| | [ Return to Bugs & Features | Roadmap 1.1 | SVN ⇄ GIT ]
STR #723
| Application: | FLTK Library |
|---|
| Status: | 1 - Closed w/Resolution |
|---|
| Priority: | 2 - Low, e.g. a documentation error or undocumented side-effect |
|---|
| Scope: | 2 - Specific to an operating system |
|---|
| Subsystem: | WIN32 |
|---|
| Summary: | GDI leak in Win32 -- suggested fix |
|---|
| Version: | 1.1.6 |
|---|
| Created By: | greg.ercolano |
|---|
| Assigned To: | mike |
|---|
| Fix Version: | 1.1.7 (SVN: v4041) |
|---|
| Update Notification: | |
|---|
Trouble Report Files:
No files
Trouble Report Comments:
| |
| #1 | greg.ercolano
10:45 Feb 13, 2005 |
| WRT the thread on fltk.bugs "fltk-1.1.6 continuous GDI memory leak"
In his case he noticed the 'ask.exe' demo leaking a GDI
each time a window is created and destroyed.
It would appear that i->region created in Fl_Widget::damage()
is indeed not getting destroyed when Fl_Window destructs via Fl_Window::hide().
The problem /appears/ to be here in Fl.cxx..
When I add the noted line and rebuild, ask.exe no longer leaks a GDI.
In 1.1.6 Fl.cxx, the code for Fl_Window::hide(),
in the first "#ifdef WIN32", the section reads:
----
#ifdef WIN32
if (ip->private_dc) ReleaseDC(ip->xid,ip->private_dc);
if (ip->xid == fl_window && fl_gc) {
ReleaseDC(fl_window, fl_gc);
fl_window = (HWND)-1;
fl_gc = 0;
}
if ( ip->region ) XDestroyRegion(ip->region); <-- ADD THIS TO FIX
#elif defined(__APPLE_QD__)
if ( ip->xid == fl_window )
fl_window = 0;
#elif defined(__APPLE_QUARTZ__)
Fl_X::q_release_context(ip);
if ( ip->xid == fl_window )
fl_window = 0;
#else
if (ip->region) XDestroyRegion(ip->region);
#endif
----
| |
| |
| #2 | mike
03:12 Feb 14, 2005 |
| Fixed in Subversion repository. | |
[ Return to Bugs & Features ]
|
| |
