Re: [fltk.bugs] [HIGH] STR #2632: FL_EXPORT that should not exists

2011-05-19 Thread Mathieu Peyréga 

DO NOT REPLY TO THIS MESSAGE.  INSTEAD, POST ANY RESPONSES TO THE LINK BELOW.

[STR New]

Link: http://www.fltk.org/str.php?L2632
Version: 1.3-current


Hello,

- here is the error message answer :

To get the following messages (localized in french), i only reverted to
genuine file for Fl_Multiline_Output, let the modification for other files
(but messages are then very similar)

1>PopUpMessageWidgetClass.obj : error LNK2001: symbole externe non résolu
"__declspec(dllimport) public: virtual __thiscall
Fl_Multiline_Output::~Fl_Multiline_Output(void)"
(__imp_??1Fl_Multiline_Output@@UAE@XZ)

1>PopUpMessageWidgetClass.obj : error LNK2001: symbole externe non résolu
"__declspec(dllimport) public: __thiscall
Fl_Multiline_Output::Fl_Multiline_Output(int,int,int,int,char const *)"
(__imp_??0Fl_Multiline_Output@@QAE@PBD@Z)

1>C:\Users\Mathieu\Documents\Devel\Incub\..\bin\Incub\vc100\Win32\Release\LibIncubSoftware.dll
: fatal error LNK1120: 2 externes non résolus

without FL_EXPORT : it builds (and works) perfectly

- here is the answer about the program structure :

the build errors occurs in a custom DLL that is using FLTK code, and were
i am deriving from FLTK classes  and exporting my custom widgets.

- here is the answer about problematic classes :

Fl_Input_Int
Fl_Multiline_Input
Fl_Multiline_Output

These derived class are located inside my custom DLL, and exported from
this DLL (with mecanism similar to the FL_EXPORT one in order to have a
compile time define switching from dllimport to dllexport)

I am having those error messages only for class were i'm deriving other
classes from.

- I'll try today to build a small solution to demonstrate and share the
behaviour.


Link: http://www.fltk.org/str.php?L2632
Version: 1.3-current

___
fltk-bugs mailing list
fltk-bugs@easysw.com
http://lists.easysw.com/mailman/listinfo/fltk-bugs

Re: [fltk.bugs] [LOW] STR #2637: Valgrind errors and their fixes for 1.3.0rc5

2011-05-19 Thread Csaba Biegl 

DO NOT REPLY TO THIS MESSAGE.  INSTEAD, POST ANY RESPONSES TO THE LINK BELOW.

[STR New]

Link: http://www.fltk.org/str.php?L2637
Version: 1.3.0


Note that the patch for screen_xywh.cxx only addresses the line that gave
an error message under Valgrind. Probably the same logic should apply to
the vertical DPI calculation. This currently polls the SAME (= fl screen)
screen in each iteration of the loop. It did not generate an error
message, so I did not change it -- just in case there is some non-obvious
reason for doing it this way...

As for why now: Because today was the day when I messed up something so
badly in my own app that I needed Valgrind:-).


Link: http://www.fltk.org/str.php?L2637
Version: 1.3.0

___
fltk-bugs mailing list
fltk-bugs@easysw.com
http://lists.easysw.com/mailman/listinfo/fltk-bugs

Re: [fltk.bugs] [LOW] STR #2637: Valgrind errors and their fixes for 1.3.0rc5

2011-05-19 Thread Greg Ercolano 

DO NOT REPLY TO THIS MESSAGE.  INSTEAD, POST ANY RESPONSES TO THE LINK BELOW.

[STR New]

Link: http://www.fltk.org/str.php?L2637
Version: 1.3.0


Nice!

When committed, please comment the "+15)&(~7)" magic numbers, which appear
to be a way to allocate in 8 byte chunks, if I read that correctly..?


Link: http://www.fltk.org/str.php?L2637
Version: 1.3.0

___
fltk-bugs mailing list
fltk-bugs@easysw.com
http://lists.easysw.com/mailman/listinfo/fltk-bugs

Re: [fltk.bugs] [LOW] STR #2637: Valgrind errors and their fixes for 1.3.0rc5

2011-05-19 Thread Matthias Melcher 

DO NOT REPLY TO THIS MESSAGE.  INSTEAD, POST ANY RESPONSES TO THE LINK BELOW.

[STR New]

Link: http://www.fltk.org/str.php?L2637
Version: 1.3.0


**SIGH**. I am really really glad that someone's using Valgrid to fix
issues. I think it's a magic tool ;-) - but why in the world do those
results always come exactly 5 days *after* the Release Candidate?

I think we can put that in the final as well. 

COnsidering though that this is the fourth change for RC5, we may be
looking at an RC6 :-/


Link: http://www.fltk.org/str.php?L2637
Version: 1.3.0

___
fltk-bugs mailing list
fltk-bugs@easysw.com
http://lists.easysw.com/mailman/listinfo/fltk-bugs

[fltk.bugs] [LOW] STR #2637: Valgrind errors and their fixes for 1.3.0rc5

2011-05-19 Thread Csaba Biegl 
DO NOT REPLY TO THIS MESSAGE.  INSTEAD, POST ANY RESPONSES TO THE LINK BELOW.

[STR New]

Link: http://www.fltk.org/str.php?L2637
Version: 1.3.0


The attached patch fixes two errors reported by Valgrind. These are on a 64
bit Linux system.

First error probably only comes up in certain dual monitor configurations.
Observed on a system using the proprietary Nvidia driver.

Second error may be 64 bit specific?



==7833== Invalid read of size 4
==7833==at 0x510047: screen_init() (screen_xywh.cxx:155)
==7833==by 0x510249: Fl::screen_xywh(int&, int&, int&, int&, int, int)
(screen_xywh.cxx:197)
==7833==by 0x4F8FFE: Fl_X::make_xid(Fl_Window*, XVisualInfo*, unsigned
long) (Fl_x.cxx:1594)
==7833==by 0x4FA1D2: Fl_Window::show() (Fl_x.cxx:1876)
==7833==by 0x4C6A97: Fl_Double_Window::show()
(Fl_Double_Window.cxx:68)
==7833==by 0x4D9E4F: Fl_Overlay_Window::show()
(Fl_Overlay_Window.cxx:46)
==7833==by 0x4F3C8C: Fl_Window::show(int, char**) (Fl_arg.cxx:361)
==7833==by 0x47BC74: Fl_Utilities::showWindow(Fl_Window*, int, char**,
char const* const*) (Fl_Utilities.cxx:142)
==7833==by 0x412945: DsViewerApp::run(int, char**)
(DsViewerApp.cxx:72)
==7833==by 0x4113FB: main (DSviewer.cxx:52)
==7833==  Address 0x7a9bf60 is not stack'd, malloc'd or (recently) free'd
==7833== 
==7833== Invalid read of size 1
==7833==at 0x501FAD: xrgb_converter(unsigned char const*, unsigned
char*, int, int) (fl_draw_image.cxx:323)
==7833==by 0x502A9F: innards(unsigned char const*, int, int, int, int,
int, int, int, void (*)(void*, int, int, int, unsigned char*), v
==7833==by 0x502D14: Fl_Xlib_Graphics_Driver::draw_image(unsigned char
const*, int, int, int, int, int, int) (fl_draw_image.cxx:547)
==7833==by 0x4C7117: fl_draw_image(unsigned char const*, int, int,
int, int, int, int) (fl_draw.H:667)
==7833==by 0x4CACA9: alpha_blend(Fl_RGB_Image*, int, int, int, int,
int, int) (Fl_Image.cxx:431)
==7833==by 0x4CB2BE: Fl_Xlib_Graphics_Driver::draw(Fl_RGB_Image*, int,
int, int, int, int, int) (Fl_Image.cxx:573)
==7833==by 0x4CAD2A: Fl_RGB_Image::draw(int, int, int, int, int, int)
(Fl_Image.cxx:438)
==7833==by 0x4BDE35: Fl_Image::draw(int, int) (Fl_Image.H:160)
==7833==by 0x5004D8: fl_draw(char const*, int, int, int, int, unsigned
int, void (*)(char const*, int, int, int), Fl_Image*, int) (fl_d
==7833==by 0x500D0C: fl_draw(char const*, int, int, int, int, unsigned
int, Fl_Image*, int) (fl_draw.cxx:386)
==7833==by 0x5065F0: fl_normal_label(Fl_Label const*, int, int, int,
int, unsigned int) (fl_labeltype.cxx:46)
==7833==by 0x50674C: Fl_Label::draw(int, int, int, int, unsigned int)
const (fl_labeltype.cxx:88)
==7833==  Address 0x7d2bbc0 is 0 bytes after a block of size 720 alloc'd
==7833==at 0x4C283AD: malloc (in
/usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==7833==by 0x52600C: operator new(unsigned long) (in
/home/cbiegl/EDAS-sw/ds-devel-3.0/bin/Debug/Linux/x86_64/DSviewer)
==7833==by 0x5251A8: operator new[](unsigned long) (in
/home/cbiegl/EDAS-sw/ds-devel-3.0/bin/Debug/Linux/x86_64/DSviewer)
==7833==by 0x4CA9F9: alpha_blend(Fl_RGB_Image*, int, int, int, int,
int, int) (Fl_Image.cxx:387)
==7833==by 0x4CB2BE: Fl_Xlib_Graphics_Driver::draw(Fl_RGB_Image*, int,
int, int, int, int, int) (Fl_Image.cxx:573)
==7833==by 0x4CAD2A: Fl_RGB_Image::draw(int, int, int, int, int, int)
(Fl_Image.cxx:438)
==7833==by 0x4BDE35: Fl_Image::draw(int, int) (Fl_Image.H:160)
==7833==by 0x5004D8: fl_draw(char const*, int, int, int, int, unsigned
int, void (*)(char const*, int, int, int), Fl_Image*, int) (fl_d
==7833==by 0x500D0C: fl_draw(char const*, int, int, int, int, unsigned
int, Fl_Image*, int) (fl_draw.cxx:386)
==7833==by 0x5065F0: fl_normal_label(Fl_Label const*, int, int, int,
int, unsigned int) (fl_labeltype.cxx:46)
==7833==by 0x50674C: Fl_Label::draw(int, int, int, int, unsigned int)
const (fl_labeltype.cxx:88)
==7833==by 0x506A18: Fl_Widget::draw_label(int, int, int, int,
unsigned int) const (fl_labeltype.cxx:134)


Link: http://www.fltk.org/str.php?L2637
Version: 1.3.0diff -ur fltk-1.3.0rc5-orig//src/Fl_Image.cxx 
fltk-1.3.0rc5-new//src/Fl_Image.cxx
--- fltk-1.3.0rc5-orig//src/Fl_Image.cxx2011-04-20 09:01:04.0 
-0500
+++ fltk-1.3.0rc5-new//src/Fl_Image.cxx 2011-05-19 14:10:16.0 -0500
@@ -384,7 +384,7 @@
   uchar *srcptr = (uchar*)img->array + cy * ld + cx * img->d();
   int srcskip = ld - img->d() * W;
 
-  uchar *dst = new uchar[W * H * 3];
+  uchar *dst = new uchar[((W * H * 3) + 15) & (~7)];
   uchar *dstptr = dst;
 
   fl_read_image(dst, X, Y, W, H, 0);
diff -ur fltk-1.3.0rc5-orig//src/screen_xywh.cxx 
fltk-1.3.0rc5-new//src/screen_xywh.cxx
--- fltk-1.3.0rc5-orig//src/screen_xywh.cxx 2011-04-22 12:30:37.0 
-0500
+++ fltk-1.3.0rc5-new//src/screen_xywh.cxx  2011-05-19 14:10:16.0 
-0500
@@ -151,9 +151,12 @@
   if (XineramaIsActive(fl_display)) {
 screens = XineramaQueryScreens

Re: [fltk.bugs] [HIGH] STR #2634: fl_help_view bug fixes and new features

2011-05-19 Thread kdi...@live.ru 
Domingo Alvarez Duarte wrote:

> Ideally I would prefer to have people from Dillo to jump in and contribute
> with their html engine that is a lot superior.

I have my own `html engine' written on c++ but it is not optimized.
Maybe in a future I'll deciding it (there're 5500 strings of code nearly  in 
the 2 classes) and I'll offering to integrate the `engine' into FLTK OR yet 
something...

___
fltk-bugs mailing list
fltk-bugs@easysw.com
http://lists.easysw.com/mailman/listinfo/fltk-bugs