similar to: [LLVMdev] LLVM and the "static initialization order fiasco"

Hi Have been looking again at SWIG output pointer parameters, as: http://www.swig.org/Doc1.3/Ruby.html#Ruby_nn32 The problem is, whenever I tried to include the SWIG/Ruby ''typemaps.i'' file, it threw loads of errors. A bit more investigation pinned the problem down to some secondary files that typemaps.i includes to map ''special'' types like File. So I tried

Two patches here: First one removes virtual from a function that shouldn''t have been virtual. Second one applies a typemap to catch GetRectFromChildWindow. For reasons that are completely beyond me the standard ruby OUTPUT typemaps don''t define directorargouts. I will take a stab at patching SWIG''s default typemaps someday so we get this fixed for free on all our

This patch for typemap.i fixes several swig warnings while compiling. You may noticed that I didn''t fix all the typecheck typemaps. I suppose I should fix all of them, but they don''t cause problems with the compilation. This also has a patch to convert void *''s into ruby numbers. This fixes a problem with the get_handle call. I still am not sure what''s

This patch adds %directorout typemaps for the GetTextExtent functions _______________________________________________ wxruby-users mailing list wxruby-users@rubyforge.org http://rubyforge.org/mailman/listinfo/wxruby-users

Here''s a patch to fix Wx::Window#get_text_extent - it requires that the patches from teh last message are applied, to enable *INPUT etc typemaps. Roy - though the function works correctly when called from Ruby, it still causes a crash when called internally when clicking on a Wx::Grid cell. Saw your message on SWIG mailing list, hopefully someone there can help out. alex

(Mostly for Roy): I saw this on the SWIG list, and wonder if it might be related to the problems we have been seeing where multi-arg typemaps end up invoking the wrong wrapper methods. (Not the newly-discovered directorout stuff). Perhaps if we switched to "compactdefaultargs" things would start to work more sanely. Looking at this section of the SWIG docs, it looks like it should

Hi The attached patches enable the use of SWIG''s *INPUT *OUTPUT and *INOUT typemaps as described in the manual. It''s done by including SWIG''s own ''typemaps.i'' file so it''ll allow us to benefit from future upgrades to that. The problem was a collision between a Ruby macro _ (fixing for old compilers) and a WxWidgets macro _ (for text

Here''s an example of the problem I was talking about the other day. Take a look at Button.i. It imports from its entire heirarchy. If you look at Button.cpp you will see that no wrappers are generated for anything except the button functions. This brings up some other things that we''ll need to keep in mind: If we''ve created a typemap in a .i file in classes and

StatusBar.i -- Adds a typemap to return a wxRect structure. You''ll really want to go over this to make sure it''s correct. I don''t have a sample but it did make part of one of the bigdemo samples work a little better. If you want to put this on the back-burner until we have a chance to test it more fully I''d understand completely. One of the things

I''ve been doing some development with TreeCtrl, and am having a few problems with TreeItemId. Specifically, I''m getting ''null reference'' errors after a time - which I think means that the wrapper is tracking objects, but the underlying C++ object is deleted. Rather than trying to fix this. I''m wondering if it''s easier to do away with the

I have not worked on wxRuby2 for awhile and decided to download the latest CVS HEAD and saw that it still had a problem with wxWindow::GetTextExtent (in my case showing up when editing Grid cells). So I copied the method signature from the wxDC::GetTextExtent which seems to be working, changed 2 variable names and made the typemap the same as used for wxDC. Well in my program editing grid cells

Mehdi, Teresa, Not sure if this is caused by one of your recent commits, or by someone else's, please excuse me if that's unrelated to your work... http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/542/steps/check-llvm%20asan/logs/stdio ==26383==ERROR: AddressSanitizer: initialization-order-fiasco on address 0x000002ef41d8 at pc 0x0000009d1aa5 bp 0x7ffd0cd72b50 sp

Wow. Look what SWIG does to GetTextExtent in Grid.cpp: void SwigDirector_wxGrid::GetTextExtent(wxString const &string, int *x, int *y, int *descent, int *externalLeading, wxFont const *font) const { ... if (swig_get_up()) { wxWindow::GetTextExtent(string,x,y,descent,externalLeading,font); return; } obj0 = rb_str_new2((const char *)(&string)->mb_str());

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

There are several functions that take the parameters (int *x, int *y) In some cases x & y are simply return values. In some cases x & y are both input and output. How are we going to handle this properly? One method would be to change the header files to something like xin or xout. I don''t really like that method as it''ll break as soon as header files change.

The following two files correct problems with the ComboBox implementation. You will also need my typemap.i.patch file (forthcoming) in order for the void * typemap to work properly (I think). Roy _______________________________________________ wxruby-users mailing list wxruby-users@rubyforge.org http://rubyforge.org/mailman/listinfo/wxruby-users

Attached are patches to make the textctrl.rb sample work fully. Note the new %directorargout typemap I added to fix wxWindows calling into SWIG. I don''t know if this can fix all such problems or not. Let me know what you think. I also rubified the sample a little bit more. Also, note that I fix the << operator so the function can be called. I hope I did this right. Roy

Forwarded from the SWIG mailing list, so we have a copy in our archives. Kevin -------- Original Message -------- Subject: Re: [Swig] Re: Object return problem Date: Sun, 09 Oct 2005 18:31:40 -0400 From: Kevin Smith <wxruby@qualitycode.com> To: Charlie Savage <cfis@interserv.com> CC: Swig@cs.uchicago.edu References: <4347277E.1030700@mindspring.com>

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.1//EN" "http://www.w3.org/TR/xhtml11/DTD/xhtml11.dtd"> <html xmlns="http://www.w3.org/1999/xhtml"> <head><meta http-equiv="content-type" content="text/html; charset=utf-8" /><style type="text/css"><!-- #msg dl { border: 1px #006 solid; background: #369; padding:

Hi I'm using the llvm-c wrapper, and trying to build some recursive types (using released 2.3). I get an assert on trying to create a second opaque pointer type after refining a first. The first time through creating an opaque pointer type, a new type is created and returned from PointerType::get, but the second time, ValueType (the opaque type) is found in the PointerTypes map, which seems