wxruby development - Mar 2007 - Fix for crash at exit

Hi
This is a very small patch that seems to solve that "Exit the app and 
get a segfault in return" problem.
Another thing: bigdemo.rb:395 does
        item = event.get_item()
        if item.is_ok
but get_item typemaps TreeItemId into a plain integer, so it doesn''t 
have is_ok method anymore, is there
any wxruby2''yish solution to this kind of problem ?

A.K.

In case attachment gets scrubbed somewhere along the way:

Index: fixdeleting.rb
==================================================================---
fixdeleting.rb    (revision 889)
+++ fixdeleting.rb    (working copy)
@@ -21,6 +21,7 @@
    
     if(core_name == "App")
         found_app_constructor = false
+        wx_name = core_name
     else
         found_app_constructor = true
     end
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: wx2.patch
Url:
http://rubyforge.org/pipermail/wxruby-development/attachments/20070317/bd536f80/attachment.pl

Artur Kuptel wrote:
> This is a very small patch that seems to solve that "Exit the app and 
> get a segfault in return" problem.
Thanks very much for the patch - waiting while it recompiles (50 
mins...yawn) then will commit
> Another thing: bigdemo.rb:395 does
>        item = event.get_item()
>        if item.is_ok
> but get_item typemaps TreeItemId into a plain integer, so it
doesn''t
> have is_ok method anymore, is there
> any wxruby2''yish solution to this kind of problem ?
The recommeded  in wxruby2 is just

if not item.zero?
# or
if item > 0

We switched to using a plain Integer typemap for TreeItemId in 0.0.39 
because it didn''t seem worth having directors et al for such a simple 
class. Guess I missed this bit of the sample, thanks for spotting
it.
> In case attachment gets scrubbed somewhere along the way:
Attachments are permitted on this list

cheers
alex

Alex Fenton wrote:
> Thanks very much for the patch - waiting while it recompiles (50 
> mins...yawn) then will commit
>   
I miss the days when I could ask swig to remove at least it''s runtime 
from the generated sources .
> The recommeded  in wxruby2 is just
>
> if not item.zero?
> # or
> if item > 0
>
> We switched to using a plain Integer typemap for TreeItemId in 0.0.39 
> because it didn''t seem worth having directors et al for such a
simple
> class. Guess I missed this bit of the sample, thanks for spotting it.
>   
Problem would be solved if we could define singleton methods for Fixnums :)

A.K.

Artur Kuptel wrote:
> Alex Fenton wrote:
>   
>> Thanks very much for the patch - waiting while it recompiles (50 
>> mins...yawn) then will commit
>>   
>>     
> I miss the days when I could ask swig to remove at least it''s
runtime
> from the generated sources .
>   
Ah - so this isn''t possible any longer with current SWIG? I''ve
seen it
mentioned in the docs and had hoped we could take this approach in the 
near future to reduce the compile time and size of the binary.

a

Alex Fenton wrote:
> Ah - so this isn''t possible any longer with current SWIG?
I''ve seen it
> mentioned in the docs and had hoped we could take this approach in the 
> near future to reduce the compile time and size of the binary.
>
> a
>   
Sadly -c/-noruntime option was removed some time ago, only thing that''s
left from those good times is
''swig -ruby -external-runtime'', which sadly isn''t
very helpful.
I even thought about writing some preprocessor script to remove most of 
the runtime from .cpp files and
writing a glue for generated files but doing it right would require a 
bit of time, which is unobtainable at the moment.

A.K.

Artur Kuptel wrote:
> Hi
> This is a very small patch that seems to solve that "Exit the app and 
> get a segfault in return" problem. 
That''s worked a treat - committed, thanks.

For info I''m still getting one in the grid and caret samples. I will 
investigate further but the error I''m getting (OS X 10.3 PPC) is:

Program received signal EXC_BAD_ACCESS, Could not access memory.
0x030f0a4c in free_wxGridCellBoolRenderer(wxGridCellBoolRenderer*) ()

a