The rbs validate command only outputs a single error at a time. However, when there are multiple issues, we have to go through a cycle of resolving errors, running rbs validate, resolving errors again, and running rbs validate repeatedly.
This issue is especially prevalent when auto-generating RBS, and it slows down the improvement cycle for addressing problems.
I propose resolving this issue by enhancing the rbs validate command to provide as much error information as possible when executed.
Expected behavior.
$ cat t.rbs
class Foo
def foo: () -> voids # typo
def bar: -> Bar # no type found
def bar: -> Bar # duplicated method
end
$ bundle exec rbs -I t.rbs validate --silent
t.rbs:2:17...2:22: Could not find voids (RBS::NoTypeFoundError)
def foo: () -> voids
^^^^^
t.rbs:3:14...3:17: Could not find Bar (RBS::NoTypeFoundError)
def bar: -> Bar
^^^
t.rbs:3:2...3:17: ::Foo#bar has duplicated definitions in t.rbs:4:2...4:17 (RBS::DuplicatedMethodDefinitionError)
def bar: -> Bar
^^^^^^^^^^^^^^^
soutaro
commented
8 months ago
The
rbs validatecommand only outputs a single error at a time. However, when there are multiple issues, we have to go through a cycle of resolving errors, runningrbs validate, resolving errors again, and runningrbs validaterepeatedly.This issue is especially prevalent when auto-generating RBS, and it slows down the improvement cycle for addressing problems.
I propose resolving this issue by enhancing the
rbs validatecommand to provide as much error information as possible when executed.Expected behavior.