log read problem

[jing]

I have an application which maintain a log file, so it adds new lines into this file continously. I need to make a program which gets the newly generated lines out and send to another machine through TCP.

My problem now is, how can I get the newly generated lines each time by reading the file?

I used the file modifaction time to check whether there is any changes, but which function can I use to get the latest update log? i have read the file API, but still don't know how to do it

[ChrisL]

Like this

#include <file.au3>
$sFilePath = @WindowsDir & "\WindowsUpdate.log"

$lastline = FIleReadLine($sFilePath,_FileCountLines ( $sFilePath ) - 1)

Msgbox(0,"",$lastline)

[randallc]

Hi,

_FileCountLines [i think] reads the whole file each time, which might slow things down, depending on your needs and the size of the file.

You could have a var which keeps current file size, and whenever file size changes, only read the end of the file by the difference in size;

See "TailRW" in my sig for funcs which read the end of the file by binary read API, or refer Larry's API.

Best, randall

[jokke]

Hiya, iwe created alot log parsers latly and i have this routine i like to use.

I have yett to see it halt, but it does drain some cpu power though but it works pretty well.

#include <file.au3>

Dim $CountLines,$lastline,$string,$log

$log = 'LOGGFILE.TXT'

$lastline = _FileCountLines($log)

While 1

    $CountLines = _FileCountLines($log)

        If $CountLines > $lastline Then
            $string = FileReadLine($log, $CountLines)
            
            ConsoleWrite($string&@CRLF)             

            $lastline = $CountLines 
        EndIf

    Sleep(10) ; Just so tailer does not drain 100% cpu, if it uppdates slower, upp this value and it drains even less power.

WEnd

Every time Log updates it will write uppdated string into console.

Edited March 11, 2007 by jokke

[jing]

Thank you so much guys!

The file line count things can work definitely. It is a great idea. But it may work quite ineffeiently since it may need to scan the whole document which is not desirable.

randallc made a really good point. Thanks. I will try it.

Thanks again , guys.

[SadBunny]

I have a better solution than counting the lines of the file many times every second... Why not just pull filesize or file-last-changed-time from the file, compare it to the value gotten 10 ms ago, if it changed to something else stay looping.

/edit: And, you can always keep a line counter variable that you add one to every time it finds the file changed, then you read that line using fileReadLine.

Edited March 12, 2007 by SadBunny