State of Observers — 28 February 2010

AdamB

Yes, I can confirm its a repeatable bug on Mac OSX SketchUp 8.

I've logged a very grumpy bug report with Google.

Useful to know you can stop the crash by saving before exiting, but I am disappointed a bug like this could be missed. I know software has bugs in it etc, but this seems like any basic regression testing / smoke testing would flush this one out.

After I'm done with a Tool, I always done .pop_tool rather than .select_tool(nil) because it seems less presumptuous. ie restore what the user was doing before rather than cancel their previous selection.

However, Gaieus found out a problem/conflict with Jim Toolbar Organizer - long story short, I've switched to doing select_tool(nil) to avoid some weird race-condition with menu validation procs.

But it seems to have cured the crash on exit of SU8 on Mac when using Observers as well...

Adam

thomthom

@adamb said:

After I'm done with a Tool, I always done .pop_tool rather than .select_tool(nil) because it seems less presumptuous. ie restore what the user was doing before rather than cancel their previous selection.

However, Gaieus found out a problem/conflict with Jim Toolbar Organizer - long story short, I've switched to doing select_tool(nil) to avoid some weird race-condition with menu validation procs.

But it seems to have cured the crash on exit of SU8 on Mac when using Observers as well...

what? damn! I rely on this feature for a plugin I'm making. what kind of race condition? you got a small example?
windows, osx?

AdamB

Ok, it appears to be that if you select_tool but then pop_tool, things can get fubar.

Mostly its benign - hence I got away with it - but with the LargeToolSet showing from Jim's Tool Organizer, the additional calls to the validation procs reveal the error. In that it is wrong to be selecting a tool then popping it, but you might hope SU would just ignore it.

The attached Ruby is a simple Tool that always performs pop_tool on keypress Escape.

It can be invoked either by calling "baddoit()" which starts it using select_tool, or by calling "evian()" which starts it using push_tool.

poptool.rb

thomthom

Ok, so:

select_tool + pop_tool = fubar
but
push_tool + pop_tool = OK

Thought I'd tried pop_tool with select_tool before - without pop_tool doing anything. Thought it only worked after push_tool.

AdamB

@thomthom said:

Ok, so:

select_tool + pop_tool = fubar
but
push_tool + pop_tool = OK

Thought I'd tried pop_tool with select_tool before - without pop_tool doing anything. Thought it only worked after push_tool.

Not quite. select_tool + pop_tool = fubar isn't always terminal. Just apparently if there are really large toolbars around hence its never caused a problem for me in the past.

Thomthom, can you get Jims' Organiser thing and see if you can repro this.

thomthom

Size of toolbars matter?
Or is it toolbars with validation procs?

I have Jim's plugin installed on my Sy7 installation. Is it a particular pre-set toolbar that should be tested against?

Dan Rathbun

@adamb said:

After I'm done with a Tool, I always done .pop_tool rather than .select_tool(nil) because it seems less presumptuous. ie restore what the user was doing before rather than cancel their previous selection.

However, Gaieus found out a problem/conflict with Jim Toolbar Organizer - long story short, I've switched to doing select_tool(nil) to avoid some weird race-condition with menu validation procs.

This issue is caused by the tool stack having no tool_id (or more precisely Tools.active_tool_id == 0 ).

@unknownuser said:

(http://code.google.com/apis/sketchup/docs/ourdoc/tools.html#active_tool_name)":u0sru3py]NOTE: Calling active_tool_name on an empty Tools collection might cause a crash. Before calling this method, it is important to check whether or not the Tools collection is empty. The Tools collection is empty if the method .active_tool_id returns zero.

So always check if tools.active_tool_id==0 before using .pop_tool() or .active_tool_name()

Dan Rathbun

onQuit() callback

I have noticed that on SU 8.0M1 (at least,) that Sketchup does not wait until the callback returns, before it closes all it's child and owned popups, and it's own app window and shuts down.

You can still have code in an onQuit() callback running after all Sketchup's windows have closed. (You can try it with a messagebox.)

I was trying to save size and position of some windows, but Sketchup closes all it's windows before my onQuit() callback code can use Win32API calls to grab the window sizes. I even 'leaned' the code out as much as possible, but still to no avail, ... my code writes empty size arrays into the registry.

My next choice will be to try and use Ruby's built-in define_finalizer for one of my objects, perhaps the plugin module, or the observer object, .. and see if Sketchup will wait until the finalizer block returns.

Did anyone notice this behaviour in SU 7.x ?

Dan Rathbun

Quick Question ...

Seems I remember that AppObserver::onNewModel does not get called when Sketchup starts up.

Is this generally true.. or specific to PC or Mac ?

I see the API says that a command line skp file will not fire the onOpenModel() callback.

thomthom

@dan rathbun said:

Seems I remember that AppObserver::onNewModel does not get called when Sketchup starts up.

Is this generally true.. or specific to PC or Mac ?

Aye.