[bug #62484] [PATCH] .../tests/: use one directory name instead of "./ ../ ../.." in a for-loop

[G. Branden Robinson]

Update of bug #62484 (project groff):

                  Status:                    None => Wont Fix               
             Assigned to:                    None => gbranden               
             Open/Closed:                    Open => Closed                 

    _______________________________________________________

Follow-up Comment #1:

I'm rejecting this for the reasons I offered in bug #62478.  However I have
some further comments.

[comment #0 original submission:]
>   Use the make-variable "abs_top_builddir" or "abs_top_srcdir" instead of
"rootbuild" in a loop.

...well, which did you choose, and why?  Did you test your patch with both
in-tree and out-of-tree builds?

>   When running the test files _outside_ of make, these  variables must be
defined with the "export" shell command.

I disagree with this.  We should be making the tests _less_ sensitive to
environment variable settings, not more.

It's possible that simply using "$top_srcdir" (for shipped artifacts) and
"$top_builddir" (for generated ones) will suffice, and eliminate the need for
the loops.  I didn't learn about these variables until my heavy revisions of
"doc/doc.am" in recent months.  That would indeed be an improvement, though I
hasten to add given some of your past bug reports that I don't think it would
make the test suite measurably more performant.  The real value would be in
reduction of complexity and code size for human readers of the test scripts.

As a matter of fact, maybe we don't need to be using $abs_top_builddir to
locate executables in the first place.  I think I copy-and-pasted that from
one of Bertrand's test scripts and I've reproduced it by rote ever since. 
Maybe we can favor $top_builddir.

If you, or someone else, would like to take a stab at either of these, please
file a new bug reflecting the narrower scope.


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?62484>

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