Mailing-List: contact dev-help@apr.apache.org; run by ezmlm
Precedence: bulk
Date: Sun, 12 Aug 2001 02:09:51 -0400 (EDT)
From: Sterling Hughes <sterling@designmultimedia.com>
To: Aaron Bannert <aaron@clove.org>
Cc: <dev@apr.apache.org>
Subject: Re: use of malloc in dso.c instead of apr_palloc?
In-Reply-To: <20010811001633.A22820@clove.org>
Message-ID: <Pine.LNX.4.33L2.0108120202390.1275-100000@localhost.localdomain>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Sat, 11 Aug 2001, Aaron Bannert wrote:

> On Sat, Aug 11, 2001 at 12:25:11PM -0400, Sterling Hughes wrote:
> >
> >     Well, as the overly long title suggests, is there any reason that
> >     malloc() is used in dso.c (unix and aix)?  apr_palloc() seems
> >     fine for these purposes (temporary string).  Also, as a side
> >     note, sizeof(char) is used in the allocation, which is extreaneous,
> >     because a char is always equivalent to one base unit of memory
> >     (well, at least that's how I interpret ANSII).
>
> If I remember my compiler class correctly, sizeof() gives a constant scalar
> at compile-time. Any decent compiler will perform constant folding on
> expressions like (sizeof(char)*12) before producing object code. This means
> you should feel free to use sizeof() liberally if it helps readability.
>

    I know gcc should optimize this out (so should most other
    compilers), just like a = 12 * 12 is translated to a = 144.

    Its more of a stylistic note, as:

    malloc(sizeof(char)*(strlen(str)+2));

    doesn't look as nice as:

    malloc(strlen(str) + 2);

    since the sizeof(char) will never provide any functional benefits
    (no it won't hurt anything, but...)

    But all this was really an aside from the main point, should
    malloc() be used? :)

    -Sterling