[Bioc-devel] msPurity check failure on Windows

Hi Martin/all

It looks like the build/check [1] is still using an older version of the
code (v0.99.9) so it is still failing the check in Windows. i.e. it has
not updated to version (v0.99.12) [2] of the code.

Is this going to be a problem for the package to be accepted into the
next stable release?

Please let me know if there is anything I can do.

Kind regards,
Tom

[1] http://bioconductor.org/checkResults/devel/bioc-LATEST/msPurity/
[2] https://hedgehog.fhcrc.org/bioconductor/trunk/madman/Rpacks/msPurity/

On Thu, Oct 13, 2016 at 4:16 PM, Thomas Lawson
<thomas.nigel.lawson at gmail.com <mailto:thomas.nigel.lawson at gmail.com>>
wrote:

    Hi Martin,

    That's great thanks!

    After running the following code in my unit test:

    -------------------------------------------------------------------------------------------------------------
    msmsPths <- list.files(system.file("extdata", "lcms", "mzML",
    package="msPurityData"), full.names = TRUE, pattern = "MSMS")
    xset <- xcms::xcmsSet(msmsPths, nSlaves = 1)
    xset <- xcms::group(xset)
    xset <- xcms::retcor(xset)
    xset <- xcms::group(xset)
    -------------------------------------------------------------------------------------------------------------

    I get an error that is coming from either xcms::xcmsSet or
    xcms::group, see the bottom of the message for full error.

    -------------------------------------------------------------------------------------------------------------
    Error in signalCondition(e) :
      no function to return from, jumping to top level
    Calls: <Anonymous> -> .handleSimpleError -> h -> signalCondition
    -------------------------------------------------------------------------------------------------------------

    I was thinking it was do with how I was calling the XCMS functions
    but perhaps it is more likely to do with the "nSlaves" parameter for
    xcmsSet that is now deprecated.

    Either way I am now finding it difficult to repeat the error on my
    local windows machine.

    I think the following updated code should now pass (I have also
    removed the additional retcor and group function calls as they were
    not required for the testing).

    -------------------------------------------------------------------------------------------------------------
    library(xcms)
    msmsPths <- list.files(system.file("extdata", "lcms", "mzML",
    package="msPurityData"), full.names = TRUE, pattern = "MSMS")
    xset <- xcmsSet(msmsPths)
    xset <- group(xset)
    -------------------------------------------------------------------------------------------------------------

    My most recent code should is now on the git mirror [1] and the test
    script in question is called test.lcms.R [2]

    Kind regards,
    Tom

    [1] https://github.com/Bioconductor-mirror/msPurity/
    <https://github.com/Bioconductor-mirror/msPurity/>

    [2]
    https://github.com/Bioconductor-mirror/msPurity/blob/master/tests/testthat/test.lcms.R
    <https://github.com/Bioconductor-mirror/msPurity/blob/master/tests/testthat/test.lcms.R>


    Error message

    1: xcms::xcmsSet at testthat/test.lcms.R:7
    2: getExportedValue(pkg, name)
    3: asNamespace(ns)
    4: getNamespace(ns)
    5: tryCatch(loadNamespace(name), error = function(e) stop(e))
    6: tryCatchList(expr, classes, parentenv, handlers)
    7: tryCatchOne(expr, names, parentenv, handlers[[1L]])
    8: doTryCatch(return(expr), name, parentenv, handler)
    9: loadNamespace(name)
    10: namespaceImportFrom(ns, loadNamespace(j <- i[[1L]], c(lib.loc, .libPaths()), versionCheck = vI[[j]]),
           i[[2L]], from = package)
    11: asNamespace(ns)
    12: loadNamespace(j <- i[[1L]], c(lib.loc, .libPaths()), versionCheck = vI[[j]])
    13: methods::cacheMetaData(ns, TRUE, ns)
    14: .updateMethodsInTable(fdef, where, attach)
    15: .mergeMethodsTable(generic, mtable, tt, attach)
    16: .resetTable(table, ns, signames)
    17: as(ifelse(seqN > length(newSig), anyLabel, newSig), "signature")
    18: asMethod(object)
    19: `as<-`(`*tmp*`, "character", value = c("ANY", "ANY"))
    20: asMethod(object, Class, value)
    21: setDataPart(structure(character(0), .Names = character(0), package = character(0), class = structure("signature", package = "methods")),
           c("ANY", "ANY"))
    22: .dataSlot(names(slots))
    23: match(dataSlot, slotNames, 0)
    24: (function (x)
       x$.self$finalize())(<environment>)

    Error in signalCondition(e) :
      no function to return from, jumping to top level
    Calls: <Anonymous> -> .handleSimpleError -> h -> signalCondition
    166 229 291 354 416 479
    Retention Time Correction Groups: 351
    166 229 291 354 416 479
    164 227 289 352 414 477
    Retention Time Correction Groups: 762
    164 227 289 352 414 477
    testthat results ================================================================
    OK: 63 SKIPPED: 0 FAILED: 1
    1. Error: checking lcms based functions (@test.lcms.R#7)

    Error: testthat unit tests failed
    Execution halted




    On Thu, Oct 13, 2016 at 1:57 PM, Martin Morgan
    <martin.morgan at roswellpark.org
    <mailto:martin.morgan at roswellpark.org>> wrote:

        On 10/13/2016 08:17 AM, Thomas Lawson wrote:

            Hi all,

            Thanks for everybody who worked to get mzR to build and pass
            on Windows.

            The package msPurity, which relies on mzR, was building and
            passing all
            checks after the mzR updates but then unfortunately started
            failing on a
            unit test on the Windows server.

            The problem seems to be from when I call xcms::group in a
            unit test. The
            error message goes away on my local Windows machine when I
            load in the XCMS
            library as 'library(xcms)' rather than calling directly
            'xcms::group' in
            the test script.

            I know we are well past the package deadline for this
            release, but I was
            just checking if there is perhaps a small chance of getting
            my package into
            the stable release (if it passes the CMD check on Windows by
            tomorrow)?


        yes, your package will be included in the release.

        But your approach sounds like a workaround rather than the right
        solution to the problem. What specifically was the error on Windows?

        Martin


            Kind regards,
            Tom

                    [[alternative HTML version deleted]]