[Bioc-devel] problem when building oligo on wilson2

Dear Robert and Patrick,

Thank you very much for your time.

Patrick, I just exposed the function, thanks for the heads up.

Robert, thanks for the hint. I'll get r49067 and try to reproduce the
problem.

b

On Aug 7, 2009, at 2:17 PM, Patrick Aboyoun wrote:

Benilton,
I pushed an early report of the builds today that don't include
Windows
because those machines haven't completed the build process yet. For
today's report oligo isn't building on wilson2 (Linux), pitt (Mac  
OS X
Tiger), pelham (Mac OSX Leopard).

http://www.bioconductor.org/checkResults/2.5/bioc-LATEST/oligo/wilson2-buildsrc.html
http://www.bioconductor.org/checkResults/2.5/bioc-LATEST/oligo/pitt-buildsrc.html
http://www.bioconductor.org/checkResults/2.5/bioc-LATEST/oligo/pelham-buildsrc.html

because of the error

Error: processing vignette 'oligo.Rnw' failed with diagnostics:
chunk 3 (label=crlmm)
Error in crlmm(fullFilenames, outputDir, verbose = FALSE) :
could not find function "requireAnnotation"
Execution halted


This missing symbol could be masking the more subtler problem you
mentioned, but at least the current error should be easy to correct.



Patrick




Robert Gentleman wrote:

Hi Benilton,
It could be any one of a number of things, not necessarily the build
of R (although that could be a factor).  Someone here will likely
need
to try to run this and see if it is reproducible from the command
line
and if so, to try and find the problem.  Alternatively you could
check
out that version of R and see if the problem is reproducible on a
computer near you.

best wishes
  Robert


Benilton Carvalho wrote:

(resending the message as my initial attempts were refused)

Hi,

I'm looking for some insight on the source of the problem reported
by
'wilson2' when building oligo:

http://www.bioconductor.org/checkResults/2.5/bioc-LATEST/oligo/wilson2-buildsrc.html


This machine has R-devel r49067... The other machines have r49060
and
r49021.

So, because the pkg builds just fine on other machines but
wilson2, I
thought that the source of the problem would be on changes between
r49060 and r49067.

Looking at the log of R-devel for all these revisions, I don't see
any change that would cause the error "'dimnames' applied to
non-array" to come up.

I just installed the latest R-devel and I can't reproduce the
problem. So, is it just a matter of waiting until R-devel is  
updated
on the machines used to build the packages?

Thanks a lot,

b