"@{5E}" and "@{7E}" escapes are wrong in Z-code

[erkyrath]

You're supposed to be able to escape string characters in two ways: "@@DEC" and "@{HEX}". The latter is newer and supports any Unicode value.

However, the hex form doesn't come out right for a couple of characters:

	print "@@92 is backslash, @@64 is atsign, @@94 is caret, @@126 is tilde.^";
	print "@{5C} is backslash, @{40} is atsign, @{5E} is caret, @{7E} is tilde.^";

\ is backslash, @ is atsign, ^ is caret, ~ is tilde.
\ is backslash, @ is atsign,
is caret, " is tilde.

That is, "@{5E}" is interpreted as "^" which is then interpreted as a newline. "@{7E}" similarly gets double-interpreted as a quote character.

(In Glulx both lines print the same thing, as expected.)

This bug goes back to 6.31 (at least) so it might not be desirable to fix it. Some old Z-code games might depend on this behavior, inconsistent as it is. But it's worth noting.

[heasm66]

In an attempt to close some old issues...

Suggestion is to add a warning that points to the syntax required for printing caret and tilde with @@-syntax. This won't break behavior if some game, against all odds, depend on this inconsistency.

chars.c

@@ -1183,6 +1183,15 @@ extern int32 text_to_unicode(char *text)
        while ((text[i] != '}') && (text[i] != 0)) i++;
        if (text[i] == '}') i++;
        textual_form_length = i;
        switch (total) {
        /* Warn that ~ and ^ is being translated to double-quote and new-line. */
        case 94:   
            warning("@{5E} will print a newline instead of a caret (^), use @@94 for a caret.");
            break;
        case 126:              
            warning("@{7E} will print a double-quote (\") instead of a tilde (~), use @@126 for tilde.");
            break;
        }
        return total;
    }

[erkyrath]

The @{HEX} notation is newer and preferred in Glulx, so we should support it correctly.

[heasm66]

Because it already works for Glulx, maybe only warn for Z-code?

[erkyrath]

How about we fix the bug instead.

[heasm66]

Fine by me, but you raised concerns about backward compability above.