Use lazy arguments for when and whenNot

[neunzehnhundert97]

Hi, here is a small change/fix.

When using when as a shortcut for an if, it violated my assumption that the conditional code would not be executed at all. Instead, it is executed but its result not used. Not a problem if your code is mostly pure, but if there are lurking some side effects, it can give you a hard time finding the problem. I did set the boolean correctly, did I?

I suggest to use lazy parameters here, which makes usage of when as equivalent to ìf as possible.

[raquo]

That's a good point, I agree with the reasoning. The PR looks all good to me, thanks! This needs a major version bump to release, so I will merge it later, when I start working on v18 again.

…

On Sat., Aug. 24, 2024, 1:23 a.m. neunzehnhundert97, < ***@***.***> wrote: @neunzehnhundert97 <https://github.com/neunzehnhundert97> requested your review on: #169 <#169> Use lazy arguments for when and whenNot as a code owner. — Reply to this email directly, view it on GitHub <#169 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAECBMHROG54SS2GH6XETHLZTA7IVAVCNFSM6AAAAABNBMSNXOVHI2DSMVQWIX3LMV45UABCJFZXG5LFIV3GK3TUJZXXI2LGNFRWC5DJN5XDWMJUGAYDAMRQGY3TQNY> . You are receiving this because your review was requested.Message ID: ***@***.***>

[raquo]

It turns out that Scala 2 does not support by-name varargs.

To keep Laminar behaviour consistent, I split the implementations of when and whenNot to only accept a single argument in Scala 2, and to accept varargs in Scala 3 (both by-name). In Scala2, users can still provide Seq(mod1, mod2) or modSeq(mod1, mod2) as a single param, the implicit conversion from Seq[Mod] to Mod will take care of that.

So, I will close this PR when I actually merge my alternative implementation. Thanks for bringing up this issue!