source builder issue (unstable: expat-2.1.0-1 missing)

[Chris Johns]

On 24/03/2014 7:38 pm, Karel Gardas wrote:
> On 03/23/14 11:28 PM, Chris Johns wrote:
>>>> Interesting and thanks for using the rtems-test command. There is a
>>>> patch about for the SPARC simulator to fix the output. I am not sure if
>>>> this is included in the gdb build. What is the error ? The log should
>>>> list the details of the tests that are in error. I should mention the
>>>> number of concurrent tests running can effect the results. It is
>>>> typically safe to have a test per core and you can go higher but there
>>>> may be failures.
>>>
>>> I think I just run tests with simple make not using any -jX option.
>>> Shall I?
>>
>> The rtems-test command has a --jobs option that controls the number of
>> parallel tests. I am not sure I follow where make is involved when using
>> the rtems-test command.
>
> Indeed! My mistake, and no, I've not used --jobs so it was run in
> single-threaded mode.
>

As Gedare states the default is the number of cores you have.

>>>
>>> The error is this:
>>>
>>> [karel at centos sis-bsp]$ sparc-rtems4.11-sis
>>>
>>> SIS - SPARC instruction simulator 2.7.5, copyright Jiri Gaisler 1995
>>> Bug-reports to jgais at wd.estec.esa.nl
>>>
>>> sis> load ./sparc-rtems4.11/c/sis/testsuites/samples/hello/hello.exe
>>> sis> go
>>> resuming at 0x02000000
>>>
>>>
>>> *** HELLO WORLD TEST ***
>>> Hello World
>>> *** END OF HELLO WORLD TEST ***
>>> IU in error mode (257)
>>> 128835 02001f38 91d02000 Address 0x0000000002001f38 is out of
>>> bounds.
>>>
>>> sis>
>>>
>>> I mean this "IU in error mode (257)
>>> 128835 02001f38 91d02000 Address 0x0000000002001f38 is out of
>>> bounds." message...
>>>
>>
>> I do not think there is any support for the sis command. There is
>> support for gdb and the run command.
>
> Err, again my mistake. I used what's supported by tester so sis and
> sis-run.  Hmm, I've rerun tests now and it looks like I did another
> mistake in the past. Now I see I run with:
>
> Passed:   466
> Failed:     3
> Timeouts:   5
> Invalid:    1
> -------------
> Total:    475
>
> Failures:
>   mrfs_fstime.exe
>   spintrcritical20.exe
>   sp2038.exe
> Timeouts:
>   monitor.exe
>   fileio.exe
>   termios.exe
>   pppd.exe
>   capture.exe
> Invalid:
>   tmcontext01.exe
>
>
> while using sis
>
> and with
>
> Passed:   470
> Failed:     3
> Timeouts:   1
> Invalid:    1
> -------------
> Total:    475
>
> Failures:
>   mrfs_fstime.exe
>   spintrcritical20.exe
>   sp2038.exe
> Timeouts:
>   pppd.exe
> Invalid:
>   tmcontext01.exe
>
>
> while using sis-run.
>
> Failures/invalids are the same, the difference is only in timeouts and
> probably caused by much slower run of sis in comparison with sis-run.

This is expected with the unpatched gdb verses the run command as the 
end marker is not found. A number of simulators do not use the simulator 
framework's IO path so when using gdb in MI mode the output is not 
encapsulated in the MI protocol so things break. With the patch things 
match up. I just need to find the patch and send it upstream.

> So, Chris, I'm really sorry for all this chaos done on my side!

Hey this is all good news to me. It is nice to see rtems-test results 
being posted. It exists for just this purpose.

Chris