Path: utzoo!mnetor!uunet!vixie!paul
From: paul@vixie.UUCP (Paul Vixie Esq)
Newsgroups: comp.sources.d
Subject: Re: A perl program to change usg style dfs into Berkeley
Message-ID: <801@vixie.UUCP>
Date: 7 Feb 88 19:58:04 GMT
References: <1821621@ncrcpx.UUCP>
Reply-To: paul@vixie.UUCP (Paul Vixie Esq)
Organization: Vixie Enterprises, San Francisco
Lines: 52
Keywords: perl source code

In article <1821621@ncrcpx.UUCP> craig@ncrcpx.UUCP (R. Craig Peterson) writes:
#[...] berkeley style dfs...  There was some c source posted not too long ago
#to do the same, but there were some problems running it on my system, and
#since perl was there, I used it instead of fixing the c version.

There is a new, fixed-up version of bdf.c in testing.  I'll be sending it out
to comp.sources.somethingorother pretty soon now...

#The input (which it gets itself from df -t) looks like:
#
#/usr/acct       (/dev/dsk/1s1):    34928 blocks   13049 i-nodes
#                   total:   233070 blocks   29120 i-nodes
#/               (/dev/dsk/0s1):    28878 blocks    9506 i-nodes
#                   total:   212924 blocks   26608 i-nodes

Beware that there can be blanks and asterisks in irritating places in the
output of 'df -t' on a USG system.  Also, the /dev/dsk can be /dev/dsk/abcdef
(six letters) on some systems.  Nice to have all this stuff standardized...

#The output looks like:
#
#Filesystem     kbytes    used  avail capacity  iused   ifree %iused Mounted on
#/usr/acct      116535   99073  17462    85%    16071   13049    55% 1s1
#/              106462   92023  14439    86%    17103    9505    64% 0s1

This is backwards from Berkeley DF -- the "Filesystem" column has the device
file and the "Mounted on" has the mount point.

#(P.S.  Great work with Perl, Larry!!  It looks like I'll be using it 
#alot more in the future!  For the rest of you, if you haven't looked at
#Perl, and you process alot of strings, etc., you're missing alot!)

I think if I'd had perl when I wrote bdf.c, I'd still have made up my own.
Nice as it is, the number of patches per week make perl a little hard to keep
up with, especially for the 10-odd USG systems I administer.

Still, Perl is obviously a better choice than C for this kind of task; the
fact that your version fits in 30 very readable lines vs. my C version at
around 200 messy ones clearly shows the advantages of Perl.

I therefore echo your sentiments -- great job, Larry.

#if test `wc -l < bdf` -ne 30 ; then
#	echo 'shar: bdf was damaged during transit (should have been 30  bytes)'
#fi

Actually, it "should have been 30 *LINES*", not "*BYTES*".
-- 
Paul A Vixie Esq
paul%vixie@uunet.uu.net
{uunet,ptsfa,hoptoad}!vixie!paul
San Francisco, (415) 647-7023