[Bioc-devel] Would it make sense for BiocManager::install to cache downloaded packages?

[Vincent Carey]

Could we think about a parameter use_unused_tarballs that automates this
reduction in redownload effort?

On Wed, Nov 4, 2020 at 1:52 PM Martin Morgan <mtmorgan.bioc using gmail.com>
wrote:

> FWIW BiocManager::install() really delegates to install.packages(). The
> packages are downloaded to 'destdir=' as documented on ?install.packages --
> typically file.path(tempdir(), "downloaded_packages") -- and one path to
> troubleshooting installs is to tackle the problematic installation(s) using
> `install.packages(pkgs_in_downloaded_packages, repos = NULL)`, where
> pkgs_in_downloaded_packages are the packages that have failed to install on
> first pass.
>
> It does seem like this would be a reasonable feature for base R, that we
> could inherit, so one could ask on R-devel or submit a bug report / feature
> request (e.g., via `bug.report()`).
>
> Martin
>
> On 11/4/20, 1:19 PM, "Bioc-devel on behalf of Vincent Carey" <
> bioc-devel-bounces using r-project.org on behalf of stvjc using channing.harvard.edu>
> wrote:
>
>     I certainly have run into this situation with R generally.  It seems
> like a
>     patch to
>     install.packages would be the best approach to introduce this
>     functionality.  I wonder
>     if Henrik Bengtsson or Dirk Eddelbuettel have pondered this.
>
>     In any case, I am sympathetic with the suggestion but I believe this
> has to
>     be done
>     and tested outside the core developer team -- resources are too scant,
> and
>     the
>     event of concern seems relatively rare.  But maybe I overestimate the
>     burden.
>
>     On Wed, Nov 4, 2020 at 12:19 PM Keith Hughitt <keith.hughitt using gmail.com
> >
>     wrote:
>
>     > Sometimes I will go to update a machine / do a clean Bioconductor
> install
>     > and a large number of packages will need to be downloaded.
>     >
>     > If something goes wrong during the installation, which is not
> uncommon in
>     > cases where the number of packages being installed is large, then the
>     > process has to be restarted and the packages re-downloaded all-over
> again.
>     >
>     > Often, it may take some trial-and-error to fix the problem
> preventing the
>     > install from completing, the result of which is that you may end up
>     > re-downloading 50+ packages several times.
>     >
>     > Would it perhaps make sense for BiocManager to download
> packages/checksums
>     > to somewhere like "~/.cache/bioc/pkg" so that they could be re-used
> during
>     > future installs?
>     >
>     > Obviously, it would be good for BiocManger to have some ability to
> suggest
>     > old packages for removal (e.g. outdated/haven't been used in a long
> time)
>     > to avoid filling up people's hard-drives.
>     >
>     > Another possibility would be to build on top of the work done by
> renv,
>     > which already has a very nice caching system.
>     >
>     > In any case, it seems like it would be something (relatively)
>     > straight-forward to implement that would save a lot of bandwidth and
> time
>     > for users.
>     >
>     > Just a thought..
>     > Keith
>     >
>     >         [[alternative HTML version deleted]]
>     >
>     > _______________________________________________
>     > Bioc-devel using r-project.org mailing list
>     > https://stat.ethz.ch/mailman/listinfo/bioc-devel
>     >
>
>     --
>     The information in this e-mail is intended only for the
> ...{{dropped:18}}
>
>     _______________________________________________
>     Bioc-devel using r-project.org mailing list
>     https://stat.ethz.ch/mailman/listinfo/bioc-devel
>

-- 
The information in this e-mail is intended only for the ...{{dropped:18}}