This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [doc, committed] fix typos in docs/comments for devirtualization optimizations


Hi Sandra,

Index: gcc/ipa-devirt.c
===================================================================
     polymorphic (indirect) call
-       This is callgraph represention of virtual method call.  Every
+       This is callgraph representation of virtual method call.  Every
       polymorphic call contains otr_type and otr_token taken from
       original OBJ_TYPE_REF at callgraph construction time.

aren't some articles missing here?  "a callgraph representation
of a virtual method call"?

/* The node of type inheritance graph.  For each type unique in
- One Defintion Rule (ODR) sense, we produce one node linking all + One Definition Rule (ODR) sense, we produce one node linking all main variants of types equivalent to it, bases and derived types. */

...in the One Definition Rule (ODR) sense... ?

-  /* All derrived types with virtual methods seen in unit;
-     built only for main variants oftypes  */
+  /* All derived types with virtual methods seen in unit;
+     built only for main variants of types.  */
  vec<odr_type> GTY((skip)) derived_types;

Something missing here, too.  "in the unit"?

-/* Return TURE if type's constructors are all visible.  */
+/* Return TRUE if type's constructors are all visible.  */

"if a type's"?

-/* Dump ODR type T and all its derrived type.  INDENT specify indentation for
-   recusive printing.  */
+/* Dump ODR type T and all its derived types.  INDENT specifies indentation for
+   recursive printing.  */

"the indentation"?

-      /* Lookup BINFO with virtual table.  For normal types it is always last
+      /* Look up BINFO with virtual table.  For normal types it is always last
	 binfo on stack.  */

"with" -> "within"?

"it is always the last"

  for (i = 0; BINFO_BASE_ITERATE (binfo, i, base_binfo); i++)
-    /* Walking bases that have no virtual method is pointless excercise.  */
+    /* Walking bases that have no virtual method is pointless exercise.  */
    if (polymorphic_type_binfo_p (base_binfo))

Shouldn't this just read "Walk bases..."?

+ because the type is always known. One of entries may be + cxa_pure_virtual so look to at least two of them. */

"One of the entries"

-  /* If there are no virtual tables refering the target alive,
-     the only way the target can be called is an instance comming from other
-     compilation unit; speculative devirtualization is build around an
+  /* If there are no live virtual tables referring the target,
+     the only way the target can be called is an instance coming from other
+     compilation unit; speculative devirtualization is built around an

"from a different compilation unit"


I will be making and committing those changes, but would prefer
your second pair of eyes given that you have touched this recently.

Gerald


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]