Mirror of https://github.com/eclipse-cdt/cdt

John Camelon 57d668240a Patch for Andrew Niefer. ... For normal lookups in the symbol table, a HashMap is faster than the tree map, but for prefix lookups the TreeMap is faster. So we are now using the HashMap for normal parses, and we use the TreeMap in the parse mode used by content assist. Note that with these changes the results returned by the IASTNode.lookup function used by content assist are now in predictable order: they are first sorted in the order of the scopes visited during the lookup and then they are sorted alphabetically Core: Modified symbol table constructor to take a ParseMode as a parameter. Modified symbol table to use a TreeMap instead of HashMap when ParseMode is COMPLETION_PARSE. Modified ASTNode.lookup to throw ASTNotImplementedException if called when ParseMode is not CONTEXTUAL_PARSE. Core.tests: Moved testBug48307_FriendFunction_1 & testBug48307_FriendFunction_2 to ContextualParseTest. Updated ContextualParseTest now that the order of prefix lookup results is predictable. UI: Updated CompletionEngine to catch ASTNotImplementedException from IASTNode.lookup.		2004-01-15 23:06:58 +00:00
build	- cleanup UI	2004-01-13 18:54:39 +00:00
core	Patch for Andrew Niefer.	2004-01-15 23:06:58 +00:00
debug	Fix for bug 48682: IThread.getBreakpoints() stubbed out.	2004-01-15 22:09:35 +00:00
doc	Fix version number.	2004-01-05 20:51:56 +00:00
launch/org.eclipse.cdt.launch	removed warning	2004-01-13 18:44:43 +00:00
releng	Make failure more visible.	2004-01-14 22:03:23 +00:00