Parser Error when erb comment tags span multiple lines

[djuber]

I noticed (initially when checking 1.0.2, but confirmed still in 1.0.4) an issue where multi-line comment tags like the following

  <%# this is a multiline comment
  it's totally fine
  nothing should fail %>

are recognized correctly by ERB and render correctly (don't render at all, the entire content is commented), but a parser error is raised when running 18n tasks

 bundle exec rspec
#<Thread:0x00005634a007a9d8 /home/djuber/reproduction/vendor/cache/ruby/3.0.0/gems/i18n-tasks-1.0.4/lib/i18n/tasks/scanners/scanner_multiplexer.rb:33 run> terminated with exception (report_on_exception is true):
/home/djuber/reproduction/vendor/cache/ruby/3.0.0/gems/parser-3.1.1.0/lib/parser/diagnostic/engine.rb:72:in `process': Error scanning app/views/layouts/application.html.erb: unterminated string meets end of file (I18n::Tasks::CommandError)


     Failure/Error: raise Parser::SyntaxError, diagnostic
     
     I18n::Tasks::CommandError:
       Error scanning app/views/layouts/application.html.erb: unterminated string meets end of file

Modifying the comment tag to include a # on each comment line resolves the parser error (tests run successfully), so we can work around this if needed, but indicates a drift between the ERB syntax and that recognized by the i18n tasks.

Originally we expected this was related to #430 (since the tag starts <%# and not <% # that was also possibly related), but the same problem was seen after updating to 1.0.4.

[glebm]

Where is this syntax specified? Naively, it seems invalid, since <% indicates Ruby code and comments don't work like that in Ruby.

[glebm]

/cc @davidwessman

[glebm]

Released v1.0.5

[davidwessman]

https://ruby-doc.org/stdlib-3.1.1/libdoc/erb/rdoc/ERB.html

From Recognized tags ERB recommends not using ruby comments:

<%# comment -- ignored -- useful in testing %> (<% # doesn't work. Don't use Ruby comments.)

[djuber]

Thank you for the quick turn-around!