Clarification regarding the GitHub Actions pre-submit testing
maurizio.cimadamore at oracle.com
Thu Mar 18 17:29:25 UTC 2021
On 18/03/2021 13:18, Aleksey Shipilev wrote:
> What I am talking about is much weaker: buildability. I believe it is
> fair to ask that when the mainline integration PR is done, the
> arch-specific code is clearly isolated, new declarations in
> arch-specific headers are done, their implementations are stubbed out
> with NotImplemented(), etc.
With Panama, we ended up in the place you describe.
While that might take some "sting" off the initial integration, one
thing worth mentioning is that it's not clear how things move from a
"buildable but not working" state into a "fully working" state (to date,
Panama has not received any contribution/help in terms of porting the
foreign linker API to x86 and other platforms).
So, while I generally agree that we should try to minimize disruptions
as much as possible, I think what users want is something that builds
and that _does what it's supposed to do_ - otherwise we quickly end up
with ports which only support a subset of the features that are
supported by the "first-class" ports, which isn't great either.
More information about the jdk-dev