* src/reader.c (readgram): Make sure rules for mid-rule actions

[bison.git] / ChangeLog

diff --git a/ChangeLog b/ChangeLog
index 6bb43074922385a3f5229093a5ec61ec5547ad72..b7c5a449df5592b97c924d41fd1b32f68af04071 100644 (file)
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,48 @@
+2001-11-26  Akim Demaille  <akim@epita.fr>
+
+       * src/reader.c (readgram): Make sure rules for mid-rule actions
+       have a lineno equal to that of their host rule.
+       Reported by Hans Aberg.
+       * tests/regression.at (Rule Line Numbers): New.
+
+2001-11-26  Akim Demaille  <akim@epita.fr>
+
+       * src/LR0.c (allocate_itemsets): kernel_size contains ints, not
+       size_ts.
+
+2001-11-26  Akim Demaille  <akim@epita.fr>
+
+       * src/complain.c, src/complain.h (error): Remove, provided by
+       lib/error.[ch].
+
+2001-11-26  Akim Demaille  <akim@epita.fr>
+
+       * src/reader.c (read_declarations): Don't abort on tok_illegal,
+       issue an error message.
+       * tests/regression.at (Invalid %directive): New.
+       Reported by Hans Aberg.
+
+2001-11-26  Akim Demaille  <akim@epita.fr>
+
+       * configure.in: Invoke AC_FUNC_OBSTACK and AC_FUNC_ERROR_AT_LINE.
+       * lib/Makefile.am (libbison_a_SOURCES): Adjust.
+
+2001-11-26  Akim Demaille  <akim@epita.fr>
+
+       * src/conflicts.c (conflicts_print): Don't complain at all when
+       there are no reduce/reduce conflicts, and as many shift/reduce
+       conflicts as expected.
+       * tests/regression.at (%expect right): Adjust.
+
+2001-11-25  Marc Autret  <autret_m@epita.fr>
+
+       * doc/bison.texinfo (Mystery Conflicts): Hint '-v' use.
+       From Golubev I. N.
+
+2001-11-23  Akim Demaille  <akim@epita.fr>
+
+       * lib/alloca.c: Update, from fileutils.
+
 2001-11-23  Akim Demaille  <akim@epita.fr>
 
        * lib/Makefile.am (libbison_a_LIBADD): Add @ALLOCA@.
@@ -15,7 +60,6 @@
        * src/closure.c (itemsetsize): Likewise.
        * src/reader.c (symbol_list_new): Static.
 
-       
 2001-11-23  Akim Demaille  <akim@epita.fr>
 
        Attaching lineno to buckets is stupid, since only one copy of each