On April 9, 2023 12:51:33 PM UTC, Santiago Vila <sanv...@debian.org> wrote: >Hello. This is an interesting bug indeed, because it forces us to re-read >the specification and maybe suggest changes if we see that it's incomplete >or just suboptimal. > >https://salsa.debian.org/ci-team/autopkgtest/raw/master/doc/README.package-tests.rst > >This is the current wording: > >> The cwd of each test is guaranteed to be the root of the source package, >> which will have been unpacked but not built. > >My interpretation of that is that the cwd of each test is guaranteed >to be the root of the source package in unpacked form. > >Under this interpretation, running autopkgtest on a built tree does not follow >the specification and is not required to work. > >Somebody could also interpret it that "you should not rely on the package >being built", >as it "may or may not have been built".
This was always my interpretation, especially given there is an explicit restriction to have a *built* tree: build-needed The tests need to be run from a built source tree. The test runner will build the source tree (honouring the source package's build dependencies), before running the tests. However, the tests are *not* entitled to assume that the source package's build dependencies will be installed when the test is run. >But I would argue that if that was the intended meaning, the wording would be >quite different. > >So, I would reassign this to whatever package is running autopkgtest in a >built tree >(even if it's autopkgtest itself), as I believe the specification is not being >followed. My above test was done using autopkgtest hello_2.10-3.dsc --setup-commands='apt-get update' -- docker --remote debian:sid But I guess that virt implementation doesn't matter too much here. >On the other hand, if we are going to allow both built and not built trees, >maybe >it would be a good idea to change the wording first so that it's clear for >everybody. > >Thanks.