Re: bison-1.30c testsuite failure log

[Akim Demaille]

>>>>> "Arthur" == Arthur Haas <address@hidden> writes:

Arthur> This release tested better than 1.30b, but I still had two
Arthur> errors.

:(

I don't understand this failure at all.

Arthur> Thanks for all the work on Bison!

Thanks for your report!

This failure is bad news: it means there are failures Autotest fails
to diagnose on some machines, such as mine.  It is related to the
missing last \n: on my machine, testsuite fails to see it's lacking,
on yours it did :(

32. regression.at:247: testing %expect not enough...
regression.at:259: bison input.y -o input.c
--- -   Wed Nov 14 14:20:51 2001
+++ /home/cs004/gnu/bison-1.30c/tests/testsuite.dir/at-stderr   Wed Nov 14 14:20
:51 2001
@@ -1,3 +1,2 @@
 input.y contains 1 shift/reduce conflict.
 expected 0 shift/reduce conflicts
-
32. regression.at:247: FAILED near `regression.at:259'


I'm fixing the test suite, thanks!

Index: ChangeLog
from  Akim Demaille  <address@hidden>

        * src/conflicts.c (conflicts_print): Add a missing \n.

Index: src/conflicts.c
===================================================================
RCS file: /cvsroot/bison/bison/src/conflicts.c,v
retrieving revision 1.22.2.16
diff -u -u -r1.22.2.16 conflicts.c
--- src/conflicts.c 2001/11/19 09:26:30 1.22.2.16
+++ src/conflicts.c 2001/11/19 11:06:21
@@ -477,8 +477,8 @@
       && src_total != expected_conflicts)
     {
       complain_message_count++;
-      fprintf (stderr, ngettext ("expected %d shift/reduce conflict",
-                                "expected %d shift/reduce conflicts",
+      fprintf (stderr, ngettext ("expected %d shift/reduce conflict\n",
+                                "expected %d shift/reduce conflicts\n",
                                 expected_conflicts),
               expected_conflicts);
     }