[Bioc-devel] Access UnitTests Log-File
dtenenba at fhcrc.org
Fri Jan 25 20:59:49 CET 2013
On Thu, Jan 24, 2013 at 3:04 PM, Dan Tenenbaum <dtenenba at fhcrc.org> wrote:
> Hi Anita,
> On Thu, Jan 24, 2013 at 2:33 PM, Anita Lerch <lerch.anita at gmail.com> wrote:
>> is there a way to access the log file of the unit-tests
>> (e.g. QuasR.Rcheck/tests/QuasR_unit_tests.Rout.fail) on the build system?
>> Currently I have failing test on george2, which I can't reproduce on my
>> The 13 line output e.g.
>> * checking tests ...
>> Running ‘QuasR_unit_tests.R’
>> Running the tests in ‘tests/QuasR_unit_tests.R’ failed.
>> Last 13 lines of output:
>> Test full-methylated directed, mode=allC
>> Test files with failing tests
>> Error in QuasR:::test() : unit tests failed for package QuasR
>> of the check report
>> is not very informative, especially if the failing test can't be reproduce.
>> It would be a great help if the log file (e.g.
>> could be copied to (e.g.
>> http://bioconductor.org/checkResults/devel/bioc-LATEST/QuasR/) somehow.
> At present that log file is not accessible through the build report.
But this is a good idea and I will work on adding this to the build
system over the next few weeks.
> In your case the relevant excerpt is:
> RUNIT TEST PROTOCOL -- Thu Jan 24 03:18:17 2013
> Number of test functions: 59
> Number of errors: 0
> Number of failures: 2
> 1 Test Suite :
> QuasR RUnit Tests - 59 test functions, 0 errors, 2 failures
> FAILURE in test_bisulfit_single: Error in checkTrue(all(ifelse(readInfo[, 8] ==
> "l", readInfo[, 3], readInfo[, :
> Test not TRUE
> Test left read position: l 1 145 1
> FAILURE in test_full_meth_paired: Error in checkTrue(all(mcols(meth)[, 1] == mco
> ls(meth)[, 2]), "Test full-methylated directed, mode=allC") :
> Test not TRUE
> Test full-methylated directed, mode=allC
> Test files with failing tests
> I cannot reproduce this either, even on george2. It seems to happen
> only in the context of the build system's nightly runs. I think it has
> something to do with the parallelization in the code and may have to
> do with the fact that the builds are already parallelized. I'll look
> into it some more.
>> Sorry, if I missed something on the devel webpage.
>> [[alternative HTML version deleted]]
>> Bioc-devel at r-project.org mailing list
More information about the Bioc-devel