an issue, asking for confirmation that it does in fact work.

...and what engine he used.

[James]

...and what engine he used.

Done.

[jchd]

I suspect it's Perl. PCRE has many compile-time adjustable limits. I've asked on the PCRE list what limits is hit during compilation of this uncommonly giant RE.

BTW the topic title is wrong: the guy used a tool to "write" the RE! So yes, we all can actually write regexps and useful ones

Edited August 12, 2013 by jchd

[James]

BTW the topic title is wrong: the guy used a tool to "write" the RE! So yes, we all can actually write regexps and useful ones

He must be able to write a good regex to write a tool which generates them though

[czardas]

Okay James, let's see you come up with an algorithm that can generate every word in the English dictionary.

Edited August 12, 2013 by czardas

[JLogan3o13]

Here's my algorithm for that...

http://www.thefreedictionary.com/unabridged+dictionary

Edited August 12, 2013 by JLogan3o13