IRC channel logs


back to list of logs

<rekado>zimoun: thanks!
<rekado>zimoun: I’m reviewing your wip-bioconductor branch
<rekado>rebased it on top of master, updated r-ggtern, but this needs R 4.2.2 instead of 4.2.1
<rekado>so now I’m upgrading R
<rekado>wasn’t there something you wanted me to merge…? A patch by cbaines…?
<zimoun>rekado: thanks.
<zimoun>Hum, I do not remember something to merge…
<rekado>found it: 59073
<rekado>zimoun_: do you happen to have more details about these DRAFT commits? They look very non-drafty to me.
<zimoun_>rekado: for instance, «DRAFT gnu: r-ggplot2: Update to 3.4.0.» the updater recommended 3 packages, but it builds without.
<zimoun_>idem for «DRAFT gnu: r-enrichplot: Update to 1.18.0.».
<zimoun_>these recommended packages are commented in the patch.
<zimoun_>about «DRAFT gnu: r-rgreat: Update to 2.0.1.» well, I was not sure since the change is big.
<rekado>do you have a list of the recommended packages?
<rekado>I prefer to add them when recommended, because R code can fail at runtime when a package is missing.
<rekado>ah, commented in the patch
<rekado>good, will check again
<zimoun_>well, if the vignette is correctly designed, and so if the tests pass, then they are just extra; often more for upstream convience than true requirements. But who can more can also less. ;-)
<rekado>with propagation we often don’t see problems even when inputs are in fact missing.
<rekado>because they happen to be always installed with those packages that they declare as inputs
<rekado>this can change, though, when other packages are modified
<rekado>this has happened in the past, so I prefer to do as the importer says.