Crystal: Fetch a char from string to equal the same string(char) return false

[icyleaf]

icr(0.22.0) > "hello!"[-1] == "!"
 => false
icr(0.22.0) > "==="[0] == "="
 => false
icr(0.22.0) > "==="[0].to_s == "="
 => true
icr(0.22.0) > "==="[0] == "=".chars[0]
 => true

2.4.1 :001 > "hello!"[-1] == "!"
 => true
2.4.1 :002 > "==="[0] == "="
 => true

Is this logic correct?

[akzhan]

Yes, Char will never be equal to String.

https://play.crystal-lang.org/#/r/29tu

[akzhan]

But You can override this behavior with:

https://play.crystal-lang.org/#/r/29ty

[RX14]

Reminder that overriding such behaviour globally is likely to break other shards or the stdlib. If you want to compare a character use Char.

[akzhan]

@RX14 I suppose that's related to Frameworks. Rails actively override core things and it's OK.

IMHO, Crystal like a Ruby just the foundation for frameworks like Rails.

[RX14]

@akzhan and rails breaks things. From what I've heard it's not even rare.

[akzhan]

@RX14 Crystal follows Ruby idea to metaprogramming. And it's great.

And yes, I dislike some defaults of Crystal (like String.underscore behavior), So ability to fix it for own code is beautiful.

[straight-shoota]

It will inevitably lead to problems with 3rd party code. If you're only using your own code, that's okay. But when you rely on 3rd party shards, I wouldn't change behavior of stdlib.

[akzhan]

@straight-shoota Yes, it is. But Crystal is foundation. A lot of people will use higher level frameworks.

P.S.: I'm used Rack not Rails. And Rails is great concept built extending and overriding Ruby.

[straight-shoota]

Rails has such an important status in the Ruby ecosystem that 3rd party libraries just have to deal with the monkey patches from rails, because they are everywhere ;)

[icyleaf]

thanks @akzhan. Maybe It's better to update this behavior to document. 😄

[icyleaf]

Update this benchmark to fast-crystal#equal-substring-of-char-code

$ crystal build --release code/string/equal-substring-of-char.cr -o bin/string/equal-substring-of-char
$ ./bin/string/equal-substring-of-char

Crystal 0.22.0 (2017-04-22) LLVM 4.0.0
         "==="[0] == '=' 252.73M (  3.96ns) (±12.75%)       fastest
    "==="[0].to_s == "="  19.65M ( 50.88ns) (± 2.72%) 12.86× slower
"==="[0] == "=".chars[0]  16.97M ( 58.91ns) (± 3.48%) 14.89× slower

[akzhan]

Thanks @icyleaf.

btw, take a look at #4586. Looks like we need tests for gsub vs tr.