HI Mike, Thanks so much for taking the time to go through everything so carefully. It's very much appreciated.
I originally found the issue with BiocCheck by instantiating the bioc-devel container locally. That one's sorted now so the devel version of ngsReports builds on r-devel using linux. A few dependency issues (GenomeInfoDb) have stymied it for OSX & Windows but that's not particularly surprising at this stage in the devel cycle. At least that version is essentially sorted. As to the current release version, I've tried removing the remotes::system_requirements() step but it seems like it's a genuine problem in the container & I'll leave it for now, or at least until smarter people than me have spotted it & figured it out. Given the devel version builds successfully & the code is 100% identical in release, I'll risk the push to bioc-release. All been a bit infuriating as the only issue I'm trying to resolve was that a deprecation warning added to ggplot2 3.4.0 broke a bunch of my unit tests. The functions all worked fine but I figured I'd be better sorting it all out sooner rather than later. Would've been nice if the ggplot2 update came out before the official Bioc release, but that's the nature of the beast really. I had a look through your code as well & that's really nice! Once this is sorted out, I might start experimenting. Thanks again for taking so much time to try & help. Stevie On Thu, 10 Nov 2022 at 19:05, Mike Smith <grimbo...@gmail.com> wrote: > Hi Stevie, > > To follow up on this, BiocCheck is missing from devel because some changes > in R-devel tripped up a few unit tests, and the package hasn't been built > successfully. It looks like Marcel has fixed those issues ( > https://code.bioconductor.org/browse/BiocCheck/commits/master) and it > should be available in a day or so once the build system has next run. > > I see you also ran into the problem where calling > system_requirements("ubuntu", > "22.04") gives an error about not being able to use Ubuntu 22.04. This is > unfortunate as the fix for this was merged into remotes in April ( > https://github.com/r-lib/remotes/pull/705) but that version has never made > it to CRAN! However, since you're running in the Docker container all of > the system requirements are already installed (that's one of the key > reasons to use the container), so you can actually just remove that step > entirely. If you ever stop using the container you'd need to reinstate it. > > Cheers, > Mike > > > > On Wed, 9 Nov 2022 at 12:04, Mike Smith <grimbo...@gmail.com> wrote: > > > Hi Stevie, > > > > It might be helpful to link to the workflow runs so people can see > exactly > > what you're doing and what's happening. Hopefully I'm looking in the > right > > place at https://github.com/steveped/ngsReports/actions/runs > > > > I'm not sure about the error message regarding BiocCheck. However there > > are couple of things I noticed in your workflow files for the most recent > > failures, which might help: > > > > In the devel workflow ( > > https://github.com/steveped/ngsReports/actions/runs/3425103022/) I'm > > pretty sure r-lib/actions/setup-r has to be given the version 'devel'. > > Using 4.3 won't work because there isn't a URL for R with that version > > until it becomes the latest release. > > > > In the latest release workflow run: > > > > - At line 54 ( > > > https://github.com/steveped/ngsReports/actions/runs/3426634604/workflow#L54 > ) > > you using the RSPM for Ubuntu Focal and > > - At line 118 ( > > > https://github.com/steveped/ngsReports/actions/runs/3426634604/workflow#L118 > ) > > you're calling 'remotes::system_requirements("ubuntu", "20.04"))' > > > > but the docker containers are actually built on Ubuntu 22.04, which is > > Ubuntu Jammy. I suspect that's the reason you're getting > > the libicui18n.so.66 error, because the binary packages you're installing > > are for a different version of Ubuntu. > > > > Also, this might not simplify your workflow too much, but you could try > my > > setup-bioc action ( > > https://github.com/grimbough/bioc-actions/tree/v1-branch/setup-bioc) > > which tries to make keeping R and Bioconductor versions in sync a bit > > easier. > > > > Cheers, > > Mike > > > > > > On Wed, 9 Nov 2022 at 04:43, Stevie Pederson < > > stephen.pederson...@gmail.com> wrote: > > > >> Hi, > >> > >> I also can't get my github actions bioc-check workflow running for > devel, > >> so I checked the bioc-devel docker image & can't seem to install > BiocCheck > >> on that container. Is this likely to be available for devel soon? Error > >> from my local installation of the container below: > >> > >> BiocManager::install("BiocCheck") > >> 'getOption("repos")' replaces Bioconductor standard repositories, see > >> '?repositories' for details > >> > >> replacement repositories: > >> CRAN: https://cloud.r-project.org > >> > >> Bioconductor version 3.17 (BiocManager 1.30.19), R Under development > >> (unstable) (2022-11-07 r83308) > >> Installing package(s) 'BiocCheck' > >> Warning message: > >> package ‘BiocCheck’ is not available for Bioconductor version '3.17' > >> > >> A version of this package for your version of R might be available > >> elsewhere, > >> see the ideas at > >> > >> > https://cran.r-project.org/doc/manuals/r-devel/R-admin.html#Installing-packages > >> > >> [[alternative HTML version deleted]] > >> > >> _______________________________________________ > >> Bioc-devel@r-project.org mailing list > >> https://stat.ethz.ch/mailman/listinfo/bioc-devel > >> > > > > [[alternative HTML version deleted]] > > _______________________________________________ > Bioc-devel@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/bioc-devel > [[alternative HTML version deleted]] _______________________________________________ Bioc-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/bioc-devel