Closed krlmlr closed 1 year ago
at least for devtools::release()?
devtools::release()
Use case: a stray testthat-problems.Rds causes R CMD build . to increase (!!!) the version number in DESCRIPTION to 3.5 instead of failing the build. This is very easy to go unnoticed. Can we make this a failure here?
testthat-problems.Rds
R CMD build .
DESCRIPTION
Related: https://github.com/r-lib/testthat/pull/1292.
It's an entirely different discussion if R CMD build . should ever touch my DESCRIPTION.
This seems hard, because R CMD build runs in a separate process, and the exit status is still zero if it bumps the version number.
R CMD build
at least for
devtools::release()
?Use case: a stray
testthat-problems.Rds
causesR CMD build .
to increase (!!!) the version number inDESCRIPTION
to 3.5 instead of failing the build. This is very easy to go unnoticed. Can we make this a failure here?Related: https://github.com/r-lib/testthat/pull/1292.
It's an entirely different discussion if
R CMD build .
should ever touch myDESCRIPTION
.