Class recompilation bug...

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

Class recompilation bug...

marcel.taeumel (old)
Hi!

Try adding some instance variable to ScrollPane. During the recompilation step, the instVar pointers in subclasses seem to be off and thus open tools break.

Why?

Best,
Marcel
Reply | Threaded
Open this post in threaded view
|

Re: Class recompilation bug...

Chris Muller-3
On Sat, Apr 25, 2015 at 1:52 AM, Marcel Taeumel
<[hidden email]> wrote:
> Hi!
>
> Try adding some instance variable to ScrollPane. During the recompilation
> step, the instVar pointers in subclasses seem to be off and thus open tools
> break.
>
> Why?

Because such recompilation is not an atomic operation..?

Reply | Threaded
Open this post in threaded view
|

Re: Class recompilation bug...

marcel.taeumel (old)
Yeah, I guess so. But why not? :)
Reply | Threaded
Open this post in threaded view
|

Re: Class recompilation bug...

Levente Uzonyi-2
It's a long story. There used to be an implementation of classbuilder
which could load code atomically, but it never made it to Squeak.

Someone else had to tackle this problem before, and IIRC the solution was
to minimize all windows before recompilation, and then the windows were
restored. But anything which prevents redrawing them should work.

Levente

On Sat, 25 Apr 2015, Marcel Taeumel wrote:

> Yeah, I guess so. But why not? :)
>
>
>
> --
> View this message in context: http://forum.world.st/Class-recompilation-bug-tp4821745p4821833.html
> Sent from the Squeak - Dev mailing list archive at Nabble.com.
>
>

Reply | Threaded
Open this post in threaded view
|

Re: Class recompilation bug...

Nicolas Cellier


2015-04-25 20:09 GMT+02:00 Levente Uzonyi <[hidden email]>:
It's a long story. There used to be an implementation of classbuilder which could load code atomically, but it never made it to Squeak.

Someone else had to tackle this problem before, and IIRC the solution was to minimize all windows before recompilation, and then the windows were restored. But anything which prevents redrawing them should work.

Levente


in Graphics-nice.276.mcz

Nicolas
 

On Sat, 25 Apr 2015, Marcel Taeumel wrote:

Yeah, I guess so. But why not? :)



--
View this message in context: http://forum.world.st/Class-recompilation-bug-tp4821745p4821833.html
Sent from the Squeak - Dev mailing list archive at Nabble.com.






Reply | Threaded
Open this post in threaded view
|

Re: Class recompilation bug...

marcel.taeumel (old)
In reply to this post by Levente Uzonyi-2
Why are there so many senders of #refreshWorld or #displayWorldSafely anyway... :)

Best,
Marcel
Reply | Threaded
Open this post in threaded view
|

Re: Class recompilation bug...

Karl Ramberg
There is also the Monticello preference to log config info to disk, instead for to Transcript. That will prevent a Transcript popping up during update.

Karl

On Sun, Apr 26, 2015 at 9:26 AM, Marcel Taeumel <[hidden email]> wrote:
Why are there so many senders of #refreshWorld or #displayWorldSafely
anyway... :)

Best,
Marcel



--
View this message in context: http://forum.world.st/Class-recompilation-bug-tp4821745p4821916.html
Sent from the Squeak - Dev mailing list archive at Nabble.com.




Reply | Threaded
Open this post in threaded view
|

Re: Class recompilation bug...

Chris Muller-3
On Wed, Apr 29, 2015 at 10:17 AM, karl ramberg <[hidden email]> wrote:
> There is also the Monticello preference to log config info to disk, instead
> for to Transcript. That will prevent a Transcript popping up during update.

Yes, but I can't stand Squeak creating those files in my filesystem.
Updating the system is a process which deserves a Transcript, IMO...

Reply | Threaded
Open this post in threaded view
|

Re: Class recompilation bug...

Karl Ramberg
I mention this in the context that a update breaks because it tries to update for example ScrollPane.

Karl

On Wed, Apr 29, 2015 at 6:17 PM, Chris Muller <[hidden email]> wrote:
On Wed, Apr 29, 2015 at 10:17 AM, karl ramberg <[hidden email]> wrote:
> There is also the Monticello preference to log config info to disk, instead
> for to Transcript. That will prevent a Transcript popping up during update.

Yes, but I can't stand Squeak creating those files in my filesystem.
Updating the system is a process which deserves a Transcript, IMO...