- Feb 12, 1997
-
-
dtc authored
-
- Jan 18, 1997
-
-
ram authored
-
- Oct 31, 1994
-
-
ram authored
-
- Feb 11, 1994
-
-
cvs2git authored
-
- Aug 25, 1993
-
-
ram authored
-
- Jan 13, 1993
-
-
cvs2git authored
-
- Jun 03, 1992
-
-
ram authored
happen, since we do DFO before constraint if it is needed, and that should flush any blocks with no predecessors (or that are otherwise unreachable), but it is happening...
-
- Mar 21, 1992
-
-
wlott authored
-
- Dec 08, 1991
-
-
ram authored
prevent type inference anomalies.
-
- Feb 20, 1991
-
-
ram authored
-
- Jul 23, 1990
-
-
ram authored
-
- Jul 06, 1990
-
-
ram authored
another due to an EQL constraint.
-
- Jun 17, 1990
- Jun 14, 1990
-
-
ram authored
constraint assignment pass, but this didn't work, since new constraints (that we also kill) could be added by other blocks or in subsequent invocations of constraint propagation. Instead, we make the Block-Kill be a list of the lambda-vars killed, and dynamically determine the kill set in FLOW-PROPAGATE-CONSTRAINTS. Also, changed stuff to reinitialize the IN and OUT sets on repeated invocations. I believe that this is no more correct, but it does make the flow analysis process more understandable, since the size of these sets will only increase. Previously these sets could start oversize and decrease if we somehow lost information between this invocation and the previous one.
-
- Jun 11, 1990
-
-
ram authored
When we substitute a constant due to an EQL constraint, punt of the loop so that we don't lose due to assuming that the ref's leaf is a lambda-var. Only change a IF test to T when there is no type constraint on the continuation (e.g. ENDP).
-
- Jun 06, 1990
-
-
ram authored
-
- May 30, 1990
-
-
cvs2git authored
-
- Mar 05, 1990
-
-
cvs2git authored
-
- Feb 06, 1990
-
-
wlott authored
-