| 1 | // Copyright (C) 2016 The Qt Company Ltd. |
| 2 | // SPDX-License-Identifier: LicenseRef-Qt-Commercial OR LGPL-3.0-only OR GPL-2.0-only OR GPL-3.0-only |
| 3 | // Qt-Security score:critical reason:data-parser |
| 4 | |
| 5 | #include "qplatformdefs.h" |
| 6 | #include "qurl.h" |
| 7 | #include "private/qdataurl_p.h" |
| 8 | |
| 9 | QT_BEGIN_NAMESPACE |
| 10 | |
| 11 | using namespace Qt::Literals; |
| 12 | |
| 13 | /*! |
| 14 | \internal |
| 15 | |
| 16 | Decode a data: URL into its mimetype and payload. Returns a null string if |
| 17 | the URL could not be decoded. |
| 18 | */ |
| 19 | Q_CORE_EXPORT bool qDecodeDataUrl(const QUrl &uri, QString &mimeType, QByteArray &payload) |
| 20 | { |
| 21 | /* https://www.rfc-editor.org/rfc/rfc2397.html |
| 22 | |
| 23 | data:[<mediatype>][;base64],<data> |
| 24 | dataurl := "data:" [ mediatype ] [ ";base64" ] "," data |
| 25 | mediatype := [ type "/" subtype ] *( ";" parameter ) |
| 26 | data := *urlchar |
| 27 | parameter := attribute "=" value |
| 28 | */ |
| 29 | |
| 30 | if (uri.scheme() != "data"_L1 || !uri.host().isEmpty()) |
| 31 | return false; |
| 32 | |
| 33 | payload = uri.toEncoded(options: QUrl::RemoveScheme); |
| 34 | // parse it: |
| 35 | // percent decode after finding the `,`, to workaround parameter |
| 36 | // values containing a percent-encoded comma |
| 37 | const qsizetype pos = payload.indexOf(c: ','); |
| 38 | if (pos != -1) { |
| 39 | QByteArray contentType = payload.first(n: pos).percentDecoded(); |
| 40 | auto data = QLatin1StringView{contentType}; |
| 41 | data = data.trimmed(); |
| 42 | |
| 43 | QLatin1StringView mime; |
| 44 | QLatin1StringView charsetParam; |
| 45 | constexpr auto charset = "charset"_L1 ; |
| 46 | bool first = true; |
| 47 | for (auto part : qTokenize(h&: data, n: u';', flags: Qt::SkipEmptyParts)) { |
| 48 | part = part.trimmed(); |
| 49 | if (first) { |
| 50 | if (part.contains(c: u'/')) |
| 51 | mime = part; |
| 52 | first = false; |
| 53 | } |
| 54 | // Minimal changes, e.g. if it's "charset=;" or "charset;" without |
| 55 | // an encoding, leave it as-is |
| 56 | if (part.startsWith(s: charset, cs: Qt::CaseInsensitive)) |
| 57 | charsetParam = part; |
| 58 | |
| 59 | if (!mime.isEmpty() && !charsetParam.isEmpty()) |
| 60 | break; |
| 61 | } |
| 62 | |
| 63 | if (mime.isEmpty()) { |
| 64 | mime = "text/plain"_L1 ; |
| 65 | if (charsetParam.isEmpty()) |
| 66 | charsetParam = "charset=US-ASCII"_L1 ; |
| 67 | } |
| 68 | if (!charsetParam.isEmpty()) |
| 69 | mimeType = mime + u';' + charsetParam; |
| 70 | else |
| 71 | mimeType = mime; |
| 72 | |
| 73 | // find out if the payload is encoded in Base64 |
| 74 | constexpr auto base64 = ";base64"_L1 ; // per the RFC, at the end of `data` |
| 75 | const bool isBas64 = data.endsWith(s: base64, cs: Qt::CaseInsensitive); |
| 76 | |
| 77 | payload.slice(pos: pos + 1); |
| 78 | data = {}; |
| 79 | payload = std::move(payload).percentDecoded(); |
| 80 | |
| 81 | if (isBas64) { |
| 82 | auto r = QByteArray::fromBase64Encoding(base64: std::move(payload)); |
| 83 | if (!r) { |
| 84 | // just in case someone uses `payload` without checking the returned bool |
| 85 | payload = {}; |
| 86 | return false; // decoding failed |
| 87 | } |
| 88 | payload = std::move(r.decoded); |
| 89 | } |
| 90 | |
| 91 | return true; |
| 92 | } else { |
| 93 | payload = {}; |
| 94 | } |
| 95 | |
| 96 | return false; |
| 97 | } |
| 98 | |
| 99 | QT_END_NAMESPACE |
| 100 | |