Re: strtod sscanf on -CURRENT?
On Fri, May 24, 2002 at 12:13:37PM -0700, Alfred Perlstein wrote: You forgot to include stdlib.h. But what's wrong with my test? strtod works Ok but sscanf does not. pre-gcc3.1 sscanf and sscanf on -STABLE work as expected. -- Igor Roboul, System administrator at Speech Technology Center http://www.speechpro.com http://www.speechpro.ru To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
strtod sscanf on -CURRENT?
Hello,
I have some trouble with PostgreSQL on -CURRENT:
SELECT birth_date,date_part('year',birth_date)::varchar from employee
where id=132;
birth_date | date_part
+
1974-05-09 | 0.046113777160645
(1 row)
This works as expected on -STABLE
I have made simple test:
%cat qq.c
#include stdlib.h
main(int argc,char**argv)
{
double a;
char bb[100];
strcpy(bb,argv[1]);
printf(%lf\n,strtod(bb,NULL));
sscanf(argv[1],%lf,a);
printf(%lf\n,a);
}
%uname -a
FreeBSD sysadm.stc 5.0-CURRENT FreeBSD 5.0-CURRENT #9: Mon May 20
17:34:23 MSD 2002
[EMAIL PROTECTED]:/opt/freebsd/obj/opt/freebsd/src/sys/SYSADM i386
%./a.out 1234.3124
1234.312400
0.124862
%uname -a
FreeBSD r1.stc 4.4-STABLE FreeBSD 4.4-STABLE #0: Tue Dec 25 15:02:18
MSK 2001 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/R1 i386
%./a.out 1234.3124
1234.312400
1234.312400
Maybe I'm wrong, but I think that sscanf on -CURRENT does not work as
expected.
--
Igor Roboul, System administrator at Speech Technology Center
http://www.speechpro.com http://www.speechpro.ru
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: strtod sscanf on -CURRENT?
On Fri, 24 May 2002, Igor Roboul wrote: Maybe I'm wrong, but I think that sscanf on -CURRENT does not work as expected. Interestingly I had some aggro with both scanf and strtod returning somewhat random results recently (about 5th/6th May); I'm afraid I no longer have the results or test programs but I ended up having to hack my own routine together to get a reliable input. What I seem to recall is that scanf didn't return the correct result at all, whereas strtod was okay for the first couple of invocations within the same program, after which it began to return garbage (I'm pretty confident that my program didn't have a buffer overrun, it was a fairly short simple effort) Chris. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: strtod sscanf on -CURRENT?
Chris Hedley [EMAIL PROTECTED] writes: On Fri, 24 May 2002, Igor Roboul wrote: Maybe I'm wrong, but I think that sscanf on -CURRENT does not work as expected. Interestingly I had some aggro with both scanf and strtod returning somewhat random results recently (about 5th/6th May); I'm afraid I no longer have the results or test programs but I ended up having to hack my own routine together to get a reliable input. What I seem to recall is that scanf didn't return the correct result at all, whereas strtod was okay for the first couple of invocations within the same program, after which it began to return garbage (I'm pretty confident that my program didn't have a buffer overrun, it was a fairly short simple effort) Would it be possible for you to reproduce the source to a small program that demonstrates the problem? Best regards, Mike Barcroft To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: strtod sscanf on -CURRENT?
On Fri, 24 May 2002, Mike Barcroft wrote:
Would it be possible for you to reproduce the source to a small
program that demonstrates the problem?
Okay, here's a cheap'n'nasty program to demonstrate the problem; the
revised problem seems to be that strtod() is okay with integers but if a
period is present in the value passed to it, it returns the count of the
digits after the period, rather oddly:
#include stdio.h
main()
{ char buf[80], *eol;
double a;
while(1) {
printf( );
if(!fgets(buf, sizeof buf, stdin) ||
tolower(buf[0]) == 'q') exit(0);
if(eol = strchr(buf, '\n')) *eol = 0;
a = strtod(buf, 0);
printf(buf=%s val=%7.2f\n, buf, a);
}
}
results:
=== uname -a
FreeBSD teabag.cbhnet 5.0-CURRENT FreeBSD 5.0-CURRENT #10: Thu May 16
15:40:46 BST 2002 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/TEABAG i386
=== ./a
324
buf=324 val= 324.00
1
buf=1 val= 1.00
1.2
buf=1.2 val= 1.00
2.1
buf=2.1 val= 1.00
22.53
buf=22.53 val= 2.00
123.45
buf=123.45 val= 2.00
4216.6547
buf=4216.6547 val= 4.00
23513.63462
buf=23513.63462 val= 5.00
23.53256
buf=23.53256 val= 5.00
5432.63426abcde
buf=5432.63426abcde val= 5.00
q
Now I'm not discounting that I've overlooked something really obvious or
have done something stupid with my test and original programs, but I can't
think what it is if this is the case!
Chris.
To Unsubscribe: send mail to [EMAIL PROTECTED]
with unsubscribe freebsd-current in the body of the message
Re: strtod sscanf on -CURRENT?
On Fri, 24 May 2002, Chris Hedley wrote: Now I'm not discounting that I've overlooked something really obvious or have done something stupid with my test and original programs, but I can't think what it is if this is the case! Right, anybody in my locality is now welcome to come and slap me around the face with a wet fish for not including the header file with the function prototype in it. Obviously I needed to make a fool of myself in public before immediately discovering the error of my ways. Sorry, all. Chris. To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
Re: strtod sscanf on -CURRENT?
* Chris Hedley [EMAIL PROTECTED] [020524 10:53] wrote: On Fri, 24 May 2002, Mike Barcroft wrote: Would it be possible for you to reproduce the source to a small program that demonstrates the problem? Okay, here's a cheap'n'nasty program to demonstrate the problem; the revised problem seems to be that strtod() is okay with integers but if a period is present in the value passed to it, it returns the count of the digits after the period, rather oddly: You forgot to include stdlib.h. Please compile things with -Wall before reporting weirdness on the lists. ~ % gcc -Wall t.c t.c:4: warning: return type defaults to `int' t.c: In function `main': t.c:10: warning: implicit declaration of function `tolower' t.c:10: warning: implicit declaration of function `exit' t.c:11: warning: implicit declaration of function `strchr' t.c:11: warning: suggest parentheses around assignment used as truth value t.c:13: warning: implicit declaration of function `strtod' That's from the code you posted, scary what can happen and how much of waste of your and our time this was. :) -- -Alfred Perlstein [[EMAIL PROTECTED]] 'Instead of asking why a piece of software is using 1970s technology, start asking why software is ignoring 30 years of accumulated wisdom.' Tax deductible donations for FreeBSD: http://www.freebsdfoundation.org/ To Unsubscribe: send mail to [EMAIL PROTECTED] with unsubscribe freebsd-current in the body of the message
