summaryrefslogtreecommitdiffstats
path: root/gfx/ots/src/cff.cc
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--gfx/ots/src/cff.cc1381
1 files changed, 1381 insertions, 0 deletions
diff --git a/gfx/ots/src/cff.cc b/gfx/ots/src/cff.cc
new file mode 100644
index 0000000000..af54a4a16c
--- /dev/null
+++ b/gfx/ots/src/cff.cc
@@ -0,0 +1,1381 @@
+// Copyright (c) 2012-2017 The OTS Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "cff.h"
+
+#include <cstring>
+#include <utility>
+#include <vector>
+
+#include "maxp.h"
+#include "cff_charstring.h"
+#include "variations.h"
+
+// CFF - PostScript font program (Compact Font Format) table
+// http://www.microsoft.com/typography/otspec/cff.htm
+// http://www.microsoft.com/typography/otspec/cffspec.htm
+
+#define TABLE_NAME "CFF"
+
+namespace {
+
+enum DICT_OPERAND_TYPE {
+ DICT_OPERAND_INTEGER,
+ DICT_OPERAND_REAL,
+ DICT_OPERATOR,
+};
+
+enum DICT_DATA_TYPE {
+ DICT_DATA_TOPLEVEL,
+ DICT_DATA_FDARRAY,
+ DICT_DATA_PRIVATE,
+};
+
+enum FONT_FORMAT {
+ FORMAT_UNKNOWN,
+ FORMAT_CID_KEYED,
+ FORMAT_OTHER, // Including synthetic fonts
+};
+
+// see Appendix. A
+const size_t kNStdString = 390;
+
+typedef std::pair<uint32_t, DICT_OPERAND_TYPE> Operand;
+
+bool ReadOffset(ots::Buffer &table, uint8_t off_size, uint32_t *offset) {
+ if (off_size > 4) {
+ return OTS_FAILURE();
+ }
+
+ uint32_t tmp32 = 0;
+ for (unsigned i = 0; i < off_size; ++i) {
+ uint8_t tmp8 = 0;
+ if (!table.ReadU8(&tmp8)) {
+ return OTS_FAILURE();
+ }
+ tmp32 <<= 8;
+ tmp32 += tmp8;
+ }
+ *offset = tmp32;
+ return true;
+}
+
+bool ParseIndex(ots::Buffer &table, ots::CFFIndex &index, bool cff2 = false) {
+ index.off_size = 0;
+ index.offsets.clear();
+
+ if (cff2) {
+ if (!table.ReadU32(&(index.count))) {
+ return OTS_FAILURE();
+ }
+ } else {
+ uint16_t count;
+ if (!table.ReadU16(&count)) {
+ return OTS_FAILURE();
+ }
+ index.count = count;
+ }
+
+ if (index.count == 0) {
+ // An empty INDEX.
+ index.offset_to_next = table.offset();
+ return true;
+ }
+
+ if (!table.ReadU8(&(index.off_size))) {
+ return OTS_FAILURE();
+ }
+ if (index.off_size < 1 || index.off_size > 4) {
+ return OTS_FAILURE();
+ }
+
+ const size_t array_size = (index.count + 1) * index.off_size;
+ // less than ((64k + 1) * 4), thus does not overflow.
+ const size_t object_data_offset = table.offset() + array_size;
+ // does not overflow too, since offset() <= 1GB.
+
+ if (object_data_offset >= table.length()) {
+ return OTS_FAILURE();
+ }
+
+ for (unsigned i = 0; i <= index.count; ++i) { // '<=' is not a typo.
+ uint32_t rel_offset = 0;
+ if (!ReadOffset(table, index.off_size, &rel_offset)) {
+ return OTS_FAILURE();
+ }
+ if (rel_offset < 1) {
+ return OTS_FAILURE();
+ }
+ if (i == 0 && rel_offset != 1) {
+ return OTS_FAILURE();
+ }
+
+ if (rel_offset > table.length()) {
+ return OTS_FAILURE();
+ }
+
+ // does not underflow.
+ if (object_data_offset > table.length() - (rel_offset - 1)) {
+ return OTS_FAILURE();
+ }
+
+ index.offsets.push_back(
+ object_data_offset + (rel_offset - 1)); // less than length(), 1GB.
+ }
+
+ for (unsigned i = 1; i < index.offsets.size(); ++i) {
+ // We allow consecutive identical offsets here for zero-length strings.
+ // See http://crbug.com/69341 for more details.
+ if (index.offsets[i] < index.offsets[i - 1]) {
+ return OTS_FAILURE();
+ }
+ }
+
+ index.offset_to_next = index.offsets.back();
+ return true;
+}
+
+bool ParseNameData(
+ ots::Buffer *table, const ots::CFFIndex &index, std::string* out_name) {
+ uint8_t name[256] = {0};
+
+ const size_t length = index.offsets[1] - index.offsets[0];
+ // font names should be no longer than 127 characters.
+ if (length > 127) {
+ return OTS_FAILURE();
+ }
+
+ table->set_offset(index.offsets[0]);
+ if (!table->Read(name, length)) {
+ return OTS_FAILURE();
+ }
+
+ for (size_t i = 0; i < length; ++i) {
+ // setting the first byte to NUL is allowed.
+ if (i == 0 && name[i] == 0) continue;
+ // non-ASCII characters are not recommended (except the first character).
+ if (name[i] < 33 || name[i] > 126) {
+ return OTS_FAILURE();
+ }
+ // [, ], ... are not allowed.
+ if (std::strchr("[](){}<>/% ", name[i])) {
+ return OTS_FAILURE();
+ }
+ }
+
+ *out_name = reinterpret_cast<char *>(name);
+ return true;
+}
+
+bool CheckOffset(const Operand& operand, size_t table_length) {
+ if (operand.second != DICT_OPERAND_INTEGER) {
+ return OTS_FAILURE();
+ }
+ if (operand.first >= table_length) {
+ return OTS_FAILURE();
+ }
+ return true;
+}
+
+bool CheckSid(const Operand& operand, size_t sid_max) {
+ if (operand.second != DICT_OPERAND_INTEGER) {
+ return OTS_FAILURE();
+ }
+ if (operand.first > sid_max) {
+ return OTS_FAILURE();
+ }
+ return true;
+}
+
+bool ParseDictDataBcd(ots::Buffer &table, std::vector<Operand> &operands) {
+ bool read_decimal_point = false;
+ bool read_e = false;
+
+ uint8_t nibble = 0;
+ size_t count = 0;
+ while (true) {
+ if (!table.ReadU8(&nibble)) {
+ return OTS_FAILURE();
+ }
+ if ((nibble & 0xf0) == 0xf0) {
+ if ((nibble & 0xf) == 0xf) {
+ // TODO(yusukes): would be better to store actual double value,
+ // rather than the dummy integer.
+ operands.push_back(std::make_pair(static_cast<uint32_t>(0),
+ DICT_OPERAND_REAL));
+ return true;
+ }
+ return OTS_FAILURE();
+ }
+ if ((nibble & 0x0f) == 0x0f) {
+ operands.push_back(std::make_pair(static_cast<uint32_t>(0),
+ DICT_OPERAND_REAL));
+ return true;
+ }
+
+ // check number format
+ uint8_t nibbles[2];
+ nibbles[0] = (nibble & 0xf0) >> 8;
+ nibbles[1] = (nibble & 0x0f);
+ for (unsigned i = 0; i < 2; ++i) {
+ if (nibbles[i] == 0xd) { // reserved number
+ return OTS_FAILURE();
+ }
+ if ((nibbles[i] == 0xe) && // minus
+ ((count > 0) || (i > 0))) {
+ return OTS_FAILURE(); // minus sign should be the first character.
+ }
+ if (nibbles[i] == 0xa) { // decimal point
+ if (!read_decimal_point) {
+ read_decimal_point = true;
+ } else {
+ return OTS_FAILURE(); // two or more points.
+ }
+ }
+ if ((nibbles[i] == 0xb) || // E+
+ (nibbles[i] == 0xc)) { // E-
+ if (!read_e) {
+ read_e = true;
+ } else {
+ return OTS_FAILURE(); // two or more E's.
+ }
+ }
+ }
+ ++count;
+ }
+}
+
+bool ParseDictDataEscapedOperator(ots::Buffer &table,
+ std::vector<Operand> &operands) {
+ uint8_t op = 0;
+ if (!table.ReadU8(&op)) {
+ return OTS_FAILURE();
+ }
+
+ if ((op <= 14) ||
+ (op >= 17 && op <= 23) ||
+ (op >= 30 && op <= 38)) {
+ operands.push_back(std::make_pair((12U << 8) + op, DICT_OPERATOR));
+ return true;
+ }
+
+ // reserved area.
+ return OTS_FAILURE();
+}
+
+bool ParseDictDataNumber(ots::Buffer &table, uint8_t b0,
+ std::vector<Operand> &operands) {
+ uint8_t b1 = 0;
+ uint8_t b2 = 0;
+ uint8_t b3 = 0;
+ uint8_t b4 = 0;
+
+ switch (b0) {
+ case 28: // shortint
+ if (!table.ReadU8(&b1) ||
+ !table.ReadU8(&b2)) {
+ return OTS_FAILURE();
+ }
+ operands.push_back(std::make_pair(
+ static_cast<uint32_t>((b1 << 8) + b2), DICT_OPERAND_INTEGER));
+ return true;
+
+ case 29: // longint
+ if (!table.ReadU8(&b1) ||
+ !table.ReadU8(&b2) ||
+ !table.ReadU8(&b3) ||
+ !table.ReadU8(&b4)) {
+ return OTS_FAILURE();
+ }
+ operands.push_back(std::make_pair(
+ static_cast<uint32_t>((b1 << 24) + (b2 << 16) + (b3 << 8) + b4),
+ DICT_OPERAND_INTEGER));
+ return true;
+
+ case 30: // binary coded decimal
+ return ParseDictDataBcd(table, operands);
+
+ default:
+ break;
+ }
+
+ uint32_t result;
+ if (b0 >=32 && b0 <=246) {
+ result = b0 - 139;
+ } else if (b0 >=247 && b0 <= 250) {
+ if (!table.ReadU8(&b1)) {
+ return OTS_FAILURE();
+ }
+ result = (b0 - 247) * 256 + b1 + 108;
+ } else if (b0 >= 251 && b0 <= 254) {
+ if (!table.ReadU8(&b1)) {
+ return OTS_FAILURE();
+ }
+ result = -(b0 - 251) * 256 + b1 - 108;
+ } else {
+ return OTS_FAILURE();
+ }
+
+ operands.push_back(std::make_pair(result, DICT_OPERAND_INTEGER));
+ return true;
+}
+
+bool ParseDictDataReadNext(ots::Buffer &table,
+ std::vector<Operand> &operands) {
+ uint8_t op = 0;
+ if (!table.ReadU8(&op)) {
+ return OTS_FAILURE();
+ }
+ if (op <= 24) {
+ if (op == 12) {
+ return ParseDictDataEscapedOperator(table, operands);
+ }
+ operands.push_back(std::make_pair(
+ static_cast<uint32_t>(op), DICT_OPERATOR));
+ return true;
+ } else if (op <= 27 || op == 31 || op == 255) {
+ // reserved area.
+ return OTS_FAILURE();
+ }
+
+ return ParseDictDataNumber(table, op, operands);
+}
+
+bool OperandsOverflow(std::vector<Operand>& operands, bool cff2) {
+ // An operator may be preceded by up to a maximum of 48 operands in CFF1 and
+ // 513 operands in CFF2.
+ if ((cff2 && operands.size() > ots::kMaxCFF2ArgumentStack) ||
+ (!cff2 && operands.size() > ots::kMaxCFF1ArgumentStack)) {
+ return true;
+ }
+ return false;
+}
+
+bool ParseDictDataReadOperands(ots::Buffer& dict,
+ std::vector<Operand>& operands,
+ bool cff2) {
+ if (!ParseDictDataReadNext(dict, operands)) {
+ return OTS_FAILURE();
+ }
+ if (operands.empty()) {
+ return OTS_FAILURE();
+ }
+ if (OperandsOverflow(operands, cff2)) {
+ return OTS_FAILURE();
+ }
+ return true;
+}
+
+bool ValidCFF2DictOp(uint32_t op, DICT_DATA_TYPE type) {
+ if (type == DICT_DATA_TOPLEVEL) {
+ switch (op) {
+ case (12U << 8) + 7: // FontMatrix
+ case 17: // CharStrings
+ case (12U << 8) + 36: // FDArray
+ case (12U << 8) + 37: // FDSelect
+ case 24: // vstore
+ return true;
+ default:
+ return false;
+ }
+ } else if (type == DICT_DATA_FDARRAY) {
+ if (op == 18) // Private DICT
+ return true;
+ } else if (type == DICT_DATA_PRIVATE) {
+ switch (op) {
+ case (12U << 8) + 14: // ForceBold
+ case (12U << 8) + 19: // initialRandomSeed
+ case 20: // defaultWidthX
+ case 21: // nominalWidthX
+ return false;
+ default:
+ return true;
+ }
+ }
+
+ return false;
+}
+
+bool ParsePrivateDictData(
+ ots::Buffer &table, size_t offset, size_t dict_length,
+ DICT_DATA_TYPE type, ots::OpenTypeCFF *out_cff) {
+ ots::Buffer dict(table.buffer() + offset, dict_length);
+ std::vector<Operand> operands;
+ bool cff2 = (out_cff->major == 2);
+ bool blend_seen = false;
+ int32_t vsindex = 0;
+
+ // Since a Private DICT for FDArray might not have a Local Subr (e.g. Hiragino
+ // Kaku Gothic Std W8), we create an empty Local Subr here to match the size
+ // of FDArray the size of |local_subrs_per_font|.
+ // For CFF2, |vsindex_per_font| gets a similar treatment.
+ if (type == DICT_DATA_FDARRAY) {
+ out_cff->local_subrs_per_font.push_back(new ots::CFFIndex);
+ if (cff2) {
+ out_cff->vsindex_per_font.push_back(vsindex);
+ }
+ }
+
+ while (dict.offset() < dict.length()) {
+ if (!ParseDictDataReadOperands(dict, operands, cff2)) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().second != DICT_OPERATOR) {
+ continue;
+ }
+
+ // got operator
+ const uint32_t op = operands.back().first;
+ operands.pop_back();
+
+ if (cff2 && !ValidCFF2DictOp(op, DICT_DATA_PRIVATE)) {
+ return OTS_FAILURE();
+ }
+
+ bool clear_operands = true;
+ switch (op) {
+ // hints
+ case 6: // BlueValues
+ case 7: // OtherBlues
+ case 8: // FamilyBlues
+ case 9: // FamilyOtherBlues
+ if ((operands.size() % 2) != 0) {
+ return OTS_FAILURE();
+ }
+ break;
+
+ // array
+ case (12U << 8) + 12: // StemSnapH (delta)
+ case (12U << 8) + 13: // StemSnapV (delta)
+ if (operands.empty()) {
+ return OTS_FAILURE();
+ }
+ break;
+
+ // number
+ case 10: // StdHW
+ case 11: // StdVW
+ case 20: // defaultWidthX
+ case 21: // nominalWidthX
+ case (12U << 8) + 9: // BlueScale
+ case (12U << 8) + 10: // BlueShift
+ case (12U << 8) + 11: // BlueFuzz
+ case (12U << 8) + 17: // LanguageGroup
+ case (12U << 8) + 18: // ExpansionFactor
+ case (12U << 8) + 19: // initialRandomSeed
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ break;
+
+ // Local Subrs INDEX, offset(self)
+ case 19: {
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().second != DICT_OPERAND_INTEGER) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().first >= 1024 * 1024 * 1024) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().first + offset >= table.length()) {
+ return OTS_FAILURE();
+ }
+ // parse "16. Local Subrs INDEX"
+ table.set_offset(operands.back().first + offset);
+ ots::CFFIndex *local_subrs_index = NULL;
+ if (type == DICT_DATA_FDARRAY) {
+ if (out_cff->local_subrs_per_font.empty()) {
+ return OTS_FAILURE(); // not reached.
+ }
+ local_subrs_index = out_cff->local_subrs_per_font.back();
+ } else { // type == DICT_DATA_TOPLEVEL
+ if (out_cff->local_subrs) {
+ return OTS_FAILURE(); // two or more local_subrs?
+ }
+ local_subrs_index = new ots::CFFIndex;
+ out_cff->local_subrs = local_subrs_index;
+ }
+ if (!ParseIndex(table, *local_subrs_index, cff2)) {
+ return OTS_FAILURE();
+ }
+ break;
+ }
+
+ // boolean
+ case (12U << 8) + 14: // ForceBold
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().second != DICT_OPERAND_INTEGER) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().first >= 2) {
+ return OTS_FAILURE();
+ }
+ break;
+
+ case 22: { // vsindex
+ if (!cff2) {
+ return OTS_FAILURE();
+ }
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().second != DICT_OPERAND_INTEGER) {
+ return OTS_FAILURE();
+ }
+ if (blend_seen) {
+ return OTS_FAILURE();
+ }
+ vsindex = operands.back().first;
+ if (vsindex < 0 ||
+ vsindex >= (int32_t)out_cff->region_index_count.size()) {
+ return OTS_FAILURE();
+ }
+ out_cff->vsindex_per_font.back() = vsindex;
+ break;
+ }
+
+ case 23: { // blend
+ if (!cff2) {
+ return OTS_FAILURE();
+ }
+ if (operands.size() < 1) {
+ return OTS_FAILURE();
+ }
+ if (vsindex >= (int32_t)out_cff->region_index_count.size()) {
+ return OTS_FAILURE();
+ }
+ uint16_t k = out_cff->region_index_count.at(vsindex);
+ uint16_t n = operands.back().first;
+ if (operands.size() < n * (k + 1) + 1) {
+ return OTS_FAILURE();
+ }
+ size_t operands_size = operands.size();
+ // Keep the 1st n operands on the stack for the next operator to use
+ // and pop the rest. There can be multiple consecutive blend operator,
+ // so this makes sure the operands of all of them are kept on the
+ // stack.
+ while (operands.size() > operands_size - ((n * k) + 1))
+ operands.pop_back();
+ clear_operands = false;
+ blend_seen = true;
+ break;
+ }
+
+ default:
+ return OTS_FAILURE();
+ }
+ if (clear_operands) {
+ operands.clear();
+ }
+ }
+
+ return true;
+}
+
+bool ParseVariationStore(ots::OpenTypeCFF& out_cff, ots::Buffer& table) {
+ uint16_t length;
+
+ if (!table.ReadU16(&length)) {
+ return OTS_FAILURE();
+ }
+
+ // Empty VariationStore is allowed.
+ if (!length) {
+ return true;
+ }
+
+ if (length > table.remaining()) {
+ return OTS_FAILURE();
+ }
+
+ if (!ParseItemVariationStore(out_cff.GetFont(),
+ table.buffer() + table.offset(), length,
+ &(out_cff.region_index_count))) {
+ return OTS_FAILURE();
+ }
+
+ return true;
+}
+
+bool ParseDictData(ots::Buffer& table, ots::Buffer& dict,
+ uint16_t glyphs, size_t sid_max, DICT_DATA_TYPE type,
+ ots::OpenTypeCFF *out_cff);
+
+bool ParseDictData(ots::Buffer& table, const ots::CFFIndex &index,
+ uint16_t glyphs, size_t sid_max, DICT_DATA_TYPE type,
+ ots::OpenTypeCFF *out_cff) {
+ for (unsigned i = 1; i < index.offsets.size(); ++i) {
+ size_t dict_length = index.offsets[i] - index.offsets[i - 1];
+ ots::Buffer dict(table.buffer() + index.offsets[i - 1], dict_length);
+
+ if (!ParseDictData(table, dict, glyphs, sid_max, type, out_cff)) {
+ return OTS_FAILURE();
+ }
+ }
+ return true;
+}
+
+bool ParseDictData(ots::Buffer& table, ots::Buffer& dict,
+ uint16_t glyphs, size_t sid_max, DICT_DATA_TYPE type,
+ ots::OpenTypeCFF *out_cff) {
+ bool cff2 = (out_cff->major == 2);
+ std::vector<Operand> operands;
+
+ FONT_FORMAT font_format = FORMAT_UNKNOWN;
+ bool have_ros = false;
+ bool have_charstrings = false;
+ bool have_vstore = false;
+ size_t charset_offset = 0;
+ bool have_private = false;
+
+ if (cff2) {
+ // Parse VariationStore first, since it might be referenced in other places
+ // (e.g. FDArray) that might be parsed after it.
+ size_t dict_offset = dict.offset();
+ while (dict.offset() < dict.length()) {
+ if (!ParseDictDataReadOperands(dict, operands, cff2)) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().second != DICT_OPERATOR) continue;
+
+ // got operator
+ const uint32_t op = operands.back().first;
+ operands.pop_back();
+
+ if (op == 18 && type == DICT_DATA_FDARRAY) {
+ have_private = true;
+ }
+
+ if (op == 24) { // vstore
+ if (type != DICT_DATA_TOPLEVEL) {
+ return OTS_FAILURE();
+ }
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (!CheckOffset(operands.back(), table.length())) {
+ return OTS_FAILURE();
+ }
+ // parse "VariationStore Data Contents"
+ table.set_offset(operands.back().first);
+ if (!ParseVariationStore(*out_cff, table)) {
+ return OTS_FAILURE();
+ }
+ break;
+ }
+ operands.clear();
+ }
+ operands.clear();
+ dict.set_offset(dict_offset);
+
+ if (type == DICT_DATA_FDARRAY && !have_private) {
+ return OTS_FAILURE(); // CFF2 FD must have PrivateDICT entry (even if 0, 0)
+ }
+
+ }
+
+ while (dict.offset() < dict.length()) {
+ if (!ParseDictDataReadOperands(dict, operands, cff2)) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().second != DICT_OPERATOR) continue;
+
+ // got operator
+ const uint32_t op = operands.back().first;
+ operands.pop_back();
+
+ if (cff2 && !ValidCFF2DictOp(op, type)) {
+ return OTS_FAILURE();
+ }
+
+ switch (op) {
+ // SID
+ case 0: // version
+ case 1: // Notice
+ case 2: // Copyright
+ case 3: // FullName
+ case 4: // FamilyName
+ case (12U << 8) + 0: // Copyright
+ case (12U << 8) + 21: // PostScript
+ case (12U << 8) + 22: // BaseFontName
+ case (12U << 8) + 38: // FontName
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (!CheckSid(operands.back(), sid_max)) {
+ return OTS_FAILURE();
+ }
+ break;
+
+ // array
+ case 5: // FontBBox
+ case 14: // XUID
+ case (12U << 8) + 7: // FontMatrix
+ case (12U << 8) + 23: // BaseFontBlend (delta)
+ if (operands.empty()) {
+ return OTS_FAILURE();
+ }
+ break;
+
+ // number
+ case 13: // UniqueID
+ case (12U << 8) + 2: // ItalicAngle
+ case (12U << 8) + 3: // UnderlinePosition
+ case (12U << 8) + 4: // UnderlineThickness
+ case (12U << 8) + 5: // PaintType
+ case (12U << 8) + 8: // StrokeWidth
+ case (12U << 8) + 20: // SyntheticBase
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ break;
+ case (12U << 8) + 31: // CIDFontVersion
+ case (12U << 8) + 32: // CIDFontRevision
+ case (12U << 8) + 33: // CIDFontType
+ case (12U << 8) + 34: // CIDCount
+ case (12U << 8) + 35: // UIDBase
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (font_format != FORMAT_CID_KEYED) {
+ return OTS_FAILURE();
+ }
+ break;
+ case (12U << 8) + 6: // CharstringType
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if(operands.back().second != DICT_OPERAND_INTEGER) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().first != 2) {
+ // We only support the "Type 2 Charstring Format."
+ // TODO(yusukes): Support Type 1 format? Is that still in use?
+ return OTS_FAILURE();
+ }
+ break;
+
+ // boolean
+ case (12U << 8) + 1: // isFixedPitch
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().second != DICT_OPERAND_INTEGER) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().first >= 2) {
+ return OTS_FAILURE();
+ }
+ break;
+
+ // offset(0)
+ case 15: // charset
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().first <= 2) {
+ // predefined charset, ISOAdobe, Expert or ExpertSubset, is used.
+ break;
+ }
+ if (!CheckOffset(operands.back(), table.length())) {
+ return OTS_FAILURE();
+ }
+ if (charset_offset) {
+ return OTS_FAILURE(); // multiple charset tables?
+ }
+ charset_offset = operands.back().first;
+ break;
+
+ case 16: { // Encoding
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().first <= 1) {
+ break; // predefined encoding, "Standard" or "Expert", is used.
+ }
+ if (!CheckOffset(operands.back(), table.length())) {
+ return OTS_FAILURE();
+ }
+
+ table.set_offset(operands.back().first);
+ uint8_t format = 0;
+ if (!table.ReadU8(&format)) {
+ return OTS_FAILURE();
+ }
+ if (format & 0x80) {
+ // supplemental encoding is not supported at the moment.
+ return OTS_FAILURE();
+ }
+ // TODO(yusukes): support & parse supplemental encoding tables.
+ break;
+ }
+
+ case 17: { // CharStrings
+ if (type != DICT_DATA_TOPLEVEL) {
+ return OTS_FAILURE();
+ }
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (!CheckOffset(operands.back(), table.length())) {
+ return OTS_FAILURE();
+ }
+ // parse "14. CharStrings INDEX"
+ table.set_offset(operands.back().first);
+ ots::CFFIndex *charstring_index = out_cff->charstrings_index;
+ if (!ParseIndex(table, *charstring_index, cff2)) {
+ return OTS_FAILURE();
+ }
+ if (charstring_index->count < 2) {
+ return OTS_FAILURE();
+ }
+ if (have_charstrings) {
+ return OTS_FAILURE(); // multiple charstring tables?
+ }
+ have_charstrings = true;
+ if (charstring_index->count != glyphs) {
+ return OTS_FAILURE(); // CFF and maxp have different number of glyphs?
+ }
+ break;
+ }
+
+ case 24: { // vstore
+ if (!cff2) {
+ return OTS_FAILURE();
+ }
+ if (have_vstore) {
+ return OTS_FAILURE(); // multiple vstore tables?
+ }
+ have_vstore = true;
+ // parsed above.
+ break;
+ }
+
+ case (12U << 8) + 36: { // FDArray
+ if (type != DICT_DATA_TOPLEVEL) {
+ return OTS_FAILURE();
+ }
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (!CheckOffset(operands.back(), table.length())) {
+ return OTS_FAILURE();
+ }
+
+ // parse Font DICT INDEX.
+ table.set_offset(operands.back().first);
+ ots::CFFIndex sub_dict_index;
+ if (!ParseIndex(table, sub_dict_index, cff2)) {
+ return OTS_FAILURE();
+ }
+ if (!ParseDictData(table, sub_dict_index,
+ glyphs, sid_max, DICT_DATA_FDARRAY,
+ out_cff)) {
+ return OTS_FAILURE();
+ }
+ if (out_cff->font_dict_length != 0) {
+ return OTS_FAILURE(); // two or more FDArray found.
+ }
+ out_cff->font_dict_length = sub_dict_index.count;
+ break;
+ }
+
+ case (12U << 8) + 37: { // FDSelect
+ if (type != DICT_DATA_TOPLEVEL) {
+ return OTS_FAILURE();
+ }
+ if (operands.size() != 1) {
+ return OTS_FAILURE();
+ }
+ if (!CheckOffset(operands.back(), table.length())) {
+ return OTS_FAILURE();
+ }
+
+ // parse FDSelect data structure
+ table.set_offset(operands.back().first);
+ uint8_t format = 0;
+ if (!table.ReadU8(&format)) {
+ return OTS_FAILURE();
+ }
+ if (format == 0) {
+ for (uint16_t j = 0; j < glyphs; ++j) {
+ uint8_t fd_index = 0;
+ if (!table.ReadU8(&fd_index)) {
+ return OTS_FAILURE();
+ }
+ (out_cff->fd_select)[j] = fd_index;
+ }
+ } else if (format == 3) {
+ uint16_t n_ranges = 0;
+ if (!table.ReadU16(&n_ranges)) {
+ return OTS_FAILURE();
+ }
+ if (n_ranges == 0) {
+ return OTS_FAILURE();
+ }
+
+ uint16_t last_gid = 0;
+ uint8_t fd_index = 0;
+ for (unsigned j = 0; j < n_ranges; ++j) {
+ uint16_t first = 0; // GID
+ if (!table.ReadU16(&first)) {
+ return OTS_FAILURE();
+ }
+
+ // Sanity checks.
+ if ((j == 0) && (first != 0)) {
+ return OTS_FAILURE();
+ }
+ if ((j != 0) && (last_gid >= first)) {
+ return OTS_FAILURE(); // not increasing order.
+ }
+ if (first >= glyphs) {
+ return OTS_FAILURE(); // invalid gid.
+ }
+
+ // Copy the mapping to |out_cff->fd_select|.
+ if (j != 0) {
+ for (auto k = last_gid; k < first; ++k) {
+ if (!out_cff->fd_select.insert(
+ std::make_pair(k, fd_index)).second) {
+ return OTS_FAILURE();
+ }
+ }
+ }
+
+ if (!table.ReadU8(&fd_index)) {
+ return OTS_FAILURE();
+ }
+ last_gid = first;
+ }
+ uint16_t sentinel = 0;
+ if (!table.ReadU16(&sentinel)) {
+ return OTS_FAILURE();
+ }
+ if (last_gid >= sentinel) {
+ return OTS_FAILURE();
+ }
+ if (sentinel > glyphs) {
+ return OTS_FAILURE(); // invalid gid.
+ }
+ for (auto k = last_gid; k < sentinel; ++k) {
+ if (!out_cff->fd_select.insert(
+ std::make_pair(k, fd_index)).second) {
+ return OTS_FAILURE();
+ }
+ }
+ } else if (cff2 && format == 4) {
+ uint32_t n_ranges = 0;
+ if (!table.ReadU32(&n_ranges)) {
+ return OTS_FAILURE();
+ }
+ if (n_ranges == 0) {
+ return OTS_FAILURE();
+ }
+
+ uint32_t last_gid = 0;
+ uint16_t fd_index = 0;
+ for (unsigned j = 0; j < n_ranges; ++j) {
+ uint32_t first = 0; // GID
+ if (!table.ReadU32(&first)) {
+ return OTS_FAILURE();
+ }
+
+ // Sanity checks.
+ if ((j == 0) && (first != 0)) {
+ return OTS_FAILURE();
+ }
+ if ((j != 0) && (last_gid >= first)) {
+ return OTS_FAILURE(); // not increasing order.
+ }
+ if (first >= glyphs) {
+ return OTS_FAILURE(); // invalid gid.
+ }
+
+ // Copy the mapping to |out_cff->fd_select|.
+ if (j != 0) {
+ for (auto k = last_gid; k < first; ++k) {
+ if (!out_cff->fd_select.insert(
+ std::make_pair(k, fd_index)).second) {
+ return OTS_FAILURE();
+ }
+ }
+ }
+
+ if (!table.ReadU16(&fd_index)) {
+ return OTS_FAILURE();
+ }
+ last_gid = first;
+ }
+ uint32_t sentinel = 0;
+ if (!table.ReadU32(&sentinel)) {
+ return OTS_FAILURE();
+ }
+ if (last_gid >= sentinel) {
+ return OTS_FAILURE();
+ }
+ if (sentinel > glyphs) {
+ return OTS_FAILURE(); // invalid gid.
+ }
+ for (auto k = last_gid; k < sentinel; ++k) {
+ if (!out_cff->fd_select.insert(
+ std::make_pair(k, fd_index)).second) {
+ return OTS_FAILURE();
+ }
+ }
+ } else {
+ // unknown format
+ return OTS_FAILURE();
+ }
+ break;
+ }
+
+ // Private DICT (2 * number)
+ case 18: {
+ if (operands.size() != 2) {
+ return OTS_FAILURE();
+ }
+ if (operands.back().second != DICT_OPERAND_INTEGER) {
+ return OTS_FAILURE();
+ }
+ const uint32_t private_offset = operands.back().first;
+ operands.pop_back();
+ if (operands.back().second != DICT_OPERAND_INTEGER) {
+ return OTS_FAILURE();
+ }
+ const uint32_t private_length = operands.back().first;
+ if (private_offset > table.length()) {
+ return OTS_FAILURE();
+ }
+ if (private_length >= table.length()) {
+ return OTS_FAILURE();
+ }
+ if (private_length + private_offset > table.length()) {
+ return OTS_FAILURE();
+ }
+ // parse "15. Private DICT data"
+ if (!ParsePrivateDictData(table, private_offset, private_length,
+ type, out_cff)) {
+ return OTS_FAILURE();
+ }
+ break;
+ }
+
+ // ROS
+ case (12U << 8) + 30:
+ if (font_format != FORMAT_UNKNOWN) {
+ return OTS_FAILURE();
+ }
+ font_format = FORMAT_CID_KEYED;
+ if (operands.size() != 3) {
+ return OTS_FAILURE();
+ }
+ // check SIDs
+ operands.pop_back(); // ignore the first number.
+ if (!CheckSid(operands.back(), sid_max)) {
+ return OTS_FAILURE();
+ }
+ operands.pop_back();
+ if (!CheckSid(operands.back(), sid_max)) {
+ return OTS_FAILURE();
+ }
+ if (have_ros) {
+ return OTS_FAILURE(); // multiple ROS tables?
+ }
+ have_ros = true;
+ break;
+
+ default:
+ return OTS_FAILURE();
+ }
+ operands.clear();
+
+ if (font_format == FORMAT_UNKNOWN) {
+ font_format = FORMAT_OTHER;
+ }
+ }
+
+ // parse "13. Charsets"
+ if (charset_offset) {
+ table.set_offset(charset_offset);
+ uint8_t format = 0;
+ if (!table.ReadU8(&format)) {
+ return OTS_FAILURE();
+ }
+ switch (format) {
+ case 0:
+ for (uint16_t j = 1 /* .notdef is omitted */; j < glyphs; ++j) {
+ uint16_t sid = 0;
+ if (!table.ReadU16(&sid)) {
+ return OTS_FAILURE();
+ }
+ if (!have_ros && (sid > sid_max)) {
+ return OTS_FAILURE();
+ }
+ // TODO(yusukes): check CIDs when have_ros is true.
+ }
+ break;
+
+ case 1:
+ case 2: {
+ uint32_t total = 1; // .notdef is omitted.
+ while (total < glyphs) {
+ uint16_t sid = 0;
+ if (!table.ReadU16(&sid)) {
+ return OTS_FAILURE();
+ }
+ if (!have_ros && (sid > sid_max)) {
+ return OTS_FAILURE();
+ }
+ // TODO(yusukes): check CIDs when have_ros is true.
+
+ if (format == 1) {
+ uint8_t left = 0;
+ if (!table.ReadU8(&left)) {
+ return OTS_FAILURE();
+ }
+ total += (left + 1);
+ } else {
+ uint16_t left = 0;
+ if (!table.ReadU16(&left)) {
+ return OTS_FAILURE();
+ }
+ total += (left + 1);
+ }
+ }
+ break;
+ }
+
+ default:
+ return OTS_FAILURE();
+ }
+ }
+ return true;
+}
+
+} // namespace
+
+namespace ots {
+
+bool OpenTypeCFF::ValidateFDSelect(uint16_t num_glyphs) {
+ for (const auto& fd_select : this->fd_select) {
+ if (fd_select.first >= num_glyphs) {
+ return Error("Invalid glyph index in FDSelect: %d >= %d\n",
+ fd_select.first, num_glyphs);
+ }
+ if (fd_select.second >= this->font_dict_length) {
+ return Error("Invalid FD index: %d >= %d\n",
+ fd_select.second, this->font_dict_length);
+ }
+ }
+ return true;
+}
+
+bool OpenTypeCFF::Parse(const uint8_t *data, size_t length) {
+ Buffer table(data, length);
+
+ Font *font = GetFont();
+
+ this->m_data = data;
+ this->m_length = length;
+
+ // parse "6. Header" in the Adobe Compact Font Format Specification
+ uint8_t major = 0;
+ uint8_t minor = 0;
+ uint8_t hdr_size = 0;
+ uint8_t off_size = 0;
+ if (!table.ReadU8(&major) ||
+ !table.ReadU8(&minor) ||
+ !table.ReadU8(&hdr_size) ||
+ !table.ReadU8(&off_size)) {
+ return Error("Failed to read table header");
+ }
+
+ if (off_size < 1 || off_size > 4) {
+ return Error("Bad offSize: %d", off_size);
+ }
+
+ if (major != 1 || minor != 0) {
+ return Error("Unsupported table version: %d.%d", major, minor);
+ }
+
+ this->major = major;
+
+ if (hdr_size != 4 || hdr_size >= length) {
+ return Error("Bad hdrSize: %d", hdr_size);
+ }
+
+ // parse "7. Name INDEX"
+ table.set_offset(hdr_size);
+ CFFIndex name_index;
+ if (!ParseIndex(table, name_index)) {
+ return Error("Failed to parse Name INDEX");
+ }
+ if (name_index.count != 1 || name_index.offsets.size() != 2) {
+ return Error("Name INDEX must contain only one entry, not %d",
+ name_index.count);
+ }
+ if (!ParseNameData(&table, name_index, &(this->name))) {
+ return Error("Failed to parse Name INDEX data");
+ }
+
+ // parse "8. Top DICT INDEX"
+ table.set_offset(name_index.offset_to_next);
+ CFFIndex top_dict_index;
+ if (!ParseIndex(table, top_dict_index)) {
+ return Error("Failed to parse Top DICT INDEX");
+ }
+ if (top_dict_index.count != 1) {
+ return Error("Top DICT INDEX must contain only one entry, not %d",
+ top_dict_index.count);
+ }
+
+ // parse "10. String INDEX"
+ table.set_offset(top_dict_index.offset_to_next);
+ CFFIndex string_index;
+ if (!ParseIndex(table, string_index)) {
+ return Error("Failed to parse String INDEX");
+ }
+ if (string_index.count >= 65000 - kNStdString) {
+ return Error("Too many entries in String INDEX: %d", string_index.count);
+ }
+
+ OpenTypeMAXP *maxp = static_cast<OpenTypeMAXP*>(
+ font->GetTypedTable(OTS_TAG_MAXP));
+ if (!maxp) {
+ return Error("Required maxp table missing");
+ }
+ const uint16_t num_glyphs = maxp->num_glyphs;
+ const size_t sid_max = string_index.count + kNStdString;
+ // string_index.count == 0 is allowed.
+
+ // parse "9. Top DICT Data"
+ this->charstrings_index = new ots::CFFIndex;
+ if (!ParseDictData(table, top_dict_index,
+ num_glyphs, sid_max,
+ DICT_DATA_TOPLEVEL, this)) {
+ return Error("Failed to parse Top DICT Data");
+ }
+
+ // parse "16. Global Subrs INDEX"
+ table.set_offset(string_index.offset_to_next);
+ CFFIndex global_subrs_index;
+ if (!ParseIndex(table, global_subrs_index)) {
+ return Error("Failed to parse Global Subrs INDEX");
+ }
+
+ // Check if all fd and glyph indices in FDSelect are valid.
+ if (!ValidateFDSelect(num_glyphs)) {
+ return Error("Failed to validate FDSelect");
+ }
+
+ // Check if all charstrings (font hinting code for each glyph) are valid.
+ if (!ValidateCFFCharStrings(*this, global_subrs_index, &table)) {
+ return Error("Failed validating CharStrings INDEX");
+ }
+
+ return true;
+}
+
+bool OpenTypeCFF::Serialize(OTSStream *out) {
+ if (!out->Write(this->m_data, this->m_length)) {
+ return Error("Failed to write table");
+ }
+ return true;
+}
+
+OpenTypeCFF::~OpenTypeCFF() {
+ for (size_t i = 0; i < this->local_subrs_per_font.size(); ++i) {
+ delete (this->local_subrs_per_font)[i];
+ }
+ delete this->charstrings_index;
+ delete this->local_subrs;
+}
+
+bool OpenTypeCFF2::Parse(const uint8_t *data, size_t length) {
+ Buffer table(data, length);
+
+ Font *font = GetFont();
+
+ this->m_data = data;
+ this->m_length = length;
+
+ // parse "6. Header"
+ uint8_t major = 0;
+ uint8_t minor = 0;
+ uint8_t hdr_size = 0;
+ uint16_t top_dict_size = 0;
+ if (!table.ReadU8(&major) ||
+ !table.ReadU8(&minor) ||
+ !table.ReadU8(&hdr_size) ||
+ !table.ReadU16(&top_dict_size)) {
+ return Error("Failed to read table header");
+ }
+
+ if (major != 2 || minor != 0) {
+ return Error("Unsupported table version: %d.%d", major, minor);
+ }
+
+ this->major = major;
+
+ if (hdr_size >= length) {
+ return Error("Bad hdrSize: %d", hdr_size);
+ }
+
+ if (top_dict_size == 0 || hdr_size + top_dict_size > length) {
+ return Error("Bad topDictLength: %d", top_dict_size);
+ }
+
+ OpenTypeMAXP *maxp = static_cast<OpenTypeMAXP*>(
+ font->GetTypedTable(OTS_TAG_MAXP));
+ if (!maxp) {
+ return Error("Required maxp table missing");
+ }
+ const uint16_t num_glyphs = maxp->num_glyphs;
+ const size_t sid_max = kNStdString;
+
+ // parse "7. Top DICT Data"
+ ots::Buffer top_dict(data + hdr_size, top_dict_size);
+ table.set_offset(hdr_size);
+ this->charstrings_index = new ots::CFFIndex;
+ if (!ParseDictData(table, top_dict,
+ num_glyphs, sid_max,
+ DICT_DATA_TOPLEVEL, this)) {
+ return Error("Failed to parse Top DICT Data");
+ }
+
+ // parse "9. Global Subrs INDEX"
+ table.set_offset(hdr_size + top_dict_size);
+ CFFIndex global_subrs_index;
+ if (!ParseIndex(table, global_subrs_index, true)) {
+ return Error("Failed to parse Global Subrs INDEX");
+ }
+
+ // Check if all fd and glyph indices in FDSelect are valid.
+ if (!ValidateFDSelect(num_glyphs)) {
+ return Error("Failed to validate FDSelect");
+ }
+
+ // Check if all charstrings (font hinting code for each glyph) are valid.
+ if (!ValidateCFFCharStrings(*this, global_subrs_index, &table)) {
+ return Error("Failed validating CharStrings INDEX");
+ }
+
+ return true;
+}
+
+bool OpenTypeCFF2::Serialize(OTSStream *out) {
+ if (!out->Write(this->m_data, this->m_length)) {
+ return Error("Failed to write table");
+ }
+ return true;
+}
+
+} // namespace ots
+
+#undef TABLE_NAME
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-01 00:39:56 +0000