[Openmcl-devel] hemlock, mac yosemite, and observing contentLayoutRect

[Paul Krueger]

I certainly don’t have a definitive answer, but one thing that I’ve observed over several releases is that Apple changes the internal workings of various window widgets such that touching or modifying them in any way requires that it be done on the main thread. Not doing so results in the sort of failure that you experienced. As a matter of course I now force EVERYTHING related to a user interface to occur on the main thread. If you’re not already doing that, you might force all of your code to run on the main thread to see if that fixes the problem.

On Nov 1, 2015, at 11:49 AM, Arthur Cater <arthur.cater at ucd.ie> wrote:
> 
> Hi,
> 
> I regret recently upgrading my macbook pro from OS 10.6.8 to yosemite 10.10.5
> 
> But now I’ve done it I’ve run into a problem. I’d previously tinkered with Hemlock
> in order to make a new kind of window where I can click on sections of text and get
> lisp actions performed in response. This still works for me, using now ccl 1.10 (16196)
> dx86cl64.
> 
> But when I close such a window, I get an exception that brings up the AltConsole
> window, where it complains that it cannot remove NSScrollViewxxx as an observer
> of the contentLayoutRect property of the NSWindow - because it isn’t an observer.
> 
> This doesn’t happen with regular Hemlock windows, just the ones whose behavior
> I have tinkered with. Doubtless I’ve done something wrong. But I wonder, when
> making ccl 1.10 work on Yosemite, did anything similar happen and if so how was
> it fixed?
> 
> I’ve searched the ccl sources for things like 'ns-window' ‘nswindow' 'ns-scroll-view' 
> ‘observe’ ‘contentlayoutrect’ and others, and found no inspiration. I’ve also found
> with google other complaints about similar-sounding issues in at least one other app,
> but I wonder if anybody with more experience than me with, and a better understanding
> than mine of, Apple’s KVO as it pertains to windows and scrollviews could drop me
> a hint.
> 
> Thanks, Arthur
> 
> _______________________________________________
> Openmcl-devel mailing list
> Openmcl-devel at clozure.com
> https://lists.clozure.com/mailman/listinfo/openmcl-devel