Issue #4 - sealed classes

[bernerbrau]

No description provided.

[ntrrgc]

This is too complex.

[ntrrgc]

It's hard to follow what has changed and what hasn't even using git diff -w.

[bernerbrau]

Yeah, I moved too much around. Let me revert and try again.

…

On Thu, Mar 29, 2018, 5:12 PM Alicia Boya García ***@***.***> wrote: It's hard to follow what has changed and what hasn't even using git diff -w. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#5 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AC4SgGBmRJ8f8fVyV79ek52GIqE56JLlks5tjWpxgaJpZM4S9y3X> .

[ntrrgc]

My worry is that the patch to implement this feature is so big and far reaching that it becomes not worth it for maintainability sake.

[ntrrgc]

This was my attempt. It does not handle non-nested sealed class children, but it's much simpler. a12e9f2

[bernerbrau]

Right, agreed. In addition to your initial concerns, I realized that typescript's type erasure makes it difficult at runtime to determine which member of the union is being used; hence why I found it necessary to include code generation that does runtime type validation. To do this, I had to make the class visitor repeatable, and that was enough to confuse git diff beyond hopefulness. So it may simply be that sealed classes are too complicated to be worth including. :(

…

On Fri, Mar 30, 2018, 8:43 AM Alicia Boya García ***@***.***> wrote: This was my attempt. It does not handle non-nested sealed class children, but it's much simpler. a12e9f2 <a12e9f2> — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#5 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AC4SgF56ucpxHY3aaH3MDISkBHEP0AFMks5tjkScgaJpZM4S9y3X> .