[bug #64509] [bjarnigroff] new tests are not reproducible, e.g. tmac/tests/an_MR-works.sh

[G. Branden Robinson]

Update of bug #64509 (project groff):

                  Status:               Need Info => Invalid                
             Open/Closed:                    Open => Closed                 

    _______________________________________________________

Follow-up Comment #2:

[https://lists.gnu.org/archive/html/groff/2023-08/msg00046.html Bjarni replied
to the mailing list instead of this ticket], but I am replying here as is the
only justifiable procedure given the existence of a ticket tracker.


An example is "an_MR-works.sh".

  After adding

.ds MF I

  to the INPUT in a line before '.TH ...'
I get the expected result.

  Tests should be independent of the surroundings and should provide the
data, on which the results are based.

  Tests are not to test the value of some used variable(!),
but the code!

  Tests also need to be tested!

----- End of original mail -----

Brandon's answer is in comment #1

"1.  I do not require this lecture."

  This is in your mind.

  What is this in other readers mind?

  In mine it is a reminder, and a criticism, as I have earlier pointed
out tests that produce a false result caused by an alternative
environment (surroundings).

"1.  I do not require this lecture.  Please refrain from uttering it."

  I do not censor myself.

  You are actually asking me to do so!

  You do not give any (valid) reason why I should stop telling a truth,
as you keep producing some tests that fail wrongly.

"2.  What surroundings?  Please attach the tmac/man.local file from
your build tree."

  Should you not be aware of your (programs) surroundings?

  Why should (must) content of files influence the results of tests,
instead of them being immune to it.

  Why should the content of files have to make a test deliver wrong
results?

N.B.
  people on this list should also check
http://lists.gnu.org/archive/html/bug-groff/
to see what decisions are made.

-.-.-.

  "For what one has forbidden so far as a matter of principle
has always been - truth alone.".  These words are valid for
the history of humankind as well as for Nietzsche's family.

Alice Miller "The Untouched key".  Virago Press 1990,
Page 122.
#
A new scientific truth does not triumph by convincing its opponents and
making them see the light, but rather because its opponents eventually
die, and a new generation grows up that is familiar with it.

Max Planck, "A Scientific Autobiography", 1949.
#


Your argument is equivalent

>   After adding
>
> .ds MF I
>
>  to the INPUT in a line before '.TH ...'
> I get the expected result.

Then you have changed something elsewhere such that this _default_ setting is
not redundant as it is in _groff_ 1.23.0 and the Git master branch.

>   Tests should be independent of the surroundings and should provide the
> data, on which the results are based.

Your application of this principle is absurd.  You are writing yourself a
license to change _groff_'s _implementation_ arbitrarily and insisting that
you should not have to reflect changes in the test cases.  That is not how
software development works.

> Tests are not to test the value of some used variable(!),
> but the code!

The test isn't checking the value of "some used variable" [sic] (!!!), but
*the output of the _troff_ program.

This fact is not difficult to observe.  Read the test script.

https://git.savannah.gnu.org/cgit/groff.git/tree/tmac/tests/an_MR-works.sh?h=1.23.0#n36

You have failed to learn as much from luminaries like Max Planck as even a
jaded teacher might hope.

Closing as invalid.


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?64509>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/