Yep, works on my machine as well. Seems like it would be a pain to have
to do it manually every time csaw (or deepSNV, or diffHic) gets
re-installed, though.
I should stress that I didn't do anything special to my R installation
to get it to install packages in a different place. As far as I know,
that's just what R decided to do by itself. (I never run R with sudo
permissions, so it wouldn't be able to put packages in the central R
installation folder anyway.)
So I think it's worth seriously thinking about whether this scenario
should be supported; a user (i.e., me) would reasonably expect that BioC
packages would work off the bat on this "standard" setting. Perhaps the
csaw.so library can be built with static linkage to the libhts.a?
Or maybe I'm just spouting rubbish here. I don't know, I never do any
serious work on a Mac.
Cheers,
Aaron
On 24/05/16 18:42, Dan Tenenbaum wrote:
----- Original Message -----
From: "Dan Tenenbaum" <dtene...@fredhutch.org>
To: "Aaron Lun" <a...@wehi.edu.au>
Cc: "bioc-devel" <bioc-devel@r-project.org>
Sent: Tuesday, May 24, 2016 9:38:01 AM
Subject: Re: [Bioc-devel] problems linking to Rhtslib on Mac OSX
----- Original Message -----
From: "Dan Tenenbaum" <dtene...@fredhutch.org>
To: "Aaron Lun" <a...@wehi.edu.au>
Cc: "bioc-devel" <bioc-devel@r-project.org>
Sent: Tuesday, May 24, 2016 8:51:30 AM
Subject: Re: [Bioc-devel] problems linking to Rhtslib on Mac OSX
----- Original Message -----
From: "Aaron Lun" <a...@wehi.edu.au>
To: "bioc-devel" <bioc-devel@r-project.org>
Sent: Tuesday, May 24, 2016 8:06:13 AM
Subject: Re: [Bioc-devel] problems linking to Rhtslib on Mac OSX
Dear Martin and List,
I have a problem with linking to Rhtslib on Mac OSX when my R installation
directory differs from the package installation directory. Trying to load csaw
(Bioc release version) gives me:
require(csaw)
# ... whole lot of dependencies...
Loading required package: csaw
Error in dyn.load(file, DLLpath = DLLpath, ...) :
unable to load shared object
'/Users/lun01/Library/R/3.3/library/csaw/libs/csaw.so':
dlopen(/Users/lun01/Library/R/3.3/library/csaw/libs/csaw.so, 6): Library not
loaded:
/Library/Frameworks/R.framework/Versions/3.3/Resources/library/Rhtslib/lib/libhts.0.dylib
Referenced from: /Users/lun01/Library/R/3.3/library/csaw/libs/csaw.so
Reason: image not found
... which is because my Rhtslib shared library is sitting instead at
/Users/lun01/Library/R/3.3/library/Rhtslib/lib/libhts.0.dylib (along with all
my other non-recommended, non-base packages). I presume this linking failure
occurs because the Makevars file (which would normally indicate where the
shared library is living) doesn't get run when installing a prebuilt binary?
The same issue occurs with deepSNV, which is the other package that links to
Rhtslib.
I can't reproduce this issue after installing the binaries of csaw and Rhtslib.
I believe that htslib is statically linked inside of Rhtslib so you don't need
your own installation of htslib on your machine in order to use it.If you look
at the Mac tarball
(https://bioconductor.org/packages/release/bioc/bin/macosx/mavericks/contrib/3.3/Rhtslib_1.4.2.tgz)
untar it and go to the Rhtslib/lib directory you see the following files:
libhts.0.dylib
libhts.a
libhts.dylib
...in addition to Rhtslib.so which is in Rhtslib/libs.
Actually I can induce issues if I install Rhtslib in a different library
directory, which is what you were saying. Will look into this a little.
The following seems to fix it:
cd ~/Library/R/3.3/csaw/libs
# change dtenenba to your username in the below
sudo install_name_tool -change
"/Library/Frameworks/R.framework/Versions/3.3/Resources/library/Rhtslib/lib/libhts.0.dylib"
"/Users/dtenenba/Library/R/3.3/Rhtslib/lib/libhts.0.dylib" csaw.so
(the sudo may not be necessary)
Then I can
require(csaw)
and it works.
I'm not sure if there is something that the bioc group should be doing about
this....we'll discuss it when we meet soon.
Dan
Dan
Dan
Any thoughts? I don't usually use Macs, so I don't know exactly what goes on
during package installation, or whether my setup (i.e., with different
package/R locations) is atypical or not.
Aaron
sessionInfo()
R version 3.3.0 (2016-05-03)
Platform: x86_64-apple-darwin13.4.0 (64-bit)
Running under: OS X 10.10.5 (Yosemite)
locale:
[1] en_GB.UTF-8/en_GB.UTF-8/en_GB.UTF-8/C/en_GB.UTF-8/en_GB.UTF-8
attached base packages:
[1] stats4 parallel stats graphics grDevices utils datasets
[8] methods base
other attached packages:
[1] SummarizedExperiment_1.2.2 Biobase_2.32.0
[3] GenomicRanges_1.24.0 GenomeInfoDb_1.8.1
[5] IRanges_2.6.0 S4Vectors_0.10.1
[7] BiocGenerics_0.18.0 BiocInstaller_1.22.2
loaded via a namespace (and not attached):
[1] AnnotationDbi_1.34.3 XVector_0.12.0 edgeR_3.14.0
[4] GenomicAlignments_1.8.0 zlibbioc_1.18.0 BiocParallel_1.6.2
[7] tools_3.3.0 DBI_0.4-1 Rhtslib_1.4.2
[10] rtracklayer_1.32.0 bitops_1.0-6 RCurl_1.95-4.8
[13] biomaRt_2.28.0 RSQLite_1.0.0 limma_3.28.5
[16] GenomicFeatures_1.24.2 Biostrings_2.40.1 Rsamtools_1.24.0
[19] XML_3.98-1.4
______________________________________________________________________
The information in this email is confidential and inte...{{dropped:6}}
_______________________________________________
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel
_______________________________________________
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel
_______________________________________________
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel
