gnash-commit
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Gnash-commit] gnash ChangeLog server/sprite_instance.cpp


From: Sandro Santilli
Subject: [Gnash-commit] gnash ChangeLog server/sprite_instance.cpp
Date: Fri, 07 Sep 2007 13:57:47 +0000

CVSROOT:        /sources/gnash
Module name:    gnash
Changes by:     Sandro Santilli <strk>  07/09/07 13:57:47

Modified files:
        .              : ChangeLog 
        server         : sprite_instance.cpp 

Log message:
                * server/sprite_instance.cpp (restoreDisplayList): don't use
                  a copy of the display_list to optimize the call to 
set_invalidated.
                  Removing characters from the copy would shift unloaded 
characters
                  to a different depth, which in turn would fail an assertion
                  (which is *needed* or set_invalidated would not include 
invalidated
                  bounds of just-unloaded characters).
                  BTW, this suggests me we need a lot of tests for inv.bounds. 
as the
                  latest changes for depth swap on unload did likely break a few
                  cases.

CVSWeb URLs:
http://cvs.savannah.gnu.org/viewcvs/gnash/ChangeLog?cvsroot=gnash&r1=1.4250&r2=1.4251
http://cvs.savannah.gnu.org/viewcvs/gnash/server/sprite_instance.cpp?cvsroot=gnash&r1=1.332&r2=1.333

Patches:
Index: ChangeLog
===================================================================
RCS file: /sources/gnash/gnash/ChangeLog,v
retrieving revision 1.4250
retrieving revision 1.4251
diff -u -b -r1.4250 -r1.4251
--- ChangeLog   7 Sep 2007 12:45:55 -0000       1.4250
+++ ChangeLog   7 Sep 2007 13:57:46 -0000       1.4251
@@ -1,5 +1,17 @@
 2007-09-07 Sandro Santilli <address@hidden>
 
+       * server/sprite_instance.cpp (restoreDisplayList): don't use
+         a copy of the display_list to optimize the call to set_invalidated.
+         Removing characters from the copy would shift unloaded characters
+         to a different depth, which in turn would fail an assertion
+         (which is *needed* or set_invalidated would not include invalidated
+         bounds of just-unloaded characters).
+         BTW, this suggests me we need a lot of tests for inv.bounds. as the
+         latest changes for depth swap on unload did likely break a few
+         cases.
+
+2007-09-07 Sandro Santilli <address@hidden>
+
        * server/sprite_instance.cpp: disable the too many debugging
          prints left over from the big action order commit.
 

Index: server/sprite_instance.cpp
===================================================================
RCS file: /sources/gnash/gnash/server/sprite_instance.cpp,v
retrieving revision 1.332
retrieving revision 1.333
diff -u -b -r1.332 -r1.333
--- server/sprite_instance.cpp  7 Sep 2007 12:45:55 -0000       1.332
+++ server/sprite_instance.cpp  7 Sep 2007 13:57:47 -0000       1.333
@@ -2388,15 +2388,8 @@
        //  2.3 Remove all non-script-referencable instances, suboptimal!
 
        // TODO: try to optize by avoid calling set_invalidated
-       DisplayList newList = m_display_list;
-       assert( newList == m_display_list );
-       newList.reset(*m_def, tgtFrame, true);
-       if ( newList != m_display_list )
-       {
-               //cout << "Modified DisplayList: " << newList << endl;
                set_invalidated();
-               m_display_list = newList;
-       }
+       m_display_list.reset(*m_def, tgtFrame, true);
 
        // 3. Execute all displaylist tags from first to target frame, with
        //    target frame tag execution including ACTION tags




reply via email to

[Prev in Thread] Current Thread [Next in Thread]