https://pantsbuild.org/ logo
Join Slack
Powered by
<@U021C96KUGJ>, <@UB2J9BQA0>: have posted two chan...
# development
w
once those are in, Chris and I can break up the migration work
a
Reviewing
w
thanks: the commits should be useful to review independently.
a
there’s a lot of them on 179
😉
@witty-crayon-22786 is your assumption in #17179 that there’ll always be a local environment of some sort?
w
yea, that is an assumption. it’s baked into the existence of the 
ChosenLocalEnvironmentName
type though
but it was already an assumption, because the 
ChosenLocalEnvironmentName
is used as the default environment for all goals (until #17155)
a
OK, there’s some code in the environment-aware options support that could be simplified if there’s always an environment around (i.e. 
EnvironmentTarget.val
is never 
None
). Is that something we’re planning to keep?
w
yes. because it is how we will always bootstrap plugins to use, for example
a
Does that mean we synthesise an empty 
local_environment
in the absence of one that’s defined?
w
but… i don’t think that that necessarily means that there is always an 
EnvironmentTarget
: i had heard from @hundreds-father-404 that you two agreed that we didn’t need to require full coverage of 
local_environment
targets
Does that mean we synthesise an empty 
local_environment
in the absence of one that’s defined?
no: see above
a
OK. SO there’s always an environment, but only a target if one is defined
that makes sense
w
yea
h
indeed, there may be no env target even when environments are in use
a
@witty-crayon-22786 reviewed; the docs may need some improved obviousness, but nothing that’s blocking merging this
w
thanks!
as to how to break them up: it seems like the only way to step on toes is to change helper APIs… but we should probably do that as we go, so.
one way might be to divide them between goals which use 
TargetRootsToFieldSets
and those which don’t…? that API could maybe assist with environment calculations
👍 2
hm. there are apparently only 4 or so that use 
TargetRootsToFieldSets
. probably still a good divider, but we’ll need more dividers.
@ancient-vegetable-10556: since there are fewer unknowns with the 
TargetRootsToFieldSets
goals, do you want to start with those tomorrow? that would replace https://github.com/pantsbuild/pants/pull/17106 as well probably
a
Sounds good. I'll talk to you if I find any questions
w
thanks: i’ll close #17106 and let you re-open
Open in Slack
PreviousNext
Powered by