------- Comment #23 from ghazi at gcc dot gnu dot org 2010-02-05 17:26 ------- (In reply to comment #22) > Kaveh, you are comparing apples to oranges: in the first case the GNU locale > model is used, a complete set of locale data is installed, thus the testcase > is > run and it fails. In the second case, evidently one of the first two > assumptions does not hold, thus the testcase is not run at all.
The failure on the branches happens on the same box (CFARM gcc14). So the locale data is identical across these cases vs trunk. E.g.: 4.4: http://gcc.gnu.org/ml/gcc-testresults/2010-02/msg00420.html 4.3: http://gcc.gnu.org/ml/gcc-testresults/2010-02/msg00421.html -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31413