[squeak-dev] re: weird image state

Bert Freudenberg bert at freudenbergs.de
Mon Dec 29 10:23:34 UTC 2014

On 28.12.2014, at 23:29, Eliot Miranda <eliot.miranda at gmail.com> wrote:
> On Sun, Dec 28, 2014 at 9:20 AM, Craig Latta <craig at netjam.org> wrote:
> Hi Eliot--
>> > my image has got into that weird state where clicking on a window
>> > not at the front causes slow redraws of the restacked windows, I
>> > suspect because the execution simulator (a.k.a. the debugger) is
>> > running the drawing code.  If anyone recognises this state do they
>> > also know how to get out of it?
>>      I can get out of it by starting a new UI process, by halting with
>> meta-period.
> Yes, that works form me.  But not always, or perhaps it got back into the same state again.  At least now the system is working well again.
>>      FYI, I can get into it by forcing a "syntax error" window (e.g., by
>> adding an instance variable to a class which already uses the instance
>> variable name as a temporary variable), editing the offending method to
>> get rid of the error, and continuing.
> Ah, that's v good to know.  Of course, now its working I'm not motivated to fix this.  But some time I'll try and take a look.  At least I'll flag this thread to come back to.

I have never really tracked it down, but I think this happens when the defer-display-updates VM flag is not enabled, which is why you see every single drawing operation. Since regular display updates from the Morphic main loop normally manage that flag, I suspect that somehow this drawing must happen outside that loop. The flag is cleared when an error happens so that debug output gets to the screen ASAP. And IIRC it always starts to happen after an error occurred.

- Bert -

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4142 bytes
Desc: not available
Url : http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20141229/c6e9f758/smime.bin

More information about the Squeak-dev mailing list