patch 8.2.0867: using \{xxx} for encoding a modifier is not nice
Problem: Using \{xxx} for encoding a modifier is not nice.
Solution: Use \<*xxx> instead, since it's the same as \<xxx> but producing a
different code.
diff --git a/src/misc2.c b/src/misc2.c
index 61c997d..02fd418 100644
--- a/src/misc2.c
+++ b/src/misc2.c
@@ -2772,13 +2772,14 @@
int modifiers;
int bit;
int key;
- int endchar = (flags & FSK_CURLY) ? '}' : '>';
uvarnumber_T n;
int l;
src = *srcp;
- if (src[0] != ((flags & FSK_CURLY) ? '{' : '<'))
+ if (src[0] != '<')
return 0;
+ if (src[1] == '*') // <*xxx>: do not simplify
+ ++src;
// Find end of modifier list
last_dash = src;
@@ -2796,15 +2797,15 @@
// Anything accepted, like <C-?>.
// <C-"> or <M-"> are not special in strings as " is
// the string delimiter. With a backslash it works: <M-\">
- if (!(in_string && bp[1] == '"') && bp[l + 1] == endchar)
+ if (!(in_string && bp[1] == '"') && bp[l + 1] == '>')
bp += l;
else if (in_string && bp[1] == '\\' && bp[2] == '"'
- && bp[3] == endchar)
+ && bp[3] == '>')
bp += 2;
}
}
if (bp[0] == 't' && bp[1] == '_' && bp[2] && bp[3])
- bp += 3; // skip t_xx, xx may be '-' or '>'/'}'
+ bp += 3; // skip t_xx, xx may be '-' or '>'
else if (STRNICMP(bp, "char-", 5) == 0)
{
vim_str2nr(bp + 5, NULL, &l, STR2NR_ALL, NULL, NULL, 0, TRUE);
@@ -2818,7 +2819,7 @@
}
}
- if (*bp == endchar) // found matching '>' or '}'
+ if (*bp == '>') // found matching '>'
{
end_of_name = bp + 1;
@@ -2864,7 +2865,7 @@
l = mb_ptr2len(last_dash + off);
else
l = 1;
- if (modifiers != 0 && last_dash[l + off] == endchar)
+ if (modifiers != 0 && last_dash[l + off] == '>')
key = PTR2CHAR(last_dash + off);
else
{