On Mon, Aug 8, 2016 at 5:53 PM, Jim Meyering <jim@meyering.net> wrote:
> On Mon, Aug 8, 2016 at 4:29 PM, Norihiro Tanaka <noritnk@kcn.ne.jp> wrote:
...
>> Thanks for adjusting.  I agree the all changes.
>
> Pushed.

Oops. I had not run ASAN-enabled tests with this patch.
Doing that exposed many test failures like this:

LC_ALL=C  ./sed/sed -f ./testsuite/madding.sed \
        < ./testsuite/madding.inp | LC_ALL=C tr -d \\r > madding.out
=================================================================
==8797==ERROR: AddressSanitizer: heap-buffer-overflow on address
0x622000019646 at pc 0x0000004e98ce bp 0x7ffee5d23130 sp
0x7ffee5d23128
READ of size 1 at 0x622000019646 thread T0
    #0 0x4e98cd in dfaexec_main sed/dfa.c:3078
    #1 0x4e98cd in dfaexec_sb sed/dfa.c:3242
    #2 0x4ece9f in dfaexec sed/dfa.c:3263
    #3 0x4f64ba in match_regex sed/regexp.c:270
    #4 0x4f46ac in do_subst sed/execute.c:1024
    #5 0x4f46ac in execute_program sed/execute.c:1510
    #6 0x4f5997 in process_files sed/execute.c:1678
    #7 0x4f6f64 in main sed/sed.c:363
    #8 0x7fe665f89730 in __libc_start_main (/lib64/libc.so.6+0x20730)
    #9 0x407118 in _start (/home/j/w/co/sed/sed/sed+0x407118)

which I am fixing with this patch: