LibIMAP+Mail: Implement RFC2047 message header encoding

This enables us to display email subject fields with non-ASCII
characters in Mail :^)

Userland/Libraries/LibIMAP/MessageHeaderEncoding.cpp

@@ -0,0 +1,97 @@
+/*
+ * Copyright (c) 2023, Valtteri Koskivuori <vkoskiv@gmail.com>
+ *
+ * SPDX-License-Identifier: BSD-2-Clause
+ */
+
+#include "QuotedPrintable.h"
+#include <AK/Base64.h>
+#include <AK/GenericLexer.h>
+#include <AK/StringBuilder.h>
+#include <LibIMAP/MessageHeaderEncoding.h>
+#include <LibTextCodec/Decoder.h>
+
+namespace IMAP {
+
+ErrorOr<ByteBuffer> decode_rfc2047_encoded_words(StringView input)
+{
+    GenericLexer lexer(input);
+    StringBuilder output;
+
+    while (!lexer.is_eof()) {
+        auto ascii_view = lexer.consume_until("=?"sv);
+        DeprecatedString ascii = ascii_view.replace("\r"sv, " "sv, ReplaceMode::All);
+        ascii = ascii.replace("\n"sv, " "sv, ReplaceMode::All);
+        TRY(output.try_append(ascii));
+        if (lexer.is_eof())
+            break;
+        lexer.consume_specific("=?"sv);
+        auto charset = lexer.consume_until('?');
+        lexer.consume();
+        auto encoding = lexer.consume_until('?');
+        lexer.consume();
+        auto encoded_text = lexer.consume_until("?=");
+        lexer.consume_specific("?="sv);
+
+        // RFC 2047 Section 6.2, "...any 'linear-white-space' that separates a pair of adjacent 'encoded-word's is ignored."
+        // https://datatracker.ietf.org/doc/html/rfc2047#section-6.2
+        bool found_next_start = false;
+        int spaces = 0;
+        for (size_t i = 0; i < lexer.tell_remaining(); ++i) {
+            if (lexer.peek(i) == ' ' || lexer.peek(i) == '\r' || lexer.peek(i) == '\n') {
+                spaces++;
+                if (lexer.peek(i + 1) == '=' && lexer.peek(i + 2) == '?') {
+                    found_next_start = true;
+                    break;
+                }
+            } else {
+                break;
+            }
+        }
+        if (found_next_start) {
+            for (int i = 0; i < spaces; i++) {
+                lexer.consume();
+            }
+        }
+
+        ByteBuffer first_pass_decoded;
+        if (encoding == 'Q' || encoding == 'q') {
+            auto maybe_decoded_data = decode_quoted_printable(encoded_text);
+            if (maybe_decoded_data.is_error()) {
+                dbgln("Failed to decode quoted-printable rfc2047 text, skipping.");
+                continue;
+            }
+            // RFC 2047 Section 4.2.2, https://datatracker.ietf.org/doc/html/rfc2047#section-4.2
+            auto decoded_data = maybe_decoded_data.release_value();
+            for (auto character : decoded_data.bytes()) {
+                if (character == '_')
+                    first_pass_decoded.append(' ');
+                else
+                    first_pass_decoded.append(character);
+            }
+        } else if (encoding == 'B' || encoding == 'b') {
+            auto maybe_decoded_data = AK::decode_base64(encoded_text);
+            if (maybe_decoded_data.is_error()) {
+                dbgln("Failed to decode base64-encoded rfc2047 text, skipping.");
+                continue;
+            }
+            first_pass_decoded = maybe_decoded_data.release_value();
+        } else {
+            dbgln("Unknown encoding \"{}\" found, skipping, original string: \"{}\"", encoding, input);
+            continue;
+        }
+        if (first_pass_decoded.is_empty())
+            continue;
+        auto maybe_decoder = TextCodec::decoder_for(charset);
+        if (!maybe_decoder.has_value()) {
+            dbgln("No decoder found for charset \"{}\", skipping.", charset);
+            continue;
+        }
+        auto decoded_text = TRY(maybe_decoder->to_utf8(first_pass_decoded));
+        TRY(output.try_append(decoded_text));
+    }
+
+    return output.to_byte_buffer();
+}
+
+}