[Issues] Problems with Gastrometric

[John P. McFarland]

Hi Philippe,

The bugs in asctoldac are fixed and the current version of LDAC shows no 
problems with the unittests.  LDAC should be updated if necessary.

Cheers,


-=John


On Thu, 13 Sep 2007, John P. McFarland wrote:

> Phillipe,
>
> Indeed, Ewout is working to make the unittests more consistent.  The
> software versions are currently not as synced as we would like throughout
> the nodes.  This should change in the near future and will be announced in
> the News list, so stay tuned.
>
> As for the GAstrometry problem in Bonn and the related problem in Leiden, I
> have traced it to a some bugs in asctoldac.  I have asked Erik to take a
> look and see if he can figure out what happened.
>
> Please still run the unitests with your current binaries and send the
> results like I described earlier so I can identify this and any other
> issues.
>
> Cheers,
>
>
> -=John
>
>
> On Thu, 13 Sep 2007, Ewout M. Helmich wrote:
>
>> Hi,
>>
>> We're in the process of moving checks on versions of external programs
>> into dependency_test. The test version has such a check, but the AWBASE
>> version does not. The SWarp processing test also does a check by itself
>> at this point, but this will be removed. All nodes should upgrade to the
>> latest version of SWarp; that's what's actually wrong. We're running
>> 2.15.5 in Groningen.
>>
>> Ewout
>>
>> Philippe Heraudeau wrote:
>>> Hi John,
>>>
>>>
>>> Thanks for your email.
>>> 1) First I had a older version of swarp (2.15.6 instead of 2.16.0) so I
>>> installed the version 2.16.0 (at  least locally!)
>>> There are actually 2 more recents version i.e.:
>>>
>>> -  /07/18/2007/: Version *2.16.4* fixes more bugs
>>> <ftp://ftp.iap.fr/pub/from_users/bertin/swarp/ChangeLog> and computes
>>> output saturation level.
>>> -  /06/08/2007/: Version *2.16.3* fixes a bunch of bugs and adds support
>>> for image lists.
>>> When do we update? could you please send a email on the issues list when
>>> we have to upgrade external softwares?
>>>
>>> 2) There is a discrepancy between the AWBASE version and the TEST
>>> version of aw for the dependency test which I don't understand because I
>>> don't change the swarp version in the mean time.
>>> and one complains, the other doesnot.
>>> I run the dependency test for both version:
>>> env database_engine= awe /user2/heraud/opipe/astro/test/dependency_test.py
>>> a) for the test version I get:
>>> /user2/heraud/opipe
>>> usno.astro-wise.org 5556
>>> ======================================================================
>>> FAIL: Test versions of external programs
>>> ----------------------------------------------------------------------
>>> Traceback (most recent call last):
>>>  File "/user2/heraud/opipe_test/astro/test/dependency_test.py", line
>>> 143, in test_versions
>>>    self.failUnless(version.startswith(expected_version), 'Expected %s
>>> version %s, got %s' % (program_name, expected_version, version))
>>>  File "unittest.py", line 278, in failUnless
>>>    if not expr: raise self.failureException, msg
>>> AssertionError: Expected swarp version 2.16.0, got 2.15.6
>>>
>>> ----------------------------------------------------------------------
>>> Ran 5 tests in 2.090s
>>>
>>> FAILED (failures=1)
>>> =======================================================================
>>>
>>> b) for the awbase version I get no problems:
>>> =======================================================================
>>> /user2/heraud/opipe
>>> usno.astro-wise.org 5556
>>> ....
>>> ----------------------------------------------------------------------
>>> Ran 4 tests in 0.395s
>>>
>>> OK
>>> =======================================================================
>>> my local version is :
>>> swarp -v
>>> SWarp version 2.16.0 (2007-09-13)
>>>
>>> =======================================================================
>>> But this error comes back later in the AWBASE version with:
>>> env database_engine= awe ${AWEPIPE}/astro/test/processing/Swarp_test.py
>>> AssertionError: Incorrect SWarp version. Expected version string to
>>> start with "SWarp version 2.16.0", found "SWarp version 2.15.6
>>> (2006-01-28)".
>>> and I want to solve this before running the other tests.
>>> Can you explain this difference?
>>> It looks like the version of swarp used by  AW in Bonn is 2.15.6 instead
>>> of 2.16.0 but for some reasons the awbase version does NOT detect this
>>> old version at the dependency_test.py but it causes
>>> problems later...
>>>
>>> Cheers,
>>> Philippe
>>>
>>>
>>> John P. McFarland wrote:
>>>
>>>> Hi Philippe,
>>>>
>>>> I assume you are getting these errors while running from Boon.  I have
>>>> recently made a test installation of Astro-WISE in Leiden in
>>>> preparation for a full installation there, but I am currently having
>>>> some issues with LDAC. I tried this command of yours there and
>>>> reproduce the problem exactly. Because I cannot reproduce your problem
>>>> in Groningen, I assume there may be an issue similar to the one in
>>>> Leiden in Bonn.
>>>>
>>>> To test if this is the case, could you please run the unittests there
>>>> and return errors here?  The process would be to run the dependency
>>>> and regression tests as a unit (this will take 10-20 minutes):
>>>>
>>>> # dependency tests
>>>> env database_engine= awe ${AWEPIPE}/astro/test/dependency_test.py
>>>> # functional tests (individual tests in astro/test/functional)
>>>> env database_engine= awe ${AWEPIPE}/astro/test/regression_test.py -f
>>>> # processing tests (individual tests in astro/test/processing)
>>>> env database_engine= awe ${AWEPIPE}/astro/test/regression_test.py -p
>>>>
>>>> When you have an error, for example:
>>>>
>>>> AstrometricParameters_test...  1 ERRORS, 0 FAILURES, 2 Successes
>>>>
>>>> run that test individually (please ignore any GalPhot and TinyTim
>>>> errors, and any version errors for SExtractor and SWarp):
>>>>
>>>> env database_engine= awe
>>>> ${AWEPIPE}/astro/test/processing/AstrometricParameters_test.py
>>>>
>>>> and return the error output.
>>>>
>>>> All this will hopefully allow us to pin down the causes of the errors
>>>> and resolve them quickly.
>>>>
>>>> Cheers,
>>>>
>>>>
>>>> -=John
>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>>
>>
>>
> _______________________________________________
> Issues mailing list
> Issues at astro-wise.org
> http://listman.astro-wise.org/mailman/listinfo/issues
>
>
>
> ** CRM114 Whitelisted by: From: "John P. McFarland" <mcfarland at astro.rug.nl **
>
> ** ACCEPT: CRM114 Whitelisted by: From: "John P. McFarland" <mcfarland at astro.rug.nl **
>
>