[Bioc-devel] deprecation guidelines
Dear Bioconductor, We are deprecating part of a package. I am following the guidelines in the "Function Deprecation Guidelines" page at bioconductor.org and I have a few questions: (i) Deprecating a S4 method which has the same name in another S4 class. We have two DelayedArray backends. Functions like dim and dimnames are deprecated in one but not in the other. Should those functions be listed in the ThePkg-deprecated document? With the name of the backend that's being deprecated? With a note about the backend that is *not* being deprecated? (ii) Deprecating a S4 class. Should this look just like a function in ThePkg-deprecated? (iii) subscript, double-subscript, assign-to-property (property<-) operators. Should they be included in ThePkg-deprecated.Rd? In favor: they are being deprecated for the class they pertain to. Against: we certainly aren't deprecating the subscript operator in general. thanks, - Sam ___ Bioc-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/bioc-devel
[Bioc-devel] celaya2 and merida2 are running different versions of R
Dear Bioconductor, With your very helpful assistance, for which we are extremely grateful, we have succeeded in removing the errors from the devel branch of our bumphunter package. More or less, anyway. We are passing on celaya2 and failing on merida2. The reason is that celaya2 is running R-devel r76245 and merida2 is running R-devel r75683. The random number generator changed between r75683 and r76245. This is a problem because our unit tests compare the output of bumphunter to a stashed .rda file. So there is no way the .rda file can make both r75683 and r76245 happy. Will this cause any problems for the package's continuation in Bioconductor 3.9? It seems probable but not guaranteed that R 3.6 will ultimately use the new RNG. thanks, - Sam ___ Bioc-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/bioc-devel
[Bioc-devel] error in bumphunter unit test
Dear Bioconductor, Our bumphunter package is failing its unit tests fairly spectacularly: the pval algorithm seems to be returning a data frame full of zeroes. Unfortunately, I have not been able to reproduce this behavior. I have instantiated an EC2 instance with the devel AMI, updated all the packages, installed bumphunter and its dependencies, and run the unit tests without errors. My sessionInfo indicates all R dependencies are identical to the snapshot on the devel build report. However, two things may be causing the disparity in results: (i) bumphunter uses 'doParallel' and 'foreach' at the point that it is failing; is it possible for me to confirm that I am using the same version of these as the build system used. (Note: we are failing even when doParallel defaults to sequential execution, so the low-level interface does not seem to be relevant.) (ii) The devel AMI is running R 3.6 release 75543 and the build system is running release 76205. Could this be the discrepancy we're looking for? We are reluctant to reinstall all of R+Bioc unless it looks relevant. Any other suggestions would be very helpful. We haven't touched the package in months and are totally buffaloed why it has decided to fail just days before the new release. thanks in advance for any advice you can offer, - Sam ___ Bioc-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/bioc-devel
Re: [Bioc-devel] package RBGL requires CRAN dependency on devel branch
Could be an issue with the configure script. Will check. - Sam On 3/25/19 7:32 AM, Vincent Carey wrote: > > External Email - Use Caution > > The error on linux for 3.9: > > ## > ## > ### > ### Running command: > ### > ### /home/biocbuild/bbs-3.9-bioc/R/bin/R CMD INSTALL RBGL > ### > ## > ## > > > * installing to library ‘/home/biocbuild/bbs-3.9-bioc/R/library’ > * installing *source* package ‘RBGL’ ... > ** using staged installation > checking R package BH ... no > configure: error: R package BH not found. > ERROR: configuration failed for package ‘RBGL’ > * removing ‘/home/biocbuild/bbs-3.9-bioc/R/library/RBGL’ > * restoring previous ‘/home/biocbuild/bbs-3.9-bioc/R/library/RBGL’ > Note that BiocParallel also uses BH and succeeds > configure: creating ./config.status > config.status: creating src/Makevars > ** libs > g++ -std=gnu++11 -I"/home/biocbuild/bbs-3.9-bioc/R/include" -DNDEBUG > -I"/home/biocbuild/bbs-3.9-bioc/R/library/BH/include" -I/usr/local/include > -fpic -g -O2 -Wall -c ipcmutex.cpp -o ipcmutex.o > In file included from > /home/biocbuild/bbs-3.9-bioc/R/library/BH/include/boost/random/detail/integer_log2.hpp:19:0, > from > /home/biocbuild/bbs-3.9-bioc/R/library/BH/include/boost/random/detail/large_arithmetic.hpp:19, > from > /home/biocbuild/bbs-3.9-bioc/R/library/BH/include/boost/random/detail/const_mod.hpp:23, > from > /home/biocbuild/bbs-3.9-bioc/R/library/BH/include/boost/random/detail/seed_impl.hpp:26, > from > /home/biocbuild/bbs-3.9-bioc/R/library/BH/include/boost/random/mersenne_twister.hpp:30, > from > /home/biocbuild/bbs-3.9-bioc/R/library/BH/include/boost/uuid/random_generator.hpp:17, > So could it be an issue with the configure script? > > On Mon, Mar 25, 2019 at 7:22 AM Samuela Pollack > mailto:spoll...@jimmy.harvard.edu>> wrote: > > Dear Bioconductor, > > The devel version of package RBGL is flunking build. > > This package has been modified to include header files in the CRAN > package 'BH' instead of using a local tarball of the header files. We > consider this an improvement because the 'BH' maintainers update > their > package every time a new version of boost is released, but rebuilding > the included tarball is unreliable. > > Is it possible to request inclusion of a CRAN package dependency in a > Bioconductor package? If not, how would Bioconductor recommend we > handle > this? > > (Full disclosure: the BH package has all the header files for all of > boost. This is a lot more disk space then we need, because we only > need > the BGL.) > > thanks, > > - Sam > > > > On 3/22/19 7:28 AM, Morgan, Martin wrote: > > External Email - Use Caution > > > > Thanks for the explanation and exemplary detective work. Will > you push these changes to the release and devel branches of RBGL? > I've given you permissions... Martin > > > > On 3/22/19, 5:15 AM, "Samuela Pollack" > mailto:spoll...@jimmy.harvard.edu>> > wrote: > > > > We have a temporary fix which we believe will alleviate the > difficulty. > > > > The only routine in RBGL that does not compile under the > LLVM-9.0 system > > is the routine that implements betweenness connectivity > clustering. This > > seems to be a rarely-used routine, so we have temporarily > removed it > > from RBGL and put in a message explaining how users may > replace it easily. > > > > I have contacted Professor Jeremy Siek at in Bloomington to > discuss how > > best to proceed. It is possible this will have to be fixed > in BGL, > > probably not in time for boost 1.70. > > > > For clarity: the incompatibility is not in clang. The > incompatibility is > > in the llvm c++ standard library. Mac users who build with > clang will be > > fine, as long as they don't link in the llvm C++ stdlib, > which most > > wouldn't think to do and Apple does not encourage. > > > > We hope this will be sufficient for no
[Bioc-devel] package RBGL requires CRAN dependency on devel branch
Dear Bioconductor, The devel version of package RBGL is flunking build. This package has been modified to include header files in the CRAN package 'BH' instead of using a local tarball of the header files. We consider this an improvement because the 'BH' maintainers update their package every time a new version of boost is released, but rebuilding the included tarball is unreliable. Is it possible to request inclusion of a CRAN package dependency in a Bioconductor package? If not, how would Bioconductor recommend we handle this? (Full disclosure: the BH package has all the header files for all of boost. This is a lot more disk space then we need, because we only need the BGL.) thanks, - Sam On 3/22/19 7:28 AM, Morgan, Martin wrote: External Email - Use Caution Thanks for the explanation and exemplary detective work. Will you push these changes to the release and devel branches of RBGL? I've given you permissions... Martin On 3/22/19, 5:15 AM, "Samuela Pollack" wrote: We have a temporary fix which we believe will alleviate the difficulty. The only routine in RBGL that does not compile under the LLVM-9.0 system is the routine that implements betweenness connectivity clustering. This seems to be a rarely-used routine, so we have temporarily removed it from RBGL and put in a message explaining how users may replace it easily. I have contacted Professor Jeremy Siek at in Bloomington to discuss how best to proceed. It is possible this will have to be fixed in BGL, probably not in time for boost 1.70. For clarity: the incompatibility is not in clang. The incompatibility is in the llvm c++ standard library. Mac users who build with clang will be fine, as long as they don't link in the llvm C++ stdlib, which most wouldn't think to do and Apple does not encourage. We hope this will be sufficient for now. - Sam On 3/22/19 4:18 AM, Morgan, Martin wrote: > External Email - Use Caution > > What's the status for RBGL? > > Thanks, Martin > > On 3/22/19, 3:57 AM, "Prof Brian Ripley" wrote: > > clang 8.0.0 is now released, and these fail there too (and break about > 50 CRAN packages which depend on them). > > Reminder: this is with clang's native libc++, as documented at > https://www.stats.ox.ac.uk/pub/bdr/Rconfig/r-devel-linux-x86_64-fedora-clang > > We really need fixes in both 3.8 and 3.9. > > BDR > > On 28/02/2019 18:52, Prof Brian Ripley wrote: > > These packages fail to install with clang 8.0.0rc3, which is supposed to > > be near-final (it is overdue for release). (AFAIR they did install with > > rc1.) In both cases the problems are with Boost. > > > > RBGL 1.95.1 > > > > In file included from bbc.cpp:1: > > In file included from ./RBGL.hpp:7: > > In file included from /usr/local/clang8/bin/../include/c++/v1/iostream:38: > > In file included from /usr/local/clang8/bin/../include/c++/v1/ios:216: > > In file included from /usr/local/clang8/bin/../include/c++/v1/__locale:15: > > In file included from /usr/local/clang8/bin/../include/c++/v1/string:505: > > In file included from > > /usr/local/clang8/bin/../include/c++/v1/string_view:176: > > In file included from /usr/local/clang8/bin/../include/c++/v1/__string:57: > > /usr/local/clang8/bin/../include/c++/v1/algorithm:2494:5: error: > > static_assert > >failed due to requirement > > > > '__is_forward_iterator > > >long, boost::property >boost::property > boost::no_property> > >, > >void *>, boost::detail::edge_desc_impl > unsigned > >long>, long> >::value' "std::max_element requires a ForwardIterator" > > static_assert(__is_forward_iterator<_ForwardIterator>::value, > > ^ ~~ > > rbgl_trimmed_boost_1_61_0/boost/graph/bc_clustering.hpp:132:26: note: in > >instantiation of function template specialization > > > > 'std::__1::max_element > > >long, boost::property >boost::property > boost::no_property> > >, > >void *>, boost::detail::edg
Re: [Bioc-devel] GenomeInfoDb is breaking bumphunter
OK, thanks Bioconductor, we'll sit tight and wait for it to propagate. - Sam On 2/13/19 10:06 AM, Shepherd, Lori wrote: > > External Email - Use Caution > > > Herve implemented a fix in release and devel as indicated by the > support site question > > https://support.bioconductor.org/p/117808/ > > It has yet to propagate to the system but��should be up later today. > > > > Lori Shepherd > > Bioconductor Core Team > > Roswell Park Cancer Institute > > Department of Biostatistics & Bioinformatics > > Elm & Carlton Streets > > Buffalo, New York 14263 > > -------- > *From:* Bioc-devel on behalf of > Samuela Pollack > *Sent:* Wednesday, February 13, 2019 10:00:18 AM > *To:* Rafael Irizarry; Leonardo Collado Torres; Bioc-devel > *Subject:* [Bioc-devel] GenomeInfoDb is breaking bumphunter > Dear Bioconductor, > > The GenomeInfoDb package is generating an error in 3.8 and causing our > package bumphunter to crash too. > > Our test calls GenomeInfoDb::fetchExtendedChromInfoDb and crashes with > an error that says some UCSC seqlevels in hg38 couldn't be mapped to > NCBI seqlevels. The release version of GenomeInfoDb is crashing in the > same function with the same error message. > > The github repo for GenomeInfoDb indicates some work was done on this > routine two days ago. It looks like maybe we need a way to set the > drop_unmapped flag? Is that possible? What would you recommend? > > thanks, > > - Sam > > > > > > On 2/12/19 12:55 PM, Rafael Irizarry wrote: > > see below > > > > -- Forwarded message - > > From: > > Date: Tue, Feb 12, 2019 at 12:48 PM > > Subject: bumphunter problems reported in the Multiple platform > > build/check report for BioC 3.8 > > To: > > > > > >� External Email - Use Caution > > > > [This is an automatically generated email. Please don't reply.] > > > > Hi bumphunter maintainer, > > > > According to the Multiple platform build/check report for BioC 3.8, > > the bumphunter package has the following problem(s): > > > >��� o ERROR for 'R CMD check' on malbec1. See the details here: > > > https://master.bioconductor.org/checkResults/3.8/bioc-LATEST/bumphunter/malbec1-checksrc.html > > > > Please take the time to address this by committing and pushing > > changes to your package at git.bioconductor.org > > > > Notes: > > > >��� * This was the status of your package at the time this email was > sent to you. > >� Given that the online report is updated daily (in normal > conditions) you > >� could see something different when you visit the URL(s) above, > especially if > >� you do so several days after you received this email. > > > >��� * It is possible that the problems reported in this report are false > > positives, > >� either because another package (from CRAN or Bioconductor) > breaks your > >� package (if yours depends on it) or because of a Build System > problem. > >� If this is the case, then you can ignore this email. > > > >��� * Please check the report again 24h after you've committed your > changes to the > >� package and make sure that all the problems have gone. > > > >��� * If you have questions about this report or need help with the > >� maintenance of your package, please use the Bioc-devel mailing > list: > > > > https://bioconductor.org/help/mailing-list/ > > > >� (all package maintainers are requested to subscribe to this list) > > > > For immediate notification of package build status, please > > subscribe to your package's RSS feed. Information is at: > > > > https://bioconductor.org/developers/rss-feeds/ > > > > Thanks for contributing to the Bioconductor project! > > ___ > Bioc-devel@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/bioc-devel > > This email message may contain legally privileged and/or confidential > information. If you are not the intended recipient(s), or the employee > or agent responsible for the delivery of this message to the intended > recipient(s), you are hereby notified that any disclosure, copying, > distribution, or use of this email message is prohibited. If you have > received this message in error, please notify the sender immediately > by e-mail and delete this email message from your computer. Thank you. [[alternative HTML version deleted]] ___ Bioc-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/bioc-devel
[Bioc-devel] GenomeInfoDb is breaking bumphunter
Dear Bioconductor, The GenomeInfoDb package is generating an error in 3.8 and causing our package bumphunter to crash too. Our test calls GenomeInfoDb::fetchExtendedChromInfoDb and crashes with an error that says some UCSC seqlevels in hg38 couldn't be mapped to NCBI seqlevels. The release version of GenomeInfoDb is crashing in the same function with the same error message. The github repo for GenomeInfoDb indicates some work was done on this routine two days ago. It looks like maybe we need a way to set the drop_unmapped flag? Is that possible? What would you recommend? thanks, - Sam On 2/12/19 12:55 PM, Rafael Irizarry wrote: see below -- Forwarded message - From: Date: Tue, Feb 12, 2019 at 12:48 PM Subject: bumphunter problems reported in the Multiple platform build/check report for BioC 3.8 To: External Email - Use Caution [This is an automatically generated email. Please don't reply.] Hi bumphunter maintainer, According to the Multiple platform build/check report for BioC 3.8, the bumphunter package has the following problem(s): o ERROR for 'R CMD check' on malbec1. See the details here: https://master.bioconductor.org/checkResults/3.8/bioc-LATEST/bumphunter/malbec1-checksrc.html Please take the time to address this by committing and pushing changes to your package at git.bioconductor.org Notes: * This was the status of your package at the time this email was sent to you. Given that the online report is updated daily (in normal conditions) you could see something different when you visit the URL(s) above, especially if you do so several days after you received this email. * It is possible that the problems reported in this report are false positives, either because another package (from CRAN or Bioconductor) breaks your package (if yours depends on it) or because of a Build System problem. If this is the case, then you can ignore this email. * Please check the report again 24h after you've committed your changes to the package and make sure that all the problems have gone. * If you have questions about this report or need help with the maintenance of your package, please use the Bioc-devel mailing list: https://bioconductor.org/help/mailing-list/ (all package maintainers are requested to subscribe to this list) For immediate notification of package build status, please subscribe to your package's RSS feed. Information is at: https://bioconductor.org/developers/rss-feeds/ Thanks for contributing to the Bioconductor project! ___ Bioc-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/bioc-devel
Re: [Bioc-devel] Request upstream pull --force of sanitized master branch
Hi Nitesh, I have fixed the commit history for rafalab/bumphunter RELEASE_3_8 branch, and request synchronization with Bioconductor RELEASE_3_8. We need this, because bumphunter is not building on RELEASE_3_8, due to an obsolete URL in one of the tests. thanks, - Sam On 11/5/18 7:56 AM, Turaga, Nitesh wrote: External Email - Use Caution Hi Sam, I have updated your “master” branch on git.bioconductor.com to reflect what you have on GitHub, at https://github.com/rafalab/bumphunter. You will not get the duplicate commit error when you push to the master branch anymore. However, RELEASE_3_8 still has duplicate commits in the history. You could follow the same procedure as you did for the master branch, and de-duplicate the commit history for RELEASE_3_8. You should invest time in this only if you think there will be any changes made to the release branch. If not, then this is not an issue. If you do choose to follow through with RELEASE_3_8 as well, please reply to this thread when you have made the de-duplication changes and added the branch to your GitHub repository. Best regards, Nitesh On Nov 2, 2018, at 5:09 PM, Samuela Pollack wrote: Dear Bioconductor, I believe I have removed all duplicate commits from the bumphunter package. The main branch in rafalab/bumphunter incorporates the new commit history. It passes 'library(devtools); build("."); install(); check("."); BiocCheck(".")' I am contacting you in accordance with item 6 of the instructions at this URL: http://bioconductor.org/developers/how-to/git/resolve-duplicate-commits/ - Sam ___ Bioc-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/bioc-devel This email message may contain legally privileged and/or confidential information. If you are not the intended recipient(s), or the employee or agent responsible for the delivery of this message to the intended recipient(s), you are hereby notified that any disclosure, copying, distribution, or use of this email message is prohibited. If you have received this message in error, please notify the sender immediately by e-mail and delete this email message from your computer. Thank you. ___ Bioc-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/bioc-devel
[Bioc-devel] Request upstream pull --force of sanitized master branch
Dear Bioconductor, I believe I have removed all duplicate commits from the bumphunter package. The main branch in rafalab/bumphunter incorporates the new commit history. It passes 'library(devtools); build("."); install(); check("."); BiocCheck(".")' I am contacting you in accordance with item 6 of the instructions at this URL: http://bioconductor.org/developers/how-to/git/resolve-duplicate-commits/ - Sam ___ Bioc-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/bioc-devel