bugGNU Octave - Bugs: bug #38679, make check fails on stemleaf.m

 
 

bug #38679: make check fails on stemleaf.m

Submitter:  Michael Godfrey <godfrey>
Submitted:  Sun 07 Apr 2013 05:57:39 AM UTC
   
 
Category:  Configuration and Build System Severity:  3 - Normal
Priority:  5 - Normal Item Group:  Incorrect Result
Status:  Fixed Assigned to:  None
Originator Name:  Michael Godfrey Open/Closed:  * Closed
Release:  * dev Operating System:  * GNU/Linux
Fixed Release:  None Planned Release:  None
* Mandatory Fields

Add a New Comment Rich Markup
   

Jump to the original submission

Sun 14 Jul 2013 12:53:17 AM UTC, comment #6: 

Rik,

It is pretty likely that you corrected the original error.
But, it has bee several weeks since stemleaf stopped failing
during make check.  Something else had the side effect of
allowing it to work.

In any case, the build with you patch runs make check
correctly.

Michael Godfrey <godfrey>
Group Member
Sat 13 Jul 2013 11:29:35 PM UTC, comment #5: 

'make check' should now properly make it pass the stemleaf.m tests.  Try updating to this changeset http://hg.savannah.gnu.org/hgweb/octave/rev/0d5d3e53255f.

Rik <rik5>
Group administrator
Tue 18 Jun 2013 11:07:47 PM UTC, comment #4: 

Sorry.  Spoke too soon.
As of build a few minutes ago, fail
in make check in stemleaf is back.

Michael Godfrey <godfrey>
Group Member
Tue 18 Jun 2013 06:54:37 PM UTC, comment #3: 

make check now runs without an error
in stemleaf.  However, since there does not
appear to have been any relevant patch applied
it may be that the problem went away due to a
change in other tests which run prior to stemleaf.

In any case, since the problem is gone this
report could be closed.

Michael Godfrey <godfrey>
Group Member
Sun 07 Apr 2013 06:37:23 PM UTC, comment #2: 

Part of the problem is that warnings are not cleared before starting testing so the lastwarn() function call can contain garbage from previous tests.  This is the problem reported in bug #35850 which I have added as a dependency here.

I haven't fixed and close the first report because there is something else going on with function contexts.  Sometimes even when a warning has been issued with warning(), lastwarn() reports the previous warning.  I think this must have something to do with function stacks, but haven't concluded where yet.

Rik <rik5>
Group administrator
Sun 07 Apr 2013 02:24:26 PM UTC, comment #1: 

It is the execution of the test in copyobj.m
that causes the %!warning test in stemleaf.m to fail.

In copyobj.m I changed line 108 from

%!testif HAVE_MAGICK

to

%!testif HAVE_MAGICKxxx

Making this change causes the test in stemleaf.m to
run correctly.  There are about 30 files with tests
between copyobj and stemleaf, but something affects the
environment for stemleaf so that the #!warning test
%!warning <X truncated to integer values> tmp = stemleaf ([0 0.5 1],"");

fails.


Michael Godfrey <godfrey>
Group Member
Sun 07 Apr 2013 05:57:39 AM UTC, original submission:  

Since the function stemleaf.m was introduced one of its
tests has failed when running make check.  If the tests
are run by:
./run-octave
test stemleaf

the tests all pass.

I found a possibly useful fact about this.  If the Octave
system was built without ImageMagick and therefore the variable
HAVE_MAGICK is not set in make check, the stemleaf test passes.
There are 2 references to HAVE_MAGICK  in the make check tests:
imread.m and copyobj.m.  I am pretty sure that the cause of the
stemleaf failure is something left over from the test in copyobj.

When I asked Rik about this problem he said there was something
strange happening in the make check process.  I hope that this
will be a useful clue.

Michael



Michael Godfrey <godfrey>
Group Member

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach Files:
   
   
Comment:
   

No files currently attached

 

Digest:
   bug dependencies.

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by rik5 (Posted a comment)
  • -email is unavailable- added by godfrey (Submitted the item)
  •  

    There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.

    Only group members can vote.

     

    Follow 3 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2013-07-13 rik5 StatusNone Fixed
        Open/ClosedOpen Closed
    2013-04-07 rik5 Dependencies- Depends on bugs #35850

    Back to the top

    Powered by Savane 3.13-758e.
    Corresponding source code