From mboxrd@z Thu Jan  1 00:00:00 1970
From: Gabriel Dos Reis <gdr@integrable-solutions.net>
To: Neil Booth <neil@daikokuya.co.uk>
Cc: Zack Weinberg <zack@codesourcery.com>, Per Bothner <per@bothner.com>, Jason Merrill <jason@redhat.com>, gcc@gcc.gnu.org
Subject: Re: Language-independent functions-as-trees representation
Date: Mon, 26 Aug 2002 03:18:00 -0000
Message-id: <m3hehiexl0.fsf@soliton.integrable-solutions.net>
References: <wvllm86k1m1.fsf@prospero.cambridge.redhat.com> <wvl65y1px1j.fsf@prospero.cambridge.redhat.com> <20020823191620.GA11032@tornado.toronto.redhat.com> <wvlznvdkspt.fsf@prospero.cambridge.redhat.com> <20020824004247.GG1074@codesourcery.com> <3D6718BA.2050207@bothner.com> <20020826070718.GB28065@codesourcery.com> <20020826100650.GB1377@daikokuya.co.uk>
X-SW-Source: 2002-08/msg01581.html

Neil Booth <neil@daikokuya.co.uk> writes:

| Zack Weinberg wrote:-
| 
| > TREE_LIST nodes, incidentally, have ridiculous overhead if being used
| > as plain old cons chains, or even alists.  Four pointers (chain, type,
| > purpose, value) plus a whole word of flag bits which are generally
| > unused.
| 
| I really hope this all gets fixed by moving to more "definite" structs,
| rather than a catch-all union for trees.

I agree with you.  I even hope that some silly type-errors will be
caught. 

-- Gaby