[Bioc-devel] affxparser: Core dump with R 2.14.0 on OSX

Kasper Daniel Hansen kasperdanielhansen at gmail.com
Tue Nov 8 21:21:32 CET 2011


Reading Dan's error I realize I did something else: he points to the zip archive
  pathname <- "~/Downloads/250k_nsp_libraryfile_rev4.zip"
whereas I unzip the archive, and point it directly to the cdf file
inside some subdirectory of the unpacked zip file.

We do not expect that readCdfHeader on a zip archive works.  We are
aware that reading "wrong" files may lead to a core dump which is
rather bad and we should try to fix this.

Kasper


On Tue, Nov 8, 2011 at 3:04 PM, Dan Tenenbaum <dtenenba at fhcrc.org> wrote:
> Hi Henrik,
>
> On Mon, Nov 7, 2011 at 2:41 PM, Henrik Bengtsson <hb at biostat.ucsf.edu> wrote:
>> Hi,
>>
>> over at the aroma.affymetrix mailing list, a user (cc:ed) reports that
>> affxparser v1.26.1 core dumps when trying to read a CDF file.  I'm
>> moving it to BioC-devel, and asks if others can reproduce this.  I
>> don't have access to R v2.14.0 on OSX.  It does not core dump on
>> Windows 7 with the same setup and versions.  The OP says he got the
>> problems after updated to R v2.14.0.  The CDF file
>> (Mapping250K_Nsp.cdf) is available in (34Mb download after login):
>>
>>  http://www.affymetrix.com/Auth/support/downloads/library_files/250k_nsp_libraryfile_rev4.zip
>>
>
> I am able to reproduce the problem:
>
>> library(digest)
>> library(affxparser)
>> pathname <- "~/Downloads/250k_nsp_libraryfile_rev4.zip"
>> fi <- file.info(pathname)
>> rownames(fi) <- NULL
>> print(fi)
>      size isdir mode               mtime               ctime
> 1 45043262 FALSE  644 2011-11-08 11:48:45 2011-11-08 11:48:45
>                atime uid gid    uname grname
> 1 2011-11-08 11:54:09 502  20 dtenenba  staff
>> print(digest(file=pathname))
> [1] "cf10a53b36108315542276875d43c24a"
>> print(sessionInfo())
> R version 2.14.0 (2011-10-31)
> Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)
>
> locale:
> [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
>
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
>
> other attached packages:
> [1] affxparser_1.26.1   digest_0.5.1        BiocInstaller_1.2.0
>
> loaded via a namespace (and not attached):
> [1] tools_2.14.0
>> hdr <- readCdfHeader(pathname)
> Abort trap: 6
>
> It happens in 32-bit R as well.
>
> That's on a Lion machine. On Leopard, I get a different error:
>
>> library(affxparser)
>> pathname <- "~/Downloads/250k_nsp_libraryfile_rev4.zip"
>> hdr <- readCdfHeader(pathname)
> Failed to read the CDF file header for:
> /Users/biocbuild/Downloads/250k_nsp_libraryfile_rev4.zip
> Error in basename(res$chiptype) : a character vector argument expected
>> sessionInfo()
> R version 2.14.0 (2011-10-31)
> Platform: i386-apple-darwin9.8.0/i386 (32-bit)
>
> locale:
> [1] en_US.US-ASCII/en_US.US-ASCII/en_US.US-ASCII/C/en_US.US-ASCII/en_US.US-ASCII
>
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
>
> other attached packages:
> [1] affxparser_1.26.1   BiocInstaller_1.2.0
>
> loaded via a namespace (and not attached):
> [1] tools_2.14.0
>> library(digest)
>> fi <- file.info(pathname)
>> rownames(fi) <- NULL
>> fi
>      size isdir mode               mtime               ctime
> 1 45043262 FALSE  644 2011-11-08 11:58:53 2011-11-08 11:58:53
>                atime uid gid     uname grname
> 1 2011-11-08 12:01:17 502  20 biocbuild  staff
>> digest(file=pathname)
> [1] "cf10a53b36108315542276875d43c24a"
>
>
> Dan
>
>
>
>> More details below.
>>
>> BTW, Peter, could you please check if you have issues with any other
>> CDF files, such as the Mapping10K_Xba142.cdf?
>>
>> /Henrik
>>
>> On Mon, Nov 7, 2011 at 2:27 PM, Peter Kang wrote:
>>> Henrik,
>>> Thanks for your advice. This is what I get:
>>>> library("affxparser");
>>>> library("digest");
>>>>
>>>> pathname <-
>>>> "annotationData/chipTypes/Mapping250K_Nsp/Mapping250K_Nsp.cdf";
>>>> fi <- file.info (pathname);
>>>> rownames(fi) <- NULL;
>>>> print(fi);
>>>        size isdir mode               mtime               ctime
>>> 1 194455495 FALSE  755 2005-08-24 16:01:42 2011-11-07 09:53:41
>>>                 atime uid gid  uname grname
>>> 1 2011-11-07 11:47:58 503  20 hpkang  staff
>>>> print(digest(file=pathname));
>>> [1] "59ae263311a2cf63b8d1b9b4cc7d663b"
>>>>
>>>> print(sessionInfo());
>>> R version 2.14.0 (2011-10-31)
>>> Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)
>>> locale:
>>> [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
>>> attached base packages:
>>> [1] stats     graphics  grDevices utils     datasets  methods   base
>>> other attached packages:
>>> [1] digest_0.5.1      affxparser_1.26.1
>>>> hdr <- readCdfHeader(pathname);
>>> zsh: abort (core dumped)  R
>>
>> [...]
>>
>>> On Mon, Nov 7, 2011 at 12:35 PM, Henrik Bengtsson wrote:
>>>>
>>>> Hi,
>>>>
>>>> you can still do sessionInfo() just before you call the core-dumping
>>>> method; sessionInfo() is so key to solving problems, especially core
>>>> dumps so we cannot be without them.  What do you get when you do:
>>>>
>>>> library("affxparser");
>>>> library("digest");
>>>>
>>>> pathname <-
>>>> "annotationData/chipTypes/Mapping250K_Nsp/Mapping250K_Nsp.cdf";
>>>> fi <- file.info(pathname);
>>>> rownames(fi) <- NULL;
>>>> print(fi);
>>>> print(digest(file=pathname));
>>>>
>>>> print(sessionInfo());
>>>> hdr <- readCdfHeader(pathname);
>>>> str(hdr);
>>>>
>>>> Also, if you get some core dump messages, what are they?
>>>>
>>>> After you reply here, I'm most likely going to move this to the
>>>> Bioconductor devel mailing list, because that is most likely where the
>>>> real help will come from.
>>>>
>>>> /Henrik
>>>>
>>>> On Mon, Nov 7, 2011 at 11:54 AM, Peter Kang wrote:
>>>> > Sorry, I couldn't do sessionInfo() or traceback() because of the core
>>>> > dump.
>>>> > The core dump was with all of the latest packages in R 2.14.0. The funny
>>>> > thing is it worked perfectly well before I upgraded R and reinstalled
>>>> > aroma.
>>>> > I tried Henrik's suggestion, and get the 'core dumped' error after
>>>> > hdr <- readCdfHeader(pathname);
>>>> > I also redownloaded the cdf files from the Affy website, no difference.
>>>> > Thank you.
>>>> >
>>>> > On Mon, Nov 7, 2011 at 11:19 AM, Henrik Bengtsson
>>>> > <henrik.bengtsson at aroma-project.org> wrote:
>>>> >>
>>>> >> On Mon, Nov 7, 2011 at 9:35 AM, Pierre Neuvial wrote:
>>>> >> > What is your sessionInfo() ?
>>>> >> >
>>>> >> > Perhaps you need to upgrade aroma.affymetrix instead, although I
>>>> >> > don't
>>>> >> > see why you would get this core dump.
>>>> >>
>>>> >> Yes, sessionInfo() and showing that you've tried with the latest
>>>> >> version speeds up troubleshooting and response time.
>>>> >>
>>>> >> I the past, basically all core dumps have been attributed to the
>>>> >> affxparser package *and* almost all of those have been due to corrupt
>>>> >> Affymetrix files.   I suspect that affxparser is the player here to,
>>>> >> and maybe even a corrupt file.  So, in a fresh R session try with:
>>>> >>
>>>> >> library("affxparser");
>>>> >> pathname <-
>>>> >> "annotationData/chipTypes/Mapping250K_Nsp/Mapping250K_Nsp.cdf";
>>>> >> hdr <- readCdfHeader(pathname);
>>>> >> str(hdr);
>>>> >> cdfList <- readCdf(pathname);
>>>> >> str(cdfList);
>>>> >>
>>>> >> Core dump?
>>>> >>
>>>> >> /Henrik
>>>> >>
>>>> >> >
>>>> >> > Pierre
>>>> >> >
>>>> >> > On Mon, Nov 7, 2011 at 6:20 PM, Peter Kang wrote:
>>>> >> >> I just upgraded to 2.14.0 (Mac OSX), and I'm getting a 'core dumped'
>>>> >> >> error when trying to access the cdf file.
>>>> >> >>
>>>> >> >>> cdf <- AffymetrixCdfFile$byChipType("Mapping250K_Nsp")
>>>> >> >>> print(cdf)
>>>> >> >> zsh: abort (core dumped)  R
>>>> >> >>
>>>> >> >> Do I need to downgrade R?
>>>> >> >>
>>>> >> >> Thank you.
>>
>> [...]
>>
>> _______________________________________________
>> Bioc-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>
>
> _______________________________________________
> Bioc-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>



More information about the Bioc-devel mailing list