Re: rcs-5.9.4: t810 test failed

On 10/26/2017 07:07 PM, Thien-Thi Nguyen wrote:

> Thanks for the bug report.  In the associated tarball, files x
> and x,v have the same mtime AFAICT, and on my system, i see:
> 
>  $ test x -nt x,v ; echo $?
>  1

$ ls -l
total 12
-rw------- 1 xose xose 226 Oct 26 19:11 out
-rw------- 1 xose xose  27 Oct 26 19:11 x
-r-------- 1 xose xose 428 Oct 26 19:11 x,v
$ test x -nt x,v ; echo $?
0

> for both that tarball and in my development tree (test invoked
> w/ "make check TESTS=t810 VERBOSE=1 KEEPD=1").  This is the
> expected result, and the expected consequence in the test should
> be control passing over (skipping) the ‘problem’ command; i.e.,
> no error, no test failure.
> 
> What do you see on your end?

[...]
make  btdt
make[2]: Entering directory '/home/xose/tmp/rcs-5.9.4/tests'
make[2]: 'btdt' is up to date.
make[2]: Leaving directory '/home/xose/tmp/rcs-5.9.4/tests'
make  check-TESTS
make[2]: Entering directory '/home/xose/tmp/rcs-5.9.4/tests'
FAIL: t810
================================
1 of 1 test failed
Please report to [email protected]
================================
make[2]: *** [Makefile:1502: check-TESTS] Error 1
make[2]: Leaving directory '/home/xose/tmp/rcs-5.9.4/tests'
make[1]: *** [Makefile:1626: check-am] Error 2
make[1]: Leaving directory '/home/xose/tmp/rcs-5.9.4/tests'
make: *** [Makefile:1392: check-recursive] Error 1



> If same as above, that indicates a weirdness in the ‘test’ of
> /bin/sh (or whatever -- depends on $(SHELL) value in Makefile),
> or perhaps the ‘&&’ operator that separates the ‘test’ from the
> ‘problem’ commands, or perhaps something else entirely... hmm.

> If different, perhaps you sent the wrong tarball?
Resent + config.log