Faking OS X text field behaviour (was: Good books on Cocoa dev?)

[Richard Gaskin]

Peter Brigham wrote:

> On Mar 3, 2010, at 7:37 PM, Richard Gaskin wrote:
...
>> According to the dictionary RevTalk allows backgroundColor for
>> chunks, and backgroundPattern for fields, and textPattern/
>> foregroundPattern for chunks, so I was a little optimistic about the
>> orthogonality of the implementation.
>>
>> But in trying it here, you're right, it seems the backgroundPattern
>> cannot be applied to chunks.
>>
>> This inconsistency seems worth addressing, since doing so has
>> practical application as you've noted, so I logged it as a request:
>>
>> <http://quality.runrev.com/qacenter/show_bug.cgi?id=8645>
>
> Is this one of those "why not just do it, looks simple" things that
> runs into the apparently huge complexity of the field object? Which I
> seem to recall Scott Raney describing as the "monster" it terms of
> complexity.

Hard to say without confirmation from the Mother Ship, but according to 
Dr. Raney, historically many such property limitations were the result 
of having a certain number of bits set aside for built-in object 
properties, so that adding another would require a format change.

Text chunks are the odd man out, since they don't have the same sort of 
record structure objects have.   It may be harder than adding properties 
for objects, or it may be easier.

Either way, the field object is due for an overhaul anyway 
(paragraph-level formatting and independent column alignment are among 
the most voted-for items at the RQCC), so it seems worth raising the 
visibility of this request to see it implemented.

Not only is it useful, but the absence of backgroundPattern for chunks 
in a system that already supports backgroundColor just means one more 
"gotcha" inconsistency newcomers need to learn.

--
  Richard Gaskin
  Fourth World
  Rev training and consulting: http://www.fourthworld.com
  Webzine for Rev developers: http://www.revjournal.com
  revJournal blog: http://revjournal.com/blog.irv