[Rcpp-devel] mzR and Rcpp - could not find function "errorOccured"

[Neumann, Steffen]

Hi,

the mzR almost always worked on BioC build machines.
It was the users who had reported failures on their systems out there.

So there is not really sth. we need to do on the BioC side,
but to prevent users from messing up somehow.
That's why I need to understand what went wrong where.

Could it be that BioC changed the linker options for their official builds, 
and regardless of the Rcpp and mzR versions,
you must not mix and match an Rcpp from before and an mzR 
from after the linker switiching ?

Yours,
Steffen

________________________________________
From: Dirk Eddelbuettel [edd at debian.org]
Sent: 29 June 2012 15:52
To: Neumann, Steffen
Cc: Laurent Gatto; Davor Cubranic; rcpp-devel at r-forge.wu-wien.ac.at
Subject: Re: [Rcpp-devel] mzR and Rcpp - could not find function        "errorOccured"

On 29 June 2012 at 13:43, Neumann, Steffen wrote:
| Hi,
|
| what is the outcome of this ? In which constellation does the problem occur ?
| Do we need to bump the minimum Rcpp version in mzR dependenceis ?

Thanks for following up.  I actually did grep through your sources when I
checked that it did in fact work just fine on Linux.

And as mzR does not use the symbol, I think we are victims of a mild compiler
or linker error here: The symbol was there (defined static, but, IIRC,
private) and is no more.

So the internals changed, but the interface did not. A rebuild should
certainly help.

If there is something else we ought to do on the Rcpp side, let me know.

Dirk

PS Shall we talk to some of the BioC folks?

| Yours,
| Steffen
| ________________________________________
| From: rcpp-devel-bounces at lists.r-forge.r-project.org [rcpp-devel-bounces at lists.r-forge.r-project.org] on behalf of Laurent Gatto [lg390 at cam.ac.uk]
| Sent: 27 June 2012 10:09
| To: Davor Cubranic
| Cc: rcpp-devel at r-forge.wu-wien.ac.at
| Subject: Re: [Rcpp-devel] mzR and Rcpp - could not find function        "errorOccured"
|
| On 26 June 2012 21:42, Davor Cubranic <cubranic at stat.ubc.ca> wrote:
| > Doesn't Rcpp 2.11 require 2.15.1, and the computers on which you had problems had 2.15?
| >
| > Also, you don't say if you tried rebuilding your package, or if you just continued using the same binary. It is possible that since Rcpp API changed, the missing symbol is detected by the dynamic link loader and so aborts loading mzR. (Not sure why it would work on some platforms and not others -- maybe dynamic linking is more tolerant than on others.)
|
| I have never had this myself on linux, building from source. These
| issues have been reported by users (one of which was using 2.15 but
| did not specify the package versions), and I have been able to
| reproduce getting the windows binary packages build on the
| Bioconductor servers.
|
| I will try with fresh binaries. Thank you very much for your help!
|
| Laurent
|
|
| > Davor
| >
| >
| > On 2012-06-26, at 12:31 PM, Laurent Gatto wrote:
| >
| >> On 26 June 2012 20:17, Dirk Eddelbuettel <edd at debian.org> wrote:
| >>>
| >>> Laurent,
| >>>
| >>> On 26 June 2012 at 19:39, Laurent Gatto wrote:
| >>> | Dear all,
| >>
| >> [...]
| >>
| >>> I am confused.  So in general no error on Linux and on some versions of
| >>> Windows and OS X, but an error on other versions of OS X and Windows?
| >>>
| >>> Could this be as simple as required a rebuild of the related packaged (eg
| >>> Rcpp and mzR) ?
| >>>
| >>> | I am assuming that this is related to Rcpp, but am pretty much stuck.
| >>> | Could this be related to "errorOccuredSym never used and commented
| >>> | out" from the ChangeLog?
| >>>
| >>> From what I recall I simply followed a warning from g++ about an unused var
| >>> which I commented out in src/Evaluator.cpp.  You could always fetch an older
| >>> src/Evaluator.cpp and drop it in and see what happens.
| >>>
| >>> We'd need something reproducible, preferable on a platform I have access too.
| >>
| >> I understand. I am confused too and not sure what to do about these
| >> sporadic reports. Will hopefully find more compelling data and report
| >> back.
| >>
| >> Thanks,
| >>
| >> Laurent
| >>
| >>> Dirk
| >>>
| >>>
| >>> | Any help greatly appreciated.
| >>> |
| >>> | Best wishes,
| >>> |
| >>> | Laurent
| >>> |
| >>> | [0] http://bioconductor.org/packages/2.10/bioc/html/mzR.html
| >>> | [1] http://stackoverflow.com/questions/11181691/r-bioconductor-mzr-library-load-error
| >>> | [2] https://stat.ethz.ch/pipermail/bioconductor/2012-June/046496.html
| >>> | [3] http://bioconductor.org/checkResults/2.10/bioc-LATEST/moscato2-NodeInfo.html
| >>> | [4] http://bioconductor.org/checkResults/2.10/bioc-LATEST/petty-NodeInfo.html
| >>> | [5] http://bioconductor.org/checkResults/2.10/bioc-LATEST/lamb2-NodeInfo.html
| >>> | _______________________________________________
| >>> | Rcpp-devel mailing list
| >>> | Rcpp-devel at lists.r-forge.r-project.org
| >>> | https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
| >>>
| >>> --
| >>> Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com
| >> _______________________________________________
| >> Rcpp-devel mailing list
| >> Rcpp-devel at lists.r-forge.r-project.org
| >> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
| >
|
|
|
| --
| Laurent Gatto
| - http://proteome.sysbiol.cam.ac.uk/lgatto/
| Cambridge Centre for Proteomics
| - http://www.bio.cam.ac.uk/proteomics
| _______________________________________________
| Rcpp-devel mailing list
| Rcpp-devel at lists.r-forge.r-project.org
| https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel
| _______________________________________________
| Rcpp-devel mailing list
| Rcpp-devel at lists.r-forge.r-project.org
| https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/rcpp-devel

--
Dirk Eddelbuettel | edd at debian.org | http://dirk.eddelbuettel.com