[Warning - don't try this in an image without saving first]
To crash your 5.1.4 image, set the Debugger preferences for "showWalkbacks" to false. And then evaluate the following: (ResourceIdentifier class: Presenter name: 'Default view') loadWithContext: DeafObject current. Looking in the errors log file we see that it caused the following exception: Unhandled exception - an InvalidExternalCall('Invalid arg 9: Cannot coerce a DeafObject to handle') But in trying to open the debugger it generates: Unhandled exception - an Error('Primitive View>>primHookWindowCreate: failed (1)') Which tries to open a debugger, which causes another failure, etc., etc., etc. :-( Yeah, yeah, I know. I shouldn't be trying to load a resource into a DeafObject, but it really wasn't that hard to get into that situation (accidentally doing some updates to a window that had been closed). It sure would be nice if the image didn't crash as a result. I find it interesting that by showing the walkback first, that there's some sort of separation that insulates the debugger that subsequently opens from having a problem. But while writing and running tests, it's really nice to have the shortcut of going directly into the debugger. Any ideas for a fix? -Bill ------------------------------------------- Bill Dargel [hidden email] Shoshana Technologies 100 West Joy Road, Ann Arbor, MI 48105 USA |
"Bill Dargel" <[hidden email]> wrote in message
news:[hidden email]... > [Warning - don't try this in an image without saving first] > To crash your 5.1.4 image, set the Debugger preferences for > "showWalkbacks" to false. And then evaluate the following: > > (ResourceIdentifier class: Presenter name: 'Default view') > loadWithContext: DeafObject current. > > Looking in the errors log file we see that it caused the following > exception: > Unhandled exception - an InvalidExternalCall('Invalid arg 9: Cannot > coerce a DeafObject to handle') > > But in trying to open the debugger it generates: > Unhandled exception - an Error('Primitive > View>>primHookWindowCreate: failed (1)') > > Which tries to open a debugger, which causes another failure, etc., > etc., etc. :-( > > Yeah, yeah, I know. I shouldn't be trying to load a resource into a > DeafObject, but it really wasn't that hard to get into that situation > (accidentally doing some updates to a window that had been closed). It > sure would be nice if the image didn't crash as a result. > > I find it interesting that by showing the walkback first, that there's > some sort of separation that insulates the debugger that subsequently > opens from having a problem. But while writing and running tests, it's > really nice to have the shortcut of going directly into the debugger. > > Any ideas for a fix? Yes: 1) Create a new subclass of ShellView, call it DebuggerShell. 2) Drag & drop (copy) the #primHookWindowCreate: method from WalkbackDialog into DebuggerShell. 3) Edit the 'Default view' (or the 'Classic view' if you use it) of the Debugger using the ViewComposer, and mutate the shell into a DebuggerShell. Now try your test expression. Works for me (although I would still recommend against running with #showWalkbacks=false). Regards Blair |
Blair McGlashan wrote:
> 1) Create a new subclass of ShellView, call it DebuggerShell. > 2) Drag & drop (copy) the #primHookWindowCreate: method from WalkbackDialog > into DebuggerShell. > 3) Edit the 'Default view' (or the 'Classic view' if you use it) of the > Debugger using the ViewComposer, and mutate the shell into a DebuggerShell. Thanks Blair. This change seems to be just what was needed. BTW, I scripted the mutate so that I could incorporate it into the automatic build of my working image: | vc | vc := (ViewComposer show: 'Default view') openOn: (ResourceIdentifier class: Debugger name: 'Default view'); yourself. vc mutateTo: DebuggerShellView. vc fileSave. vc exit. > > Now try your test expression. Works for me (although I would still recommend > against running with #showWalkbacks=false). Hmm. Care to elaborate on your recommendation of showing the walkbacks? Thanks, -Bill ------------------------------------------- Bill Dargel [hidden email] Shoshana Technologies 100 West Joy Road, Ann Arbor, MI 48105 USA |
"Bill Dargel" <[hidden email]> wrote in message
news:[hidden email]... > Blair McGlashan wrote: >> (... I would still recommend > > against running with #showWalkbacks=false). > > Hmm. Care to elaborate on your recommendation of showing the walkbacks? > http://groups.google.com/groups?hl=en&lr=lang_en&ie=UTF-8&oe=UTF-8&selm=s0uT5.7929%24ea7.156595%40news2-win.server.ntlworld.com (or search Google groups for 'self halt in queryCommand') Regards Blair |
Free forum by Nabble | Edit this page |