gcc4.7 can't build codeblocks

[MortenMacFly]

@ Morten. Only with the patches from here + -fpermissive in the compiler vars.

What patches are left? The help-plugin is up-to-date. Is there something else missing?

[xunxun]

@ Morten. Only with the patches from here + -fpermissive in the compiler vars.

What patches are left? The help-plugin is up-to-date. Is there something else missing?

You can see

http://forums.next.codeblocks.org/index.php/topic,16039.msg109945.html#msg109945

[MortenMacFly]

http://forums.next.codeblocks.org/index.php/topic,16039.msg109945.html#msg109945

Thats what I meant with:

The help-plugin is up-to-date.

So this is in SVN already, isn't it? So nothing to do here.

[reckless]

patched before trunk update so didnt know. Ok nothing to report then.

[Jenna]

-fpermissive can't fix all things (especially in contrib plugins), maybe you can try SharkCZ's patches.

Hope Jens will fix soon.

Long time ago:
After many fixes from killerbot to silent warnings on gcc4.7 and a  pm from him, it looks like I found a solution for the "-fpermissive"-desaster:

Code Select Expand

commit e70c8a008439072718760876b048deae1ec74c27
Author: Jens Lody <jens@codeblocks.org>
Date:   Mon Aug 6 23:30:49 2012 +0200

    * gcc 4.7 fixes

Index: src/include/scripting/bindings/sc_base_types.h
===================================================================
--- src/include/scripting/bindings/sc_base_types.h
+++ src/include/scripting/bindings/sc_base_types.h
@@ -33,6 +33,7 @@ DECLARE_INSTANCE_TYPE(wxPoint);
DECLARE_INSTANCE_TYPE(wxSize);
DECLARE_INSTANCE_TYPE(wxString);

+using SqPlus::GetTypeName;
// C::B primitives and types
DECLARE_INSTANCE_TYPE(ConfigManager);
DECLARE_INSTANCE_TYPE(EditorManager);
@@ -59,6 +60,7 @@ namespace SqPlus \
     inline T Get(TypeWrapper<T>,HSQUIRRELVM v,int idx) { SQInteger i; SQPLUS_CHECK_GET(sq_getinteger(v,idx,&i)); return (T)i; } \
}

+using SqPlus::Push;
DECLARE_ENUM_TYPE(wxPathFormat);
DECLARE_ENUM_TYPE(wxPathNormalize);
DECLARE_ENUM_TYPE(PrintColourMode);
Index: src/plugins/contrib/codesnippets/codesnippets.cpp
===================================================================
--- src/plugins/contrib/codesnippets/codesnippets.cpp
+++ src/plugins/contrib/codesnippets/codesnippets.cpp
@@ -93,7 +93,6 @@ CodeSnippets::CodeSnippets()
     m_bMouseLeftKeyDown = false;
     m_bMouseIsDragging = false;
     m_bDragCursorOn = false;
-    m_pDragCursor = false;
     m_MouseDownX = m_MouseDownY = 0;
     m_MouseUpX = m_MouseUpY = 0;
     #if !wxCHECK_VERSION(2, 8, 12)
Index: src/plugins/contrib/help_plugin/defs.h
===================================================================
--- src/plugins/contrib/help_plugin/defs.h
+++ src/plugins/contrib/help_plugin/defs.h
@@ -297,7 +297,7 @@ class QMap : public std::map<Key, T>

         int remove(const Key &k)
         {
-            return erase(k);
+            return this->erase(k);
         }
};

Index: src/plugins/debuggergdb/gdb_driver.cpp
===================================================================
--- src/plugins/debuggergdb/gdb_driver.cpp
+++ src/plugins/debuggergdb/gdb_driver.cpp
@@ -82,6 +82,7 @@ static wxRegEx reInferiorExited2(wxT("^\\[[Ii]nferior[ \\t].+[ \\t]exited[ \\t]w
                                  wxRE_EXTENDED);

// scripting support
+using SqPlus::Push;
DECLARE_INSTANCE_TYPE(GDB_driver);

GDB_driver::GDB_driver(DebuggerGDB* plugin)



Any feedback welcome!

[MortenMacFly]

After many fixes from killerbot to silent warnings on gcc4.7 and a  pm from him, it looks like I found a solution for the "-fpermissive"-desaster:

Is this the same you posted in the other thread?

[Jenna]

After many fixes from killerbot to silent warnings on gcc4.7 and a  pm from him, it looks like I found a solution for the "-fpermissive"-desaster:

Is this the same you posted in the other thread?

Yes, but this one is for the public to test it.

It works fine here on fedora 64-bit and gcc4.7, but I did not test it on other platforms and with older compilers.
I don't think it can break anything, just want to be sure, before committing it.

[killerbot]

see my comment in the other thread

EDIT : which not everyone can see :
duplicating here :

works fine for me too, however I would change one thing to the patch:

you just removed :     m_pDragCursor = false;

Now the member is not init-ed ==>

Code Select Expand


   m_pDragCursor = 0;


[Jenna]

you just removed :     m_pDragCursor = false;

Now the member is not init-ed ==>

Code Select Expand


    m_pDragCursor = 0;


It was initialised at the end of the constructor anyway (that's why I just removed the wrong initialisation):

Code Select Expand

    m_pDragCursor = new wxCursor(wxCURSOR_HAND);

To be sure it's 0 in case of a not working initialisation, we can preset it to 0.

[killerbot]

no you are absolutely correct, no need to set it. I overlooked that line.

And to be really nice : all of them could have been done in the initializer list :-)
But that's another story.

Let's give people time to object to the patch till this evening ?

[Jenna]

And to be really nice : all of them could have been done in the initializer list :-)
But that's another story.

And we could do this in many places I guess.

Let's give people time to object to the patch till this evening ?

Yes, and I will try to compile it with an older gcc (4.2 if I remember correctly 4.4, I do not use my windows installation on my laptop often) on windows.

[Alpha]

Patch builds and runs with no apparent issues (GCC 4.6, Windows XP).

[Jenna]

I committed at to trunk last evening  (svn r8206).
C::B should compile without -fpermissive now.

[MortenMacFly]

C::B should compile without -fpermissive now.

Cools stuff. Now all thats needed on windows is a stable GCC 4.7.x compiler... :-)

Maybe we should nag TDM all together... Harhar...

[killerbot]

I already kindly did last week 
Maybe we should send him some nice beers ...