Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/5/84; site umd5.UUCP
Path: utzoo!watmath!clyde!bonnie!akgua!gatech!seismo!umcp-cs!cvl!umd5!zben
From: zben@umd5.UUCP
Newsgroups: net.arch
Subject: Re: Unix C benchmarks
Message-ID: <810@umd5.UUCP>
Date: Sat, 7-Dec-85 23:00:57 EST
Article-I.D.: umd5.810
Posted: Sat Dec  7 23:00:57 1985
Date-Received: Tue, 10-Dec-85 05:44:20 EST
References: <342@ncr-sd.UUCP> <6203@utzoo.UUCP>
Reply-To: zben@umd5.UUCP (Ben Cranston)
Organization: U of Md, CSC, College Park, Md
Lines: 22
Summary: Question about portable code

In article <6203@utzoo.UUCP> henry@utzoo.UUCP (Henry Spencer) writes:

>Sigh, would it be too much to ask that "lseek(fd,0,0);" be rendered as
>"lseek(fd,0L,0);" to make it portable?

Sigh, isn't there a portable way to do this?  Something on the order of:

#include <grail.h>
...
   lseek( fd , (t_seek_second_arg) 0 , 0 );

I thought I saw something like that for "t_ino" for inode number types...
Is the problem that grail.h is different under every Unix dialect?  If so,
would it be worthwhile to attempt to standardize this?

I'm getting tired of typing:

#define SEEKTYPE long

myself...
-- 
Ben Cranston  ...{seismo!umcp-cs,ihnp4!rlgvax}!cvl!umd5!zben  zben@umd2.umd.edu