source: trunk/abcl/src/org/armedbear/lisp/compiler-pass2.lisp

Revision Log Mode:


Legend:

Added
Modified
Copied or renamed
Diff Rev Age Author Log Message
(edit) @11795   14 years Mark Evenson Ant-based build works again, fixing error introduced in r11792
(edit) @11793   14 years vvoutilainen Combine load/store resolvers.
(edit) @11792   14 years vvoutilainen Little combination fix for pool-long and pool-double.
(edit) @11783   14 years ehuelsmann Fix fasl reader special bindings leak. * Bind the …
(edit) @11782   14 years ehuelsmann Remove code duplication.
(edit) @11781   14 years ehuelsmann Instead of interning symbols over and over, use the ones already interned.
(edit) @11780   14 years ehuelsmann Clean up the known symbols cache: * instead of returning only the …
(edit) @11764   14 years ehuelsmann Don't use the implementation details in WITH-COMPILATION-UNIT to …
(edit) @11763   14 years ehuelsmann Fix COMPILE and COMPILE-FILE secondary and tertiary return values in …
(edit) @11722   14 years ehuelsmann Make sure to create Fixnums when the LispInteger/Bignum? created is in …
(edit) @11719   14 years ehuelsmann Fix COMPILE called inside COMPILE-FILE: COMPILE doesn't compile to a …
(edit) @11709   14 years ehuelsmann Remove unnecessary saving/restoring of dynamic environment from …
(edit) @11708   14 years ehuelsmann Save/Restore? dynamic environment for functions which take a special …
(edit) @11707   14 years ehuelsmann r11706 follow up: Jump over the exception handler.
(edit) @11706   14 years ehuelsmann Protect lastSpecialBinding from unexpected PROGV body termination. See …
(edit) @11705   14 years ehuelsmann Make MULTIPLE-VALUE-BIND restore its specials environment upon …
(edit) @11704   14 years ehuelsmann Restore lastSpecialBinding upon exception-driven LET/LET* block exit.
(edit) @11703   14 years ehuelsmann DOTIMES loop variable efficiency: if the derived type of the limit var …
(edit) @11652   14 years ehuelsmann Check cast takes a class name as its argument, not a class reference. …
(edit) @11651   14 years ehuelsmann Followup from last commit: add the Symbol type to …
(edit) @11650   14 years ehuelsmann Generate Symbol-typed fields if we expect to be loading off one. Also: …
(edit) @11649   14 years ehuelsmann Indentation: make opcode-enumerations look like elsewhere.
(edit) @11648   14 years ehuelsmann Strict checking of representations delivered vs requested - inspired …
(edit) @11646   14 years ehuelsmann Fix compilation of Maxima: the wrong representation was returned. …
(edit) @11642   14 years ehuelsmann Remove commented-out code ("emit-unbox-*") which is superseeded by …
(edit) @11641   14 years ehuelsmann Integer bounds derivation for MIN and MAX. Compilation of (*) -> 1 …
(edit) @11640   14 years ehuelsmann Type derivation for multi-argument +/-/*; Fixed type-arguments for …
(edit) @11638   14 years ehuelsmann Fix P2-MIN/MAX backquoting/unquoting error.
(edit) @11637   14 years ehuelsmann Extend EMIT-DUP to be able to duplicate past the top-most stack value.
(edit) @11636   14 years ehuelsmann Generalize code generation paths for :INT/:LONG representations in …
(edit) @11634   14 years ehuelsmann Better in-lining compilation of MIN and MAX.
(edit) @11633   14 years ehuelsmann Clean up DERIVE-TYPE-{MIN,MAX,ASH} using the new …
(edit) @11632   14 years ehuelsmann Make our (static) fields FINAL, as we only write to them once: during …
(edit) @11631   14 years ehuelsmann Fix DERIVE-VARIABLE-REPRESENTATION in case the LIMIT-VARIABLE resolves …
(edit) @11630   14 years ehuelsmann Optimize (and fix) CONVERT-REPRESENTATION for the case of 2 eql …
(edit) @11624   14 years ehuelsmann Wider use of CONVERT-REPRESENTATION shows an issue: …
(edit) @11622   14 years ehuelsmann Eliminate NEW-FIXNUM and EMIT-FIXNUM-INIT in favor of …
(edit) @11621   14 years ehuelsmann Implement P2-COMPILAND-UNBOX-VARIABLE in terms of new primitives. …
(edit) @11620   14 years ehuelsmann Kill long code repetitions in COMPILE-VAR-REF and P2-SETQ - making the …
(edit) @11619   14 years ehuelsmann Reduce code duplication: move variable representation deduction to …
(edit) @11617   14 years ehuelsmann More CONVERT-REPRESENTATIONs.
(edit) @11614   14 years ehuelsmann Inline all compiled subtractions instead of only the 2 and 3 argument …
(edit) @11613   14 years ehuelsmann Fix #'+ compilation with FEWER than 2 arguments.
(edit) @11611   14 years ehuelsmann Clean up COMPILE-CONSTANT: there's no reason to cast from one type to …
(edit) @11610   14 years ehuelsmann Add 2 used - but not enabled - opcodes.
(edit) @11609   14 years ehuelsmann Don't enumerate representation conversion inline: we have a generic …
(edit) @11608   14 years ehuelsmann Support inline comparisons for many types of compiler types (including …
(edit) @11607   14 years ehuelsmann Implement inline float and double calculations for P2-TIMES. Cleanup …
(edit) @11604   14 years ehuelsmann Smarter type derivation: start *using* the float and double storage …
(edit) @11603   14 years ehuelsmann Use generic representation conversions instead of enumerating in line.
(edit) @11602   14 years ehuelsmann Implement generic type-representation derivations and conversions; …
(edit) @11601   14 years ehuelsmann Only compile one or the other argument to an :int.
(edit) @11600   14 years ehuelsmann Add more opcodes to the list; update stack effect information.
(edit) @11594   14 years ehuelsmann Make DERIVE-TYPE support SINGLE-FLOAT and DOUBLE-FLOAT.
(edit) @11593   14 years ehuelsmann Optimize unboxing of booleans.
(edit) @11592   14 years ehuelsmann Generic representation conversion (from one JVM type to another) and …
(edit) @11588   14 years ehuelsmann Optimize MIN/MAX inline calculations: with the right stack use, we can …
(edit) @11584   14 years ehuelsmann Handle both 'long' as well as 'double' argument and return types as …
(edit) @11583   14 years ehuelsmann Revert change to p2-minus: we can't do this without adding a new …
(edit) @11582   14 years ehuelsmann Use additional opcodes: don't store the "obvious" constants in the …
(edit) @11580   14 years ehuelsmann Commit some of the changes required for FLOAT and DOUBLE support …
(edit) @11578   14 years ehuelsmann Miscelanious: - Merge [within p2-plus] (fixnum-type-p type2) case …
(edit) @11571   14 years ehuelsmann Implement some building blocks for compilation of float math to byte …
(edit) @11569   14 years ehuelsmann Eliminate float-serializing ambiguities: if you need a float/double, …
(edit) @11564   14 years ehuelsmann Fix compiler issue found by compiling AP5: Instead of calculating the …
(edit) @11554   14 years ehuelsmann Followup to the introduction of compile-forms-and-maybe-emit-clear-values.
(edit) @11545   14 years vvoutilainen Another small helper for p2-plus/minus.
(edit) @11544   14 years ehuelsmann Add bounds checking and prepare for support for 'wide' instruction prefix.
(edit) @11543   14 years vvoutilainen Little helper for p2-plus/minus/times.
(edit) @11542   14 years vvoutilainen Helper macro for declare-* functions that use hashtables.
(edit) @11541   14 years vvoutilainen Look, I can do conditionals in the middle of a backquote-form! :) …
(edit) @11540   14 years vvoutilainen Helper macro for defining inlining functions.
(edit) @11534   14 years vvoutilainen Helper function for creating a new fixnum and emitting dup immediately …
(edit) @11533   14 years vvoutilainen Helper macro for p2-test-minusp/plusp/zerop/oddp/evenp.
(edit) @11532   14 years vvoutilainen Helper function for p2-test-minusp/plusp/zerop.
(edit) @11531   14 years vvoutilainen In preparation for further refactorings, a tiny change to p2-plus.
(edit) @11525   14 years vvoutilainen Helper function for fixnum initializations.
(edit) @11524   14 years vvoutilainen Tiny helper for checking that class file is loadable.
(edit) @11523   14 years vvoutilainen Macro for temp files in p2-flet/labels-process-compiland. At the same …
(edit) @11522   14 years vvoutilainen Helper function for p2-flet-process-compiland and …
(edit) @11521   14 years vvoutilainen Helper macro for derive-compiler type, when checking for parameter …
(edit) @11520   14 years ehuelsmann Make the compiler recognize subtypes while compiling THE forms. Note: …
(edit) @11519   14 years vvoutilainen Remove code repetition in the beginning of p2-compiland.
(edit) @11518   14 years ehuelsmann Eliminate dead code.
(edit) @11517   14 years ehuelsmann Replace Java type indicator with pre-existing constant with the same …
(edit) @11509   14 years vvoutilainen Remove Cons special case from car/cdr, it doesn't seem to make a big …
(edit) @11508   14 years ehuelsmann Cleanup: `, -> (nothing); (EMIT 'LABEL ...) -> (LABEL ...)
(edit) @11507   14 years ehuelsmann Emit the most efficient ALOAD and ASTORE instructions.
(edit) @11506   14 years ehuelsmann Consolidate code-path with existing function EMIT-CALL-EXECUTE.
(edit) @11505   14 years ehuelsmann Consolidate code-path with existing function PROCESS-ARGS.
(edit) @11504   14 years vvoutilainen Further combine car/cdr.
(edit) @11503   14 years vvoutilainen Small refactoring combining common bits in p2-car/p2-cdr. This is in …
(edit) @11502   14 years ehuelsmann Introduce mnemonic for (emit 'getstatic ....) to load a symbol naming …
(edit) @11501   14 years ehuelsmann Efficiency/correctness of generated code: choose opcodes based on …
(edit) @11500   14 years ehuelsmann EMIT-LOAD-VARIABLE was actually renamed to EMIT-LOAD-LOCAL-VARIABLE.
(edit) @11498   14 years ehuelsmann Follow up to r11494: +lisp-stream-abstract-class+ doesn't exist; …
(edit) @11497   14 years ehuelsmann Factor out some code to load a local variable onto the stack.
(edit) @11494   14 years Mark Evenson (dmiles) Fix SLIME compatibilit via STREAM type check value.
(edit) @11493   14 years ehuelsmann Fix READ-BYTE.ERROR.5 (interpreted version) which was broken since r11482.
(edit) @11482   14 years ehuelsmann Followup to r11467: STREAM wasn't in the original list of types. …
Note: See TracRevisionLog for help on using the revision log.