Path: csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!weretis.net!feeder4.news.weretis.net!newsreader4.netcologne.de!news.netcologne.de!newsfeed.arcor.de!newsspool3.arcor-online.net!news.arcor.de.POSTED!not-for-mail
Content-Type: text/plain; charset="UTF-8"
Message-ID: <1936289.aqSPkdTlGX@PointedEars.de>
From: Thomas 'PointedEars' Lahn <PointedEars@web.de>
Reply-To: Thomas 'PointedEars' Lahn <cljs@PointedEars.de>
Organization: PointedEars Software (PES)
Date: Wed, 08 Feb 2012 14:43:42 +0100
User-Agent: KNode/4.4.11
Content-Transfer-Encoding: 8Bit
Subject: Re: Blackberry-specific problem
Newsgroups: comp.lang.javascript
References: <3969324.4567.1328673767667.JavaMail.geo-discussion-forums@vbbfd4>
Followup-To: comp.lang.javascript
MIME-Version: 1.0
Lines: 212
NNTP-Posting-Date: 08 Feb 2012 14:43:44 CET
NNTP-Posting-Host: 8ae7dd05.newsspool4.arcor-online.net
X-Trace: DXC=k9W2NhMdOQ4_A0jCfgHO6>4IUK<Cl32<14Fo<]lROoR18kF<OcfhCO;IHeiH@`A^P6DZm8W4\YJN<b@mF9jNikd5PcfN;IC;G;3=IC:8HI^Zc2
X-Complaints-To: usenet-abuse@arcor.de
Xref: x330-a1.tempe.blueboxinc.net comp.lang.javascript:11501

For goodness' sake, people, either you learn about Google Groups' 
limitations and work around them every time, or you just DO NOT USE Google 
Groups for posting.  The decoded lines in the precursor are up to 331 
characters long.  This is unacceptable.  The acceptable, compatible limit is 
80 characters or less (see the FAQ).  If it wasn't for the word-wrapping 
capabilities of some newsreaders (like mine), the posting would be simply 
unreadable.

Jason C wrote:

> […]
> The problem, though, is that with Blackberry users, one of the values of
> the checkboxes (seemingly random) will come through as the value of the
> subcategory.

That is a non-description, which will probably result in a non-answer (from 
a limited point of view).

<http://jibbering.com/faq/#posting> pp.

> I've tried to duplicate this with IE, FF, Chrome, Safari, Opera, iPhone,
> and Android, with no luck.

There are hundreds of versions of those browsers and dozens of platforms 
they can run on.  Even on Android for which there are several versions and 
ROMs, there are several browsers with several versions.  I am using Dolphin 
Browser HD 7.3.0 most of the time, for example.

<http://catb.org/esr/faqs/smart-questions.html#beprecise>

> The only similarity between the submissions with errors is that the user
> is using a Blackberry, so I'm guessing that it's either something with the
> phone, or an error somewhere in the code that's only being picked up by
> Blackberry.

Again, which one *exactly*?

> FWIW, the page "almost" validates via W3.org. The only error that it finds
> is that I use <p></p> to let the user increase the font size of the page,
> but it's used within several inline tags. This should be irrelevant for
> this error, though.

With invalid markup, all bets are off.  Especially with DOM scripting like 
this.

<http://web.archive.org/web/20110514122344/http://diveintomark.org/archives/2003/05/05/why_we_wont_help_you>

> The relevant code is below. Do you guys see a flaw anywhere that could be
> causing this?
> 
> // The JavaScript

Please get informed about what JavaScript is and is not.  Useful keywords 
include "Netscape", "Mozilla.org", "JavaScript", "JScript", "Google V8", 
"JavaScriptCore", and "ECMAScript".  Avoid books, take blogs and Wikipedia 
with a handful of salt.

> // this is used later in the script, but I'm
> // leaving it here in case it's somehow relevant
> var has_subcat = false;

Avoid global variables unless you absolutely need them.  Post the code that 
you use, not something else.  And do not use code that you do not understand 
(code that in this case is obviously written by someone who does not 
understand these programming languages either).

> // Set subcategories
> function getSubcat(cat) {
>   var list_cat = new Array();

You do not want an Array instance (see below), you want an Object instance:

  var list_cat = new Object();

or 

  var list_cat = {};

>   var show_msg;
> 
>   // Modified for this post
>   // The actual script creates several arrays, where the
>   // list_cat[index] represents a category (like "Autos"),
>   // and the values represent a subcategory (like "Cars")
>   list_cat['Autos'] = new Array (
             ^^^^^^^
>     "Cars",
>     "Trucks, Vans, and SUVs",
>     "Motorcycles / ATVs",
>     "Campers / Boats",
>     "Farm Equipment",
>     "Parts and Accessories"
>   );
> 
>   if (cat && (list_cat[cat].length > 0)) {

This throws an exception if `list_cat[cat]' evaluates to `undefined'.

>     has_subcat = true;

That assigns to the global variable, unchanging unless modified elsewhere.

>     show_msg = "<select name='subcat' size='1'>";
>     show_msg += "<option value=''>Choose a subcategory<\/option>";

If you must, e. g.

    show_msg = "<select name='subcat' size='1'>"
             +   "<option value=''>Choose a subcategory<\/option>";

>     for (n=0; n < list_cat[cat].length; n++)

`n' is not declared (with `var'), so the assignment will leak into outer 
scopes (side-effects include throwing exceptions).  Use

  "use strict";

on the top of the code or the function body to find such mistakes earlier in 
recent browsers (implementing ECMAScript Ed. 5).  Use semantic validation 
elsewhere.

It is possible that this is the cause of the problem that *you* have found.

An optimization:

    for (var n = 0, len = list_cat[cat].length; n < len; ++n)

>       show_msg += "<option value='" + list_cat[cat][n] + "'>"

Always escape attribute values.  Never access the same property twice (DRY).

>         + list_cat[cat][n] + "<\/option>";
                                ^^^
The one good thing about this script (which does not help it to win any good 
award, though) is that the ETAGO delimiter (`</') is always escaped (in the 
most efficient way possible) so that the code can be copied easily into an 
HTML document without causing a markup syntax error.
 
>     show_msg += "<\/select>";
>   }
> 
>   if (document.getElementById)
>     document.getElementById('subcategory').innerHTML = show_msg;

Avoid innerHTML (especially avoid mixing standards-compliant methods like 
getElementById with the proprietary innerHTML), use standards-compliant DOM 
properties and methods, only complemented with proprietary approaches if the 
latter are better compatible.

>   else if (document.all)
>     document.all[subcategory].innerHTML = show_msg;
      ^^^^^^^^^^^^
Really, Internet Explorer 4 (which does not support CSS floats)?
 
>   else if (document.layers) {
             ^^^^^^^^^^^^^^^

Really, Netscape 4 (which does not support CSS floats)?

>     document.layers[subcategory].document.write(show_msg);

In theory, there should be

  document.layers[subcategory].document.open();

before, further optimized to

  var doc = document.layers[subcategory].document;
  doc.open();

aso.  But you probably want to get rid of this altogether.

>     document.layers[subcategory].document.close();
>   }
> }
> 
> 
> // The HTML part of the SELECT menu
> <select name="category" size="1" onChange="getSubcat(this.value)">

Avoid *unconditional* `onchange' for `select' elements; that is _not_ user-
friendly.

> // The HTML part of the checkboxes; roughly 40 of these
> // are actually shown
> //   "option" is intended as a variable here,

IIUC: Don't, this is proprietary.

> // so the live page doesn't actually say "option"

And *what* *does* it say?

> <div style="float:left">
>   <input type="checkbox" name="more" id="option" value="option">

What do you need the ID for?  And should the value not be, say, "1"?

>   Option

Use `label' elements so that `Option' triggers the related checkbox.
Use the `accesskey' attribute if possible.

> </div>


PointedEars
-- 
> If you get a bunch of authors […] that state the same "best practices"
> in any programming language, then you can bet who is wrong or right...
Not with javascript. Nonsense propagates like wildfire in this field.
  -- Richard Cornford, comp.lang.javascript, 2011-11-14