Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uflorida!haven!adm!smoke!gwyn
From: gwyn@smoke.BRL.MIL (Doug Gwyn)
Newsgroups: comp.lang.c
Subject: Re: forward references in typedefs
Message-ID: <10605@smoke.BRL.MIL>
Date: 25 Jul 89 18:10:36 GMT
References: <55480@tut.cis.ohio-state.edu> <1989Jul20.152935.14872@utzoo.uucp> <24CB9E07.9547@marob.masa.com> <18729@mimsy.UUCP> <686@ftp.COM>
Reply-To: gwyn@brl.arpa (Doug Gwyn)
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 14

In article <686@ftp.COM> wjr@ftp.UUCP (Bill Rust) writes:
>The only way to tell if a jump table is better than compare and jump is to
>see what the range of the switch variable is and how many values it
>actually assumes. This is very difficult to do in a one pass compiler.

I don't see that it's particularly hard to do.  The compiler generates
a preamble that branches to a postamble to be generated later (after the
entire switch has been seen), and for each case as it is encountered the
case value is stashed in a compiler internal table (only 257 labels need
be accommodated) and an entry point label output followed by the case code
then a jump to a second deferred label which will follow the postamble.
The postamble can be a jump table, or a sequence of comparisons (hopefully
ordered as a binary tree instead of linear), or any other algorithm the
compiler decided was best for the particular set of case labels.