HI Kasper, Thanks for that info, it seems the mess is not confined to Rsamtools But I have a solution "of sorts" --------------------------------- if I use install.packages() and choose packages interactively: FOR R version 2.13.0 Under development (unstable) (2010-11-01 r53513) I get the following : trying URL 'http://www.bioconductor.org/packages/2.7/bioc/src/contrib/IRanges_1.8 .2.tar.gz' (1.9.6 should be used) trying URL 'http://www.bioconductor.org/packages/2.7/bioc/src/contrib/Rsamtools_1 .2.0.tar.gz' (1.3.3 should be used) trying URL 'http://www.bioconductor.org/packages/2.7/bioc/src/contrib/ShortRead_1 .8.0.tar.gz' (1.9.2 ....) are downloaded with update.packages() I can install Rsamtools_1.3.3.tar.gz but needs the latest IRanges and I'm then on the dependency treadmill.... Problem is for both 2010-10-26 r53428 and r53513 installed from svn checkout https://svn.r-project.org/R/trunk/ /media/Bioinform-D/Research/R_latest_install etc ... BOTH source("http://bioconductor.org/biocLite.R") update.packages(repos=biocinstallRepos(), ask=FALSE) OR source("http://bioconductor.org/biocLite.R") pkgs <- rownames(installed.packages()) biocLite(pkgs) Do not help it all otherwise appears up-to date $$$$$$$$ HOWEVER if I use > biocLite("Rsamtools") I get trying URL 'http://www.bioconductor.org/packages/2.8/bioc/src/contrib/Rsamtools_1 .3.3.tar.gz' and the correct version of ShortRead etc etc So its just install.packages() and using the Tcl/Tk interface to install that is causing the problem....? Not sure how that gets fixed.... Cheers Paul -----Original Message----- From: Kasper Daniel Hansen <kasperdanielhansen@gmail.com> To: Paul Leo <p.leo@uq.edu.au> Cc: Dan Tenenbaum <dtenenba@fhcrc.org>, BioC <bioconductor@stat.math.ethz.ch> Subject: Re: [BioC] Rsamtools from biocLite using R-devel Date: Mon, 1 Nov 2010 23:00:46 -0400 Hi Paul The issue was (on my end at least) fixed this afternoon (EST time). I now get the new version of Rsamtools (and a lot of other core packages) when I use biocLite under R-devel. But earlier today the source tarball was broken. I have no idea why a svn checkout would not work. As far as I know, that has been working for some days. Did you remember to rebuild the source tarball before installing, like R-devel CMD build Rsamtools R-devel CMD INSTALL Rsamtools_1.3.3.tar.gz Note that your svn checkout version should certainly be 1.3.3. I would recommend R> source("http://www.bioconductor.org/biocLite.R") R> update.packages(repos = biocinstallrepos()) in a fresh R-devel session. Double-check that the version of Rsamtools it is downloading is something like 1.3.3 (and not 1.1.17). Given the problems there has been the last couple of days it might be nice if you report back whether or not you got this fixed. Kasper On Mon, Nov 1, 2010 at 9:43 PM, Paul Leo <p.leo@uq.edu.au> wrote: It appears this is a known issue, was the fix time just optimistic? I did a svn checkout today abs R and Rsamtools do not play nice. ShortRead will not load without Rsamtools either.... ** preparing package for lazy loading Attaching package: 'IRanges' The following object(s) are masked from 'package:base': Map, cbind, eval, mapply, order, paste, pmax, pmax.int, pmin, pmin.int, rbind, rep.int , table Error : object ‘normalizePath’ is not exported by 'namespace:utils' ERROR: lazy loading failed for package ‘Rsamtools’ * removing ‘/home/pleo/R_latest/library/Rsamtools’ The downloaded packages are in ‘/tmp/RtmpuHAod7/downloaded_packages’ Updating HTML index of packages in '.Library' Warning message: In install.packages() : installation of package 'Rsamtools' had non-zero exit status > sessionInfo() R version 2.13.0 Under development (unstable) (2010-11-01 r53513) Platform: x86_64-unknown-linux-gnu (64-bit) locale: [1] LC_CTYPE=en_AU.UTF-8 LC_NUMERIC=C [3] LC_TIME=en_AU.UTF-8 LC_COLLATE=en_AU.UTF-8 [5] LC_MONETARY=C LC_MESSAGES=en_AU.UTF-8 [7] LC_PAPER=en_AU.UTF-8 LC_NAME=C [9] LC_ADDRESS=C LC_TELEPHONE=C [11] LC_MEASUREMENT=en_AU.UTF-8 LC_IDENTIFICATION=C attached base packages: [1] stats graphics grDevices utils datasets methods base loaded via a namespace (and not attached): [1] tcltk_2.13.0 tools_2.13.0 -----Original Message----- From: Dan Tenenbaum <dtenenba@fhcrc.org> To: Kasper Daniel Hansen <kasperdanielhansen@gmail.com> Cc: BioC <bioconductor@stat.math.ethz.ch> Subject: Re: [BioC] Rsamtools from biocLite using R-devel Date: Thu, 28 Oct 2010 07:41:22 -0700 On Thu, Oct 28, 2010 at 7:06 AM, Kasper Daniel Hansen < kasperdanielhansen@gmail.com> wrote: > Just a note about this: so far I have not seen any resolution. And > according to the devel build logs the R-devel on lamb2 (which I assume > is being used for generating the source tarballs, but I could be > completely wrong) is from 9-30 which is almost a month old (which kind > of surprised me) (pelham is 1.5 months old). > > Kasper, R and Rsamtools should play well together now on lamb2. R 2.13 has been updated on all build nodes in the last couple of days. The build report is out of date due to other issues which will be resolved today. Thanks Dan > Kasper > > On Tue, Oct 26, 2010 at 4:11 PM, Martin Morgan <mtmorgan@fhcrc.org> wrote: > > On 10/26/2010 12:47 PM, Kasper Daniel Hansen wrote: > >> When I try to use biocLite to upgrade my Rsamtools under R-devel, > >> biocLite reports that it tries to use the 2.8 repository (correct), > >> but it tries to download Rsamtools 1.1.17 whereas the version of > >> Rsamtools installed using R-2.12 is version 1.2.0 (later than 1.1.17) > >> and the version accessible from subversion is version 1.3.2. > >> > >> I suspect this has something to do with the recent problem with > >> R-devel and biocLite and also by the fact that Rsamtools currently > >> does not seem to build under any platform. > >> > >> Besides the obvious problem, I am running into problems with my > >> existing Rsamtools because normalizePath has been moved from utils to > >> base (which is commented on in NAMESPACE). Coincidently, it seems > >> that the issue that makes Rsamtools 1.3.2 fail to build is also > >> related to the use of normalizePath that may have its arguments > >> changed, > > > > yes you and I are ahead of the build system -- Rsamtools wants > > normalizePath to be in base, but R-devel on the build system was not (it > > has been updated very recently) current enough for that. So you were > > getting the last Rsamtools that the 'devel' portion of the build system > > correctly built, v. 1.1.17. > > > > This is being straightened out, and should be back to normal over the > > next day or so. The short-term solution is to install Rsamtools from > > source, from an svn checkout > > > > http://bioconductor.org/developers/source-control/ > > > > Martin > > > >> > >> Kasper > >> > >> _______________________________________________ > >> Bioconductor mailing list > >> Bioconductor@stat.math.ethz.ch > >> https://stat.ethz.ch/mailman/listinfo/bioconductor > >> Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > > > > > > -- > > Computational Biology > > Fred Hutchinson Cancer Research Center > > 1100 Fairview Ave. N. PO Box 19024 Seattle, WA 98109 > > > > Location: M1-B861 > > Telephone: 206 667-2793 > > > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > > [[alternative HTML version deleted]] _______________________________________________ Bioconductor mailing list Bioconductor@stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/bioconductor Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor [[alternative HTML version deleted]]

Hi Kasper, The "GUI" is: setRepositories() (default launches a tcltk window to select repositories) install.packages() (default launches a tcltk window to select libraries) I see now that setRepositories() uses > tools:::.read_repositories function (file) { db <- utils::read.delim(file, header = TRUE, comment.char = "#", colClasses = crep.int("character", 3L), rep.int("logical", 4L))) db[, "URL"] <- .expand_BioC_repository_URLs(db[, "URL"]) db } <environment: namespace:tools=""> which provides the incorrect version of Bioc at .expand_BioC_repository_URLs(db[, "URL"]) I think. The first time I have encountered that in using the development version for 3 years... I see now that BiocLite has the bioC version hard-wired so avoids the issue, as you have said. Guess I was unaware that setRepositories() were unsafe on the development version. I promise not to stray from the path again ;-) Versions and platform as per my previous email Thanks for you help Best wishes Paul -----Original Message----- From: Kasper Daniel Hansen <kasperdanielhansen@gmail.com> To: Paul Leo <p.leo@uq.edu.au> Cc: Dan Tenenbaum <dtenenba@fhcrc.org>, BioC <bioconductor@stat.math.ethz.ch> Subject: Re: [BioC] Rsamtools from biocLite using R-devel Date: Tue, 2 Nov 2010 09:19:18 -0400 Hi Paul A couple of immediate comments on your output below (1) Yes, upgrading Rsamtools means upgrading IRanges means ..., but that is unavoidable. If you are using R-devel, this is what happens. If you want stability, you should use R-2.12 and Bioconductor 2.7. (2) Your GUI (which I have no idea what is, please tell us (and also the platform)), is broken. If you look at the URLs below you will see that it tries to pull tarballs from Bioconductor 2.7 under R-devel, which is a big no-no. This is why we have biocLite and why there are thousands of posts on this email list that says always use biocLite. If you have packages installed under R-devel from the GUI you should expect that your entire Bioconductor installation has problems. Of course, since we are close to release Biconductor 2.7 is approximately equal to Bioconductor 2.8, but the discrepancy will increase as times go on. You are already experiencing the first big discrepancy. To be precise, the only way to get Rsamtools to work under current R-devel is to get the newest version which depends on the newest IRanges. There is no way around it. Now, I would simply do as I said in my previous post (source biocLite, use update.packages(repos = biocinstallrepos()). This might update quite a few packages, but there is no way around it. And if you are using a fresh R-devel (as I can see you are), you are on the bleeding edge anyway. Trying to deal with this by selectively updating one package at a time will only lead to frustration. Kasper On Tue, Nov 2, 2010 at 1:57 AM, Paul Leo <p.leo@uq.edu.au> wrote: HI Kasper, Thanks for that info, it seems the mess is not confined to Rsamtools But I have a solution "of sorts" --------------------------------- if I use install.packages() and choose packages interactively: FOR R version 2.13.0 Under development (unstable) (2010-11-01 r53513) I get the following : trying URL 'http://www.bioconductor.org/packages/2.7/bioc/src/contrib/IRa nges_1.8.2.tar.gz' (1.9.6 should be used) trying URL 'http://www.bioconductor.org/packages/2.7/bioc/src/contrib/Rsa mtools_1.2.0.tar.gz' (1.3.3 should be used) trying URL 'http://www.bioconductor.org/packages/2.7/bioc/src/contrib/Sho rtRead_1.8.0.tar.gz' (1.9.2 ....) are downloaded with update.packages() I can install Rsamtools_1.3.3.tar.gz but needs the latest IRanges and I'm then on the dependency treadmill.... Problem is for both 2010-10-26 r53428 and r53513 installed from svn checkout https://svn.r-project.org/R/trunk/ /media/Bioinform-D/Research/R_latest_install etc ... BOTH source("http://bioconductor.org/biocLite.R") update.packages(repos=biocinstallRepos(), ask=FALSE) OR source("http://bioconductor.org/biocLite.R") pkgs <- rownames(installed.packages()) biocLite(pkgs) Do not help it all otherwise appears up-to date $$$$$$$$ HOWEVER if I use > biocLite("Rsamtools") I get trying URL 'http://www.bioconductor.org/packages/2.8/bioc/src/contrib/Rsa mtools_1.3.3.tar.gz' and the correct version of ShortRead etc etc So its just install.packages() and using the Tcl/Tk interface to install that is causing the problem....? Not sure how that gets fixed.... Cheers Paul -----Original Message----- From: Kasper Daniel Hansen <kasperdanielhansen@gmail.com> To: Paul Leo <p.leo@uq.edu.au> Cc: Dan Tenenbaum <dtenenba@fhcrc.org>, BioC <bioconductor@stat.math.ethz.ch> Subject: Re: [BioC] Rsamtools from biocLite using R-devel Date: Mon, 1 Nov 2010 23:00:46 -0400 Hi Paul The issue was (on my end at least) fixed this afternoon (EST time). I now get the new version of Rsamtools (and a lot of other core packages) when I use biocLite under R-devel. But earlier today the source tarball was broken. I have no idea why a svn checkout would not work. As far as I know, that has been working for some days. Did you remember to rebuild the source tarball before installing, like R-devel CMD build Rsamtools R-devel CMD INSTALL Rsamtools_1.3.3.tar.gz Note that your svn checkout version should certainly be 1.3.3. I would recommend R> source("http://www.bioconductor.org/biocLite.R") R> update.packages(repos = biocinstallrepos()) in a fresh R-devel session. Double-check that the version of Rsamtools it is downloading is something like 1.3.3 (and not 1.1.17). Given the problems there has been the last couple of days it might be nice if you report back whether or not you got this fixed. Kasper On Mon, Nov 1, 2010 at 9:43 PM, Paul Leo <p.leo@uq.edu.au> wrote: It appears this is a known issue, was the fix time just optimistic? I did a svn checkout today abs R and Rsamtools do not play nice. ShortRead will not load without Rsamtools either.... ** preparing package for lazy loading Attaching package: 'IRanges' The following object(s) are masked from 'package:base': Map, cbind, eval, mapply, order, paste, pmax, pmax.int, pmin, pmin.int, rbind, rep.int , table Error : object ‘normalizePath’ is not exported by 'namespace:utils' ERROR: lazy loading failed for package ‘Rsamtools’ * removing ‘/home/pleo/R_latest/library/Rsamtools’ The downloaded packages are in ‘/tmp/RtmpuHAod7/downloaded_packages’ Updating HTML index of packages in '.Library' Warning message: In install.packages() : installation of package 'Rsamtools' had non-zero exit status > sessionInfo() R version 2.13.0 Under development (unstable) (2010-11-01 r53513) Platform: x86_64-unknown-linux-gnu (64-bit) locale: [1] LC_CTYPE=en_AU.UTF-8 LC_NUMERIC=C [3] LC_TIME=en_AU.UTF-8 LC_COLLATE=en_AU.UTF-8 [5] LC_MONETARY=C LC_MESSAGES=en_AU.UTF-8 [7] LC_PAPER=en_AU.UTF-8 LC_NAME=C [9] LC_ADDRESS=C LC_TELEPHONE=C [11] LC_MEASUREMENT=en_AU.UTF-8 LC_IDENTIFICATION=C attached base packages: [1] stats graphics grDevices utils datasets methods base loaded via a namespace (and not attached): [1] tcltk_2.13.0 tools_2.13.0 -----Original Message----- From: Dan Tenenbaum <dtenenba@fhcrc.org> To: Kasper Daniel Hansen <kasperdanielhansen@gmail.com> Cc: BioC <bioconductor@stat.math.ethz.ch> Subject: Re: [BioC] Rsamtools from biocLite using R-devel Date: Thu, 28 Oct 2010 07:41:22 -0700 On Thu, Oct 28, 2010 at 7:06 AM, Kasper Daniel Hansen < kasperdanielhansen@gmail.com> wrote: > Just a note about this: so far I have not seen any resolution. And > according to the devel build logs the R-devel on lamb2 (which I assume > is being used for generating the source tarballs, but I could be > completely wrong) is from 9-30 which is almost a month old (which kind > of surprised me) (pelham is 1.5 months old). > > Kasper, R and Rsamtools should play well together now on lamb2. R 2.13 has been updated on all build nodes in the last couple of days. The build report is out of date due to other issues which will be resolved today. Thanks Dan > Kasper > > On Tue, Oct 26, 2010 at 4:11 PM, Martin Morgan <mtmorgan@fhcrc.org> wrote: > > On 10/26/2010 12:47 PM, Kasper Daniel Hansen wrote: > >> When I try to use biocLite to upgrade my Rsamtools under R-devel, > >> biocLite reports that it tries to use the 2.8 repository (correct), > >> but it tries to download Rsamtools 1.1.17 whereas the version of > >> Rsamtools installed using R-2.12 is version 1.2.0 (later than 1.1.17) > >> and the version accessible from subversion is version 1.3.2. > >> > >> I suspect this has something to do with the recent problem with > >> R-devel and biocLite and also by the fact that Rsamtools currently > >> does not seem to build under any platform. > >> > >> Besides the obvious problem, I am running into problems with my > >> existing Rsamtools because normalizePath has been moved from utils to > >> base (which is commented on in NAMESPACE). Coincidently, it seems > >> that the issue that makes Rsamtools 1.3.2 fail to build is also > >> related to the use of normalizePath that may have its arguments > >> changed, > > > > yes you and I are ahead of the build system -- Rsamtools wants > > normalizePath to be in base, but R-devel on the build system was not (it > > has been updated very recently) current enough for that. So you were > > getting the last Rsamtools that the 'devel' portion of the build system > > correctly built, v. 1.1.17. > > > > This is being straightened out, and should be back to normal over the > > next day or so. The short-term solution is to install Rsamtools from > > source, from an svn checkout > > > > http://bioconductor.org/developers/source-control/ > > > > Martin > > > >> > >> Kasper > >> > >> _______________________________________________ > >> Bioconductor mailing list > >> Bioconductor@stat.math.ethz.ch > >> https://stat.ethz.ch/mailman/listinfo/bioconductor > >> Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > > > > > > -- > > Computational Biology > > Fred Hutchinson Cancer Research Center > > 1100 Fairview Ave. N. PO Box 19024 Seattle, WA 98109 > > > > Location: M1-B861 > > Telephone: 206 667-2793 > > > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > > [[alternative HTML version deleted]] _______________________________________________ Bioconductor mailing list Bioconductor@stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/bioconductor Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor [[alternative HTML version deleted]]