[PATCH] master: deterministic user-token-number redeclaration errors.

[Akim Demaille]

Address nondeterminism reported by Joel E. Denny.
http://lists.gnu.org/archive/html/bison-patches/2009-05/msg00023.html

        * src/uniqstr.h: Comment changes.
        * src/location.h (boundary_cmp, location_cmp): New.
        * src/symtab.c (user_token_number_redeclaration): New.
        (symbol_translation): Use it.
        * tests/input.at (Numbered tokens): Adjust the expected output.
---
 ChangeLog      |   12 ++++++++++++
 src/location.h |   24 ++++++++++++++++++++++++
 src/symtab.c   |   28 ++++++++++++++++++++++++----
 src/uniqstr.h  |    6 +++---
 tests/input.at |    6 ++++--
 5 files changed, 67 insertions(+), 9 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 14eb499..36dd87a 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,15 @@
+2009-06-10  Akim Demaille  <address@hidden>
+
+       deterministic user-token-number redeclaration errors.
+       Address nondeterminism reported by Joel E. Denny.
+       http://lists.gnu.org/archive/html/bison-patches/2009-05/msg00023.html
+
+       * src/uniqstr.h: Comment changes.
+       * src/location.h (boundary_cmp, location_cmp): New.
+       * src/symtab.c (user_token_number_redeclaration): New.
+       (symbol_translation): Use it.
+       * tests/input.at (Numbered tokens): Adjust the expected output.
+
 2009-05-25  Akim Demaille  <address@hidden>
 
        build: avoid ignored errors.
diff --git a/src/location.h b/src/location.h
index a99232b..ec48164 100644
--- a/src/location.h
+++ b/src/location.h
@@ -55,6 +55,19 @@ boundary_set (boundary *b, const char *f, int l, int c)
   b->column = c;
 }
 
+/* Return -1, 0, 1, depending whether a is before, equal, or
+   after b.  */
+static inline int
+boundary_cmp (boundary a, boundary b)
+{
+  int res = strcmp (a.file, b.file);
+  if (!res)
+    res = a.line - b.line;
+  if (!res)
+    res = a.column - b.column;
+  return res;
+}
+
 /* Return nonzero if A and B are equal boundaries.  */
 static inline bool
 equal_boundaries (boundary a, boundary b)
@@ -87,6 +100,17 @@ void location_compute (location *loc,
 
 void location_print (FILE *out, location loc);
 
+/* Return -1, 0, 1, depending whether a is before, equal, or
+   after b.  */
+static inline int
+location_cmp (location a, location b)
+{
+  int res = boundary_cmp (a.start, b.start);
+  if (!res)
+    res = boundary_cmp (a.end, b.end);
+  return res;
+}
+
 /* LOC_STR must be formatted as `file:line.column', it will be modified.  */
 void boundary_set_from_string (boundary *bound, char *loc_str);
 
diff --git a/src/symtab.c b/src/symtab.c
index c956e72..e68e473 100644
--- a/src/symtab.c
+++ b/src/symtab.c
@@ -193,6 +193,7 @@ semantic_type_redeclaration (semantic_type *s, const char 
*what, location first,
 }
 
 
+
 /*-----------------------------------------------------------------.
 | Set the TYPE_NAME associated with SYM.  Does nothing if passed 0 |
 | as TYPE_NAME.                                                    |
@@ -582,7 +583,26 @@ symbol_pack_processor (void *this, void *null 
ATTRIBUTE_UNUSED)
 }
 
 
+static void
+user_token_number_redeclaration (int num, symbol *first, symbol *second)
+{
+  /* User token numbers are not assigned during the parsing, but in a
+     second step, via a (nondeterministic) traversal of the symbol
+     hash table.
 
+     Make errors deterministic: keep the first declaration first.  */
+  if (location_cmp (first->location, second->location) > 0)
+    {
+      symbol* tmp = first;
+      first = second;
+      second = tmp;
+    }
+  complain_at (second->location,
+               _("user token number %d redeclaration for %s"),
+               num, second->tag);
+  complain_at (first->location, _("previous declaration for %s"),
+               first->tag);
+}
 
 /*--------------------------------------------------.
 | Put THIS in TOKEN_TRANSLATIONS if it is a token.  |
@@ -597,10 +617,10 @@ symbol_translation (symbol *this)
     {
       /* A token which translation has already been set? */
       if (token_translations[this->user_token_number] != undeftoken->number)
-       complain_at (this->location,
-                    _("tokens %s and %s both assigned number %d"),
-                    symbols[token_translations[this->user_token_number]]->tag,
-                    this->tag, this->user_token_number);
+       user_token_number_redeclaration
+          (this->user_token_number,
+           symbols[token_translations[this->user_token_number]],
+           this);
 
       token_translations[this->user_token_number] = this->number;
     }
diff --git a/src/uniqstr.h b/src/uniqstr.h
index c986df0..04600e9 100644
--- a/src/uniqstr.h
+++ b/src/uniqstr.h
@@ -1,6 +1,6 @@
 /* Keeping a unique copy of strings.
 
-   Copyright (C) 2002, 2003, 2008 Free Software Foundation, Inc.
+   Copyright (C) 2002, 2003, 2008, 2009 Free Software Foundation, Inc.
 
    This file is part of Bison, the GNU Compiler Compiler.
 
@@ -32,8 +32,8 @@ uniqstr uniqstr_new (char const *str);
 /* Two uniqstr values have the same value iff they are the same.  */
 #define UNIQSTR_EQ(USTR1, USTR2) ((USTR1) == (USTR2))
 
-/* Compare two uniqstr a la strlen: negative for <, nul for =, and
-   positive for >.  */
+/* Compare two uniqstr a la strcmp: negative for <, nul for =, and
+   positive for >.  Undefined order, relies on addresses.  */
 #define UNIQSTR_CMP(USTR1, USTR2) ((USTR1) - (USTR2))
 
 /*--------------------------------------.
diff --git a/tests/input.at b/tests/input.at
index c43aeb1..6de0e1a 100644
--- a/tests/input.at
+++ b/tests/input.at
@@ -686,8 +686,10 @@ start: HEXADECIMAL_1 HEXADECIMAL_2
 ]])
 
 AT_BISON_CHECK([input.y], [1], [],
-[[input.y:12.12-20: tokens HEXADECIMAL_2 and DECIMAL_2 both assigned number 
16702650
-input.y:9.8-20: tokens DECIMAL_1 and HEXADECIMAL_1 both assigned number 
11259375
+[[input.y:10.12-20: user token number 11259375 redeclaration for DECIMAL_1
+input.y:9.8-20: previous declaration for HEXADECIMAL_1
+input.y:12.12-20: user token number 16702650 redeclaration for DECIMAL_2
+input.y:11.8-20: previous declaration for HEXADECIMAL_2
 ]])
 
 AT_DATA_GRAMMAR([input.y],
-- 
1.6.3.2