Path: csiph.com!news.mixmin.net!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Tim Rentsch <tr.17687@z991.linuxsc.com>
Newsgroups: comp.lang.c
Subject: Re: How About Disallowing Assignments In Expressions?
Date: Wed, 14 Feb 2024 21:54:24 -0800
Organization: A noiseless patient Spider
Lines: 46
Message-ID: <86jzn61ddb.fsf@linuxsc.com>
References: <uq3s76$28dsr$1@dont-email.me> <uq5h7r$2mgb3$2@dont-email.me> <uq5i0r$2mkc4$1@dont-email.me> <uqg5do$25lge$1@dont-email.me> <20240213192257.00002da6@yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Injection-Info: dont-email.me; posting-host="d6a0776853173adee93a90f6ecdad10b"; logging-data="3342924"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1+fOaw0OSTwrpaR/X9GMrtJh8OVm3Gne6E="
User-Agent: Gnus/5.11 (Gnus v5.11) Emacs/22.4 (gnu/linux)
Cancel-Lock: sha1:R6o2onLzXDwze4Wt43tn4ieBuTI= sha1:EVPgbidpjre+L/pkVnN6wIWKnMc=
Xref: csiph.com comp.lang.c:382504

Michael S <already5chosen@yahoo.com> writes:

[.. should  'a += b += c;'  be disallowed for style reasons? ..]

> It's actually highly confusing for me as well.
> For reasons I am not fully understand, it is interpreted as
> a += (b += c)
> which is obviously legal.
> I don't understand why it isn't interpreted as
> (a += b) += c
> which is [a little less obviously] illegal.

Here is another way to think about it.

C (and many or most other languages) are fairly strict about what
is allowed on the left-hand side of an assignment.  For example,

    a + b += c;

is not allowed.  Neither is

    (a + b) += c;

although the reasoning is a little different in the two cases,
neither is allowed, because neither one makes sense.

By analogy with the second example, the statement

    (a += b) += c;

isn't allowed, because the left-hand side is wrong.  So if we
see

    a += b += c;

it would be silly to parse it as

    (a += b) += c;

because that's illegal.  The only other way to parse it, viz.,

    a += (b += c);

passes the semantic prerequisites.  Since it's the only way that
makes sense, it seems natural that (without any parentheses) the
syntax should reflect that.