Re: Implement dragging and resizing of tty child frames

[Gerd Möllmann]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>>  1674         (or (and str
>>  1675                  (< (cdr str) (length (car str)))
>>  1676                  (get-text-property (cdr str) property (car str)))
>>  1677             ;; Mouse clicks in the fringe come with a position in
>>  1678             ;; (nth 5).  This is useful but is not exactly where we 
>> clicked, so
>>  1679             ;; don't look up that position's properties!
>>  1680             (and pt (not (memq (posn-area pos)
>>  1681                                '(left-fringe right-fringe
>>  1682                                  left-margin right-margin tab-bar)))
>>  1683                  (get-char-property pt property w))))
>>
>> In the list at line 1681, I think there is menu-bar missing.
>
> Quite likely.  BTW, I've recently seen a fair number of patches fly by
> which change those same kinds of hard-coded lists.  We should probably
> move those to separate variables with names and docstrings that say what
> they mean (and hopefully that will let us consolidate some of those
> lists).

Good idea. I've done that.

>
>> From the history of the function I'd conjecture that mode-line,
>> header-line, vertical-line are probably okay because the C code
>> doesn't add positions for these events, but that's only a conjecture.
>
> A comment mentioning that those symbols would be harmless in the list
> would be helpful as well (e.g. might help consolidate such lists).

I've gone the other route and made the variable contain all areas with no
buffer positions, as listed in lispref. Looked cleaner that way to me.

>
>> Adding Stefan Monnier in CC. Maybe he remembers something about this.
>> It's only been 11 years since he changed that list :-).
>
> I'm pretty sure it was "sharply fuzzy" back then.  I'll let you guess
> whether it got sharper of fuzzier in the mean time.  🙂

😀