This pull request is intentionally incomplete and just serves to make the work-in-progress more visible.
External dependencies are very limited on anything XS-like at present because we don't know what we can use, if anything, due to EUMM depending on us in some way some of the time, which may be a bootstrap problem when
EUMM -> EUM -> EUM.test -> Time::HiRes.configure -> EUMM
(And that might be a problem during perl bootstrap )
Either way, the tempdir consumption is expressly written such that
it doesn't matter how it works from the code side
the code side just uses runtemp
runtemp can be implemented however we want it to be
.... which could be T:TD:T ... who knows.
The initial commit factors out tests specific to mkmanifest invocation and there is one remanining call in the main .t yet to be isolated ( Because I don't know where to put it, mostly ).
Doing my best to migrate changes from Manifest.t to the relevant test files in steps so its obvious that our test coverage stays the same as it was ( or better ).
The new tests are going to be much larger by default due to the levels of isolation vs the existing one where every test is inextricably dependent on all the ones before it.
This pull request is intentionally incomplete and just serves to make the work-in-progress more visible.
External dependencies are very limited on anything XS-like at present because we don't know what we can use, if anything, due to EUMM depending on us in some way some of the time, which may be a bootstrap problem when
EUMM -> EUM -> EUM.test -> Time::HiRes.configure -> EUMM
(And that might be a problem during perl bootstrap )
Either way, the tempdir consumption is expressly written such that
runtempruntempcan be implemented however we want it to beThe initial commit factors out tests specific to
mkmanifestinvocation and there is one remanining call in the main .t yet to be isolated ( Because I don't know where to put it, mostly ).Doing my best to migrate changes from
Manifest.tto the relevant test files in steps so its obvious that our test coverage stays the same as it was ( or better ).The new tests are going to be much larger by default due to the levels of isolation vs the existing one where every test is inextricably dependent on all the ones before it.