[Javascript] Re: strange gecko (firefox/mozilla) appendchild behavior

[Robert Van Dam]

Update:

The problem seems to be specifically that an empty tbody breaks the
border-collapse: collapse; algorithm.

I should have realized that border-collapse was part of the problem
because the other css was still working.

Try this in your favorite gecko based browser

<style type="text/css">
table, td { border-collapse: collapse; border: 1px solid black; }
</style>

<table>
<tbody><tr><td>1</td><td>2</td></tr></tbody>
<tbody></tbody>
<tbody><tr><td>3</td><td>4</td></tr></tbody>
</table>

On 7/25/07, Robert Van Dam <rvandam00 at gmail.com> wrote:
> Take a look at these and tell me why the first one acts funny in
> firefox/mozilla:
>
> http://rvandam.com/test_appendchild.html
>
> http://rvandam.com/test_appendchild2.html
>
> I have a table (of actual raw data) that is being generated
> dynamically but is relatively expensive to generate.  I wanted to make
> it possible to sort the rows by clicking on the column header but
> without going back to the server or requiring me to cache the entire
> page.
>
> So I dynamically generate js objects containing all the relevant data
> and I got sorting working very cleaning using
> mytable.appendChild(myrow) to rearrange the rows after I had resorted
> them.  That all worked great, but here's the problem:
>
> After the _second_ time you sort, firefox (2.0 on linux and 1.5 on
> windows) and mozilla (1.7) mysteriously lose the borders on all of the
> newly rearranged cells (plus some of the header ones that didn't
> move).
>
> I finally tracked down (thanks to firefox's dom inspector) that I was
> generating a new tbody each time (because otherwise ie doesn't show
> any of the rows) and for some reason firefox/mozilla don't seem to
> apply stylesheet rules of the form "table td" to dynamically generated
> rows that are in the third or later tbody.  However, it works fine in
> konqueror and ie (6 and 7).
>
> Since I've already fixed it I don't know if anyone will bother
> responding but the "why" still has me perplexed.  Anyone have any
> clues?  I'm not getting anything useful out of google at the moment.
>
> (After a little more experimentation, it is only dynamically generated
> tbodies that have this problem, a table with several static tbodies
> has css applied just fine).
>