[Javascript] table collapse bug in IE6? - border-spacing

[Paul Novitski]

Håkan,

Thanks for mentioning CSS border-spacing.  Doesn't it map directly onto 
HTML's table cellspacing attribute?  If so, perhaps Microsoft omitted the 
"borderSpacing" because they do support cellSpacing.

If so, there are at least four ways to stipulate a 
borderSpacing/cellSpacing of zero, the CSS method being non-IE-compatible:

CSS: table {border-spacing: 0}
HTML: <table cellspacing="0" ...>
client script: document.createElement('table cellspacing="0"');
client script: oTable.cellSpacing = 0;

Paul




At 07:54 AM 3/28/2004, Håkan Magnusson wrote:
>Note that the style border-spacing ("cellspacing" if defined in the markup 
>of the table) is not supported by any IE version, so in order to create a 
>table through the DOM without cellspacing in IE is to
>
>document.createElement('table cellspacing="0"');
>
>This actually works, while it IS ugly. :)
>
>Regards,
>H
>
>David Lovering wrote:
>
>>Dear Paul:
>>You're welcome.  Also, I apologize for not mentioning the "border-collapse:
>>collapse" attribute fix -- the original code I wrote was intended for a
>>browser which was a near-IE lookalike, but that had some missing CSS
>>support, including the "border-collapse" thingy.  My newest code does in
>>fact use the border-collapse option, but I'd forgotten that I'd amended it
>>when the other browser project tanked.
>>On the timing issue -- I think it depends to some extent on where the data
>>for the rows is coming from.  If it is from a slow database (like DB2,
>>Access, MSSQL), or if it done with a slow pointer/row SQL call, you're
>>absolutely right -- hiding may be the best way of dealing with this issue.
>>On the other hand, if you're working with a speed-optimized database, such
>>as MySQL, Postgres/SQL, etc., or you want only up-to-the-second real-time
>>data mapping to your screen, then the delete and rebuild row method may work
>>better for you.  Also, if you're using ASP or .NET stuff inline as part of
>>your window build, prepare yourself for a HUGE slowdown, which only gets
>>worse as your server(s) get loaded.  Anything that requires a computational
>>rebuild of thousands of cells will amplify the inherent slownesses (is that
>>a word?) of whatever system you pick.
>>-- Dave Lovering
>>
>>----- Original Message ----- From: "Paul Novitski" <paul at dandemutande.org>
>>To: "[JavaScript List]" <javascript at LaTech.edu>
>>Sent: Saturday, March 27, 2004 11:05 PM
>>Subject: Re: [Javascript] table collapse bug in IE6?
>>
>>
>>>Dave,
>>>
>>>At 12:44 PM 3/27/2004, you wrote:
>>>...
>>>
>>>>If you REALLY want to hide the cells, and not merely nuke/rebuild the
>>rows,
>>
>>>>I'd suggest building into your routine a number of CSS attribute calls
>>which
>>
>>>>set the border width to 0 on each of the affected cells/rows,
>>>
>>>*Blush!*  That works perfectly; I must be overworked & sleep-deprived not
>>>to have seen something so obvious.  Thanks!  You win the kewpie doll.
>>>
>>>
>>>
>>>>making sure
>>>>that you still preserve one top/bottom border to "join" the rows which
>>>>remain undeleted above and below the one you're taking out.
>>>
>>>I've given the table the CSS style "border-collapse: collapse" so I don't
>>>have to worry about doubled border thicknesses between adjacent cells.
>>>
>>>
>>>>Personally, nuking the rows makes sense in another way as well -- every
>>>>screen refresh still exercises the "hidden" objects, taking longer than
>>>>necessary to paint the window.  I've seen an appreciable speed up when I
>>>>trash the unwanted rows, rebuilding them only when necessary.
>>>
>>>In this case, it feels like 6:half-dozen.  It appears to take slightly
>>>longer to rebuild the display table with selected rows from the hidden
>>>table than it does to hide & unhide selected rows, which makes sense
>>>(changing visibility being faster than adding branches to the DOM tree).
>>>
>>>Thanks,
>>>Paul
>>>
>>>
>>>_______________________________________________
>>>Javascript mailing list
>>>Javascript at LaTech.edu
>>>https://lists.LaTech.edu/mailman/listinfo/javascript
>>>
>>
>>_______________________________________________
>>Javascript mailing list
>>Javascript at LaTech.edu
>>https://lists.LaTech.edu/mailman/listinfo/javascript
>_______________________________________________
>Javascript mailing list
>Javascript at LaTech.edu
>https://lists.LaTech.edu/mailman/listinfo/javascript