Open deemp opened 4 months ago
@deemp thanks for the report, here is a feedback:
The expected behavior and actual result are entangled in the description. They need to be distinctly separated for a clearer understanding.
{
⟦
org ↦ ⟦
eolang ↦ ⟦
...
⟧,
λ ⤍ Package
⟧
⟧
}
{
⟦
org ↦ ⟦
eolang ↦ ⟦
...
⟧,
λ ⤍ Package
⟧,
λ ⤍ Test
⟧
}
This will help anyone reviewing the bug report to quickly understand what the problem is and how it deviates from the expected behavior.
Please fix the bug report in order it to get resolved faster.
Analyzed with gpt-4
@yegor256 @deemp why do we need this? Does not this help?
@maxonfjvipon, I updated the issue description
@deemp in general, it's better to submit a new issue instead of editing description/title of existing one -- this will help us keep discussions more concise.
@yegor256, sure. However, there's no discussion in this issue apart from the current meta-discussion.
Problem
When a formation has a
λ ⤍ Package
, thenormalizer
dataizes formations in sibling bindings. Otherwise, thenormalizer
doesn't dataize formations in sibling bindings. However, there's no top-levelλ ⤍ Package
binding in generated PHI programs.pipeline/phi/as-phi-tests.phi
(see this artefact):Suggestion
Make
eoc
optionally keep the information about tests. That is, botheoc phi
andeoc unphi
should have a--tests
flag.For
eoc phi --tests
, write aλ ⤍ Test
binding in the top-level formation.For
eoc unphi --tests
, handle this binding too.