widgets(terminal): read full codepoints
Only read full codepoints. Some programs will split codepoints between reads, which can mess with our parser (looking at you, btop)
This commit is contained in:
parent
eeb69f012a
commit
37b0d5c69d
1 changed files with 19 additions and 0 deletions
|
@ -66,6 +66,17 @@ pub fn parseReader(self: *Parser, buffered: *BufferedReader) !Event {
|
|||
|
||||
inline fn parseGround(self: *Parser, reader: *BufferedReader) !Event {
|
||||
var buf: [1]u8 = undefined;
|
||||
{
|
||||
std.debug.assert(self.buf.items.len > 0);
|
||||
// Handle first byte
|
||||
const len = try std.unicode.utf8ByteSequenceLength(self.buf.items[0]);
|
||||
var i: usize = 1;
|
||||
while (i < len) : (i += 1) {
|
||||
const read = try reader.read(&buf);
|
||||
if (read == 0) return error.EOF;
|
||||
try self.buf.append(buf[0]);
|
||||
}
|
||||
}
|
||||
while (true) {
|
||||
if (reader.start == reader.end) return .{ .print = self.buf.items };
|
||||
const n = try reader.read(&buf);
|
||||
|
@ -78,6 +89,14 @@ inline fn parseGround(self: *Parser, reader: *BufferedReader) !Event {
|
|||
},
|
||||
else => {
|
||||
try self.buf.append(b);
|
||||
const len = try std.unicode.utf8ByteSequenceLength(b);
|
||||
var i: usize = 1;
|
||||
while (i < len) : (i += 1) {
|
||||
const read = try reader.read(&buf);
|
||||
if (read == 0) return error.EOF;
|
||||
|
||||
try self.buf.append(buf[0]);
|
||||
}
|
||||
},
|
||||
}
|
||||
}
|
||||
|
|
Loading…
Reference in a new issue