v 1.9.1, bug reporting.

[Hugh Sasse Staff Elec Eng]

I have been running tests on this (1.9.1) version of automake, and
would like to report result when it has finished.  This raised a
number of matters:

1) I see from the archives and the web pages that 1,9.2 seems to be
out.  It doesn't seem to have reached any of the UK mirrors yet. I
don't know whom one should contact about this, but I raise it here
in case somebody else does...

2) I've had a look in the tests directory and see that the makefile
is derived from Makefile.am (which I especially expect for this
project).  Reading the file I see that the tests are created from a
list which is looped though using $$test.  This has the disadvantage
that one cannot take advantage of GNU Make's --jobs=N flag to get
through all the tests more quickly.  Since the file is generated
rather than constructed entirely by hand, there would seem to be
scope for creating many make targets for the tests, so GNU make can
parallelize the tests.  Is there a good reason for avoiding this 
approach?

3) The code for running the tests doesn't do much with the results.
One gets a PASS, XPASS, FAIL, XFAIL message, but that is it, and for
reporting one must run the tests by hand.  I'm not sufficiently
familiar with the code base to do this myself, but would it be
possible to send the results of /X?FAIL/ed tests to files within,
say, ./tests/log/, so that a script can pick them up and mail them
if desired?  It would seem to fit in with the spirit of a project
for automating things, I think.

        Thank you,
        Hugh