Path: utzoo!utgpu!water!watmath!clyde!att-cb!att-ih!ihnp4!occrsh!uokmax!rmtodd
From: rmtodd@uokmax.UUCP (Richard Michael Todd)
Newsgroups: comp.os.minix
Subject: Re: BUNDLE for Minix (was: Backup for HD asked)
Message-ID: <1072@uokmax.UUCP>
Date: 29 Feb 88 15:57:34 GMT
References: <46400003@ztivax.UUCP> <10579@santra.UUCP> <665@ast.cs.vu.nl>
Reply-To: rmtodd@uokmax.UUCP ()
Organization: University of Oklahoma, Norman, OK
Lines: 17

In article <665@ast.cs.vu.nl> ast@cs.vu.nl (Andy Tanenbaum) writes:
>It is not a bug.  It is a feature.  Having pipes larger than
>7K would mean not using "small" files, but files with indirect blocks as pipe
>inodes.  V7 does not allow this; neither does MINIX.  There is no easy fix.
I think it's a bug.  When you attempt to write a block greater than 7K to
a file and it fails for no apparent reason, it's a violation of the Principle
of Least Astonishment.  Why should MINIX perpetuate all of V7's bugs?  Does
Sys5 have this bug? (BSD pipes aren't implemented using spare inodes, so this
doesn't apply to them, and BSD doesn't freak out when you write >7k to a pipe.)
Some time ago I found a way to patch MINIX FS to allow writes >7K to a pipe.
It required changes to two lines, I think.  Nothing else seemed to break.
Unfortunately I'm at school right now and the MINIX machine's at home, so
it'll be at least this weekend before I can post them.
--------------------------------------------------------------------------
Richard Todd
USSnail:820 Annie Court,Norman OK 73069
UUCP: {allegra!cbosgd|ihnp4}!occrsh!uokmax!rmtodd