develooper Front page | perl.perl5.porters | Postings from January 2014

[perl #114126] Pod::Html: t/htmldir3.t test produces spurious pass

From:
James E Keenan via RT
Date:
January 8, 2014 02:22
Subject:
[perl #114126] Pod::Html: t/htmldir3.t test produces spurious pass
Message ID:
rt-4.0.18-12898-1389147718-1702.114126-15-0@perl.org
On Wed Jan 01 12:35:53 2014, jkeenan wrote:

> Two calendar years later ... and this just got weirder!
> 
> I created a branch on the dromedary test server today and added the
> two test files (with minor updatings) originally posted in this
> ticket.
> 
> I then configure, built and tested perl in that branch.  As expected,
> the two additional files failed.
> 
> I then repeatedly ran this set of commands:
> 
> #####
> cd t;./perl harness -v ../ext/Pod-Html/t/*.t; cd-
> #####
> 
> ... and got different results each time!
> 
> t/xhtmldir3.t and t/yhtmldir3.t failed each time, as expected.  But
> each repetition generated failures in 0 to 2 *other*, previously
> passing files!  The output is in the attachment.
> 

More data.  Tonight I built blead on the dromedary server.  All tests, including those in ext/Pod-Html/t/, passed.  (The command I run there is:
TEST_JOBS=${TEST_JOBS} make -j${TEST_JOBS} test_harness
... with TEST_JOBS=8.)

I then ran those same tests with the harness 8 times -- and got 8 different results.

#####
cd t;./perl harness -v ../ext/Pod-Html/t/*.t; cd-
#####

Results attached.

It is interesting to note that many, but not all, of the test failures in the attachment are failures like these two:

#####
# -<p><a href="../testdir/test.lib/var-copy.html">var-copy</a></p>
# +<p><a>var-copy</a></p>
#  
# -<p><a href="../testdir/test.lib/var-copy.html#pod">&quot;$&quot;&quot; in var-copy</a></p>
# +<p><a>&quot;$&quot;&quot; in var-copy</a></p>
#####

That is, the output expected -- the '-' lines in the diff excerpt above -- is an <a> tag where something like the following has been interpolated in:

#####
 href="../testdir/test.lib/var-copy.html
#####

But what is actually generated is simply:  <a>

In my attempt to make Pod-Html CPANnable this type of test failure is one I'm getting in many files -- but at least I'm getting them consistently.  And it's the only type of failure I'm still getting.  So, while there may well be errors in my refactoring of the code, the fundamental problem may have been there all along but simply have been masked.

Thank you very much.
Jim Keenan

---
via perlbug:  queue: perl5 status: open
https://rt.perl.org/Ticket/Display.html?id=114126



nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About