tomoya ishida

I think it shuold be fixed as ```diff - def self.respond_to?(name, include_private = false) + def respond_to?(name, include_private = false) ```

`IRB.conf[:EVAL_HISTORY]` should be number or nil. Do you think IRB needs type validation for configs? Or `EVAL_HISTORY=true` should enable EVAL_HISTORY with some default size? (maybe about 10?) These are some...

The document is wrong. `$XDG_CONFIG_HOME/irb/irbrc` and `$HOME/.config/irb/irbrc` part is based on [XDG Base Directory Specification](https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html) > $XDG_CONFIG_HOME defines the base directory relative to which user-specific configuration files should be stored....

We can check command-like assignment by regexp like `/\A#{COMMAND_NAME} (=|\+=|-=|\*=|...)/` ```ruby info = 123 info += 1 info

Looks like this is fixed in irb-1.3.5

`show_source =~ -s` is a valid option for show_source command, and also a valid ruby code. `info = -1` will be still interpreted as command because it matches `COMMAND_WITH_ARGS_AND_FLAGS_REGEXP`. I...

I like #809 idea :+1: Although I have some concern about regexp including flag and args. It restricts not only command's implementation but also user's input. Input like `show_source class...

I agree local_variable defined or not will make implementation complicated and harder to understand the behavior. I think we can use a simple regexp to handle it. I made a...

How about something like this? It can be implemented with regexp. ```ruby if input.match?(/\A#{command_name_regexp}( |\z)/) # Maybe command if input.match?(/\A#{command_name_regexp} #{not_a_command_operators_regexp} /) # Not a command. `info = `, `info...

> it's about we only treat input as either command or Ruby code, which will miss the opportunity to warn users about incorrect forms of commands? Yes, that's what I'm...