GNU bug report logs - #14650
coreutils' getlimits fails to represent float limits correctly

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 14650 in the body.
You can then email your comments to 14650 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: bugdal <at> aerifal.cx
To: bug-coreutils <at> gnu.org
Subject: coreutils' getlimits fails to represent float limits correctly
Date: Tue, 18 Jun 2013 02:10:21 -0400

I noticed a bug in the getlimits utility's printing of DBL_MAX, etc.
The precision at which the limits are printed is grossly insufficient
to represent the actual limits; if the output is converted back to a
floating point number of the appropriate type, the result will differ
greatly from the actual limit.

The format specifier used for printing these values is is %Le, which
gives 7 decimal digits. The number of digits needed to faithfully
represent the limits varies by type, but a safe fix would be replacing

    printf (#TYPE"_MIN=%Le\n", (long double)TYPE##_MIN); \
    printf (#TYPE"_MAX=%Le\n", (long double)TYPE##_MAX);

with:

    printf (#TYPE"_MIN=%.*Le\n", DECIMAL_DIG, (long double)TYPE##_MIN); \
    printf (#TYPE"_MAX=%.*Le\n", DECIMAL_DIG, (long double)TYPE##_MAX);

It would be possible to use appropriate precisions for each type, but
unfortunately I'm not aware of a convenient way to compute the number
of digits needed based on other properties of the type. FLT_DIG and
DBL_DIG are not the correct values for this; they deal with the
opposite round-trip. One solution would be to simply hard-code the
number of digits for float and double when they're IEEE types, and
otherwise use DECIMAL_DIG for all three types (for obscure systems
where float/double are not IEEE single/double). I believe the correct
number of digits for IEEE single or double would be 10 and 17,
respectively.

Rich

Message #8 received at 14650 <at> debbugs.gnu.org (full text, mbox):

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: bugdal <at> aerifal.cx
Cc: 14650 <at> debbugs.gnu.org
Subject: Re: bug#14650: coreutils' getlimits fails to represent float limits
 correctly
Date: Tue, 18 Jun 2013 00:47:03 -0700

On 06/17/2013 11:10 PM, bugdal <at> aerifal.cx wrote:
> I'm not aware of a convenient way to compute the number
> of digits

coreutils already can do that; see lib/ftoastr.h.
Presumably coreutils should use the ftoastr module hear.

Message #13 received at 14650-done <at> debbugs.gnu.org (full text, mbox):

From: Pádraig Brady <P <at> draigBrady.com>
To: Paul Eggert <eggert <at> cs.ucla.edu>
Cc: 14650-done <at> debbugs.gnu.org, bugdal <at> aerifal.cx
Subject: Re: bug#14650: coreutils' getlimits fails to represent float limits
 correctly
Date: Tue, 18 Jun 2013 10:29:32 +0100

On 06/18/2013 08:47 AM, Paul Eggert wrote:
> On 06/17/2013 11:10 PM, bugdal <at> aerifal.cx wrote:
>> I'm not aware of a convenient way to compute the number
>> of digits
> 
> coreutils already can do that; see lib/ftoastr.h.
> Presumably coreutils should use the ftoastr module hear.

Something like the patch below?
which changes the output from:

 FLT_MIN=1.175494e-38
 FLT_MAX=3.402823e+38
 DBL_MIN=2.225074e-308
 DBL_MAX=1.797693e+308
 LDBL_MIN=3.362103e-4932
 LDBL_MAX=1.189731e+4932

to...

 FLT_MIN=1.1754944e-38
 FLT_MAX=3.4028235e+38
 DBL_MIN=2.2250738585072014e-308
 DBL_MAX=1.7976931348623157e+308
 LDBL_MIN=3.3621031431120935063e-4932
 LDBL_MAX=1.189731495357231765e+4932

Note the difference lengths of LDBL_MIN and LDBL_MAX.

cheers,
Pádraig.

diff --git a/src/getlimits.c b/src/getlimits.c
index 7c1fbe2..dfc4b12 100644
--- a/src/getlimits.c
+++ b/src/getlimits.c
@@ -21,6 +21,7 @@
 #include <sys/types.h>
 #include <float.h>

+#include "ftoastr.h"
 #include "system.h"
 #include "long-options.h"

@@ -97,6 +98,19 @@ decimal_absval_add_one (char *buf)
   return result;
 }

+#define PRINT_FLOATTYPE(N, T, FTOASTR, BUFSIZE)                         \
+static void                                                             \
+N (T x)                                                                 \
+{                                                                       \
+  char buf[BUFSIZE];                                                    \
+  FTOASTR (buf, sizeof buf, FTOASTR_LEFT_JUSTIFY, 0, x);                \
+  puts (buf);                                                           \
+}
+
+PRINT_FLOATTYPE (print_FLT, float, ftoastr, FLT_BUFSIZE_BOUND)
+PRINT_FLOATTYPE (print_DBL, double, dtoastr, DBL_BUFSIZE_BOUND)
+PRINT_FLOATTYPE (print_LDBL, long double, ldtoastr, LDBL_BUFSIZE_BOUND)
+
 int
 main (int argc, char **argv)
 {
@@ -127,8 +141,8 @@ main (int argc, char **argv)
     }

 #define print_float(TYPE)                                                \
-  printf (#TYPE"_MIN=%Le\n", (long double)TYPE##_MIN);                   \
-  printf (#TYPE"_MAX=%Le\n", (long double)TYPE##_MAX);
+  printf (#TYPE"_MIN="); print_##TYPE(TYPE##_MIN);                       \
+  printf (#TYPE"_MAX="); print_##TYPE(TYPE##_MAX);

   /* Variable sized ints */
   print_int (CHAR);

Message #16 received at 14650-done <at> debbugs.gnu.org (full text, mbox):

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: Pádraig Brady <P <at> draigBrady.com>
Cc: 14650-done <at> debbugs.gnu.org, bugdal <at> aerifal.cx
Subject: Re: bug#14650: coreutils' getlimits fails to represent float limits
 correctly
Date: Tue, 18 Jun 2013 08:00:46 -0700

On 06/18/2013 02:29 AM, Pádraig Brady wrote:
> Something like the patch below?

Yes, thanks, that was fast and looks good.

Did you forget to push it?

Message #19 received at 14650 <at> debbugs.gnu.org (full text, mbox):

From: Pádraig Brady <P <at> draigBrady.com>
To: Paul Eggert <eggert <at> cs.ucla.edu>
Cc: 14650 <at> debbugs.gnu.org, bugdal <at> aerifal.cx
Subject: Re: bug#14650: coreutils' getlimits fails to represent float limits
 correctly
Date: Tue, 18 Jun 2013 16:21:29 +0100

On 06/18/2013 04:00 PM, Paul Eggert wrote:
> On 06/18/2013 02:29 AM, Pádraig Brady wrote:
>> Something like the patch below?
> 
> Yes, thanks, that was fast and looks good.
> 
> Did you forget to push it?

Adjusted to pass make syntax-check and now pushed.

thanks for the review!
Pádraig.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.