Are completion and prompt expansion 8-bit clean?

Are completion and prompt expansion 8-bit clean?

[Alain Caron]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 714 bytes --]



Hi everyone,

Being a francophone, I like using filenames with french characters.
I usually have no problem with it except in two situations.

The first one is with completion:

touch élément1 élément2

ls é<TAB>							  (expands into)
ls élément<TAB>                 (lists \M-il\Miment1 \M-il\M-iment2)


The second one is with prompt expansion:

PS1="%~ %h> "
mkdir élément
cd élément

The prompt is

\M-il\M-iment 4>

instead of

élément 4>


As far as I can tell, these are the only cases where characters with
the 8th bit set are not printed correctly.

I have seen this behaviour in zsh-3.0.2 and zsh-3.1.2


Regards,

Alain Caron
Sanga Research Corporation
alainc@sangacorp.com
514-288-4498 ext. 225

Re: Are completion and prompt expansion 8-bit clean?

[Zefram]

Alain Caron wrote:
>Subject: Are completion and prompt expansion 8-bit clean?

In one sense, yes: they can handle 8-bit data.  In fact, they can handle
arbitrary binary data.  In another sense, no: they do not treat all 8-bit
data as printable, even when the display will interpret them as Latin-1
or whatever.  Unfortunately, Unix does not provide a way to find out
what characters are printable on a particular display device, so we have
to choose in the source which characters zsh will treat as printable.
No choice can be right for all displays; the current choice is a safe one.

The bug is in Unix; please send the bug report to K.Thompson and
D.Ritchie, twenty years ago.  This bug will be fixed in a future OS[1].

-zefram

[1] If I ever find the time to write the OS I've been planning, this is
one of the things that will be fixed.  Character set and printability
are a function of the tty.

Re: Are completion and prompt expansion 8-bit clean?

[Bart Schaefer]

On Jun 23, 10:15pm, Zefram wrote:
> Subject: Re: Are completion and prompt expansion 8-bit clean?
> 
> In one sense, yes: they can handle 8-bit data.  In fact, they can handle
> arbitrary binary data.  In another sense, no: they do not treat all 8-bit
> data as printable, even when the display will interpret them as Latin-1
> or whatever.  Unfortunately, Unix does not provide a way to find out
> what characters are printable on a particular display device, so we have
> to choose in the source which characters zsh will treat as printable.

Sounds like a job for a module ... supply several localized tty-output
modules, and select one to load based on $TERM or a capability of $TERM
or some other environment setting ($TERM combined with $LANG, perhaps).

Re: Are completion and prompt expansion 8-bit clean?

[Andrej Borsenkow]

On Mon, 23 Jun 1997, Zefram wrote:

> Alain Caron wrote:
> >Subject: Are completion and prompt expansion 8-bit clean?
> 
> In one sense, yes: they can handle 8-bit data.  In fact, they can handle
> arbitrary binary data.  In another sense, no: they do not treat all 8-bit
> data as printable, even when the display will interpret them as Latin-1
> or whatever.  Unfortunately, Unix does not provide a way to find out
> what characters are printable on a particular display device, so we have
> to choose in the source which characters zsh will treat as printable.
> No choice can be right for all displays; the current choice is a safe one.
> 
> The bug is in Unix; please send the bug report to K.Thompson and
> D.Ritchie, twenty years ago.  This bug will be fixed in a future OS[1].
> 

Sorry? What about LANG and LC_CTYPE? They existed long ago.

I don't know, which OS is in use; but check your LANG/LC_CTYPE/LC_ALL
settings - they should be appropriate for French locale. If it doesn't
help, it your OS (probably, implementation of ctype functions or
setlocale()) that is broken. 

BTW it is documeneted in zshparam(1) - at least for 3.0.4 and 3.1.2.

-------------------------------------------------------------------------
Andrej Borsenkow 		Fax:   +7 (095) 252 01 05
SNI ITS Moscow			Tel:   +7 (095) 252 13 88

NERV:  borsenkow.msk		E-Mail: borsenkow.msk@sni.de
-------------------------------------------------------------------------

Re: Are completion and prompt expansion 8-bit clean?

[Peter Stephenson]

Alain Caron wrote:
> Being a francophone, I like using filenames with french characters.
> I usually have no problem with it except in two situations.

If this is any help, I've now set

LC_ALL=en_US

for my X environment and ISO-8859-1 characters are appearing OK
(setting LC_CTYPE would have done), which they didn't in the default
locale.  The setup here is a bit stingy with locales, there's only the
two.  mileage/milage [i.e., your mileage may vary].

I may add some such suggestion to the FAQ --- if anyone has better
suggestions for particular machines, please mail me.

-- 
Peter Stephenson <pws@ifh.de>       Tel: +49 33762 77366
WWW:  http://www.ifh.de/~pws/       Fax: +49 33762 77413
Deutsches Elektronen-Synchrotron --- Institut fuer Hochenergiephysik Zeuthen
DESY-IfH, 15735 Zeuthen, Germany.

Re: Are completion and prompt expansion 8-bit clean?

[raul segura acevedo]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: TEXT/PLAIN; charset=US-ASCII, Size: 697 bytes --]

On Mon, 23 Jun 1997, Alain Caron wrote:

> Being a francophone, I like using filenames with french characters.
> I usually have no problem with it except in two situations.

> The first one is with completion:

> touch élément1 élément2

> ls é<TAB>							  (expands into)
> ls élément<TAB>                 (lists \M-il\Miment1 \M-il\M-iment2)

use "nlsinfo" to see which "locales" are installed in your
machine, set CTYPE to an iso88591 one, for example

export LC_CTYPE=english.iso88591

-- 
raul segura acevedo  Chemistry Institute  raul@hplara.iquimica.unam.mx
......................................................................
and god typed
M-x let-there-be-light