Question about Regex...

[JScript]

Hello guys!

I have a doubt that for many can be quite simple to solve, but unfortunately I do not understand almost nothing of regex!

I wanted to know how to identify only the character _ at the end of the line indicating that it continues in the next line...

Example:

MsgBox(4096, "Title _ v1", "This program is free software: you can redistribute it and/or modify" & _
        'it under the terms of the GNU General Public License " _"' & _ ; Only comment!
        "as published by the Free Software Foundation, either version 3" & _
        "of the License, or (at your option) any later version.")

If anyone can help me I'll be very grateful!

Edit: Another example:

$aRtl = DllCall("ntdll.dll", "int", "RtlCompressBuffer", _
        "ushort", 2, _ ;------> CompressFormat
        "ptr", $pInput, _ ;---> ptrSrceBuffer
        "dword", $iSize, _ ;--> SrceBufferSize
        "ptr", $pBuffer, _ ;--> ptrDestBuffer
        "dword", $iBufLen, _ ;> DestBufferSize
        "dword", 4096, _ ;----> ChunkSize
        "dword*", 0, _ ;------> CompressedSize
        "ptr", $pWrkSp) ;-----> WorkspaceBuffer

thankful,

JS

Edited October 31, 2012 by JScript

[BrewManNH]

Look for a "_" followed by a CRLF.

[JScript]

And in this case:

_ ; Only comment!

??

JS

[BrewManNH]

Two vastly different questions, if you have alternate requirements, then you should have stated them up front.

You can use the pipe as an OR statement to see if the underscore is followed by a space and either a CRLF or a semi-colon.

[iamtheky]

why not just regexp for "& _"

$Aunderscores = StringRegExp ($string , "& _", 3)

[czardas]

This isn't quite working and I don't know why. I decided to post my attemps anyway, in case it turns out to be useful and that I might also learn something.

#include <Array.au3>

$sString = "This program is free software: you can redistribute it and/or modify & _" & @CRLF & _
        'it under the terms of the GNU General Public License " _" & _ ; Only comment!' & @CRLF & _ ; Only comment!
        "as published by the Free Software Foundation, either version 3 & _" & @CRLF & _
        "of the License, or (at your option) any later version."

$regExp = "(&h+_h*(;.*)?)"

$aArray = StringRegExp($sString, $regExp, 3)

_ArrayDisplay($aArray)

Anyway I believe it is customary to strip away all comments before parsing. I would just concatenate strings ending with underscore using StringReplace.

Edit

I believe the above RegExp approach is not suitable in most circumstances. It is inefficient and my example (besides being broken) does not cater for matches found within strings. There may be simpler ways than using RegExp, depending on what you want to achieve.

Edited October 31, 2012 by czardas

[Bert]

have you thought of trying the RegEx tool that is part of the help file? Look for the Tutorial - Regular Expression in the help file and at the bottom is the script button for the tool.

[JScript]

Two vastly different questions, if you have alternate requirements, then you should have stated them up front.

I guess you did not pay attention to the example I posted!

MsgBox(4096, "Title _ v1", "This program is free software: you can redistribute it and/or modify" & _
        'it under the terms of the GNU General Public License " _"' & _ ; Only comment!
        "as published by the Free Software Foundation, either version 3" & _
        "of the License, or (at your option) any later version.")

You would have seen this sequence: " _"' & _ ; Only comment!

And would not have given this answer:

Look for a "_" followed by a CRLF.

@boththose

Sorry, but sometimes can be as follows: , _

@czardas

Excellent, but as I said above, sometimes can be as follows: , _

@scullion

That phrase / advice was for me? If yes, thank you!

JS

Edited October 31, 2012 by JScript

[PhoenixXL]

This might help

Func MatchSingleScores($nString)
Local $nPattern = "['" & '"]' & ".*" & "['" & '"]'

;~ ConsoleWrite($nPattern & @CR)

;Remove String enclosed in Quotes
Local $nRegEx = StringRegExpReplace($nString, $nPattern, '')
If @error Then Exit @extended

;Remove Comments
$nPattern = ";.*"
$nRegEx = StringRegExpReplace($nRegEx, $nPattern, '')

;Return Match
Return StringRegExp($nRegEx,' _')
EndFunc ;==>ReplaceScores

If MatchSingleScores(' "Hello there" _ ;This is a comment' & @CRLF _
& ' "This is the Second Line"') Then Exit ConsoleWrite('Underscore is There!'&@CR)=0

Exit ConsoleWrite('No Match Found')=0

Edited October 31, 2012 by PhoenixXL

[Varian]

This captures each line that ends with an underscore or ends with an underscore & a comment; This returns the entire matched line up to the underscore

#include <Array.au3>

Local $iString
$iString &= 'MsgBox(4096, "Title _ v1", "This program is free software: you can redistribute it and/or modify" & _' & @LF
$iString &= ' "it under the terms of the GNU General Public License " _" & _ ; Only comment!' & @LF
$iString &= ' "as published by the Free Software Foundation, either version 3" & _' & @LF
$iString &= ' "of the License, or (at your option) any later version.")' & @LF
$iString &= '$aRtl = DllCall("ntdll.dll", "int", "RtlCompressBuffer", _' & @LF
$iString &= ' "ushort", 2, _ ;------> CompressFormat' & @LF
$iString &= ' "ptr", $pInput, _ ;---> ptrSrceBuffer' & @LF
$iString &= ' "dword", $iSize, _ ;--> SrceBufferSize' & @LF
$iString &= ' "ptr", $pBuffer, _ ;--> ptrDestBuffer' & @LF
$iString &= ' "dword", $iBufLen, _ ;> DestBufferSize' & @LF
$iString &= ' "dword", 4096, _ ;----> ChunkSize' & @LF
$iString &= ' "dword*", 0, _ ;------> CompressedSize' & @LF
$iString &= ' "ptr", $pWrkSp) ;-----> WorkspaceBuffer' & @LF
$iString &= '_ ; Only comment!'
MsgBox(262208, 'Original String', $iString)
RegEx($iString)

Func RegEx($nString)
Local $nPattern = "(?m)(^.*_)(?:s*$|s*;.*s*$)"
ConsoleWrite($nPattern & @CR)
Local $nRegex = StringRegExp($nString, $nPattern, 3)
If @error Then Exit @extended
_ArrayDisplay($nRegex)
EndFunc ;==>RegEx

Edited October 31, 2012 by Varian

[jchd]

Detecting AutoIt statement continuation requires parsing from the left hand and implies actual parsing of the syntax.

Ad hoc patterns can do it in a restricted number of cases, but certainly not in all cases.

I don't believe a simple regexp can be produced to this effect and, should one ever pop up, it would be fairly complex and unmaintainable.

This is a job for a lexer.

[JScript]

@PhoenixXL

Very nice, thank you!!!

@Varian

Of all the tests I did was his example was what went right

but as was said in the post above, I believe that may fail in some cases, although it has not failed with me!

@jchd

You hit the point...

I thank everyone for their efforts in helping me!

JS

Edited October 31, 2012 by JScript

[BrewManNH]

identify only the character _ at the end of the line

From your description of what you were looking for in the first post, and me not looking at the whole string, you were asking how to identify the underscore at the end of the line. Having a further requirement to identify a comment at the end of the line with an underscore before it isn't in the description of what you were looking for. As I said, explain the full objective upfront and it won't require you having to refine it as the thread continues. A line with an underscore at the end IS vastly different from a continued line with a comment at the end.Just my 2 cents.

[JScript]

No problem, I'm sorry if I have not explained well!

JS

[PhoenixXL]

modified a bit varian's code

-->Now Replaces the Underscores with the Next Line

Local $iString
$iString = 'DllCall("ntdll.dll", "int", "RtlCompressBuffer", _' & @LF & _
'"ushort", 2, _ ;------> CompressFormat' & @LF & _
'"ptr", $pInput, _ ;---> ptrSrceBuffer' & @LF & _
'"dword", $iSize, _ ;--> SrceBufferSize' & @LF & _
'"ptr", $pBuffer, _ ;--> ptrDestBuffer' & @LF & _
'"dword", $iBufLen, _ ;> DestBufferSize' & @LF & _
'"dword", 4096, _ ;----> ChunkSize' & @LF & _
'"dword*", 0, _ ;------> CompressedSize' & @LF & _
'"ptr", $pWrkSp) ;-----> WorkspaceBuffer'

ConsoleWrite('Returned: ' & RegEx($iString) & @CR)

;Replaces the Underscores with the NextLine
Func RegEx($sString)
Local $sPattern = "(?m)(^.*)( _)(?:s*$|s*;.*s*$)"
;Remove the Comments
Local $sRegex = StringRegExpReplace($sString, $sPattern, '1')
;Now Lets Join them
Return StringRegExpReplace($sRegex, '(?m)([rn])', '')
EndFunc   ;==>RegEx

[jchd]

As I said, that simple-minded approach fails if the substring ' _' is present in either a literal string in the active statement part or in a mono-line comment.

The general case needs a lexer. Period.

[PhoenixXL]

As I said, that simple-minded approach fails if the substring ' _' is present in either a literal string in the active statement part or in a mono-line comment.

Can you please give a correct Autoit Syntax where this fails

Thanks for your time

[jchd]

Here you are:

Local $sString
$sString = '$s = "This a short example."' & @CRLF & _
'ConsoleWrite($s & @LF)'
ConsoleWrite('Returned: ' & RegEx($sString) & @CR)

$sString = '$s = "Another example." ; with a short comment _' & @CRLF & _
'ConsoleWrite($s & @LF)'
ConsoleWrite('Returned: ' & RegEx($sString) & @CR)

$sString = '$s = "Yet another example _ ; tada."' & @CRLF & _
'ConsoleWrite($s & @LF)'
ConsoleWrite('Returned: ' & RegEx($sString) & @CR)

;Replaces the Underscores with the NextLine
Func RegEx($sString)
Local $sPattern = "(?m)(^.*)( _)(?:s*$|s*;.*s*$)"
;Remove the Comments
Local $sRegex = StringRegExpReplace($sString, $sPattern, '1')
;Now Lets Join them
Return StringRegExpReplace($sRegex, '(?m)([rn])', '')
EndFunc ;==>RegEx

[PhoenixXL]

right its very complex

[jchd]

Please don't take it bad.

The issue here is that we can have a number of cases, the most delicate being that we need to decide if we're in a string literal, a true continuation operator or inside a comment. We may encounter the ' _' sequence everywhere, or not at all.

To make things harder, a string can use single- or double-quotes delimiters ('abc' = "abc"), yet include double- or single-quotes respectively, or escaped (doubled) quote delimiter, non exclusively ("Mac'Do" = 'Mac''Do', 'Hi "folks"' = "Hi ""folks"""). That only by itself is far from completely trivial. But realize that what looks like a valid string literal with any quoting might in fact be inside a comment, as a line by itself or after a partial or final statement.

The only way out is to parse the syntax from the left hand and that requires a gross-grain parser. I fully acknowledge this _can_ in theory be done in a single PCRE regexp (after all PCRE is as close to a Turing machine as one can dream). BUT --and this is a big but-- doing so is very far from elementary and would be both ugly, very large and unmaintainable.