http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html Against the grain software and startups Wed, 04 Apr 2012 19:38:00 +0000 hourly 1 http://wordpress.org/?v=3.3.2 http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-94 Gabriel Claramunt Sun, 14 Feb 2010 15:29:14 +0000 http://ilya.cobrio.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-94 <p>I think everybody agrees that null is a big wart in the Scala's type system but is there for Java compatibility. As you said, idiomatic Scala uses Option[T] instead. Nulls are a runtime event, so the compiler cannot infer much, ( unless you use Option :)) In this particular case, maybe WrappedString should throw a "null parameter exception" or check for null in isEmpty, but I agree that implicits can have nasty interactions with nulls. bottom line: nulls are bad, mmmmkay?</p> I think everybody agrees that null is a big wart in the Scala’s type system but is there for Java compatibility. As you said, idiomatic Scala uses Option[T] instead. Nulls are a runtime event, so the compiler cannot infer much, ( unless you use Option ) In this particular case, maybe WrappedString should throw a “null parameter exception” or check for null in isEmpty, but I agree that implicits can have nasty interactions with nulls. bottom line: nulls are bad, mmmmkay?

]]> http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-229 Gabriel Claramunt Sun, 14 Feb 2010 15:29:00 +0000 http://ilya.cobrio.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-229 <p>I think everybody agrees that null is a big wart in the Scala's type system but is there for Java compatibility. As you said, idiomatic Scala uses Option[T] instead. Nulls are a runtime event, so the compiler cannot infer much, ( unless you use Option :)) In this particular case, maybe WrappedString should throw a "null parameter exception" or check for null in isEmpty, but I agree that implicits can have nasty interactions with nulls. bottom line: nulls are bad, mmmmkay?</p> I think everybody agrees that null is a big wart in the Scala’s type system but is there for Java compatibility. As you said, idiomatic Scala uses Option[T] instead. Nulls are a runtime event, so the compiler cannot infer much, ( unless you use Option ) In this particular case, maybe WrappedString should throw a “null parameter exception” or check for null in isEmpty, but I agree that implicits can have nasty interactions with nulls. bottom line: nulls are bad, mmmmkay?

]]> http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-93 Ilya Sterin Thu, 11 Feb 2010 08:54:05 +0000 http://ilya.cobrio.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-93 <p>Yes that's a good option, but being that anyone can use implicits I wonder if this check shouldn't be done by the compiler? I can't think of any reason why one would want an implicit to wrap a null. With this being a compiler check it will prevent any surprises from 3rd party defined implicits.</p> Yes that’s a good option, but being that anyone can use implicits I wonder if this check shouldn’t be done by the compiler? I can’t think of any reason why one would want an implicit to wrap a null. With this being a compiler check it will prevent any surprises from 3rd party defined implicits.

]]> http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-228 Ilya Sterin Thu, 11 Feb 2010 08:54:00 +0000 http://ilya.cobrio.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-228 <p>Yes that's a good option, but being that anyone can use implicits I wonder if this check shouldn't be done by the compiler? I can't think of any reason why one would want an implicit to wrap a null. With this being a compiler check it will prevent any surprises from 3rd party defined implicits.</p> Yes that’s a good option, but being that anyone can use implicits I wonder if this check shouldn’t be done by the compiler? I can’t think of any reason why one would want an implicit to wrap a null. With this being a compiler check it will prevent any surprises from 3rd party defined implicits.

]]> http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-92 Martin Thu, 11 Feb 2010 07:58:12 +0000 http://ilya.cobrio.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-92 <p>I tend to agree; implicit wrappers need to check that they apply only to non-null values. That needs to be done in the classes defining those wrappers.</p> I tend to agree; implicit wrappers need to check that they apply only to non-null values. That needs to be done in the classes defining those wrappers.

]]> http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-227 Martin Thu, 11 Feb 2010 07:58:00 +0000 http://ilya.cobrio.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-227 <p>I tend to agree; implicit wrappers need to check that they apply only to non-null values. That needs to be done in the classes defining those wrappers.</p> I tend to agree; implicit wrappers need to check that they apply only to non-null values. That needs to be done in the classes defining those wrappers.

]]> http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-91 mr.x Thu, 11 Feb 2010 07:21:31 +0000 http://ilya.cobrio.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-91 <p>That implicit conversion makes no sense.</p> That implicit conversion makes no sense.

]]> http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-226 mr.x Thu, 11 Feb 2010 07:21:00 +0000 http://ilya.cobrio.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-226 <p>That implicit conversion makes no sense.</p> That implicit conversion makes no sense.

]]> http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-90 Sakuraba Thu, 11 Feb 2010 05:47:27 +0000 http://ilya.cobrio.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-90 <p>In my mind the implicit conversion is the bug. A null-reference String should not be turned into a WrappedString that is no longer null. If that was the case, there would not be any unexpected behaviour.</p> In my mind the implicit conversion is the bug. A null-reference String should not be turned into a WrappedString that is no longer null. If that was the case, there would not be any unexpected behaviour.

]]> http://ilyasterin.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-225 Sakuraba Thu, 11 Feb 2010 05:47:00 +0000 http://ilya.cobrio.com/blog/2010/02/avoid-using-nulls-in-scala.html#comment-225 <p>In my mind the implicit conversion is the bug. A null-reference String should not be turned into a WrappedString that is no longer null. If that was the case, there would not be any unexpected behaviour.</p> In my mind the implicit conversion is the bug. A null-reference String should not be turned into a WrappedString that is no longer null. If that was the case, there would not be any unexpected behaviour.

]]>