Re: test failure with current tip

[Ben Abbott]

On Jan 13, 2010, at 7:39 PM, Ben Abbott wrote:

> On Jan 13, 2010, at 7:26 PM, John W. Eaton wrote:
> 
>> On 13-Jan-2010, Ben Abbott wrote:
>> 
>> | My tip is current, but "make check" still results in
>> | 
>> |      test_bc_overloads.m .................................... PASS  
>> 225/451  FAIL 226
>> | 
>> | However, if I run octave and cd to the test directory ...
>> | 
>> |    octave:1> test test_bc_overloads
>> |    PASSES 451 out of 451 tests
>> | 
>> | Is this another MacOS X specific thing? ... Is anyone else seeing this?
>> 
>> I'm not seeing this failure.  So what tests are failing, and how do
>> they fail?
>> 
>> jwe
> 
> I just tried "make check" from the tests directory and ...
> 
>       test_bc_overloads.m .................................... PASS  451/451 
> 
> But from the root of my archive I still get ...
> 
>       test_bc_overloads.m .................................... PASS  225/451  
> FAIL 226
> 
> I'd guess that if one test can be resolved the others will be as well.
> 
> The first failure is ...
> 
> 378 >>>>> processing test_bc_overloads
> 379   ***** assert ((@tbcover) (ex.double, ex.double), "double")
> 380 !!!!! test failed
> 381 assert ((@tbcover) (ex.double, ex.double),"double") expected
> 382 double
> 383 but got
> 384 none
> 385 shared variables {
> 386   ex =
> 387   {
> 388     double =  1
> 389     single =  1
> 390     logical =  1
> 391     char = char
> 392     int8 = 1
> 393     int16 = 1
> 394     int32 = 1
> 395     int64 = 1
> 396     uint8 = 1
> 397     uint16 = 1
> 398     uint32 = 1
> 399     uint64 = 1
> 400     cell = {}(0x0)
> 401     struct =
> 402     {
> 403       1x1 struct array containing the fields:
> 404 
> 405     }
> 406 
> 407     function_handle =
> 408 
> 409 numel
> 410 
> 411   }
> 412 
> 413 }
> 
> For the other tests, I attached the log to a prior email. While it hasn't 
> posted to the list yet, both you and Jaroslav should have received it.
> 
> Ben

John,

I pulled the change below ...

        http://hg.savannah.gnu.org/hgweb/octave/rev/29959c705df2

... and now all tests pass. I don't understand how this impacts the tests. Was 
that the intent, or am I seeing some instability in the build process?

Ben