My two cents:
Le Tue, Dec 27, 2022 at 03:44:29PM -0500, Chet Ramey a écrit :
>
> So maybe the better thing to do is to list the set of valid format
> specifiers from the standard printf(3) set.
>
I'm not a C programmer, so I didn't hear about `printf %n` before this thread.
Re-reading all `man {1,3,3pl,3avr,...} printf' present on my host, I'v finally
found one short explanation, abount %n in printf(3), but with a lot of other
informations useless, regarding memory size and other. For sample, I'd never
used `%ld`, `%lld` nor`%qd` ( where `%q` is now binded to another bashism...)
(And a bug/remark about security flaw regarding memory misallocation when
using a format script from non validated user input, but if some flaws still
exists under bash, things are very differents)
The first use of this i've imaginated, was for differentiating char length
with byte length:
For this my previous version was: https://stackoverflow.com/a/31009961/1765658
strU8DiffLen0() { local _cLen=${#1} LC_ALL=C;return $((${#1}-_cLen));}
for string in Généralités Language "Yin Yang ☯";do
strU8DiffLen0 "$string";
printf " - %-*s is %2d chars length, but uses %2d bytes\n" \
$((14+$?)) "'$string'" ${#string} $((${#string}+$?));
done
- 'Généralités' is 11 chars length, but uses 14 bytes
- 'Language' is 8 chars length, but uses 8 bytes
- 'Yin Yang ☯' is 10 chars length, but uses 12 bytes
My new version of this became:
strU8DiffLen1() { local _bl;printf -v _ %s%n "$1" _bl;return $((_bl-${#1}));}
This version is near two time quicker!!
time for i in {1..10000};do strU8DiffLen0 'Généralité';ret=$?;done
real 0m0.260s
user 0m0.252s
sys 0m0.005s
time for i in {1..10000};do strU8DiffLen1 'Généralité';ret=$?;done
real 0m0.148s
user 0m0.144s
sys 0m0.000s
But I have to confess:
printf -v _ %s%n "$string" targetVar
This syntax has someting counter-intuitive.
So yes I think personely that printf chapter in bash man page could be expanded.
Thanks for all, regards!
--
Félix
