Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!aplcen!samsung!munnari.oz.au!comp.vuw.ac.nz!kaukau.comp.vuw.ac.nz!andrew
From: Andrew.Vignaux@comp.vuw.ac.nz (Andrew Vignaux)
Newsgroups: comp.lang.perl
Subject: Re: Bug in Perl 3.0/6 under SunOS 4.0x on Sun3
Message-ID: <1989Dec31.003420.18182@kaukau.comp.vuw.ac.nz>
Date: 31 Dec 89 00:34:20 GMT
References: <291@fwi.uva.nl>
Sender: news@kaukau.comp.vuw.ac.nz (News Admin)
Reply-To: andrew@comp.vuw.ac.nz (Andrew Vignaux)
Organization: Comp Sci, Victoria University, Wellington, New Zealand.
Lines: 24

In article <291@fwi.uva.nl> casper@fwi.uva.nl (Casper H.S. Dik) writes:
>sun3% perl -e 'printf("%x\n", -1);'
>0
>doarg.c:549:	    (void)sprintf(buf,s,(unsigned int)str_gnum(*(sarg++)));

The unsigned casts went in after my bug report about
	printf("%x\n", 0x82c3060c);
core dumping on some machines (HP9000/300 & /800) due to the bad
double->int conversion.

I was never very happy about my suggestion to just put in the casts, because
	printf("%x\n", 12345678901234567890);
can still dump core or give bogus results.

You can't do very much about protecting perl from this sort of thing.
The best method probably is to examine the format string and perform
the appropriate conversion -- which is exactly what happens now (bar
any core-dump prevention).  Unfortunately, it means that -1 can't be
"%x"d.  Anyway, core dumping is better than printing some other number
(unless Larry wants to put BIGNUMs into perl ;-)

Andrew
-- 
Domain address: Andrew.Vignaux@comp.vuw.ac.nz