Path: csiph.com!x330-a1.tempe.blueboxinc.net!usenet.pasdenom.info!news.albasani.net!feeder.erje.net!news.internetdienste.de!noris.net!newsfeed.arcor.de!newsspool2.arcor-online.net!news.arcor.de.POSTED!not-for-mail
Content-Type: text/plain; charset="UTF-8"
Message-ID: <1784154.EMZun36R47@PointedEars.de>
From: Thomas 'PointedEars' Lahn <PointedEars@web.de>
Reply-To: Thomas 'PointedEars' Lahn <cljs@PointedEars.de>
Organization: PointedEars Software (PES)
Date: Mon, 31 Oct 2011 02:19:21 +0100
User-Agent: KNode/4.4.11
Content-Transfer-Encoding: 8Bit
Subject: Re: Changing letters in boxes
Newsgroups: comp.lang.javascript
References: <1cb16076-87c8-4bc6-8007-1efddfb7530b@s10g2000yqa.googlegroups.com> <j8ji7m$fj8$1@dont-email.me> <3024323.SPkdTlGXAF@PointedEars.de> <j8kne2$1e3$1@dont-email.me>
Followup-To: comp.lang.javascript
MIME-Version: 1.0
Lines: 59
NNTP-Posting-Date: 31 Oct 2011 02:19:22 CET
NNTP-Posting-Host: 7782107e.newsspool3.arcor-online.net
X-Trace: DXC=`L<_QU_[o`\E47KDAk81NWMcF=Q^Z^V3X4Fo<]lROoRQ8kF<OcfhCO[Vji1_n_VeiYDZm8W4\YJN\;?f@h5gMfb\m>IFKdLPRgS3?AEY=\RkWQ
X-Complaints-To: usenet-abuse@arcor.de
Xref: x330-a1.tempe.blueboxinc.net comp.lang.javascript:7889

Matt McDonald wrote:

> Thomas 'PointedEars' Lahn wrote:
>> Unreliable.  Especially MSHTML (the layout engine that Internet Explorer
>> uses) is prone to errors with `innerHTML' accesses in table structures.
>>
>> A safer approach is to remove all cells of the row except one, then
>> replace the content of the first cell and add more cells...
> 
> Agreed. Given the Table API that the DOM provides, usage of innerHTML
> here is erroneous, especially in IE. There's also, of course,
> createElement + appendChild if desired.
> 
> I wrote an article earlier today on this topic, except utilizing the DOM
> Table API (insertCell + deleteCell):
> 
> http://www.fortybelow.ca/hosted/comp-lang-javascript/table-cell-
replacement/
> 
> Comments are welcome.

One problem which what you dubbed "DOM Table Cell API" (there really is no 
official name for that part of the W3C DOM Level 2 HTML Specification) is 
that the W3C DOM API Specification and the MSHTML DOM API definition 
disagree (again, cf. HTMLSelectElement::add() [1]) with regard to the 
meaning of the method's arguments and its return value:

Both support the value -1 to append a new cell; but while the standards-
compliant implementation requires the number of existing cells to be passed 
for an alternative of that (tr.cells.length), MSHTML is at least documented 
to require tr.cells.length - 1.  Also, while the standards-compliant 
implementation is supposed to throw DOMException INDEX_SIZE_ERR on error, 
MSHTML is at least documented to throw no exception at all then, but to 
return "null" instead (which, for the ECMAScript binding, is the `null' 
value of the JScript implementation of the Null type). [2]

The same problem exists with 
HTMLTableElement/HTMLTableSectionElement::insertRow(). [3]

Those differences suggest the possibility of other differences between 
standards-compliant and other implementations than the MSHTML DOM in this 
area.  Therefore, I have intentionally used DOM Level 2+ Core methods in my 
example – for which I am not aware of such quirks – and made no mention of 
that "DOM Table Cell API".


PointedEars
___________
[1] <http://www.w3.org/TR/DOM-Level-2-HTML/html.html#ID-14493106>
    <http://msdn.microsoft.com/en-us/library/ms535921(VS.85).aspx>
[2] <http://www.w3.org/TR/DOM-Level-2-HTML/html.html#ID-68927016>
    <http://msdn.microsoft.com/en-us/library/ms536455(VS.85).aspx>
[3] <http://www.w3.org/TR/DOM-Level-2-HTML/html.html#ID-93995626>
    <http://msdn.microsoft.com/en-us/library/ms536457(VS.85).aspx>
-- 
var bugRiddenCrashPronePieceOfJunk = (
    navigator.userAgent.indexOf('MSIE 5') != -1
    && navigator.userAgent.indexOf('Mac') != -1
)  // Plone, register_function.js:16