13:05 <mps> Blacksmoke16: re: xml to json https://github.com/mamantoha/xml_converter

13:05 <mps> works fine in some of my tests

13:06 <FromGitter> <Blacksmoke16> oh nice!

13:06 <FromGitter> <Blacksmoke16> only in some? 😬

13:07 <mps> well, didn't tested it extensively, don't need to process big xmls

13:07 <mps> actually worked in all small tests I made

13:08 <FromGitter> <Blacksmoke16> 👍 good to hear

21:24 <FromGitter> <nanobowers> kind of a general question about crystal APIs (vs ruby): where in ruby it might be common to return nil from a function when something failed - this seems like a pain Crystal from a usage standpoint. One could make the function return `Foo?` instead of `Foo`, but then the user would have to either handle the Nil case on their side using `.try` or otherwise. It seems the preferred approach is to raise

21:24 <FromGitter> ... Exceptions? Particular case in point: parsing a string to generate a Foo, but parsing fails. ⏎ ⏎ Relatedly, i'd be interested to buy a book about "The Crystal Way", b/c the more I spend time in Crystal I can see the Ruby way is not the right way for crystal. The documentation is great and a lot can be gleaned from reading s ... [https://gitter.im/crystal-lang/crystal?at=6141130a63dca818915d1d51]

22:00 <FromGitter> <Blacksmoke16> you'd still have to handle `nil` in ruby, it just doesn't warn you where it's possible it *could* be `nil`

22:00 <FromGitter> <Blacksmoke16> Crystal way would be to have a `#parse` and `#parse?` method where the first raises and the latter returns nil

22:03 <FromGitter> <Blacksmoke16> if you find yourself doing like ⏎ ⏎ ```if value = some_method ⏎ value ⏎ else ⏎ some_default ⏎ end``` [https://gitter.im/crystal-lang/crystal?at=61411c2bf3449e13a47291f3]

22:03 <FromGitter> <Blacksmoke16> as using exceptions for control flow isn't great