Assorted clarifications.

python · markshannon · May 7, 2021 · Apr 19, 2021 · Apr 21, 2021 · Apr 21, 2021
commit 43c199dd8581cf095462fe47a0a2896103d79775
diff --git a/Misc/NEWS.d/next/Core and Builtins/2021-04-30-15-48-36.bpo-40222.j3VxeX.rst b/Misc/NEWS.d/next/Core and Builtins/2021-04-30-15-48-36.bpo-40222.j3VxeX.rst
@@ -1,6 +1,7 @@
 "Zero cost" exception handling.
 
-* Uses a lookup table to determine how to handle exceptions
+* Uses a lookup table to determine how to handle exceptions.
 * Removes SETUP_FINALLY and POP_TOP block instructions, eliminating the runtime overhead of try statements.
-* Reduces the size of the fram object by about 60%
+* Reduces the size of the frame object by about 60%.
 
+Patch by Mark Shannon
diff --git a/Objects/exception_table_notes.txt b/Objects/exception_table_notes.txt
@@ -1,7 +1,11 @@
 Description of the internal format of the exception table.
 
 Conceptually, the exception table consists of a sequence of 5-tuples:
-    start-offset (inclusive), end-offset (exclusive), target, stack-depth, push-lasti (boolean)
+    1. start-offset (inclusive)
+    2. end-offset (exclusive)
+    3. target
+    4. stack-depth
+    5. push-lasti (boolean)
 
 All offsets and lengths are in instructions, not bytes.
 
@@ -31,19 +35,19 @@ The 8 bits of a bit are (msb left) SXdddddd where S is the start bit. X is the e
 In addition, we will combine depth and lasti into a single value, ((depth<<1)+lasti), before encoding.
 
 For example, the exception entry:
-    start: 20
-    end:   28
-    target 100
-    depth  3
-    lasti  False
+    start:  20
+    end:    28
+    target: 100
+    depth:  3
+    lasti:  False
 
 is encoded first by converting to the more compact four value form:
-    start:        20
-    size:         8
-    target        100
-  depth<<1+lasti  6
+    start:         20
+    size:          8
+    target:        100
+  depth<<1+lasti:  6
 
-Where is then encoded as:
+which is then encoded as:
     148 (MSB + 20 for start)
     8   (size)
     65  (Extend bit + 1)

@@ -285,7 +285,7 @@ explain_incompatible_block_stack(int64_t to_stack)
         return "stack is too deep to analyze";
     }
     if (to_stack == UNINITIALIZED) {
-        return "can't jump into an exception handler, or code may be unreachable.";
+        return "can't jump into an exception handler, or code may be unreachable";
     }
     Kind target_kind = top_block(to_stack);
     switch(target_kind) {

@@ -4070,7 +4070,7 @@ _PyEval_EvalFrameDefault(PyThreadState *tstate, PyFrameObject *f, int throwflag)
                - (FOURTH, FIFTH, SIXTH) = previous exception
                - SEVENTH: lasti of exception in exc_info()
                - EIGHTH: the context.__exit__ bound method
-               We call SEVENTH(TOP, SECOND, THIRD).
+               We call EIGHTH(TOP, SECOND, THIRD).
                Then we push again the TOP exception and the __exit__
                return value.
             */
@@ -4772,9 +4772,13 @@ positional_only_passed_as_keyword(PyThreadState *tstate, PyCodeObject *co,
 
 }
 
+/* Exception table parsing code.
+ * See Objects/exception_table_notes.txt for details.
+ */
+
 static inline unsigned char *
 parse_varint(unsigned char *p, int *result) {
-    int val = p[0]&63;
+    int val = p[0] & 63;
     while (p[0] & 64) {
         p++;
         val = (val << 6) | (p[0] & 63);
@@ -4788,11 +4792,13 @@ scan_back_to_entry_start(unsigned char *p) {
     for (; (p[0]&128) == 0; p--);
     return p;
 }
+
 static inline unsigned char *
 skip_to_next_entry(unsigned char *p) {
     for (; (p[0]&128) == 0; p++);
     return p;
 }
+
 static inline unsigned char *
 parse_range(unsigned char *p, int *start, int*end)
 {

@@ -122,7 +122,6 @@ is_relative_jump(struct instr *i)
 static inline int
 is_jump(struct instr *i)
 {
-
     return i->i_opcode >= SETUP_WITH || is_bit_set_in_table(_PyOpcode_Jump, i->i_opcode);
 }
 
@@ -1138,14 +1137,21 @@ stack_effect(int opcode, int oparg, int jump)
         case LOAD_GLOBAL:
             return 1;
 
-        /* Exception handling */
+        /* Exception handling pseudo-instructions */
         case SETUP_FINALLY:
             /* 0 in the normal flow.
-             * Restore the stack position and push 6 values before jumping to
+             * Restore the stack position and push 3 values before jumping to
              * the handler if an exception be raised. */
             return jump ? 3 : 0;
         case SETUP_CLEANUP:
+            /* As SETUP_FINALLY, but pushes lasti as well */
             return jump ? 4 : 0;
+        case SETUP_WITH:
+            /* 0 in the normal flow.
+             * Restore the stack position to the position before the result
+             * of __(a)enter__ and push 4 values before jumping to the handler
+             * if an exception be raised. */
+            return jump ? -1 + 4 : 0;
 
         case RERAISE:
             return -3;
@@ -1197,12 +1203,6 @@ stack_effect(int opcode, int oparg, int jump)
         /* Iterators and generators */
         case GET_AWAITABLE:
             return 0;
-        case SETUP_WITH:
-            /* 0 in the normal flow.
-             * Restore the stack position to the position before the result
-             * of __aenter__ and push 7 values before jumping to the handler
-             * if an exception be raised. */
-            return jump ? -1 + 4 : 0;
 
         case BEFORE_ASYNC_WITH:
         case BEFORE_WITH:

diff --git a/Tools/scripts/generate_opcode_h.py b/Tools/scripts/generate_opcode_h.py
@@ -18,6 +18,8 @@
 footer = """
 #define HAS_ARG(op) ((op) >= HAVE_ARGUMENT)
 
+/* Reserve some bytecodes for internal use in the compiler.
+ * The value of 240 is arbitrary. */
 #define IS_ARTIFICIAL(op) ((op) > 240)
 
 #ifdef __cplusplus