dfa.c problem with dfacomp

[Aharon Robbins]

Hello.

This is based on a bug report submitted for gawk, but I have reproduced
it with the dfa.c in grep 2.14.

Attached are the following files:

testdfa.c:      A harness that pulls in all the gawk stuff needed
                to call dfa the same way gawk does.
dfa_test_re1:   A regexp that causes dfa to assert.
dfa_test_re2:   A similar regexp that doesn't. The difference is the
                [..] which are not present.
dfa_test.in:    Input data

Compile testdfa.c with dfa.o and run as follows:

        LC_ALL=en_US.UTF-8 testdfa "$(cat dfa_test_re1)" < dfa_test.in
        LC_ALL=en_US.UTF-8 testdfa "$(cat dfa_test_re2)" < dfa_test.in

The first will assert. The second will not.

There is no problem if LC_ALL=C.

Thanks!

Arnold
-- 
Aharon (Arnold) Robbins                         arnold AT skeeve DOT com
P.O. Box 354            Home Phone: +972 8 979-0381
Nof Ayalon
D.N. Shimshon 9978500   ISRAEL

testdfa.c
Description: Text Data

dfa_test_re1
Description: Text document

dfa_test_re2
Description: Text document

dfa_test.in
Description: Text document