German umlaute in ESS, Emacs & R

[enno_sandkasten]

Hi list,

I use GNU Emacs 22.2.1 (i386-mingw-nt5.1.2600) of 2008-03-26 on Windows XP and ESS Version 5.3.7, both downloaded from http://vgoulet.act.ulaval.ca/en/ressources/emacs/, and R version 2.7.1 for Windows.

My problem is that Emacs cannot handle german umlaute as R variable names:

> ä <- 3
Error: unexpected input in "\201"

This does not occur in RGui.exe. I think this is more an emacs than ess problem, but maybe there are any suggestions.

Thanks, Sven

______________________________________________
ESS-help@... mailing list
https://stat.ethz.ch/mailman/listinfo/ess-help

[Martin Maechler]

>>>>> "es" == enno sandkasten <enno_sandkasten@...>
>>>>>     on Thu, 31 Jul 2008 11:17:38 +0200 (CEST) writes:

    es> Hi list,
    es> I use GNU Emacs 22.2.1 (i386-mingw-nt5.1.2600) of 2008-03-26 on Windows XP and ESS Version 5.3.7, both downloaded from http://vgoulet.act.ulaval.ca/en/ressources/emacs/, and R version 2.7.1 for Windows.

    es> My problem is that Emacs cannot handle german umlaute as R variable names:

    >> ä <- 3
    es> Error: unexpected input in "\201"

    es> This does not occur in RGui.exe. I think this is more an emacs than ess problem, but maybe there are any suggestions.

Yes, it seems an Emacs rather than ESS problem, also visible on
Linux.
I can reproduce the problem also by starting R from the *shell* (M-x
shell) buffer, hence it's not related to ESS,
but most probably related to how emacs uses the 'comint-*'
functions to communicate with the "inferior" process.
I see the problem in all cases (within Emacs),
both when then emacs-input (and -output method of the process is set to
utf-8 or iso-latin-1 respectively {C-x RET p ...;
C-x RET C-h shows you the list} and both for my default unicode
utf-8 locale and with the C locale.

OTOH, I can use 'ä' in the emacs shell more less fine:
I can use it for input, but it does not correctly render for
output.

  $ touch ä   << shows Umlaut
  $ ls -l ä   << shows Umlaut
  -rw-r--r-- 1 maechler sfsstaff 0 Jul 31 18:25 ? << shows question mark
  "?" (
  $ . /u/sfs/adm/locale-deCH-UTF8.sh
  $ ls -l ä <
  -rw-r--r-- 1 maechler sfsstaff 0 Jul 31 18:25 ?
  $ echo BLA >> ä
  $ ls -l ä
  -rw-r--r-- 1 maechler sfsstaff 4 Jul 31 18:25 ?


BTW: I would never use non-ASCII characters in variable names,
      but that's really a separate issue;
 and I can imagine people having "good" reasons to do so.
 Note that it *can* dangerous: I see different files created by
 'touch ä' depending on the locale (ISO-latin-1 vs Unicode UTF-8)


Martin

______________________________________________
ESS-help@... mailing list
https://stat.ethz.ch/mailman/listinfo/ess-help

[Tyler Smith]

Martin Maechler <maechler@...> writes:

>>>>>> "es" == enno sandkasten <enno_sandkasten@...>
>>>>>>     on Thu, 31 Jul 2008 11:17:38 +0200 (CEST) writes:
>
>     es> Hi list, I use GNU Emacs 22.2.1 (i386-mingw-nt5.1.2600) of
>     es> 2008-03-26 on Windows XP and ESS Version 5.3.7, both
>     es> downloaded from
>     es> http://vgoulet.act.ulaval.ca/en/ressources/emacs/, and R
>     es> version 2.7.1 for Windows.
>
>     es> My problem is that Emacs cannot handle german umlaute as R
>     es> variable names:
>
>     >> ä <- 3
>     es> Error: unexpected input in "\201"
>
>     es> This does not occur in RGui.exe. I think this is more an emacs
>     es> than ess problem, but maybe there are any suggestions.
>
> Yes, it seems an Emacs rather than ESS problem, also visible on Linux.
> I can reproduce the problem also by starting R from the *shell* (M-x
> shell) buffer, hence it's not related to ESS, but most probably
> related to how emacs uses the 'comint-*' functions to communicate with
> the "inferior" process.

For what it's worth, it works fine for me in Emacs 23.0.60.1, running on
Debian Lenny, R version 2.7.1, ESS 5.3.8, and also with R running in an
Emacs shell buffer:

> ä <- 3
> ä
[1] 3

So upgrading to Emacs23, which boasts improved unicode support, may fix
the problem. This is not yet a stable release, so you may run into other
issues if you do (I haven't yet though).

Cheers,

Tyler

--
Philosophy of science is about as useful to scientists as ornithology is to
birds.                                 --Richard Feynman

______________________________________________
ESS-help@... mailing list
https://stat.ethz.ch/mailman/listinfo/ess-help

[svga]

I agree, using non-ASCII characters in variable names is a hassle. Unfortunately, these variable names are common in my working area. I find this issue interesting, because german umlauts as a value in a variable is no problem:

> g <- "Säge"; plot(1:10, ylab=g)
> g
[1] "Säge"

works, but

> ä <- 2
Error: unexpected input in "\201"

Btw.: I have no problems with these under Linux.

Best, Sven

----- Original Nachricht ----
Von:     Martin Maechler <maechler@...>
An:      enno_sandkasten@...
Datum:   31.07.2008 18:30
Betreff: Re: [ESS] German umlaute in ESS, Emacs & R

> >>>>> "es" == enno sandkasten <enno_sandkasten@...>
> >>>>>     on Thu, 31 Jul 2008 11:17:38 +0200 (CEST) writes:
>
>     es> Hi list,
>     es> I use GNU Emacs 22.2.1 (i386-mingw-nt5.1.2600) of 2008-03-26 on
> Windows XP and ESS Version 5.3.7, both downloaded from
> http://vgoulet.act.ulaval.ca/en/ressources/emacs/, and R version 2.7.1 for
> Windows.
>
>     es> My problem is that Emacs cannot handle german umlaute as R variable
> names:
>
>     >> ä <- 3
>     es> Error: unexpected input in "\201"
>
>     es> This does not occur in RGui.exe. I think this is more an emacs than
> ess problem, but maybe there are any suggestions.
>
> Yes, it seems an Emacs rather than ESS problem, also visible on
> Linux.
> I can reproduce the problem also by starting R from the *shell* (M-x
> shell) buffer, hence it's not related to ESS,
> but most probably related to how emacs uses the 'comint-*'
> functions to communicate with the "inferior" process.
> I see the problem in all cases (within Emacs),
> both when then emacs-input (and -output method of the process is set to
> utf-8 or iso-latin-1 respectively {C-x RET p ...;
> C-x RET C-h shows you the list} and both for my default unicode
> utf-8 locale and with the C locale.
>
> OTOH, I can use 'ä' in the emacs shell more less fine:
> I can use it for input, but it does not correctly render for
> output.
>
>   $ touch ä   << shows Umlaut
>   $ ls -l ä   << shows Umlaut
>   -rw-r--r-- 1 maechler sfsstaff 0 Jul 31 18:25 ? << shows question mark
>   "?" (
>   $ . /u/sfs/adm/locale-deCH-UTF8.sh
>   $ ls -l ä <
>   -rw-r--r-- 1 maechler sfsstaff 0 Jul 31 18:25 ?
>   $ echo BLA >> ä
>   $ ls -l ä
>   -rw-r--r-- 1 maechler sfsstaff 4 Jul 31 18:25 ?
>
>
> BTW: I would never use non-ASCII characters in variable names,
>       but that's really a separate issue;
>  and I can imagine people having "good" reasons to do so.
>  Note that it *can* dangerous: I see different files created by
>  'touch ä' depending on the locale (ISO-latin-1 vs Unicode UTF-8)
>
>
> Martin
>

______________________________________________
ESS-help@... mailing list
https://stat.ethz.ch/mailman/listinfo/ess-help

[Martin Maechler]

>>>>> "s" == svga  <svga@...>
>>>>>     on Fri, 1 Aug 2008 13:51:16 +0200 (CEST) writes:

    s> I agree, using non-ASCII characters in variable names is
    s> a hassle. Unfortunately, these variable names are common
    s> in my working area. I find this issue interesting,
    s> because german umlauts as a value in a variable is no
    s> problem:
    >> g <- "Säge"; plot(1:10, ylab=g) g
    s> [1] "Säge"

Yes I know, and I'd say that's  *MUCH* more important.
``everyone'' want's to be able to have strings in thei native
language, notbaly used in plots and labels (e.g. factor levels).

    s> works, but

    >> ä <- 2
    s> Error: unexpected input in "\201"

    s> Btw.: I have no problems with these under Linux.

It seems to depend on the version of Emacs too:

As Tyler has noted, the problem is no longer there in the future
emacs 23  ;-)
i.e., I can confirm that the problem is not there for me, if I
use debian/ubuntu's "emacs-snapshot" which is a version of emacs
23 (but I won't use that on a regular basis: of course, I would
do that when I was 20, even when 30, but now, I restrict
to work with unreleased versions of R (and some of its
extensions) and ESS, but for everything else, I'd like to keep
to released SW.

Martin


    s> ----- Original Nachricht ---- Von: Martin Maechler
    s> <maechler@...> An: enno_sandkasten@...
    s> Datum: 31.07.2008 18:30 Betreff: Re: [ESS] German umlaute
    s> in ESS, Emacs & R

    >> >>>>> "es" == enno sandkasten <enno_sandkasten@...>
    >> >>>>> on Thu, 31 Jul 2008 11:17:38 +0200 (CEST) writes:
    >>
    es> Hi list, I use GNU Emacs 22.2.1 (i386-mingw-nt5.1.2600)
    es> of 2008-03-26 on
    >> Windows XP and ESS Version 5.3.7, both downloaded from
    >> http://vgoulet.act.ulaval.ca/en/ressources/emacs/, and R
    >> version 2.7.1 for Windows.
    >>
    es> My problem is that Emacs cannot handle german umlaute as
    es> R variable
    >> names:
    >>
    >> >> ä <- 3
    es> Error: unexpected input in "\201"
    >>
    es> This does not occur in RGui.exe. I think this is more an
    es> emacs than
    >> ess problem, but maybe there are any suggestions.
    >>
    >> Yes, it seems an Emacs rather than ESS problem, also
    >> visible on Linux.  I can reproduce the problem also by
    >> starting R from the *shell* (M-x shell) buffer, hence
    >> it's not related to ESS, but most probably related to how
    >> emacs uses the 'comint-*' functions to communicate with
    >> the "inferior" process.  I see the problem in all cases
    >> (within Emacs), both when then emacs-input (and -output
    >> method of the process is set to utf-8 or iso-latin-1
    >> respectively {C-x RET p ...; C-x RET C-h shows you the
    >> list} and both for my default unicode utf-8 locale and
    >> with the C locale.
    >>
    >> OTOH, I can use 'ä' in the emacs shell more less fine: I
    >> can use it for input, but it does not correctly render
    >> for output.
    >>
    >> $ touch ä << shows Umlaut $ ls -l ä << shows Umlaut
    >> -rw-r--r-- 1 maechler sfsstaff 0 Jul 31 18:25 ? << shows
    >> question mark "?" ( $ . /u/sfs/adm/locale-deCH-UTF8.sh $
    >> ls -l ä < -rw-r--r-- 1 maechler sfsstaff 0 Jul 31 18:25 ?
    >> $ echo BLA >> ä $ ls -l ä -rw-r--r-- 1 maechler sfsstaff
    >> 4 Jul 31 18:25 ?
    >>
    >>
    >> BTW: I would never use non-ASCII characters in variable
    >> names, but that's really a separate issue; and I can
    >> imagine people having "good" reasons to do so.  Note that
    >> it *can* dangerous: I see different files created by
    >> 'touch ä' depending on the locale (ISO-latin-1 vs Unicode
    >> UTF-8)
    >>
    >>
    >> Martin
    >>

______________________________________________
ESS-help@... mailing list
https://stat.ethz.ch/mailman/listinfo/ess-help