[BioC] RWebServices failing on mkserver (ant) target.

[Michael P Shipway]

Thank you Martin, your suggestion got me pointed in the correct direction.
It seems it was an Axis feature that was tripping me up.
I found an Axis mailing list archive entry which seem to apply.
http://mail-archives.apache.org/mod_mbox/ws-axis-user/200512.mbox/%3c43A3AE62.8080006@jp.fujitsu.com%3e

Simply put, Axis doesn't like the package name to be the same as an included function.
Therefore, it appended "_PortType" to one of them, but didn't use the changed name in all places.

When I changed the spelling of the R package name, my problem went away.


Michael Shipway
mshipway at jhu.edu
Center for Cardiovascular Bioinformatics and Modeling 
Institute for Computational Medicine 
Johns Hopkins University 
214 Computational Science & Engineering Building
3400 N. Charles St. 
Baltimore, MD 21218 
(410) 516-8616 


----- Original Message -----
From: Martin Morgan <mtmorgan at fhcrc.org>
Date: Monday, May 18, 2009 10:43 pm
Subject: Re: [BioC] RWebServices failing on mkserver (ant) target.
To: mshipway at jhu.edu
Cc: bioconductor at stat.math.ethz.ch


> Hi Michael --
>  
>  Glad to hear you're giving RWebServices a try! I think the problem is
>  occuring in the ant gen-wsdl stage, and suspect that the problem is
>  use of axis other than 1.4. This is just from googling '_PortType'; I
>  don't see these issues with my set-up. Does that help? If not it might
>  help to get the version information on all the tools that are involved
>  -- library(RWebServices); sessionInfo() in R, ant -version, and axis,
>  at least.
>  
>  Also, I'd encourage use of the 'devel' version of R, RWebServices, and
>  SJava, e.g.,
>  
>    source('
>    biocLite('RWebServices')
>  
>  from within the development version of R, or (not encouraged) using
>  svn from
>  
>    
>  
>  Martin
>  
>  Michael P Shipway <mshipwa1 at jhu.edu> writes:
>  
>  > I am trying to install and use RWebServices on a system running 
> CentOS 5 (Linux).
>  > I've installed all the prerequisites, including the optional ones.
>  > I've created a simple R package to test the setup
>  > which takes a single numeric parameter and returns a canned text string.
>  >
>  > I am hung up on step 4.2 of the “Enabling R packages for web or 
> grid services”
>  >  vignette document, which builds a web service server.
>  > One of the java code files which was autogenerated by an earlier 
> step implements 
>  > a java class which does not exist
>  > 
> (org.bioconductor.rserviceJms.services.helloWorld.HelloWorld_PortType 
> ). 
>  >
>  > It appears that this class should also have been autogenerated, 
>  > but I am unable to determine where this should have happened. 
>  >
>  > Can anyone suggest what I may be doing wrong?
>  >

-snip-