Stephane CHAZELAS <stephane.chaze...@gmail.com> wrote:
> Note if I remember the past discussion correctly, readdir() is
> already allowed not to return "." and ".." and in that case
> .* expansion is allowed (even required?) not to include "." and
> "..".
>
> I'd like it to go further and allow .* not to include "." and
> ".." regardless of whether they're returned by readdir. Like zsh
> or pdksh do or ksh93 used to do when FIGNORE is not empty.
Allowing this feature would be a first step before requiring it.
Then you would be able to rely on that feature.
> Playing with ltrace, for things like
>
> printf 'foo bar %d, %20s, %b\n' 1 2 3
>
> none of them I tried (GNU, bash, zsh) call
> printf("foo bar %d, %20s, %b", 1, "2", "3"). And obviously, that
> wouldn't work for %b.
The IBM implementation used on all commercial UNIXes calls printf(3).
Your expectation however does not apply as non-format-related parts
of the format string are output separately and only the format related parts
are use separately for a printf(3) call.
> > It may make sense to add a wprintf(1) ;-)
>
> All other text utilities deal with characters. We didn't get a
> wcut, wsed a wtr a wecho (which has similar problems on many
> implementations) when they were modified to support multibyte
> characters.
They work different. If printf(1) had been added after aprox. 1998, people may
have decided different and may have required printf(1) to work on wchar_t
internally.
Jörg
--
EMail:jo...@schily.net (home) Jörg Schilling D-13353 Berlin
joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/
URL: http://cdrecord.org/private/ http://sf.net/projects/schilytools/files/'
