Thanks Marc for that fix. 

How can I configure the build so that the default encoding is always UTF-8? (So that the option -:t8 need not be explicitly passed.)

I guess, after `configure`, gambit.h should contain definitions like

```
#define ___FILE_SETTINGS_INITIAL ___CHAR_ENCODING_UTF_8
#define ___TERMINAL_SETTINGS_INITIAL ___CHAR_ENCODING_UTF_8
#define ___STDIO_SETTINGS_INITIAL ___CHAR_ENCODING_UTF_8
```

Thanks,

--Vijay

On Sun, Jun 26, 2016 at 6:13 PM, Marc Feeley <feeley@iro.umontreal.ca> wrote:
Indeed the use of the -:t8 runtime option avoids this problem by allowing all Unicode characters to be output to the terminal.  However, the REPL should not go into an infinite loop when it tries to display an error message containing a character beyond the highest character encodable by the terminal.

I found that the REPL’s implementation is using the idiom (##write-string (##object->string message) port) to output to the console the Scheme objects in error messages.  But ##object->string does not know the limits of the character encoding of the port and can produce an output string with characters that can’t be encoded by the port.  So I have added an argument to ##object->string to specify the character encoding limit and fixed the REPL’s implementation accordingly.  The REPL now behaves properly:

% gsi
Gambit v4.8.5

> (println #\u221E)
*** ERROR IN ##wr-ch -- Can't convert to C char-string
(write-char #\u221e '#<input-output-port #2 (console)>)
1>

and

% gsi -:t8
Gambit v4.8.5

> (println #\u221E)
∞

Marc

> On Jun 26, 2016, at 3:12 AM, Vijay Mathew <vijay.the.lisper@gmail.com> wrote:
>
>
> ---------- Forwarded message ----------
> From: Vijay Mathew <vijay.the.lisper@gmail.com>
> Date: Sun, Jun 26, 2016 at 10:51 AM
> Subject: Re: [gambit-list] Printing unicode characters in Gambit
> To: Adam <adam.mlmb@gmail.com>
>
>
> Thanks Adam, will try what you suggested. But I really don't understand why a special switch is required to properly print Unicode characters. Even if the user is unaware of this, 'print' going into a loop is bad behavior.
>
> --Vijay
>
>
> On Jun 26, 2016 8:26 AM, "Adam" <adam.mlmb@gmail.com> wrote:
> You need to switch on unicode output for the output medium (gsc -t8) and maybe also specify the maximum unicode character number that the writer will accept to write. Please report back when you figured.
>
> 2016-06-26 10:34 GMT+08:00 Vijay Mathew <vijay.the.lisper@gmail.com>:
> How can I make gambit to print a Unicode character's symbol (or glyph)?
> For example, I want to display the Unicode infinity symbol ∞ (#\u221E).
> Calling `(display #\u221E)` or `(print #\u221E)` throws the REPL into an infinite loop.
> (This happens for most Unicode literals I tried, not just the infinity symbol :-))
>
>       $ gsc
>       Gambit v4.8.4
>
>> (print #\u221E)
> *** ERROR IN ##wr-ch -- Can't convert to C char-string
> (write-char #\*** ERROR IN ##wr-ch -- Can't convert to C char-string
> (write-char #\*** ERROR IN ##wr-ch -- Can't convert to C char-string
> (write-char #\*** ERROR IN ##wr-ch -- Can't convert to C char-string
> (write-char #\*** ERROR IN ##wr-ch -- Can't convert to C char-string
>
> Thanks,
>
> --- Vijay
>
>
> _______________________________________________
> Gambit-list mailing list
> Gambit-list@iro.umontreal.ca
> https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list
>
>
>
> _______________________________________________
> Gambit-list mailing list
> Gambit-list@iro.umontreal.ca
> https://webmail.iro.umontreal.ca/mailman/listinfo/gambit-list