Date: Thu, 19 Dec 2002 11:52:47 +0200 From: Ruslan Ermilov <ru@freebsd.org> To: Bruce Evans <bde@zeta.org.au> Cc: Garrett Wollman <wollman@lcs.mit.edu>, cvs-committers@freebsd.org, cvs-all@freebsd.org Subject: Re: cvs commit: src/lib/libc/sys _exit.2 accept.2 access.2 acct.2 adjtime.2 aio_cancel.2 aio_error.2 aio_read.2 aio_return.2 aio_suspend.2 aio_waitcomplete.2 aio_write.2 bind.2 brk.2 chdir Message-ID: <20021219095247.GA79372@sunbay.com> In-Reply-To: <20021219132023.A28577-100000@gamplex.bde.org> References: <200212181654.gBIGswYP062315@khavrinen.lcs.mit.edu> <20021219132023.A28577-100000@gamplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--XsQoSWH+UP9D9v3l Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Dec 19, 2002 at 01:23:27PM +1100, Bruce Evans wrote: > On Wed, 18 Dec 2002, Garrett Wollman wrote: >=20 > > <<On Wed, 18 Dec 2002 01:22:37 -0800 (PST), Ruslan Ermilov <ru@FreeBSD.= org> said: > > > > > Stop calling system calls "function calls". > > > > > Use "The .Fn system call" a-la "The .Nm utility". > > > > I believe that this is a mistake. System calls are a special case of > > function calls. The standards from which many of these descriptions > > are taken intentionally do not reference system calls, and neither > > should we, since the interfaces may move from one side of the boundary > > to the other. (Many of these are, strictly speaking, function calls > > in the threaded C library.) >=20 > I mostly agree, but note that system calls are still carefully separated > from functions by putting them in a separate manpage section. The thread= ed > case already bogotifies this separation. >=20 The purpose of my fixes was to unify the way we refer to syscalls. It's clear why standards that _borrowed many of BSD descriptions_ do not reference system calls -- that's an implementation detail how an interface is made available: as a library or system call. I don't consider the threaded library a special case here, because threaded library is a different library, while these describe the "Standard C Library (libc, -lc)" interfaces, as denoted by the LIBRARY section. At one point, we had the internal implementation details of wrapping system calls documented across section 2 manpages; it was decided to remove them, since they provide not much interest to a programmer. For an example, please see: cvs di -r1.20 -r1.21 accept.2 If the consensus is to refer to syscalls as functions, there is also no reason to keep section 2. Cheers, --=20 Ruslan Ermilov Sysadmin and DBA, ru@sunbay.com Sunbay Software AG, ru@FreeBSD.org FreeBSD committer, +380.652.512.251 Simferopol, Ukraine http://www.FreeBSD.org The Power To Serve http://www.oracle.com Enabling The Information Age --XsQoSWH+UP9D9v3l Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (FreeBSD) iD8DBQE+AZbvUkv4P6juNwoRAp9OAJ9+w0aCLNLR7Am8FSqCLC2IGCWgQQCgiLDx Vu8nErPdxaAERlR0OZ9tM34= =LVJj -----END PGP SIGNATURE----- --XsQoSWH+UP9D9v3l-- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20021219095247.GA79372>