| 1 | /* -*- c++ -*- |
| 2 | SPDX-FileCopyrightText: 2002 Marc Mutz <mutz@kde.org> |
| 3 | |
| 4 | SPDX-License-Identifier: LGPL-2.0-or-later |
| 5 | */ |
| 6 | |
| 7 | #include "kcodecsuuencode.h" |
| 8 | |
| 9 | #include <QDebug> |
| 10 | |
| 11 | #include <cassert> |
| 12 | |
| 13 | using namespace KCodecs; |
| 14 | |
| 15 | namespace KCodecs |
| 16 | { |
| 17 | class UUDecoder : public Decoder |
| 18 | { |
| 19 | uint mStepNo; |
| 20 | uchar mAnnouncedOctetCount; // (on current line) |
| 21 | uchar mCurrentOctetCount; // (on current line) |
| 22 | uchar mOutbits; |
| 23 | bool mLastWasCRLF : 1; |
| 24 | bool mSawBegin : 1; // whether we already saw ^begin... |
| 25 | uint mIntoBeginLine : 3; // count #chars we compared against "begin" 0..5 |
| 26 | bool mSawEnd : 1; // whether we already saw ^end... |
| 27 | uint mIntoEndLine : 2; // count #chars we compared against "end" 0..3 |
| 28 | |
| 29 | void searchForBegin(const char *&scursor, const char *const send); |
| 30 | |
| 31 | protected: |
| 32 | friend class UUCodec; |
| 33 | UUDecoder(Codec::NewlineType newline = Codec::NewlineLF) |
| 34 | : Decoder(newline) |
| 35 | , mStepNo(0) |
| 36 | , mAnnouncedOctetCount(0) |
| 37 | , mCurrentOctetCount(0) |
| 38 | , mOutbits(0) |
| 39 | , mLastWasCRLF(true) |
| 40 | , mSawBegin(false) |
| 41 | , mIntoBeginLine(0) |
| 42 | , mSawEnd(false) |
| 43 | , mIntoEndLine(0) |
| 44 | { |
| 45 | } |
| 46 | |
| 47 | public: |
| 48 | ~UUDecoder() override |
| 49 | { |
| 50 | } |
| 51 | |
| 52 | bool decode(const char *&scursor, const char *const send, char *&dcursor, const char *const dend) override; |
| 53 | // ### really needs no finishing??? |
| 54 | bool finish(char *&dcursor, const char *const dend) override |
| 55 | { |
| 56 | Q_UNUSED(dcursor); |
| 57 | Q_UNUSED(dend); |
| 58 | return true; |
| 59 | } |
| 60 | }; |
| 61 | |
| 62 | Encoder *UUCodec::makeEncoder(NewlineType newline) const |
| 63 | { |
| 64 | Q_UNUSED(newline) |
| 65 | return nullptr; // encoding not supported |
| 66 | } |
| 67 | |
| 68 | Decoder *UUCodec::makeDecoder(NewlineType newline) const |
| 69 | { |
| 70 | return new UUDecoder(newline); |
| 71 | } |
| 72 | |
| 73 | /********************************************************/ |
| 74 | /********************************************************/ |
| 75 | /********************************************************/ |
| 76 | |
| 77 | void UUDecoder::searchForBegin(const char *&scursor, const char *const send) |
| 78 | { |
| 79 | static const char begin[] = "begin\n" ; |
| 80 | static const uint beginLength = 5; // sic! |
| 81 | |
| 82 | assert(!mSawBegin || mIntoBeginLine > 0); |
| 83 | |
| 84 | while (scursor != send) { |
| 85 | uchar ch = *scursor++; |
| 86 | if (ch == begin[mIntoBeginLine]) { |
| 87 | if (mIntoBeginLine < beginLength) { |
| 88 | // found another char |
| 89 | ++mIntoBeginLine; |
| 90 | if (mIntoBeginLine == beginLength) { |
| 91 | mSawBegin = true; // "begin" complete, now search the next \n... |
| 92 | } |
| 93 | } else { // mIntoBeginLine == beginLength |
| 94 | // found '\n': begin line complete |
| 95 | mLastWasCRLF = true; |
| 96 | mIntoBeginLine = 0; |
| 97 | return; |
| 98 | } |
| 99 | } else if (mSawBegin) { |
| 100 | // OK, skip stuff until the next \n |
| 101 | } else { |
| 102 | // qWarning() << "UUDecoder: garbage before \"begin\", resetting parser"; |
| 103 | mIntoBeginLine = 0; |
| 104 | } |
| 105 | } |
| 106 | } |
| 107 | |
| 108 | // uuencoding just shifts all 6-bit octets by 32 (SP/' '), except NUL, |
| 109 | // which gets mapped to 0x60 |
| 110 | static inline uchar uuDecode(uchar c) |
| 111 | { |
| 112 | return (c - ' ') // undo shift and |
| 113 | & 0x3F; // map 0x40 (0x60-' ') to 0... |
| 114 | } |
| 115 | |
| 116 | bool UUDecoder::decode(const char *&scursor, const char *const send, char *&dcursor, const char *const dend) |
| 117 | { |
| 118 | // First, check whether we still need to find the "begin" line: |
| 119 | if (!mSawBegin || mIntoBeginLine != 0) { |
| 120 | searchForBegin(scursor, send); |
| 121 | } else if (mSawEnd) { |
| 122 | // or if we are past the end line: |
| 123 | scursor = send; // do nothing anymore... |
| 124 | return true; |
| 125 | } |
| 126 | |
| 127 | while (dcursor != dend && scursor != send) { |
| 128 | uchar ch = *scursor++; |
| 129 | uchar value; |
| 130 | |
| 131 | // Check whether we need to look for the "end" line: |
| 132 | if (mIntoEndLine > 0) { |
| 133 | static const char end[] = "end" ; |
| 134 | static const uint endLength = 3; |
| 135 | |
| 136 | if (ch == end[mIntoEndLine]) { |
| 137 | ++mIntoEndLine; |
| 138 | if (mIntoEndLine == endLength) { |
| 139 | mSawEnd = true; |
| 140 | scursor = send; // shortcut to the end |
| 141 | return true; |
| 142 | } |
| 143 | continue; |
| 144 | } else { |
| 145 | // qWarning() << "UUDecoder: invalid line octet count looks like \"end\" (mIntoEndLine =" |
| 146 | // << mIntoEndLine << ")!"; |
| 147 | mIntoEndLine = 0; |
| 148 | // fall through... |
| 149 | } |
| 150 | } |
| 151 | |
| 152 | // Normal parsing: |
| 153 | |
| 154 | // The first char of a line is an encoding of the length of the |
| 155 | // current line. We simply ignore it: |
| 156 | if (mLastWasCRLF) { |
| 157 | // reset char-per-line counter: |
| 158 | mLastWasCRLF = false; |
| 159 | mCurrentOctetCount = 0; |
| 160 | |
| 161 | // try to decode the chars-on-this-line announcement: |
| 162 | if (ch == 'e') { // maybe the beginning of the "end"? ;-) |
| 163 | mIntoEndLine = 1; |
| 164 | } else if (ch > 0x60) { |
| 165 | // ### invalid line length char: what shall we do?? |
| 166 | } else if (ch > ' ') { |
| 167 | mAnnouncedOctetCount = uuDecode(c: ch); |
| 168 | } else if (ch == '\n') { |
| 169 | mLastWasCRLF = true; // oops, empty line |
| 170 | } |
| 171 | |
| 172 | continue; |
| 173 | } |
| 174 | |
| 175 | // try converting ch to a 6-bit value: |
| 176 | if (ch > 0x60) { |
| 177 | continue; // invalid char |
| 178 | } else if (ch > ' ') { |
| 179 | value = uuDecode(c: ch); |
| 180 | } else if (ch == '\n') { // line end |
| 181 | mLastWasCRLF = true; |
| 182 | continue; |
| 183 | } else { |
| 184 | continue; |
| 185 | } |
| 186 | |
| 187 | // add the new bits to the output stream and flush full octets: |
| 188 | switch (mStepNo) { |
| 189 | case 0: |
| 190 | mOutbits = value << 2; |
| 191 | break; |
| 192 | case 1: |
| 193 | if (mCurrentOctetCount < mAnnouncedOctetCount) { |
| 194 | *dcursor++ = (char)(mOutbits | value >> 4); |
| 195 | } |
| 196 | ++mCurrentOctetCount; |
| 197 | mOutbits = value << 4; |
| 198 | break; |
| 199 | case 2: |
| 200 | if (mCurrentOctetCount < mAnnouncedOctetCount) { |
| 201 | *dcursor++ = (char)(mOutbits | value >> 2); |
| 202 | } |
| 203 | ++mCurrentOctetCount; |
| 204 | mOutbits = value << 6; |
| 205 | break; |
| 206 | case 3: |
| 207 | if (mCurrentOctetCount < mAnnouncedOctetCount) { |
| 208 | *dcursor++ = (char)(mOutbits | value); |
| 209 | } |
| 210 | ++mCurrentOctetCount; |
| 211 | mOutbits = 0; |
| 212 | break; |
| 213 | default: |
| 214 | assert(0); |
| 215 | } |
| 216 | mStepNo = (mStepNo + 1) % 4; |
| 217 | |
| 218 | // check whether we ran over the announced octet count for this line: |
| 219 | if (mCurrentOctetCount == mAnnouncedOctetCount + 1) { |
| 220 | // qWarning() |
| 221 | // << "UUDecoder: mismatch between announced (" |
| 222 | // << mAnnouncedOctetCount << ") and actual line octet count!"; |
| 223 | } |
| 224 | } |
| 225 | |
| 226 | // return false when caller should call us again: |
| 227 | return scursor == send; |
| 228 | } // UUDecoder::decode() |
| 229 | |
| 230 | } // namespace KCodecs |
| 231 | |