A fork of https://github.com/crosspoint-reader/crosspoint-reader
refactor: Use std binary search algorithms for font lookups (#1202)
## Summary
**What is the goal of this PR?**
Rewrite of font routines to use std binary search algorithms instead of
custom repeated implementations: `lookupKernClass`,
`EpdFont::getLigature`, and `EpdFont::getGlyph`.
---
### AI Usage
While CrossPoint doesn't have restrictions on AI tools in contributing,
please be transparent about their usage as it
helps set the right context for reviewers.
Did you use AI tools to help write this code? _**NO**_
authored by
Zach Nelson
and committed by
GitHub
3cc8e272
80d18563
+32
-42
+32
-42
lib/EpdFont/EpdFont.cpp
+32
-42
lib/EpdFont/EpdFont.cpp
···
80
80
if (!entries || count == 0 || cp > 0xFFFF) {
81
81
return 0;
82
82
}
83
+
83
84
const auto target = static_cast<uint16_t>(cp);
84
-
int left = 0;
85
-
int right = static_cast<int>(count) - 1;
86
-
while (left <= right) {
87
-
const int mid = left + (right - left) / 2;
88
-
const uint16_t midCp = entries[mid].codepoint;
89
-
if (midCp == target) {
90
-
return entries[mid].classId;
91
-
}
92
-
if (midCp < target) {
93
-
left = mid + 1;
94
-
} else {
95
-
right = mid - 1;
96
-
}
85
+
const auto* end = entries + count;
86
+
87
+
// lower_bound: exact-key lookup. Finds the first entry with codepoint >= target,
88
+
// then the equality check confirms an exact match exists.
89
+
const auto it = std::lower_bound(
90
+
entries, end, target, [](const EpdKernClassEntry& entry, uint16_t value) { return entry.codepoint < value; });
91
+
92
+
if (it != end && it->codepoint == target) {
93
+
return it->classId;
97
94
}
95
+
98
96
return 0;
99
97
}
100
98
···
117
115
}
118
116
119
117
const uint32_t key = (leftCp << 16) | rightCp;
120
-
int left = 0;
121
-
int right = static_cast<int>(count) - 1;
118
+
const auto* end = pairs + count;
122
119
123
-
while (left <= right) {
124
-
const int mid = left + (right - left) / 2;
125
-
const uint32_t midKey = pairs[mid].pair;
126
-
if (midKey == key) {
127
-
return pairs[mid].ligatureCp;
128
-
}
129
-
if (midKey < key) {
130
-
left = mid + 1;
131
-
} else {
132
-
right = mid - 1;
133
-
}
120
+
// lower_bound: exact-key lookup. Finds the first entry with pair >= key,
121
+
// then the equality check confirms an exact match exists.
122
+
const auto it =
123
+
std::lower_bound(pairs, end, key, [](const EpdLigaturePair& pair, uint32_t value) { return pair.pair < value; });
124
+
125
+
if (it != end && it->pair == key) {
126
+
return it->ligatureCp;
134
127
}
128
+
135
129
return 0;
136
130
}
137
131
···
154
148
}
155
149
156
150
const EpdGlyph* EpdFont::getGlyph(const uint32_t cp) const {
157
-
const EpdUnicodeInterval* intervals = data->intervals;
158
151
const int count = data->intervalCount;
159
-
160
152
if (count == 0) return nullptr;
161
153
162
-
// Binary search for O(log n) lookup instead of O(n)
163
-
// Critical for Korean fonts with many unicode intervals
164
-
int left = 0;
165
-
int right = count - 1;
154
+
const EpdUnicodeInterval* intervals = data->intervals;
155
+
const auto* end = intervals + count;
166
156
167
-
while (left <= right) {
168
-
const int mid = left + (right - left) / 2;
169
-
const EpdUnicodeInterval* interval = &intervals[mid];
157
+
// upper_bound: range lookup. Finds the first interval with first > cp, so the
158
+
// interval just before it is the last one with first <= cp. That's the only
159
+
// candidate that could contain cp. Then we verify cp <= candidate.last.
160
+
const auto it = std::upper_bound(
161
+
intervals, end, cp, [](uint32_t value, const EpdUnicodeInterval& interval) { return value < interval.first; });
170
162
171
-
if (cp < interval->first) {
172
-
right = mid - 1;
173
-
} else if (cp > interval->last) {
174
-
left = mid + 1;
175
-
} else {
176
-
// Found: cp >= interval->first && cp <= interval->last
177
-
return &data->glyph[interval->offset + (cp - interval->first)];
163
+
if (it != intervals) {
164
+
const auto& interval = *(it - 1);
165
+
if (cp <= interval.last) {
166
+
return &data->glyph[interval.offset + (cp - interval.first)];
178
167
}
179
168
}
169
+
180
170
if (cp != REPLACEMENT_GLYPH) {
181
171
return getGlyph(REPLACEMENT_GLYPH);
182
172
}