Re: Typechecking union types

Hans-Peter Diettrich <>
5 Jul 2006 15:15:42 -0400

          From comp.compilers

Related articles
Typechecking union types (Thomas Christensen) (2006-06-27)
Re: Typechecking union types (Curtis W) (2006-07-05)
Re: Typechecking union types (Hans-Peter Diettrich) (2006-07-05)
Re: Typechecking union types (2006-07-05)
Re: Typechecking union types (Thomas Christensen) (2006-07-28)
| List of all articles for this month |

From: Hans-Peter Diettrich <>
Newsgroups: comp.compilers
Date: 5 Jul 2006 15:15:42 -0400
Organization: Compilers Central
References: 06-06-072
Keywords: types
Posted-Date: 05 Jul 2006 15:15:42 EDT

Thomas Christensen wrote:

> For my master thesis, I need to get up to speed
> on the current state of the art within the somewhat
> narrow field of typechecking union types.
> What I need is some background on the various problems/issues related
> specifically to union types.

IMO the most important question is:

Is a union kind of implicit type cast, or can it hold polymorphic
(variant) information, but only one kind at a time.

As Wirth possibly had in mind with the tag field in Pascal variant
records, or as actually implemented in (Windows) Variant types, a
compiler could at least add a hidden information about the last
written "branch" of a union. When such techniques are used, the
problems should go away.


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.