Xref: utzoo comp.sources.games.bugs:1221 rec.games.hack:5153
Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!gem.mps.ohio-state.edu!ginosko!uunet!van-bc!tessera!jtc
From: jtc@tessera.uucp (J.T. Conklin)
Newsgroups: comp.sources.games.bugs,rec.games.hack
Subject: Re: Nethack 3.0 & Xenix/286
Message-ID: <1989Sep11.153747.754@tessera.uucp>
Date: 11 Sep 89 15:37:47 GMT
References: <1989Sep9.191638.23379@tessera.uucp> <997@mgse.UUCP>
Reply-To: jtc@tessera.UUCP (J.T. Conklin)
Organization: Tesseract Communications, Burnaby, B.C., Canada
Lines: 36

In article <997@mgse.UUCP> marks@.UUCP (Mark Seiffert) writes:
>In article <1989Sep9.191638.23379@tessera.uucp> jtc@tessera.UUCP (J.T. Conklin) writes:
>-I've finally been able to get the nethack level compiler to work
>-with Xenix/286.  The fix is simple: generate the lexer with flex
>-rather than lex.
>
>I have always used Flex when i am compiling Nethack 3. I fail to see
>how that is the solution. I do know that prior to patch 3, lev_comp
>failed to produce an output file, and with the level 3 patches it 
>dumps core. I am using Flex 2 what was posted to the net recently.

Ok, calm down and we'll see whats going on.

First off, I'm using using nethack 3.0 patchlevel 3, the beta release
of flex 2.11, and SCO Xenix/286 2.2.3.  It works fine.

Well, actually, you have to cast the yyleng macro to int; otherwise you
get "conversion of long address to a short address" errors and lev_comp
won't work.  I did it in flex.skel just to be done with it once and for
good.

>-If any of you are still having trouble compiling nethack on a '286,
>-drop me a line and I'll see how I can help.
>
>Ok i dropped the line. Now what?
>

I'd try flex 2.11 if you can get a hold of it.  It's availiable at a whole
lot of archive sites.  If you can't get a copy, or it still doesn't work
send email and I'll try to help.

    --jtc

-- 
J.T. Conklin
	...!{ubc-cs,uunet}!van-bc!tessera!jtc