- Sep 15, 1992
-
-
ram authored
function.
-
- Aug 01, 1992
-
-
ram authored
compiler.
-
- Dec 11, 1991
-
-
ram authored
:OPTIONAL entries, etc.) Changed to set NODE-TAIL-P to T instead of a tail set.
-
- Dec 08, 1991
-
-
ram authored
to prevent type inference anomalies.
-
- Nov 09, 1991
-
-
ram authored
no successors.
-
- Sep 23, 1991
-
-
ram authored
closure variable have been deleted after pre-analysis.) Moved variable deletion out of GET-LAMBDA-ENVIRONMENT into a new function REINIT-LAMBDA-ENVIRONMENT, which is called by ENVIRONMENT-ANALYZE. This function cleans up environment analysis assumptions that were invalidated by optimization.
-
- Sep 05, 1991
-
-
ram authored
for NLX-INFO.
-
- Aug 28, 1991
-
-
ram authored
away after environment pre-analysis.
-
- Mar 18, 1991
-
-
ram authored
cleanup generation.
-
- Mar 11, 1991
-
-
ram authored
-
- Feb 20, 1991
-
-
ram authored
-
- Jan 02, 1991
-
-
ram authored
deleted, it is not totally simple to find the home lambda. Moved unreferenced variable noting into NOTE-UNREFERENCED-VARS, since this code is now shared with function deleting code in ir1util.
-
- Nov 26, 1990
-
-
ram authored
-
- Oct 17, 1990
-
-
ram authored
components.
-
- Oct 06, 1990
-
-
ram authored
full calls.
-
- Aug 23, 1990
-
-
ram authored
"does not return" notes.
-
- Jul 23, 1990
-
-
ram authored
Moved function-does-not-return test here from ir1final. Numerous changes due to NLX changes associated with new lexenv cleanup stuff. Also, ENTRY-EXITS is now a list of EXITs, so we don't have to worry about continuations are deleted, and can find NLX points without iterating over all blocks.
-
- May 30, 1990
-
-
cvs2git authored
-
- Apr 24, 1990
-
-
ram authored
merge them into the same component when there are closure references.
-
- Apr 13, 1990
-
-
ram authored
so that this information is easily avalible for later use (and isn't trashed by TR return deletion.
-
- Mar 08, 1990
-
-
ram authored
%NLX-ENTRY function when we delete the escape function. Formerly there was no use to represent the type uncertainty of the thrown value, which could cause LTN to get confused, perhaps choosing a specialized representation when the value could in fact be of any type.
-
- Mar 05, 1990
-
-
cvs2git authored
-
- Feb 07, 1990
-
-
ram authored
FUNCALL so that the cleanup function reference will be recognized as a local call. (Which is must be, since there is no XEP for cleanup functions.)
-
- Feb 06, 1990
-
-
wlott authored
-