conda-forge / rstudio-feedstock

A conda-smithy repository for rstudio.
BSD 3-Clause "New" or "Revised" License
6 stars 9 forks source link

rstudio-desktop downgrading to gcc9 despite build being with 10? #2

Closed gdevenyi closed 2 years ago

gdevenyi commented 2 years ago

Comment:

Trying to integrate rstudio-desktop into my existing R environment I find it wants to make a lot of changes:

  Downgrade:                                                                                             
─────────────────────────────────────────────────────────────────────────────────────────────

  gcc_impl_linux-64             10.3.0  hf2f2afa_16         installed                      
  gcc_impl_linux-64              9.4.0  h03d3576_16         conda-forge/linux-64      43 MB
  gcc_linux-64                  10.3.0  hc39de41_9          installed                      
  gcc_linux-64                   9.4.0  h391b98a_9          conda-forge/linux-64      25 KB
  gfortran_impl_linux-64        10.3.0  h73f4979_16         installed                      
  gfortran_impl_linux-64         9.4.0  h0003116_16         conda-forge/linux-64      15 MB
  gfortran_linux-64             10.3.0  hb09a455_9          installed                      
  gfortran_linux-64              9.4.0  hf0ab688_9          conda-forge/linux-64      24 KB
  graphviz                       3.0.0  h5abf519_1          installed                      
  graphviz                      2.50.0  h85b4f2f_1          conda-forge/linux-64       3 MB
  gxx_impl_linux-64             10.3.0  hf2f2afa_16         installed                      
  gxx_impl_linux-64              9.4.0  h03d3576_16         conda-forge/linux-64      11 MB
  gxx_linux-64                  10.3.0  h2593f52_9          installed                      
  gxx_linux-64                   9.4.0  h0316aca_9          conda-forge/linux-64      24 KB
  harfbuzz                       4.2.1  hf9f4e7c_0          installed                      
  harfbuzz                       3.4.0  hb4a5f5f_0          conda-forge/linux-64       2 MB
  icu                             70.1  h27087fc_0          installed                      
  icu                             69.1  h9c3ff4c_0          conda-forge/linux-64     Cached
  imagemagick                 7.1.0_34  pl5321heb7c40d_0    installed                      
  imagemagick                 7.1.0_27  pl5321hae5efcc_0    conda-forge/linux-64       3 MB
  libgcc-devel_linux-64         10.3.0  he6cfe16_16         installed                      
  libgcc-devel_linux-64          9.4.0  hd854feb_16         conda-forge/linux-64       4 MB
  librsvg                       2.54.3  h7abd40a_0          installed                      
  librsvg                       2.52.5  hc3c00ef_1          conda-forge/linux-64       5 MB
  libsanitizer                  10.3.0  h26c7422_16         installed                      
  libsanitizer                   9.4.0  h79bfe98_16         conda-forge/linux-64       7 MB
  libstdcxx-devel_linux-64      10.3.0  he6cfe16_16         installed                      
  libstdcxx-devel_linux-64       9.4.0  hd854feb_16         conda-forge/linux-64      10 MB
  libxml2                       2.9.14  h22db469_0          installed                      
  libxml2                       2.9.12  h885dcf4_1          conda-forge/linux-64     Cached
  openssl                        3.0.3  h166bdaf_0          installed                      
  openssl                       1.1.1o  h166bdaf_0          conda-forge/linux-64       2 MB
  pango                         1.50.7  hbd2fdc8_0          installed                      
  pango                        1.48.10  h4dcc4a0_3          conda-forge/linux-64     399 KB
  r-base                         4.1.3  h06d3f91_1          installed                      
  r-base                         4.1.2  hde4fec0_0          conda-forge/linux-64      25 MB
  xorg-libx11                    1.7.2  h7f98852_0          installed                      
  xorg-libx11                   1.6.12  h36c2ea0_0          conda-forge/linux-64     919 KB

I'm surprised its built against older versions despite being brand new? This also seems to conflict with how its built here: https://github.com/conda-forge/rstudio-feedstock/blob/main/.ci_support/linux_64_r_base4.1.yaml#L13-L14

izahn commented 2 years ago

This is normal for new packages because they are built with current global pins. In this case the global pango pin is 1.48. That forces rstudio-desktop to depend on r-base=4.1.2 because r-base=4.1.3 has already been migrated to pango 1.50. In due course the bot will open a PR here to migrate rstudio to pango=1.50 and any other active migrations that apply to this feedstock. As those migration PRs are merged here this package will become compatible with more recent r-base builds.

gdevenyi commented 2 years ago

Is there a development guide for conda-forge or such which describes the moving parts of how all this happens? I have similar confusion regarding openssl 1/3 versions. Thanks!

izahn commented 2 years ago

Things are documented pretty well in https://conda-forge.org/docs/ but didn't really start to understand how things worked until I started maintaining packages myself.

gdevenyi commented 2 years ago

Thanks!