newline in arrayelement

[Ward Germonpré]

Hi,

From a PHP backend I generate a javascript array like this :

bevestigingen =
[{id:'47',leverancier:'VELUX',depot:'KNO',klant:'', werf:'',referentie:'Ro
els',producten:'DKL U04 1100 1st
DKL U04 1705 1st ...

The problem is that the 'producten' field contains data from a textarea
element, in the example 2 products, that were seperated by a newline. The
newline that was stored in the database, as it now emerges in html,
breaks the javascript code.

Is there a silver bullet for this, or should I look into a replacement
strategy serverside ?

thx

Ward

[Evertjan.]

Ward Germonpré wrote on 08 jun 2006 in comp.lang.javascript:

From a PHP backend I generate a javascript array like this :

bevestigingen =
[{id:'47',leverancier:'VELUX',depot:'KNO',klant:'', werf:'',referentie:'
Ro els',producten:'DKL U04 1100 1st
DKL U04 1705 1st ...

The problem is that the 'producten' field contains data from a
textarea element, in the example 2 products, that were seperated by a
newline. The newline that was stored in the database, as it now
emerges in html, breaks the javascript code.

Is there a silver bullet for this, or should I look into a replacement
strategy serverside ?

t = t.replace(/\n/g,' ')

Exchange the return for a space of whatever.
Better do this serverside if the string is from a db,
but PHP is beond the scope of this NG.

You should always validate any form input,
so that unwanted characters are replaced/deleted.

You could keep the return in a string litteral like this:

t = t.replace(/\n/g,'\\n')

Or have PHP do that.

Mee eens?

--
Evertjan.
The Netherlands.
(Please change the x'es to dots in my emailaddress)

[Richard Cornford]

Ward Germonpré wrote:

From a PHP backend I generate a javascript array like this :

bevestigingen =
[{id:'47',leverancier:'VELUX',depot:'KNO',klant:'', werf:'',referentie:'Ro
els',producten:'DKL U04 1100 1st
DKL U04 1705 1st ...

The problem is that the 'producten' field contains data from a textarea
element, in the example 2 products, that were seperated by a newline.
The newline that was stored in the database, as it now emerges in
html, breaks the javascript code.

Is there a silver bullet for this, or should I look into a replacement
strategy serverside ?

Generally, if you want to use arbitrary text data as the contents of a
javascript string literal you must _always_ transform various
characters into escape sequence equivalents of one sort or another.
These characters include: the backslash (which must be escaped first
when going in this direction as escape sequences commence with a
backslash), quote characters (either just the type being used to
delimit the string, or both types for ease/safety), all line
terminators (Line Feed, Carriage Return, and Unicode (hex) 2028 <LS>
and 2029 <PS>), backspace, form-feed, horizontal tabs and vertical tabs
(which should be escaped to Unicode or Hex escape sequences rather than
\v as IE/JScript has a bug with \v, which it considers a literal v).

There is no silver bullet, and this problem is the same as when
inserting arbitrary text into HTML when certain characters should be
substituted for entities.

Richard.

[Ward]

"Richard Cornford" <Ri*****@litotes.demon.co.uk> wrote in
news:11**********************@f6g2000cwb.googlegro ups.com:

Ward Germonpré wrote:

From a PHP backend I generate a javascript array like this :

bevestigingen
[{id:'47',leverancier:'VELUX',depot:'KNO',klant:'', werf:'',referentie:
'Ro els',producten:'DKL U04 1100 1st
DKL U04 1705 1st ...

The problem is that the 'producten' field contains data from a
textarea element, in the example 2 products, that were seperated by a
newline. The newline that was stored in the database, as it now
emerges in html, breaks the javascript code.

Is there a silver bullet for this, or should I look into a
replacement strategy serverside ?

Generally, if you want to use arbitrary text data as the contents of a
javascript string literal you must _always_ transform various
characters into escape sequence equivalents of one sort or another.
These characters include: the backslash (which must be escaped first
when going in this direction as escape sequences commence with a
backslash), quote characters (either just the type being used to
delimit the string, or both types for ease/safety), all line
terminators (Line Feed, Carriage Return, and Unicode (hex) 2028 <LS>
and 2029 <PS>), backspace, form-feed, horizontal tabs and vertical
tabs (which should be escaped to Unicode or Hex escape sequences
rather than \v as IE/JScript has a bug with \v, which it considers a
literal v).

There is no silver bullet, and this problem is the same as when
inserting arbitrary text into HTML when certain characters should be
substituted for entities.

Richard.

Thank you both for replying.

I used PHP's urlencode() en javascript's unescape() functions to get rid
of the above problem.

kind regards,

Ward

[Richard Cornford]

Ward wrote:

"Richard Cornford" <Ri*****@litotes.demon.co.uk> wrote in <snip>

Generally, if you want to use arbitrary text data as the contents of a
javascript string literal you must _always_ transform various
characters into escape sequence equivalents of one sort or another.

<snip> Thank you both for replying.
Apparently it was a waste of effort.
I used PHP's urlencode() en javascript's unescape() functions to
get rid of the above problem.

If you would rather use a ****-*** hack than do the job properly the
least you should do is use - decodeURIComponent - on the client (where
supported) so that the process is symmetrical.

Richard.

[Ward Germonpré]

"Richard Cornford" <Ri*****@litotes.demon.co.uk> wrote in
news:11**********************@h76g2000cwa.googlegr oups.com:

Ward wrote:

"Richard Cornford" <Ri*****@litotes.demon.co.uk> wrote in

<snip>

Generally, if you want to use arbitrary text data as the contents of
a javascript string literal you must _always_ transform various
characters into escape sequence equivalents of one sort or another.

<snip>

Thank you both for replying.

Apparently it was a waste of effort.

I used PHP's urlencode() en javascript's unescape() functions to
get rid of the above problem.

If you would rather use a ****-*** hack than do the job properly the
least you should do is use - decodeURIComponent - on the client (where
supported) so that the process is symmetrical.

Richard.

Thanks for pointing me to decodeURIComponent.

I do take your advice to heart, in fact I'm rewriting an intranet app
but was asked to quickly implement a new functionality on the old '0.2
framewerkz'.., the new 1.0 app will have validation.

kind regards

Ward