ision) * 0.5 : \
d + pow_10(-(p)->precision) * 0.5
where (-0 < 0.) isn't true.
And there is probably no easy solution for this:
http://en.wikipedia.org/wiki/Signed_zero
Thanks,
Petr
On 04/16/12 01:14 AM, Chet Ramey wrote:
On 4/13/12 9:57 AM, Petr Sumbera wrote:
On 04/13/12 03:18 PM, Petr Sumbera wrote:
Problem 2:
==
bash -c 'printf "x%+010.0fx\n" 123'
x00+123x
where it should be:
x+00123x
Fixed in chunk #3 (but the problem is there also for other types!).
My fix breaks following case:
bash -c 'pr
should be:
x+123 x
Fixed in chunk #5 and #1.
Problem 4:
==
bash -c 'printf "%f\n" -123'
-inf
where it should be:
-123.00
Fixed in chunk #2.
--
Petr
On 04/13/12 02:27 AM, Chet Ramey wrote:
On 4/12/12 12:25 PM, Petr Sumbera wrote:
Hi,
we are running Bas
Hi,
we are running Bash 3.2.38 on Solaris 10 where asprintf() isn't
available. Thus code in snprintf.c is used.
There is a problem with following command:
bash -c 'printf "x%10.0fx\n" 123'
x000123x
Where correct output should be:
x 123x
It seems that '0' after '.' enables zero pad
Hi,
It seems that patch
http://ftp.gnu.org/gnu/bash/bash-3.2-patches/bash32-049
was left from version 4.0.
Anyone has comment on this? Can this be radded?
Thanks,
Petr
Hi All,
Solaris doesn't implement getcwd() with support for dynamic allocation
of memory. This BASH consider as broken so that configure defines
GETCWD_BROKEN.
Later in bash-3.2/config-bot.h it disables HAVE_GETCWD completely:
/* If we have a getcwd(3), but it calls popen(), #undef HAVE_GETC
Hi All,
Solaris doesn't implement getcwd() with support for dynamic allocation
of memory. This BASH consider as broken so that configure defines
GETCWD_BROKEN.
Later in bash-3.2/config-bot.h it disables HAVE_GETCWD completely:
/* If we have a getcwd(3), but it calls popen(), #undef HAVE_GETCWD
