Narrow repeated calls to the same attr_reader-style accessor

[apiology]

Summary

Flow-sensitive typing already narrows nil-checks on local/instance variables (x = foo; return if x.nil?; x.bar), but a nil-guard on obj.attr did not narrow a later call to obj.attr in the same method body -- each call was treated as an independent, unnarrowed invocation.

FlowSensitiveTyping now recognizes receivers that are a dotted chain of simple, argument-less, blockless calls rooted in a tracked local or instance variable (e.g. pin.location, @pin.location) and records nil-narrowing facts against a synthesized pin for that chain, the same way it already does for a plain variable:

# @param pin [Pin]
def bundled_filename(pin)
  return nil unless pin.location
  pin.location.filename  # now typed Location, not Location-or-nil
end

Chain::Call#resolve looks up those facts by threading a dotted "receiver path" through Chain#define/#infer, checked before falling back to ordinary method resolution. This required adding an optional receiver_path parameter to Chain::Link#resolve and its subclasses (mechanical -- all but Chain::Call just accept and ignore it).

Also fixes a latent Pin::BaseVariable#equality_fields gap found while testing this: downcast copies of the same pin (e.g. the if-true and if-false facts from one guard) shared identical name/location/closure/source and so had identical equality_fields, despite carrying different presence/narrowed types. That let them collide as cache keys in Chain's inference cache, returning a stale, wrongly-narrowed or wrongly-unnarrowed result depending on lookup order -- this made the new narrowing intermittently flaky until fixed.

This intentionally does not cover a bare, implicit-self call (e.g. location.nil?; location.filename where location is a 0-arg method, not a local) -- narrowing that would need a way to resolve the enclosing self type at indexing time, which FlowSensitiveTyping does not currently have. Left as a natural follow-up.

Fixes #1249

Test plan

• bundle exec rspec spec/parser/flow_sensitive_typing_spec.rb -- new specs cover the truthy-guard, .nil?-guard, and ivar-rooted cases from the issue
• bundle exec rspec -- full suite, 1621 examples, 0 failures (matches baseline)
• bundle exec rubocop on changed files -- no new offenses vs baseline
• SOLARGRAPH_ASSERTS=on bundle exec solargraph typecheck --level strong -- no new problems vs baseline (repo-wide strong typecheck is not currently clean and is not a hard CI gate; this PR was checked line-by-line against a baseline run to avoid adding new debt)

🤖 Generated with Claude Code

https://claude.ai/code/session_01KGu6zb5faStTC754PxMUSA

[apiology]

@castwide - ready for review