summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* vm/asm: new prof instruction.Kaz Kylheku2018-03-273-33/+52
| | | | | | | | | | * share/txr/stdlib/asm.tl (prof): New opcode. (op-prof): New opcode class. * vm.c (vm_prof_callback, vm_prof): New static functions. (vm_execute): Handle PROF opcode via vm_prof. * vmop.h: Regenerated.
* eval: refactor op_prof to support reuse.Kaz Kylheku2018-03-272-2/+19
| | | | | | | | | | | * eval.c (prof_call): New function, contents based on op_prof. (struct prof_ctx): New struct type. (op_prof_callback): New static function. (op_prof): Reduced to call to prof_call, passing context through to callback which performs the evaluation that is timed. * eval.h (prof_call): Declared.
* compiler: implement sys:switch special op.Kaz Kylheku2018-03-271-0/+32
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle sys:switch via comp-switch method. (comp-switch): New method.
* compiler: bugfix in sys:fbind op.Kaz Kylheku2018-03-271-1/+2
| | | | | | | | | | | | When a (flet ...) form is compiled, the function arguments of the lambda are being bound in the same frame and clashing against the variables being bound by the construct. * share/txr/stdlib/compiler.tl (compiler comp-fbind): For the non-recursive case, insert the dummy empty environment eenv, and compile the forms in that environment. This raises them up to the appropriate display level without affecting what is visible in their scope.
* compiler: recognize call and apply forms.Kaz Kylheku2018-03-271-11/+23
| | | | | | | | | | * share/txr/stdlib/compiler.tl (compiler): New slots, gcallop and callop. (compiler comp-fun-form): Restructured to handle apply and call forms, turning them into better code, exploiting the call, gcall, apply and gapply instructions. (compiler comp-call): Take opcode argument so apply calls can be handled.
* regression: fix broken tagbody.Kaz Kylheku2018-03-261-19/+5
| | | | | | | | | | | | | | | | | | | | | | | | | commit 87caead8269055b4791de53be0e03afab01f1dd4, subject "macros: expand declined form in outer env" broke tagbody. The tagbody implementation uses a dubious trick of setting up local macros which return the original form, as a way of prevening an outer scoped macro from expanding the form. The above commit specifically changes the behavior in such a way that this strategy is nullified. However, the macro fallback feature introduced by the above commit is exactly what tagbody needs! * share/txr/stdlib/tagbody.tl (tagbody): Simplify the treatment of (go label) thanks to new behavior in macro expander. We no longer need an extra sys:expand pass to allow inner tagbodies to expand their go forms, and intercept any unexpanded ones. We have a simple go macrolet which performs the expansiion when the label is recognized, or else returns the form to decline expansion. The macro expander will then fall back by trying the macro in the next outer scope. Thus, every (go label) is resolved in the tagbody to which label belongs, or else lands into the top-level go macro which diagnoses undefined labels.
* compiler: implement sys:catch special op.Kaz Kylheku2018-03-261-0/+46
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle sys:catch via comp-catch method. (comp-catch): New method.
* vm/asm: new swtch instruction.Kaz Kylheku2018-03-263-18/+80
| | | | | | | | | | | | | * share/txr/stdlib/asm.tl (backpatch-low16, backpatch-high16): New struct types. (%backpatch-low16%, %backpatch-high16%): New global variables. (swtch): New opcode. (op-swtch): New opcode class. * vm.c (vm_swtch): New static function. (vm_execute): Handle SWTCH opcode via vm_swtch. * vmop.h: Regenerated.
* compiler: implement handler-bind special op.Kaz Kylheku2018-03-261-0/+14
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle defmacro via comp-handler-bind method. (comp-handler-bind): New method.
* vm: bugfix: faulty memcpy in closure.Kaz Kylheku2018-03-261-1/+1
| | | | | | * vm.c (vm_make_closure): Add missing scale factor to memcpy. This is the memcpy which relocates display frame contents from the stack to the frame.
* compiler/vm: implement sys:abscond-from special form.Kaz Kylheku2018-03-264-14/+30
| | | | | | | | | | | | | | | * share/txr/stdlib/asm.tl (abscsr): New instruction. (op-abscsr): New opcode class, derived from op-retsr. * share/txr/stdlib/compiler.tl: Handle sys:abscond-from via comp-return-from method. (compiler comp-return-from): Handle sys:abscond-from by switching to abscsr opcode instead of ret pseudo-op. * vm.c (vm_abscsr): New static function. (vm_execute): Dispatch ABSCSR opcode. * vmop.h: Regenerated.
* compiler: implement block* special form.Kaz Kylheku2018-03-261-3/+6
| | | | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Route block* to same helper method as block. (compiler comp-block): Handle block* also by compiling the name form and using the resulting value as the name operand in the block instruction.
* compiler: implement defsymacro special op.Kaz Kylheku2018-03-252-0/+14
| | | | | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle defsymacro via expand-defsymacro expander. (expand-defsymacro): New function. * eval.c (rt_defsymacro): New static function. (eval_init): register sys:rt-defsymacro intrinsic.
* compiler: implement defmacro special op.Kaz Kylheku2018-03-251-0/+14
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle defmacro via expand-defmacro expander. (expand-defmacro): New function.
* eval: remove hack of macro deffers evaled on expansion.Kaz Kylheku2018-03-255-24/+28
| | | | | | | | | | | | | | | | | | * eval.c (do_expand): When a defmacro or defsymacro form is traversed, do not evaluate it, except in backward compatibility mode. Unfortunately, this breaks some code. * tests/011/macros-1.txr: A defmacro form has to be wrapped in macro-time. * tests/011/macros-2.txr: Likewise. * tests/011/mandel.txr: Likewise. * tests/012/man-or-boy.tl (defun-cbn): This macro generates a progn which which expects that a defmacro form will come into effect for the subsequent lambda in the same form. We must wrap it in macro-time to make this happen now.
* compiler: implement defun special op.Kaz Kylheku2018-03-251-0/+20
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle defun via expand-defun expander. (expand-defun): New function.
* eval/compiler: run-time support for compiled defun.Kaz Kylheku2018-03-251-8/+18
| | | | | | | * eval.c (rt_defun, rt_defmacro): New static functions. (op_defun): Use static functions. (eval_init): Register sys:rt-defun and sys:rt-defmacro intrinsics.
* compiler: implement defvarl special op.Kaz Kylheku2018-03-251-0/+9
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle defvarl via expand-defvarl expander. (expand-defvarl): New function.
* eval/compiler: run-time support for compact defvarl.Kaz Kylheku2018-03-251-5/+20
| | | | | | * eval.c (rt_defvarl): New static function. (op_defvarl): Simplified using rt_defvarl. (eval_init): Register sys:rt-defvarl.
* compiler: implement sys:setqf special op.Kaz Kylheku2018-03-241-0/+17
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle sys:setqf via comp-setqf. (compiler comp-setqf): New method.
* compiler: setq bug: wrong return for globals.Kaz Kylheku2018-03-251-2/+2
| | | | | | | | | | * share/txr/stdlib/compiler.tl (compiler comp-setq): The output register of the returned frag must be that of the value calculating frag, not vloc, because in the dynamic variable case, vloc holds the d-reg with the variable's name, not the v-reg that receives its value. (compiler compl-lisp1-setq): Fix same bug which got propagated by copy-and-paste coding.
* compiler: implement sys:lisp1-setq special op.Kaz Kylheku2018-03-241-0/+17
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle sys:lisp1-setq via comp-lisp1-setq. (compiler comp-lisp1-setq): New method.
* compiler: implement dwim special op.Kaz Kylheku2018-03-241-0/+7
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle dwim case via comp-dwim. (compiler comp-dwim): New method.
* compiler: implement sys:lisp1-value special formKaz Kylheku2018-03-241-3/+31
| | | | | | | | | | | | * share/txr/stdlib/compiler.tl (vbinding, fbinding): New structs subtyping binding. (sys:env lookup-lisp1): New method. (sys:env extend-var, sys:env extend-var*, sys:env extend-fun): Instantiate a vbinding or fbinding as appropriate. This lets us tell whether we have a function or variable binding after doing a successful lookup-lisp1. (compiler compile): Add sys:lisp1-value case. (compiler comp-lisp1-value): New method.
* compiler: specially compile (call ...) forms.Kaz Kylheku2018-03-241-1/+13
| | | | | | | * share/txr/stdlib/compiler.tl (compiler comp-fun-form): Add a caseq form to handle certain top-level functions specially. Add a case for the call function, handled by comp-call. (compiler comp-call): New method.
* compiler: rename comp-call.Kaz Kylheku2018-03-241-2/+2
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Call comp-fun-form instead of comp-call. (compiler comp-call): Rename to comp-fun-form.
* compiler: implement fun special formKaz Kylheku2018-03-241-0/+8
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Route fun to comp-fun. (compiler comp-fun): New method.
* compiler: add fbind and lbind special formsKaz Kylheku2018-03-241-0/+32
| | | | | | | | This supports labels and flet. * share/txr/stdlib/compiler.tl (compiler compile): Route fbind and lbind to comp-fbind method. (compiler comp-fbind): New method.
* compiler: fix wrong frame level in lexical functions.Kaz Kylheku2018-03-241-1/+1
| | | | | | * share/txr/stdlib/compiler.tl (sys:env extend-fun): The v register's level is two less than the frame level. Add missing ppred call; extend-var has it already.
* compiler: fix lexical function call.Kaz Kylheku2018-03-241-2/+2
| | | | | * share/txr/stdlib/compiler.tl (compiler comp-call): Pass the location, not the binding itself, to comp-call-impl.
* compiler: implement unwind-protect.Kaz Kylheku2018-03-241-0/+25
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Wire in unwind-protect case via comp-unwind-protect method. (compiler comp-unwind-protect): New method.
* compiler: implement mac-param-bind.Kaz Kylheku2018-03-241-0/+12
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Wire in tree-bind case via comp-mac-param-bind method. (compiler comp-mac-param-bind): New method.
* compiler: hoist quoting out of bind expander.Kaz Kylheku2018-03-231-8/+12
| | | | | | | | | | | * share/txr/stdlib/compiler.tl (expand-bind-mac-params): Take the context form and error forms as separate arguments instead of calculating one from the other. Moreover, they are no longer assumed to be objects to be quoted and inserted but rather expressions to be substituted into the code directly. This gives the caller flexibility to make them calculated. (compiler comp-tree-bind, compiler comp-tree-case): Make the compensating adjustments to preserve the behavior.
* asm: report instruction count in disassembly.Kaz Kylheku2018-03-231-2/+7
| | | | | | * share/txr/stdlib/asm.tl (assembler dis-listing): Return the number of instructions that were traversed. (disassemble-c-d): Print count after instruction listing.
* tree-bind et al: don't retain full form for errors.Kaz Kylheku2018-03-231-2/+3
| | | | | | | | | * share/txr/stdlib/compiler.tl (expand-bind-mac-params): We don't need the full form for error reporting, because only the leading symbol is reported in diagnostics. We do need the full form if the :form parameter is used. If :form is not used, then a compiled function need not carry the entire form in its data table.
* compiler: avoid using var symbol.Kaz Kylheku2018-03-231-5/+5
| | | | | | | | | var is sys:var when we're working in the system package. This sometimes confuses the code walker, because (sys:var X) is the implementation of the @X notation. * share/txr/stdlib/compiler.tl (expand-bind-mac-params): Use sym instead of var.
* compiler: auto-load sys:bind-mac-error.Kaz Kylheku2018-03-231-1/+1
| | | | | | | This is run-time support needed by compiled tree-bind code. * lisplib.c (compiler_set_entries): Add bind-mac-error to sys package auto-load list.
* compiler: constant-optimize prog1Kaz Kylheku2018-03-231-1/+2
| | | | | | | | * share/txr/stdlib/compiler.tl (compiler comp-prog1): prog1 already uses progn. When using progn, though, it should append the nil form to the tail, otherwise progn will spare the last from from being eliminated, leaving an unused constant in the data table.
* compiler: basic optimization in progn.Kaz Kylheku2018-03-231-2/+6
| | | | | | * share/txr/stdlib/compiler.tl (compiler comp-progn): Eliminate any form which is a constant or a symbol, unless it appears in the last position.
* compiler: optimizations in cond and if.Kaz Kylheku2018-03-231-34/+58
| | | | | | | | | | | | We optimize based on test being constant expressions. * share/txr/stdlib/compiler.tl (compiler comp-cond): Whenever a test is a constant null, we need not evaluate the forms in that case; the entire case basically disappears so we compile it to nil. We also need not consider any cases after the first case whose test is a constant true. (compiler comp-if): Optimize all situations when the test is constant.
* compiler: last cond case: don't jmp to end.Kaz Kylheku2018-03-231-5/+11
| | | | | | | * share/txr/stdlib/compiler.tl (compiler comp-cond): Don't generate the jmp instruction at the end of the last case. Why do I bother; a simple peephole optimizer will eliminate these. I want a little better code, now! That's why.
* vm: change encoding of getv, setv and related.Kaz Kylheku2018-03-232-12/+12
| | | | | | | | | | | | | | | | | | | | | | | In most compiler uses of bindv, getv and setv, the operand which names the dynamic symbol is the data table: a d0-dff register. The source or destination register of the transfer could be anything. Therefore, the existing encoding is suboptimal: the symbol is being put into the full sized operand field of the instruction, and the source or destination register into the small, ten-bit extra field in the upper half. This breaks on cases like (set x y) where x is a deeply nested lexical variable and y is a dynamic variable. Let's reverse the encoding. * share/txr/stdlib/asm.tl (op-getv, op-setv): Reverse the operands. All derived opcodes follow this since they reuse the code via inheritance. * vm.c (vm_get_binding): Fetch the symbol from the small operand field rather than the main operand field. (vm_getsym, vm_getbind, vm_setsym, vm_bindv): Pull the destination or source from the main operand of the instruction rather than the small opreand.
* compiler: streamline tree-caseKaz Kylheku2018-03-231-20/+9
| | | | | | | | | | | | | | | | | | | | | | | | | The setting of an error variable upon destructuring mismatch is not useful; it just takes extra instructions to check for a colon return out of the body or that variable being set. Let's have expand-bind-mac-params generate code which returns that symbol itself of assigning to it. The caller can then specify : for the strict parameter. A destructuring mismatch turns into a : return value, exactly the same as the value which indicates "fall through to next case". * share/txr/stdlib/compiler.tl (compiler comp-tree-case): Don't generate the err-var; remove all references to it. Pass : to expand-bind-mac-params as the strict parameter, rather than err-var.sym. Generate much simplified code after the cfrag: just test for a colon and continue through to the next case or else branch to the end. In the last case, the fall through path precipiates to the end, so we insert an instruction to clobber the : in the return register with a nil. (expand-bind-mac-params): Eliminate assignments to strict; it doesn't function as a variable any longer. Return that symbol in the return-from forms.
* compiler: bug: dynamic var assignment.Kaz Kylheku2018-03-231-1/+1
| | | | | * share/txr/stdlib/compiler.tl (compiler comp-setq): Fix reversed setv instruction operands.
* compiler: implement tree-case.Kaz Kylheku2018-03-221-0/+52
| | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Add tree-case case, handled via comp-tree-case method. (compiler comp-tree-case): New method.
* compiler: implement tree-bind special op.Kaz Kylheku2018-03-221-0/+134
| | | | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Wire in tree-bind case via comp-tree-bind method. (compiler comp-tree-bind): New method. (expand-bind-mac-parse-params, expand-bind-mac-params): New functions.
* vm: initialize vd->funvec to nil.Kaz Kylheku2018-03-221-0/+1
| | | | | | | | * vm.c (vm_make_desc): Initialize vd->funvec to nil so that when the object is created, it doesn't have a garbage field. However, there is no risk here that the field will be traversed by the garbage collector, since immediately after allocating the objct, we initialize the fields.
* vm: bugfix: cannot access vec during gc.Kaz Kylheku2018-03-221-6/+8
| | | | | | | | | | | | | | Let's have a field in struct vm_desc that keeps track of the ftab size. The problem is when we do length_vec(vd->funvec) in vm_desc_mark, we are in the middle of garbage collection. The object's tag has a reachable bit which blows up the type check. * vm.c (struct vm_desc): New member, ftsz. (vm_make_desc): Store the length of the function table into ftsz. (vm_desc_mark): Use the stored length as the loop bound; do not access vd->funvec.
* gc bugs: incorrect struct init function assignments.Kaz Kylheku2018-03-221-2/+2
| | | | | | | | | | * struct.c (struct_set_initfun, struct_set_postinitfun): Replace incorrect direct assignments with set macro. This manifested itself as corruption. I ran into a situation in which the postinitfun of a struct type was prematurely reclaimed and the heap object was re-used for something else wreaking havoc on the postinit call when the struct was instantiated.
* compiler: implement return-from and return.Kaz Kylheku2018-03-221-0/+18
| | | | | | | * share/txr/stdlib/compiler.tl (compiler compile): Handle return and return-from special forms. (compiler comp-return-from, compiler comp-return): New methods.
generated by cgit v1.2.3 (git 2.25.1) at 2025-03-10 14:36:23 +0000