Bayse統計をつかった統計的デバッグ支援ツール

BayseRuby.patch

diff --git a/Makefile.in b/Makefile.in
index 6b12187..9a6e1b8 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -59,7 +59,7 @@ CPPFLAGS = @CPPFLAGS@ $(INCFLAGS)
 LDFLAGS = @STATIC@ $(CFLAGS) @LDFLAGS@
 EXTLDFLAGS = 
 XLDFLAGS = @XLDFLAGS@ $(EXTLDFLAGS)
-EXTLIBS = 
+EXTLIBS = -lgsl -lgslcblas
 LIBS = @LIBS@ $(EXTLIBS)
 MISSING = @LIBOBJS@ @ALLOCA@
 LDSHARED = @LIBRUBY_LDSHARED@
diff --git a/eval.c b/eval.c
index ddbe31a..ae6c0ea 100644
--- a/eval.c
+++ b/eval.c
@@ -1174,6 +1174,8 @@ rb_f_method_name(void)
     }
 }
 
+void ruby_bayse_define_class(void);
+
 void
 Init_eval(void)
 {
@@ -1213,3 +1215,121 @@ Init_eval(void)
     OBJ_TAINT(exception_error);
     OBJ_FREEZE(exception_error);
 }
+
+VALUE bayse_hash_sfn_to_ler;
+VALUE cBayse;
+
+#include <gsl/gsl_math.h>
+#include <gsl/gsl_sf.h>
+#include <gsl/gsl_rng.h>
+#include <gsl/gsl_randist.h>
+
+double 
+kl_divergence(double a, double b,
+              double c, double d)
+{
+    double val = gsl_sf_beta(a, b) / gsl_sf_beta(c, d);
+    double ret =
+        log(val) + ((c - a) * (gsl_sf_psi_int(c) - gsl_sf_psi_int(c + d)))
+        + ((d - b) * (gsl_sf_psi_int(d) - gsl_sf_psi_int(c + d)));
+    return ret;
+}
+
+VALUE 
+bayse_func_m_kl_divergence(VALUE self,
+                           VALUE a,
+                           VALUE b,
+                           VALUE c,
+                           VALUE d)
+{
+    double ret;
+    ret = kl_divergence(NUM2DBL(a), NUM2DBL(b), NUM2DBL(c), NUM2DBL(d));
+    return DBL2NUM(ret);
+}
+
+VALUE 
+bayse_func_m_results(VALUE self)
+{
+    return bayse_hash_sfn_to_ler;
+}
+
+VALUE
+bayse_func_m_clearlog(VALUE self)
+{
+    bayse_hash_sfn_to_ler = Qnil;
+    bayse_hash_sfn_to_ler = rb_hash_new();
+    rb_gc_register_mark_object(bayse_hash_sfn_to_ler);
+}
+
+VALUE 
+bayse_func_m_start_logging(VALUE self)
+{
+    ID prof_id = 0;
+    prof_id = rb_intern("BayseProf");
+}
+
+VALUE 
+bayse_func_m_stop_logging(VALUE self)
+{
+    ID prof_id = 0;
+    prof_id = rb_intern("BayseProf");
+}
+
+/*
+  a = Bayse.results
+  a.each do |bayse|
+    # do something
+  end
+*/
+
+void
+Init_Bayse(void)
+{
+    bayse_hash_sfn_to_ler = rb_hash_new();
+    rb_gc_register_mark_object(bayse_hash_sfn_to_ler);    
+    
+    cBayse = rb_define_class("Bayse", rb_cObject);
+    
+    rb_define_singleton_method(cBayse, "start", bayse_func_m_start_logging, 0);
+    rb_define_singleton_method(cBayse, "stop", bayse_func_m_stop_logging, 0);
+    rb_define_singleton_method(cBayse, "results", bayse_func_m_results, 0);
+    rb_define_singleton_method(cBayse, "clearlog", bayse_func_m_clearlog, 0);
+    rb_define_singleton_method(cBayse, "kldivergence", bayse_func_m_kl_divergence, 4);
+}
+
+/* kazushi addition */
+VALUE
+ruby_logging_branch(int unless, rb_control_frame_t *reg_cfp)
+{
+    static ID prof_id = 0;
+    VALUE sfn_to_ler, conds, ret;
+    
+    prof_id = rb_intern("BayseProf");
+
+    /* rb_const_defは遅いhashを引いている */
+    /* 例えば：ProfileをONにすると、定義されているbranchunlessを全部俺のブランチアクセスバイトコードに変える */
+    /* 2種類ある：すでに生成されているバイトコードを書き換える方法と、これから生成されるやるも書き換える。 */
+
+    if (rb_const_defined(rb_cObject, prof_id) == Qfalse ||
+        rb_const_get(rb_cObject, prof_id) == Qfalse) {
+        return;
+    } else {
+        sfn_to_ler = rb_hash_lookup(bayse_hash_sfn_to_ler, reg_cfp->iseq->filename);
+        if (sfn_to_ler == Qnil) {
+            sfn_to_ler = rb_hash_new();
+            rb_gc_register_mark_object(sfn_to_ler);
+            rb_hash_aset(bayse_hash_sfn_to_ler, reg_cfp->iseq->filename, sfn_to_ler);
+        }
+
+        conds = rb_hash_lookup(sfn_to_ler, LONG2FIX(rb_vm_get_sourceline(reg_cfp)));
+        if (conds == Qnil) {
+            conds = rb_ary_new();      
+            rb_gc_register_mark_object(conds);            
+            rb_hash_aset(sfn_to_ler, LONG2FIX(rb_vm_get_sourceline(reg_cfp)), conds);
+        }
+        
+        ret = unless ? INT2FIX(0) : INT2FIX(1);
+        rb_ary_push(conds, ret);
+    }
+}
+
diff --git a/inits.c b/inits.c
index 73b9eb4..bf81e6d 100644
--- a/inits.c
+++ b/inits.c
@@ -60,5 +60,6 @@ rb_call_inits(void)
     CALL(Rational);
     CALL(Complex);
     CALL(version);
+    CALL(Bayse);
 }
 #undef CALL
diff --git a/insns.def b/insns.def
index 2b4c3db..87149eb 100644
--- a/insns.def
+++ b/insns.def
@@ -1164,9 +1164,12 @@ branchif
 ()
 {
     if (RTEST(val)) {
+    ruby_logging_branch(0, GET_CFP());
 	RUBY_VM_CHECK_INTS();
 	JUMP(dst);
-    }
+    } else  {
+    ruby_logging_branch(1, GET_CFP());
+    }     
 }
 
 /**
@@ -1180,9 +1183,12 @@ branchunless
 (VALUE val)
 ()
 {
-    if (!RTEST(val)) {
+    if (!RTEST(val)) {   
+    ruby_logging_branch(1, GET_CFP());
 	RUBY_VM_CHECK_INTS();
 	JUMP(dst);
+    } else {
+    ruby_logging_branch(0, GET_CFP());
     }
 }
 
diff --git a/vm_core.h b/vm_core.h
index d44bd25..ba11a5d 100644
--- a/vm_core.h
+++ b/vm_core.h
@@ -749,6 +749,9 @@ extern VALUE rb_get_coverages(void);
 extern void rb_set_coverages(VALUE);
 extern void rb_reset_coverages(void);
 
+extern VALUE ruby_logging_branch(int unless, 
+                                 rb_control_frame_t *reg_cfp);
+
 #if defined __GNUC__ && __GNUC__ >= 4
 #pragma GCC visibility pop
 #endif