Re: R package CI test failures
On Fri, May 13, 2016 at 05:37:35PM +0200, Gordon Ball wrote:
> I don't have permissions for debian-med, nor am I a DM/DD, so I'm afraid
> I haven't committed any patches or uploaded them, but a couple should be
> attached here.
In any case I've added you to the Debian Med team to enable you
commiting next time.
> > A suggested patch (commited to SVN + team upload) would be great.
>
> My suggested patch for this is just to rm debian/tests/control (so the
> test is still installed for manual running, but won't be autopkgtested).
> CI will have to make do with the autodep8 load test (which should
> hopefully appear before too long - the relevant version of autodep8 (>=
> 0.5) is now in jessie-backports and will presumably make it to the CI
> installation at some point.
Done since last week in version 2.28.0-1.
> >> r-bioc-cummerbund
> >> -----------------
> >>
> >> Both vignettes are patched (suppress_test_writing_to_usr.patch) removing
> >> attempts to write to a readonly dir, but leaving subsequent code which
> >> consequently fails (data hasn't been loaded).
> >>
> >> Possibly changing the patch to copy extdata from the installed package
> >> and replace the `system.file` call with a local dir would let these run.
> >
> > A suggested patch (commited to SVN + team upload) would be great.
>
> Patch (git format-patch for the repository) is attached. It
> suppress_test_writing_to_usr.patch, which removed some of the pre-amble
> to each vignette with one which modifies the paths the vignettes check
> for their data files. They should now build, but the example data has to
> be copied to the pwd (run-unit-test has been updated to do this).
I'll check the patch soon. Upgrade of the package is pending anyway but
there are some missing preconditions for the new version.
> >> r-bioc-genefilter
> >> -----------------
> >>
> >> Two vignettes (independent_filtering{,_plots}.Rnw) require knitr
> >> (unpackaged), not Stangle for compilation.
> >
> > knitr is ITPed (#808155) but needs some dependencies itself.
Its available now.
> >> r-bioc-genomicalignments
> >> ------------------------
> >>
> >> I think this is a bug with the test from upstream.
> >> test_readGAlignments.R :: test_read_GAlignments_BamViews creates a
> >> temporary file, doesn't hold onto the file handle and then tries to read
> >> it. Hence a race condition depending on whether R has yet
> >> garbage-collected the reference and the file has consequently been
> >> deleted first.
> >
> > Hoping for upstream to respond to my first mail. I might ping at
> > end of the week with these more detailed information.
>
> I'm now less certain this is exactly the cause - it's still somehow
> related to a tempfile/tempdir race, but I misunderstood the semantics of
> tempfile() so the above is probably not exactly correct.
> >
> >> r-bioc-limma
> >> ------------
> >>
> >> Compared stdout of the test script to a saved copy. It's probably
> >> sufficient to just run the script and test the exit code, since this
> >> seems to fail on trivial differences.
> >
> > I might have a look. Needs to be updated anyway.
>
> Attached a replacement run-unit-test script which does this.
I'll check this out.
> >> r-bioc-shortread
> >> ----------------
> >>
> >> A bug somehow related to the package's class system, but I haven't
> >> managed to identify a cause.
> >
> > I might ping upstream.
>
> Error disappeared in 1.30.0
Yep.
> >> r-cran-doparallel
> >> -----------------
> >>
> >> Uses R CMD BATCH instead of R --no-save < filename to run tests, so all
> >> the output is hidden in the CI output. Doesn't copy the test files to a
> >> temporary dir so when they try and create output it fails due to read-only.
> >
> > (Hopefully) fixed in upload of new upstream version.
> >
>
> Still appears present in 1.0.10. I think the test script needs to be
> patched since it hardcodes an attempt to write a report to the directory
> in which the package is installed. I'll try and write a patch for this.
Fine. As I said you should be able to commit now.
> In addition, a re-do of the r-bioc-snpstats run-unit-test is attached
> (three vignettes should be blacklisted, but the current script only gets
> two of them).
I'll check soon.
Thanks a lot for your input
Andreas.
--
http://fam-tille.de
Reply to: