There are some cases where the required packages are not directly listed in the "R/" directory. For instance, a Rmd file in "inst/" that is run from a function of the package. If there are specific packages in the Rmd file, there are not detected by attachment::att_amend_desc()
I think I will not recommend this practice as it may lead to more packages than needed in Imports of the DESCRIPTION file. I would prefer recommend using @importFrom package function where better suited, or opening an issue here to explain where {attachment} failed to detect a dependency.
[ ] I can directly add an 'Import' during the att_amend_desc() for specific cases with extra.imports field
[ ] extra.imports is more important that extra.suggests
[ ] If I use extra.imports and the package was not already detected as "imports" by {attachment}, there is a message informing the user that this practice must be used with caution because they had hard dependencies.
[ ] Add a comment: please open an issue on {attachment} to explain the use case that forced you to use this parameter
Context
There are some cases where the required packages are not directly listed in the "R/" directory. For instance, a Rmd file in "inst/" that is run from a function of the package. If there are specific packages in the Rmd file, there are not detected by
attachment::att_amend_desc()
See: https://thinkr-open.github.io/attachment/dev/articles/a-fill-pkg-description.html#a-package-was-not-detected-by-att_amend_desc
Validation
att_amend_desc()
for specific cases withextra.imports
fieldextra.imports
is more important thatextra.suggests
extra.imports
and the package was not already detected as "imports" by {attachment}, there is a message informing the user that this practice must be used with caution because they had hard dependencies.