+2001-12-31 Havoc Pennington <hp@pobox.com>
+
+ * gtk/gtktextbtree.c (gtk_text_btree_node_view_check_consistency):
+ fix an incorrect assertion that the "valid" flag is always correct
+ in a node and add explanatory comment. This keeps me from
+ reproducing #59101, but I'm remembering that 59101 was actually
+ about incorrect pixel totals, so I'm not sure 59101 is actually
+ fixed. May just have stopped happening due to some change in
+ how C-k works.
+
2001-12-31 Havoc Pennington <hp@pobox.com>
* gtk/gtktextview.c (gtk_text_view_start_selection_drag): extend
+2001-12-31 Havoc Pennington <hp@pobox.com>
+
+ * gtk/gtktextbtree.c (gtk_text_btree_node_view_check_consistency):
+ fix an incorrect assertion that the "valid" flag is always correct
+ in a node and add explanatory comment. This keeps me from
+ reproducing #59101, but I'm remembering that 59101 was actually
+ about incorrect pixel totals, so I'm not sure 59101 is actually
+ fixed. May just have stopped happening due to some change in
+ how C-k works.
+
2001-12-31 Havoc Pennington <hp@pobox.com>
* gtk/gtktextview.c (gtk_text_view_start_selection_drag): extend
+2001-12-31 Havoc Pennington <hp@pobox.com>
+
+ * gtk/gtktextbtree.c (gtk_text_btree_node_view_check_consistency):
+ fix an incorrect assertion that the "valid" flag is always correct
+ in a node and add explanatory comment. This keeps me from
+ reproducing #59101, but I'm remembering that 59101 was actually
+ about incorrect pixel totals, so I'm not sure 59101 is actually
+ fixed. May just have stopped happening due to some change in
+ how C-k works.
+
2001-12-31 Havoc Pennington <hp@pobox.com>
* gtk/gtktextview.c (gtk_text_view_start_selection_drag): extend
+2001-12-31 Havoc Pennington <hp@pobox.com>
+
+ * gtk/gtktextbtree.c (gtk_text_btree_node_view_check_consistency):
+ fix an incorrect assertion that the "valid" flag is always correct
+ in a node and add explanatory comment. This keeps me from
+ reproducing #59101, but I'm remembering that 59101 was actually
+ about incorrect pixel totals, so I'm not sure 59101 is actually
+ fixed. May just have stopped happening due to some change in
+ how C-k works.
+
2001-12-31 Havoc Pennington <hp@pobox.com>
* gtk/gtktextview.c (gtk_text_view_start_selection_drag): extend
+2001-12-31 Havoc Pennington <hp@pobox.com>
+
+ * gtk/gtktextbtree.c (gtk_text_btree_node_view_check_consistency):
+ fix an incorrect assertion that the "valid" flag is always correct
+ in a node and add explanatory comment. This keeps me from
+ reproducing #59101, but I'm remembering that 59101 was actually
+ about incorrect pixel totals, so I'm not sure 59101 is actually
+ fixed. May just have stopped happening due to some change in
+ how C-k works.
+
2001-12-31 Havoc Pennington <hp@pobox.com>
* gtk/gtktextview.c (gtk_text_view_start_selection_drag): extend
+2001-12-31 Havoc Pennington <hp@pobox.com>
+
+ * gtk/gtktextbtree.c (gtk_text_btree_node_view_check_consistency):
+ fix an incorrect assertion that the "valid" flag is always correct
+ in a node and add explanatory comment. This keeps me from
+ reproducing #59101, but I'm remembering that 59101 was actually
+ about incorrect pixel totals, so I'm not sure 59101 is actually
+ fixed. May just have stopped happening due to some change in
+ how C-k works.
+
2001-12-31 Havoc Pennington <hp@pobox.com>
* gtk/gtktextview.c (gtk_text_view_start_selection_drag): extend
+2001-12-31 Havoc Pennington <hp@pobox.com>
+
+ * gtk/gtktextbtree.c (gtk_text_btree_node_view_check_consistency):
+ fix an incorrect assertion that the "valid" flag is always correct
+ in a node and add explanatory comment. This keeps me from
+ reproducing #59101, but I'm remembering that 59101 was actually
+ about incorrect pixel totals, so I'm not sure 59101 is actually
+ fixed. May just have stopped happening due to some change in
+ how C-k works.
+
2001-12-31 Havoc Pennington <hp@pobox.com>
* gtk/gtktextview.c (gtk_text_view_start_selection_drag): extend
gtk_text_btree_node_compute_view_aggregates (node, nd->view_id,
&width, &height, &valid);
+
+ /* valid aggregate not checked the same as width/height, because on
+ * btree rebalance we can have invalid nodes where all lines below
+ * them are actually valid, due to moving lines around between
+ * nodes.
+ *
+ * The guarantee is that if there are invalid lines the node is
+ * invalid - we don't guarantee that if the node is invalid there
+ * are invalid lines.
+ */
+
if (nd->width != width ||
nd->height != height ||
- !nd->valid != !valid)
+ (nd->valid && !valid))
{
g_error ("Node aggregates for view %p are invalid:\n"
"Are (%d,%d,%s), should be (%d,%d,%s)",