remove +hard, replace with new ;;

[pilfer-pandex]

The previous commit changed the semantics of ;; to just call the mold with the result of the hoon, rather than having it macroexpand to something like hard. By "hard" is meant applying the mold and then checking whether the result is unchanged. This second commit removes the +hard function itself, with the effect that there are no remaining mold "idempotence" tests in the system.

#1055

[cgyarvin]

I believe this to be right and proper. While the equation of type, schema and validator is extremely useful, Postel’s law is much less so — you rarely want to normalize incoming network data, though schema evolution remains a thing. And what the “idempotence” (really fixpoint) test is testing is not the correctness of the data but the correctness of the schema, which really no one needs...

…

Sent from my iPhone

On May 9, 2019, at 12:49 PM, pilfer-pandex ***@***.***> wrote: The previous commit changed the semantics of ;; to just call the mold with the result of the hoon, rather than having it macroexpand to something like hard. By "hard" is meant applying the mold and then checking whether the result is unchanged. This second commit removes the +hard function itself, with the effect that there are no remaining mold "idempotence" tests in the system. #1055 You can view, comment on, or merge this pull request online at: #1166 Commit Summary remove +hard, replace with new ;; File Changes M lib/collections.hoon (4) M lib/hall-legacy.hoon (2) M lib/hood/kiln.hoon (4) M lib/oauth1.hoon (3) M lib/oauth2.hoon (3) M lib/test/runner.hoon (5) M mar/collections/action.hoon (4) M sys/arvo.hoon (20) M sys/hoon.hoon (35) M sys/vane/ames.hoon (12) M sys/vane/behn.hoon (2) M sys/vane/clay.hoon (67) M sys/vane/dill.hoon (4) M sys/vane/eyre.hoon (12) M sys/vane/ford.hoon (6) M sys/vane/gall.hoon (6) M sys/vane/jael.hoon (2) M sys/vane/xmas.hoon (8) M sys/zuse.hoon (8) Patch Links: https://github.com/urbit/arvo/pull/1166.patch https://github.com/urbit/arvo/pull/1166.diff — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.

[joemfb]

It looks like there's still a call in lib/hall-legacy. Otherwise, this LGTM

[pilfer-pandex]

You're right. My test procedure was a combo of |reset and +test, which did not catch this. What testing strategies do you use?

[pilfer-pandex]

Not a call to hard, but still wrong.