This lastest release seems to have broken something.
I just opened a .wxs file and the .cpp and .h files were changed. Not a problem.
However, the following code change seems odd:
Old code
TextCtrlPointA_X = new wxTextCtrl(PanelSideBar, ID_TEXTCTRL1, _("108.0"), wxDefaultPosition, wxSize(50,-1), 0, wxDefaultValidator, _T("ID_TEXTCTRL1"));
FlexGridSizer5->Add(TextCtrlPointA_X, 0, wxALL|wxALIGN_CENTER_HORIZONTAL|wxALIGN_CENTER_VERTICAL, 0);
New code
TextCtrlPointA_X = new wxTextCtrl(PanelSideBar, ID_TEXTCTRL1, _("108.0"), wxDefaultPosition, wxSize(50,-1), 0, wxDefaultValidator, _T("ID_TEXTCTRL1"));
FlexGridSizer5->Add(0, 0, wxALL|wxALIGN_CENTER_HORIZONTAL|wxALIGN_CENTER_VERTICAL, 0);
Notice that the Add now passes a '0' instead of the pointer 'TextCtrlPointA_X'? Is this correct??
Worse though is that now the code does not compile.
PanelMain->SetSizer(FlexGridSizer2);
was changed to PanelMain->SetSizer(PanelSideBar);
and the compiler complains that SetSizer cannot accept a wxPanel, but it can accept a wxSizer.
I'm using WinXP SP2, wxwidgets 2.8.4, minGW.
I hope the above helps track down the problem.
In the mean time, I need to rollback to the previous nightly build.
Please let me know if you require more information.
Thanks.
I reported what I thought was a general problem in the previous version but it happens to be very code-specific, rather unusual, and not only on today's build but the last two as well (I didn't test older ones). Under Windows 2000 SP4, using gcc (either 3.4 or 4.2), compile a single file with this exact code:
struct Flags {
int bits;
};
struct Base {
struct BaseFlags : Flags {};
static const BaseFlags DATA;
};
struct Derived : Base {
struct Flags : BaseFlags {};
};
char dummy[Base::DATA.bits]; // Error: array bound is not an integer constant.
Make a cosmetic change and compile again. Repeat this four or five times and Code::Blocks will suddenly abort and close without warning. This does not happen with other code: make a meaningful change (comment out something for example) and Code::Blocks has no problem re-compiling until the cows come home. But this specific code plays some weird voodoo on CB. I don't know what to make of it.