Re: problem using install-info with tar.info

[Alan Wehmann]

Today I did the exercise of running ./configure, make and then make check.  The 
abbreviated output from make check was:

============================================================================
Testsuite summary for GNU Texinfo 5.2
============================================================================
# TOTAL: 53
# PASS:  51
# SKIP:  0
# XFAIL: 0
# FAIL:  2
# XPASS: 0
# ERROR: 0
============================================================================
See install-info/tests/test-suite.log
Please report to address@hidden
============================================================================
make[4]: *** [test-suite.log] Error 1
make[3]: *** [check-TESTS] Error 2
make[2]: *** [check-am] Error 2
make[1]: *** [check-recursive] Error 1
make: *** [check-recursive] Error 1

I can supply "install-info/tests/test-suite.log" separately.

I then used the newly created "ginstall-info", as follows:

bash-3.2$ /Users/alanwehmann/Downloads/texinfo-5.2/install-info/ginstall-info 
--debug tar.info dir
debug: reading dir file dir
debug: reading input file tar.info
Abort trap: 6
bash-3.2$ pwd
/Users/alanwehmann/Downloads/tar-1.28/doc

I did all of this in the Emacs shell buffer.  I have tested running of 
"install-info" in the terminal window on my Mac & that gave a similar result, 
so the limitations of the shell buffer versus a terminal window make no 
difference.

Alan Wehmann
address@hidden


On Oct 30, 2015, at 6:11 PM, Karl Berry <address@hidden> wrote:

   * Tar: (tar).                   Making tape (or disk) archives.

   The node name is missing (after the closing parenthesis), which is
   allowed, although less common. 

Hmm.  It's standard practice for dir file entries, as far as I can see
(not counting "Individual utilities").

Maybe the second set of parens causes the error, although I see that
coreutils, gzip, bison, and others also have that, so probably not.  E.g.,
* Coreutils: (coreutils).       Core GNU (file, text, shell) utilities.

The other weird thing is that i-i has extensive tests.  I'm surprised
the problem was not found there.  If it wasn't, another test should be
added.

Alan, did you run make check?

k