[Tcsh] "Readable" Unicode in setenv

[Christos Zoulas]

The tests succeed on bsd because $echo_style=bsd there. On linux
where $echo_style=both, the new echo was unescaping backlashes
which it should not be doing. Fixed.

christos

> On Nov 14, 2021, at 5:03 PM, Kimmo Suominen <kim at netbsd.org> wrote:
> 
> All the Linux builds fail now running these two tests:
> 
> 123: Quoting of expansions in `...`                  FAILED (lexical.at:604)
> 241: $<                                              FAILED (sh.dol.at:18)
> 
> https://github.com/tcsh-org/tcsh/runs/4201252365
> 
> Interestingly FreeBSD and macOS builds succeed.
> 
> On Sun, 14 Nov 2021 at 02:16, Christos Zoulas <christos at zoulas.com> wrote:
>> 
>> Committed, thanks!
>> 
>> christos
>> 
>> On Nov 12, 2021, at 10:46 AM, H.Merijn Brand <tcsh at tux.freedom.nl> wrote:
>> 
>> Signed PGP part
>> On Fri, 12 Nov 2021 16:32:42 +0100, "H.Merijn Brand" <tcsh at tux.freedom.nl> wrote:
>> 
>> If I have an environment variable that is to contain something Unicodish,
>> I currently have to to something similar to
>> 
>> % setenv EURO_CH `perl -CO -e'print "\N{EURO SIGN}"'`
>> or
>> % setenv EURO_CH `perl -CO -e'print "\x{20ac}"'`
>> 
>> so this works
>> % echo $EURO_CH
>> €
>> 
>> I browsed the tcsh manual, but could not find anything that would hint
>> to doing this natively. Is there a (hidden) feature to set Unicode
>> characters from the command line by their name or hex value (in the
>> current encoding)? Something similar to
>> 
>> % setenv EURO_CH "\x{20ac}"
>> % setenv EURO_CH "\u20AC"
>> 
>> In my digging, I found that
>> 
>> % setenv TAB_CH "\t"
>> 
>> just sets the environment variable TAB_CH to a literal '\' followed by
>> a 't'. Which was kinda surprising to me, as I expected a TAB to be in
>> there as a literal TAB. That was done in echo in sh.func.c, so
>> 
>> % echo $TAB_CH
>> 
>> translated \t to TAB. To do the same for \x{20ac}, \xbf, and \u0020ac
>> I changed sh.func.c like below, but I eventually want those escapes
>> to end up literally in the environment. Thoughts welcome
>> 
>> --8<---
>> diff --git a/sh.func.c b/sh.func.c
>> index cdfb6d8d..cbc4ff41 100644
>> --- a/sh.func.c
>> +++ b/sh.func.c
>> @@ -1196,6 +1196,22 @@ doglob(Char **v, struct command *c)
>>    flush();
>> }
>> 
>> +static Char
>> +parse_hex_range(Char **cp, int l)
>> +{
>> +    int  ui = 0;
>> +    char ub[9];
>> +
>> +    if (l > 8) return 0; /* Unsupported length */
>> +
>> +    while (**cp && ui < l && isxdigit(**cp)) {
>> +       ub[ui++] = (char)**cp;
>> +       (*cp)++;
>> +    }
>> +    ub[ui] = (char)0;
>> +    return strtol (ub, NULL, 16);
>> +}
>> +
>> static void
>> xecho(int sep, Char **v)
>> {
>> @@ -1289,6 +1305,28 @@ xecho(int sep, Char **v)
>>                   if (*cp >= '0' && *cp < '8')
>>                       c = c * 8 + *cp++ - '0';
>>                   break;
>> +               case 'x':
>> +                   if (*cp == '{' && isxdigit(*(cp + 1))) { /* \x{20ac} */
>> +                       cp++;
>> +                       c = parse_hex_range (&cp, 8);
>> +                       if (*cp != '}')
>> +                           stderror(ERR_NAME | ERR_VARBEGIN);
>> 
>>    This needs a proper new error message of course
>> 
>> +                       cp++;
>> +                   }
>> +                   else if (isxdigit(*cp)) {   /* \x9f */
>> +                       c = parse_hex_range (&cp, 2);
>> +                   }
>> +                   else /* backward compat */
>> +                       xputchar('\\' | QUOTE);
>> +                   break;
>> +               case 'u':
>> +                   if (isxdigit(*cp)) {        /* \u0020ac */
>> +                       c = parse_hex_range (&cp, 6);
>> +                   }
>> +                   else /* backward compat */
>> +                       xputchar('\\' | QUOTE);
>> +                   break;
>> +
>>               case '\0':
>>                   c = '\\';
>>                   cp--;
>> -->8---
>> 
>> 
>> And a demo of course
>> 
>> % unsetenv EURO
>> % echo $EURO
>> EURO: Undefined variable.
>> 
>> % setenv EURO "\u20ac"
>> % env | grep EURO
>> EURO=\u20ac
>> % echo $EURO
>> €
>> 
>> % setenv EURO `echo "\u20ac"`
>> % env | grep EURO
>> EURO=€
>> % echo $EURO
>> €
>> 
>> --
>> H.Merijn Brand  https://tux.nl   Perl Monger   http://amsterdam.pm.org/
>> using perl5.00307 .. 5.33        porting perl5 on HP-UX, AIX, and Linux
>> https://tux.nl/email.html http://qa.perl.org https://www.test-smoke.org
>> 
>> 
>> 
>> 
>> --
>> Tcsh mailing list
>> Tcsh at astron.com
>> https://mailman.astron.com/mailman/listinfo/tcsh
> --
> Tcsh mailing list
> Tcsh at astron.com
> https://mailman.astron.com/mailman/listinfo/tcsh

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 235 bytes
Desc: Message signed with OpenPGP
URL: <https://mailman.astron.com/pipermail/tcsh/attachments/20211115/f2fbeb14/attachment.asc>