newlines

[digitalecartoon]

I'm having my flash form send it's data as plain text mail. When I enter a couple of paragraphs in the message field and receive the plain text mail, all paragraphs are put in a row instead that every paragraph starts at a new line. So I get this:

Bericht:
Dit is een nieuwe alineaDit zou op een nieuwe regel moeten beginnenDeze regel eigenlijk ook

Instead of this:

Bericht:
Dit is een nieuwe alinea
Dit zou op een nieuwe regel moeten beginnen
Deze regel eigenlijk ook

I think it's because I've used stripslashes, but I've use that so that entered names like O'Brien don't appear as O\'Brien in the mail message.

How can I fix this?

Actionscript:

Versturen.onRelease = function() { 
    mySendVars = new LoadVars(); 
    myLoadVars = new LoadVars(); 
    mySendVars.naam = naam.text; 
    mySendVars.email = email.text; 
    mySendVars.bericht = bericht.text; 
    mySendVars.sendAndLoad("mailform.php", myLoadVars, "POST"); 
    gotoAndStop(2); 

PHP code:

 
$naam = stripslashes(utf8_decode($_POST["naam"])); 
$email = stripslashes(utf8_decode($_POST["email"])); 
$bericht = stripslashes(utf8_decode($_POST["bericht"])); 

$message = "Naam:\r\n".$naam."\r\n\r\n"; 
$message .= "Emailadres:\r\n".$email."\r\n\r\n"; 
$message .= "Bericht:\r\n".$bericht."\r\n"; 

$headers = "MIME-Version: 1.0\r\n";  

$headers .= "Content-type: text/plain; charset=iso-8859-1\r\n"; 
$headers .= "From: ".mb_encode_mimeheader($naam, "iso-8859-1", "Q")." <".$email.">\r\n";  

mail($to, $subject, $message, $headers); 

}

[sneakyimp]

Are you certain that the line breaks are in the data when it arrives in PHP? As far as I know, stripslashes should have no effect on line breaks.

And where are you viewing the mail? If you are displaying it as HTML then the linebreaks will not appear - html parsers treat newline chars as plain old spaces.

[digitalecartoon]

By using the same php script I discovered that a html form sends it's newlines through Post differently than a flash form does through Post.

With html all paragraphs start on new lines, with flash they are put behind each other.

Don't know exactly what the difference is and what exactly a html form posts as newline characters and what a flash form does. But the sollution seems to be just using:
$message = preg_replace('~\r\n|\r|\n~', "\r\n", $message);

to make it all \r\n, no matter what the source, html or flash.

Anyone who knows the difference between html posted newlines and flash posted ones?

By the way, I've got this code from someone:
$message = preg_replace('~\r\n|\r|\n~', "\r\n", $message);
But the regex used doesn't make sense to me: why have \r\n replaced with \r\n ? Perhaps the original code replaced this complete regex with e.g. <br> and is this case \r\n was used (which mail() uses) as replacement and the first to be replaced part (\r\n) could be deleted?

[sneakyimp]

Are you saying that HTML properly sends the newline characters and flash instead sends \r or \n instead of the actual newline char? That sounds like some really stupid behavior from the flash client which frankly would not surprise me.

It would probably be helpful if you were to write the data that arrives in your PHP file to a log or something so you can get a better look at what is actually in there, then you can make the right choice of action. If the flash form submission is consistent in its behavior, you should be able to fix the problem with one replacement command as you suggested. The only problem you might have is if your PHP file is supposed to handle both flash and html submissions.

[digitalecartoon]

what should I add to my form processing php script to the posted content is logged somehow? Would it also show \r or \n and stuff like that? And this log can then be downloaded?

Yes, what I did is make a html form with the same fields (name, email, message) as the flash form has. Have it post a mail with the message containing 2 paragraphs. I got the plain text mail with the paragraphs on top of each other. When I entered 2 in the flash form it would arrive with the paragraphs next to each other. I used the same php script for processing the 3 fields. So it would seem that html indeed sends its newlines differently

[sneakyimp]

This could be a bit tricky to figure out...

To write a log file, your web server will need to have write access to some file on your server. Then you can use a logging function like this...i tries to open the file and if it can't then it tries to change the permissions and open it again. Keep in mind that flash has no idea what to do when the script it's visiting just dies with an error message.

define('LOG_FILE', '/path/to/log.txt'); // this needs to be the file system path to your log file
function log_it($msg_to_log) {
	if (!($fd = fopen(LOG_FILE, "a"))) {
		$tmparr = posix_getpwnam(CONST_SERVER_USER);
		$server_uid = $tmparr['uid'];
		chown(LOG_FILE, $server_uid);
		chmod(LOG_FILE, 0775);
		if (!($fd = fopen(LOG_FILE, "a"))) {
			die("Could not open debug file.  Mesg was: $msg_to_log");
		}
	}
	if (!fwrite($fd, $msg_to_log . "\n")) {
		die("LOG WRITE FAILED!");
	}
	fclose($fd);
	if (filesize(LOG_FILE) == 0) {
		/* Something went wrong! */
		die("LOG WRITE FAILED FILESIZE IS 0");
	}
}

If the log file lives in your web root then yes you can visit it just by visiting the file in a web browser. A web browser should interpret the file as a text file if it ends in ".txt" - at least Firefox does on my Win XP machine. This means that when I write a string with newline characters in it to my log, when I visit the log I actually see the text displayed on separate lines. If you mistakenly called your log file 'log.html' then your browser would render it as html and the line returns just look like a space. Does that make sense?

Another thing to consider is that notepad on a windows machine will only show newline characters if there is a \r AND ALSO A \n character. This is kind of a weird thing. I'm not sure what the historical reason is, but I seem to recall a unix newline char is \n and for some reason these don't get rendered as new lines in windows. Why? Wish I knew. It may be that flash is just interpreting new lines as \n (which corresponds to an ASCII code of 10) and these aren't a 'complete' line return to a windows machine. It may be that your mail client requires both the \r and the \n to show a new line. \r corresponds to ascii code 13.

In order to find out exactly what ASCII codes are in the string that gets submitted, you have to write a separate function which converts each char in a string to its ASCII code so you can see what is really going on. Something like this:

$str = "foobar\r\n";
$max = strlen($str);
echo 'max:' . $max . '<br>';
for($i=0; $i<$max; $i++) {
  $char = substr($str, $i, 1);
  echo $char .' = ' . ord($str[$i]) . '<br>';
}

As you can see when you run that script, \r and \n are whitespace chars so they don't show up but you still get the '= 13' and '= 10'.

I hope that's helpful.

[digitalecartoon]

Well, it's a bit too much effort :-) But so far I've noticed that flash sends newlines differently.

But about this code I got to switch everything to normal newlines (no matter what form is used):
$message = preg_replace('~\r\n|\r|\n~', "\r\n", $message);

Isn't the first \r\n of the regex not necessary (why replace \r\n with \r\n)? That's probably because the orignal line used something like <br> for the replacement string?

[sneakyimp]

Yes I think you're right about that being unnecessary.

By the way, rather than all the logging shiznit, you could probably just do this to get an idea of what's going on:

$new_message = '';
$max = strlen($message);
for($i=0; $i<$max; $i++) {
  $char = substr($message, $i, 1);
  echo $char .' = ' . ord($char) . "\r\n";
} 
mail($to, $subject, $new_message, $headers); 

That way it'll just report the ASCII values contained in the message to you thru email and you can see what it's doing with the newline chars. My guess is that it's just using \n for newlines.

[digitalecartoon]

With this code and using a html and a flash form I've determined the following when entering 2 paragraphcs. Each with a letter: a and b. So entering:

a
b

In the html form it outputs as:
a = 97 = 13 = 10 b = 98

In the flash form it outputs as:
a = 97 = 13 b = 98

Flash seems to post just \r as newline whereas html posts \r\n as newline code.

[sneakyimp]

Yes exactly. I was wrong...instead of \n, flash is just doing \r. I'm curious if it might be different on a Mac or Linux machine. A lot of times when you get a text file from either and open it in notepad on a windows machine, all the line returns are gone.

BTW now that I think about it, you probably do need the first \r\n in your regex because otherwise it would turn every occurrence of \r\n into TWO line breaks rather than just one because it would replace \r and then \n every time. Leave it in there.

[digitalecartoon]

So how does that work exactly with preg_replace? Is there some kind of order? Lets take part of your text as string:

occurrence of \r\n into TWO line break

so when using: $message = preg_replace('~\r\n|\r|\n~', "\r\n", $message);

it first searches for \r\n, making it \r\n so it would remain:
occurrence of \r\n into TWO line break

then it would replace \r for \r\n, making it
occurrence of \r\n\n into TWO line break

finally it would replace \n for \r\n making it
occurrence of \r\r\n\r\n into TWO line break

Thats the way I see it , but it probably doesn't work that way with preg_replace and those 3 options seperated by a pipe character?

[sneakyimp]

To be honest, I'm not entirely sure. You'd need to experiment with it. The pipe character means OR. So I guess your regular expression would check a potential match against the \r\n before bothering with the individual \r or \n. If it matches the first time, we don't bother checking it again.

In other words, I think if it matches the \r\n then it replaces that with \r\n - there's no change and it skips over the other two possibilities and moves on to the rest of the string. I'm not 100% certain of that.

And, more importantly, most text editing programs and mail clients, etc. will interpret \r\n as only ONE line break rather than two. Not 100% about that either. I do know that notepad requires both to display a line break -- at least on my winxp machine.

I think your original preg_replace works fine because it should turn:
\r into \r\n
\n into \r\n
\r\n into \r\n

which i believe is what you would want.