1 // Copyright 2014 The PDFium Authors
2 // Use of this source code is governed by a BSD-style license that can be
3 // found in the LICENSE file.
4
5 // Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com
6
7 #include "core/fpdfapi/cmaps/fpdf_cmaps.h"
8
9 #include <algorithm>
10
11 #include "third_party/base/check.h"
12
13 namespace fxcmap {
14
15 namespace {
16
17 struct SingleCmap {
18 uint16_t code;
19 uint16_t cid;
20 };
21
22 struct RangeCmap {
23 uint16_t low;
24 uint16_t high;
25 uint16_t cid;
26 };
27
FindNextCMap(const CMap * pMap)28 const CMap* FindNextCMap(const CMap* pMap) {
29 return pMap->m_UseOffset ? pMap + pMap->m_UseOffset : nullptr;
30 }
31
32 } // namespace
33
CIDFromCharCode(const CMap * pMap,uint32_t charcode)34 uint16_t CIDFromCharCode(const CMap* pMap, uint32_t charcode) {
35 DCHECK(pMap);
36 const uint16_t loword = static_cast<uint16_t>(charcode);
37 if (charcode >> 16) {
38 while (pMap) {
39 if (pMap->m_pDWordMap) {
40 const DWordCIDMap* begin = pMap->m_pDWordMap;
41 const auto* end = begin + pMap->m_DWordCount;
42 const auto* found = std::lower_bound(
43 begin, end, charcode,
44 [](const DWordCIDMap& element, uint32_t charcode) {
45 uint16_t hiword = static_cast<uint16_t>(charcode >> 16);
46 if (element.m_HiWord != hiword)
47 return element.m_HiWord < hiword;
48 return element.m_LoWordHigh < static_cast<uint16_t>(charcode);
49 });
50 if (found != end && loword >= found->m_LoWordLow &&
51 loword <= found->m_LoWordHigh) {
52 return found->m_CID + loword - found->m_LoWordLow;
53 }
54 }
55 pMap = FindNextCMap(pMap);
56 }
57 return 0;
58 }
59
60 while (pMap && pMap->m_pWordMap) {
61 switch (pMap->m_WordMapType) {
62 case CMap::Type::kSingle: {
63 const auto* begin =
64 reinterpret_cast<const SingleCmap*>(pMap->m_pWordMap);
65 const auto* end = begin + pMap->m_WordCount;
66 const auto* found = std::lower_bound(
67 begin, end, loword, [](const SingleCmap& element, uint16_t code) {
68 return element.code < code;
69 });
70 if (found != end && found->code == loword)
71 return found->cid;
72 break;
73 }
74 case CMap::Type::kRange: {
75 const auto* begin =
76 reinterpret_cast<const RangeCmap*>(pMap->m_pWordMap);
77 const auto* end = begin + pMap->m_WordCount;
78 const auto* found = std::lower_bound(
79 begin, end, loword, [](const RangeCmap& element, uint16_t code) {
80 return element.high < code;
81 });
82 if (found != end && loword >= found->low && loword <= found->high)
83 return found->cid + loword - found->low;
84 break;
85 }
86 }
87 pMap = FindNextCMap(pMap);
88 }
89
90 return 0;
91 }
92
CharCodeFromCID(const CMap * pMap,uint16_t cid)93 uint32_t CharCodeFromCID(const CMap* pMap, uint16_t cid) {
94 // TODO(dsinclair): This should be checking both pMap->m_WordMap and
95 // pMap->m_DWordMap. There was a second while() but it was never reached as
96 // the first always returns. Investigate and determine how this should
97 // really be working. (https://codereview.chromium.org/2235743003 removed the
98 // second while loop.)
99 DCHECK(pMap);
100 while (pMap) {
101 switch (pMap->m_WordMapType) {
102 case CMap::Type::kSingle: {
103 const auto* pCur =
104 reinterpret_cast<const SingleCmap*>(pMap->m_pWordMap);
105 const auto* pEnd = pCur + pMap->m_WordCount;
106 while (pCur < pEnd) {
107 if (pCur->cid == cid)
108 return pCur->code;
109 ++pCur;
110 }
111 break;
112 }
113 case CMap::Type::kRange: {
114 const auto* pCur = reinterpret_cast<const RangeCmap*>(pMap->m_pWordMap);
115 const auto* pEnd = pCur + pMap->m_WordCount;
116 while (pCur < pEnd) {
117 if (cid >= pCur->cid && cid <= pCur->cid + pCur->high - pCur->low)
118 return pCur->low + cid - pCur->cid;
119 ++pCur;
120 }
121 break;
122 }
123 }
124 pMap = FindNextCMap(pMap);
125 }
126 return 0;
127 }
128
129 } // namespace fxcmap
130