From mboxrd@z Thu Jan  1 00:00:00 1970
X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/1541
Path: news.gmane.org!not-for-mail
From: Rich Felker <dalias@aerifal.cx>
Newsgroups: gmane.linux.lib.musl.general
Subject: portable fgetln
Date: Sat, 11 Aug 2012 20:24:50 -0400
Message-ID: <20120812002450.GA17066@brightrain.aerifal.cx>
Reply-To: musl@lists.openwall.com
NNTP-Posting-Host: plane.gmane.org
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="zYM0uCDKw75PZbzx"
X-Trace: dough.gmane.org 1344731042 4504 80.91.229.3 (12 Aug 2012 00:24:02 GMT)
X-Complaints-To: usenet@dough.gmane.org
NNTP-Posting-Date: Sun, 12 Aug 2012 00:24:02 +0000 (UTC)
To: musl@lists.openwall.com
Original-X-From: musl-return-1542-gllmg-musl=m.gmane.org@lists.openwall.com Sun Aug 12 02:24:03 2012
Return-path: <musl-return-1542-gllmg-musl=m.gmane.org@lists.openwall.com>
Envelope-to: gllmg-musl@plane.gmane.org
Original-Received: from mother.openwall.net ([195.42.179.200])
	by plane.gmane.org with smtp (Exim 4.69)
	(envelope-from <musl-return-1542-gllmg-musl=m.gmane.org@lists.openwall.com>)
	id 1T0LyL-0005wY-Ru
	for gllmg-musl@plane.gmane.org; Sun, 12 Aug 2012 02:24:01 +0200
Original-Received: (qmail 26317 invoked by uid 550); 12 Aug 2012 00:24:00 -0000
Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm
Precedence: bulk
List-Post: <mailto:musl@lists.openwall.com>
List-Help: <mailto:musl-help@lists.openwall.com>
List-Unsubscribe: <mailto:musl-unsubscribe@lists.openwall.com>
List-Subscribe: <mailto:musl-subscribe@lists.openwall.com>
Original-Received: (qmail 26309 invoked from network); 12 Aug 2012 00:24:00 -0000
Content-Disposition: inline
User-Agent: Mutt/1.5.21 (2010-09-15)
Xref: news.gmane.org gmane.linux.lib.musl.general:1541
Archived-At: <http://permalink.gmane.org/gmane.linux.lib.musl.general/1541>


--zYM0uCDKw75PZbzx
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

Hi all,

Based on a discussion on #musl about the merits of fgetln as an
efficient (probably optimal) way to do line-based stdio processing if
the implementation is good, I got to wondering if it's viable to use
the interface in programs intended to be portable, i.e. if it could be
written portably for use on any stdio implementation or at least any
POSIX one. Here's the result -- a sample portable implementation that
keeps a separate buffer for each fd, or for each FILE* when the stream
lacks an associated fd. It has some limitations, but it seems to meet
the interface contract and avoids using a single static buffer that
would render if non-thread-safe.

Rich

--zYM0uCDKw75PZbzx
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment; filename="portable_fgetln.c"

#define _POSIX_C_SOURCE 200809L
#include <stdio.h>
#include <pthread.h>
#include <stdlib.h>

struct entry
{
	struct entry *next;
	int fd;
	FILE *f;
	char *buf;
};

struct entry *table[64];

#define K (sizeof table / sizeof table[0])

pthread_rwlock_t lock = PTHREAD_RWLOCK_INITIALIZER;

char *fgetln(FILE *f, size_t *l)
{
	int fd, h;
	struct entry *p;
	ssize_t cnt;
	char *ret = 0;

	flockfile(f);
	fd = fileno(f);
	h = fd % K;

	pthread_rwlock_rdlock(&lock);
	for (p=table[h]; p && (p->fd!=fd || (fd<0 && p->f!=f)); p=p->next);
	if (!p) {
		if (!(p = calloc(sizeof *p, 1))) {
			pthread_rwlock_unlock(&lock);
			funlockfile(f);
			return 0;
		}
		p->fd = fd;
		p->f = f;
		pthread_rwlock_unlock(&lock);
		pthread_rwlock_wrlock(&lock);
		p->next = table[h];
		table[h] = p;
	}
	pthread_rwlock_unlock(&lock);
	cnt = getline(&p->buf, (size_t[]){0}, f);
	if (cnt >= 0) {
		*l = cnt;
		ret = p->buf;
	}
	funlockfile(f);
	return ret;
}

--zYM0uCDKw75PZbzx--