[Debian-med-packaging] [devteam-bioc] Failures in test suites of some BioConductor packages

[Martin Morgan]

Hi Andreas --

Yes, in principle, these would be helpful, but see below

On 04/29/2016 02:28 AM, Maintainer wrote:
> Hi BioConductor maintainers,
>
> most of the BioConductor packages inside Debian are running the test
> suite regularly as part of a Continuous Integration[1] effort.  I realised
> that some packages are failing and I wonder how we could ideally deal with

In each of these, you're using the release candidate (R-3.3.0 series) of 
R, but the already released version 3.2 of Bioconductor. Bioconductor 
3.2 is meant to work with the R-3.2.* series.

Bioconductor has a release (3.3) timed immediately after R's

     http://bioconductor.org/developers/release-schedule/

The protocol that we follow is to impose a code freeze (too loosely 
observed) on trunk, and to branch immediately before release (May 3). So 
the relevant packages are those in the current Bioc-devel. You can 
compare your builds with our nightly builds on the devel branch at

     http://bioconductor.org/checkResults/devel/bioc-LATEST/

> this.  I'd like to provide you with a set of packages and the link to the
> log of the failure:
>
>     bioMart:
>       https://ci.debian.net/data/packages/unstable/amd64/r/r-bioc-biomart/20160428_233412.autopkgtest.log.gz
>

This is a result of a web service query. The web service is (or was at 
the time of the build) not responding correctly; the error is likely 
transient; it was seen in our own nightly build 
http://bioconductor.org/checkResults/devel/bioc-LATEST/biomaRt/zin2-buildsrc.html.

>     GenomicAlignments:
>       https://ci.debian.net/data/packages/unstable/amd64/r/r-bioc-genomicalignments/20160429_032733.autopkgtest.log.gz

This is a result of a change in an underlying package (BiocParallel), 
and was fixed by version 1.7.8,

------------------------------------------------------------------------
r111667 | mtmorgan at fhcrc.org | 2015-12-19 14:18:52 -0500 (Sat, 19 Dec 
2015) | 2 lines
Changed paths:
    M /trunk/madman/Rpacks/GenomicAlignments/DESCRIPTION
    M 
/trunk/madman/Rpacks/GenomicAlignments/inst/unitTests/test_readGAlignments.R

cleanup: update unit test for BiocParallel changes

------------------------------------------------------------------------


>     RSamtools:
>       https://ci.debian.net/data/packages/unstable/amd64/r/r-bioc-rsamtools/20160429_033451.autopkgtest.log.gz

This no longer occurs, although I am not sure exactly which commit fixed 
the issue; I think it was in Rsamtools 1.23.1

------------------------------------------------------------------------
r110234 | mtmorgan at fhcrc.org | 2015-11-02 19:53:29 -0500 (Mon, 02 Nov 
2015) | 4
lines
Changed paths:
    M /trunk/madman/Rpacks/Rsamtools/DESCRIPTION
    M /trunk/madman/Rpacks/Rsamtools/inst/unitTests/test_utilities.R
    M /trunk/madman/Rpacks/Rsamtools/src/samtools_patch.c

fail on knet_connect_file error, instead of warning

- introduces small memory leak in C

------------------------------------------------------------------------


>     ShortRead:
>       https://ci.debian.net/data/packages/unstable/amd64/r/r-bioc-shortread/20160429_002544.autopkgtest.log.gz
>

This was fixed in ShortRead version 1.29.1 with

------------------------------------------------------------------------
r109902 | mtmorgan at fhcrc.org | 2015-10-25 02:13:12 -0400 (Sun, 25 Oct 
2015) | 4
lines
Changed paths:
    M /trunk/madman/Rpacks/ShortRead/DESCRIPTION
    M /trunk/madman/Rpacks/ShortRead/R/readAligned.R

unname colClasses in read.table

- probably a bug introduced into R-devel

------------------------------------------------------------------------


> The log contains the packages and the versions installed and the failure
> that occured on a machine which has only these packages installed and is
> not connected to the internet.

biomaRt will always fail, then.

>
> Is this information helpful and sufficient for you to give some hints how
> to solve these issues or do you need further information?
>
> Kind regards
>
>         Andreas.
>
>
> [1] https://ci.debian.net/
>


This email message may contain legally privileged and/or confidential information.  If you are not the intended recipient(s), or the employee or agent responsible for the delivery of this message to the intended recipient(s), you are hereby notified that any disclosure, copying, distribution, or use of this email message is prohibited.  If you have received this message in error, please notify the sender immediately by e-mail and delete this email message from your computer. Thank you.