@szukw000 Winfried, you wrote that you "found where the bug resides. See the attachment". This attachment is pretty useless (at least for me) without the code and the image file that causes it. Note however that we can't test and debug your code, even if the final error happens inside FLTK code.

3. src/drivers/GDIFl_GDI_Graphics_Driver_image.cxx
=====================================================
 // uchar *newarray = new uchar[(Bpr+pad)*h];
  uchar *newarray = new uchar[512 + (Bpr+pad)*h]; 

With this change the program now works without crash.

Where did you get the constant 512 from? Adding an arbitrary size to an array is not a fix if you (or we, the FLTK Team) don't know why the array is accessed out of bounds.

How to proceed:

1. Please provide an image that fails (post it here).
2. Please provide a test scenario that we can use to test the image.
3. If possible, use one of the FLTK demo programs, e.g. test/pixmap_browser.
4. If (3.) is not possible, please provide a small demo program that shows the issue (post it here).
5. Please use the latest FLTK version (either from Git or a snapshot).
6. Please answer the questions I asked above about which FLTK version you are using.

I may repeat myself, but this is essential: the code you refer to is not build system dependent, it is used on all Windows systems. Hence I assume that the real cause of the issue lies somewhere else in your code and the effects you see are caused by random program execution addresses which can exhibit the error if you build the program with VS/Nmake versus MinGW.

Without a working example and test image we need to close this issue.

—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you are subscribed to this thread.Message ID: <fltk/fltk/issues/496/1257831122@github.com>