* Remove ambiguous code that caused problems with understanding how stuff
worked.
* Implement DerivedType::typeIsRefined right. Add a big comment talking
about it.
* The removeUserFromConcrete method already checks to see if the type is
not abstract, so no need to duplicate the test all over this code
This checkin makes all of the type parsing related failures work. All
Assembler Regression tests now pass!
llvm-svn: 2127
That was due to the StructType object not getting called multiple times
to remove itself from the ATU list for an object, because it stopped
getting queried due to this wrong test
* Improve debug output when DEBUG_MERGE_TYPES is on
llvm-svn: 2124
* Refactor a bunch of code. Types now only remove one subelement at a time
when they are told that they have changed
* Improve debugging output, add more assertions...
* FIX Bugs:
* test/Regression/Assembler/2002-04-04-PureVirtMethCall.ll
* test/Regression/Assembler/2002-04-04-PureVirtMethCall2.ll
* The change to fix these bugs was the to ValTypeBase::refineAbstractType
method. Basically we #if 0'd out the chunk of code there and make the
table get reorganized EVEN IF the type has the same pointer. Merging
opportunities were being missed!
llvm-svn: 2121
* Fix a bug in DerivedType::refineAbstractTypeTo that caused
test/Regression/Assembler/2002-04-05-TypeParsing.ll to fail.
The added comment describes the fix
llvm-svn: 2118
* Eliminate by inlining the old newTH, newTH, and TypeDone functions
* OPAQUE is now just a token that gets returned by the lexer, not a type
Parser now creates type, not lexer
llvm-svn: 2104
causing a problem before because global values with incoming edges didn't
copy the incoming edges to the node they were being merged from, causing
the poolalloc pass to die.
llvm-svn: 2079
* Implement a new form of node folding to catch cases missed in Addtree
* Add removeIndistinguishableNodePairs to merge calls (todo) and globals
llvm-svn: 2062
* Add more debugging output
* Fix problems refering to wrong versions of various graphs
* Build the pool descriptor map!
* Clear the nodemapping map after building pool descriptor to avoid assert
* Transform the NEw function body, not the old one...
* Matrix.ll now works!
llvm-svn: 2058
* TransformFunctionInfo now has call field form field mapping. May be
removed in the future.
* Moved the computation of "Scalars" into transformFunctionBody so
transformFunction didn't have to recompute it.
* Implement the node mapping calculation in preparation to calculate
PoolDescriptors to pass to transformFunctionBody
* Print out the node mapping [it looks right!]
* Other minor changes
llvm-svn: 2056
* Converts malloc instructions to poolalloc
* Converts free instructions to poolfree
* Convert calls to call the new cloned hacked up versions
Note that this does not modify hacked up stuff yet, just the top level
function
llvm-svn: 2052