Skip to content

Commit 0da8733

Browse files
adazem009adazem009
authored
Merge pull request #626 from scratchcpp/refactor_strings
Refactor strings
2 parents c3431c6 + 2a3b7c8 commit 0da8733

Some content is hidden

Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.

49 files changed

+73002
-761
lines changed

‎CMakeLists.txt

+7
Original file line numberDiff line numberDiff line change
@@ -38,6 +38,9 @@ target_sources(scratchcpp
3838
include/scratchcpp/value.h
3939
include/scratchcpp/valuedata.h
4040
include/scratchcpp/value_functions.h
41+
include/scratchcpp/stringptr.h
42+
include/scratchcpp/string_functions.h
43+
include/scratchcpp/string_pool.h
4144
include/scratchcpp/entity.h
4245
include/scratchcpp/variable.h
4346
include/scratchcpp/list.h
@@ -77,6 +80,10 @@ target_link_libraries(scratchcpp PRIVATE zip)
7780
set(UTFCPP_SRC thirdparty/utfcpp/source)
7881
target_include_directories(scratchcpp PUBLIC ${UTFCPP_SRC})
7982

83+
# cpp-unicodelib
84+
add_subdirectory(thirdparty/cpp-unicodelib)
85+
target_link_libraries(scratchcpp PRIVATE cpp-unicodelib)
86+
8087
# spimpl
8188
include_directories(thirdparty/spimpl)
8289

‎include/scratchcpp/compiler.h

+4
Original file line numberDiff line numberDiff line change
@@ -56,6 +56,8 @@ class LIBSCRATCHCPP_EXPORT Compiler
5656
CompilerValue *addTargetFunctionCall(const std::string &functionName, StaticType returnType = StaticType::Void, const ArgTypes &argTypes = {}, const Args &args = {});
5757
CompilerValue *addFunctionCallWithCtx(const std::string &functionName, StaticType returnType = StaticType::Void, const ArgTypes &argTypes = {}, const Args &args = {});
5858
CompilerConstant *addConstValue(const Value &value);
59+
CompilerValue *addStringChar(CompilerValue *string, CompilerValue *index);
60+
CompilerValue *addStringLength(CompilerValue *string);
5961
CompilerValue *addLoopIndex();
6062
CompilerValue *addLocalVariableValue(CompilerLocalVariable *variable);
6163
CompilerValue *addVariableValue(Variable *variable);
@@ -104,6 +106,8 @@ class LIBSCRATCHCPP_EXPORT Compiler
104106
CompilerValue *createExp(CompilerValue *num);
105107
CompilerValue *createExp10(CompilerValue *num);
106108

109+
CompilerValue *createStringConcat(CompilerValue *string1, CompilerValue *string2);
110+
107111
CompilerValue *createSelect(CompilerValue *cond, CompilerValue *trueValue, CompilerValue *falseValue, Compiler::StaticType valueType);
108112

109113
CompilerLocalVariable *createLocalVariable(Compiler::StaticType type);

‎include/scratchcpp/list.h

+43-24
Original file line numberDiff line numberDiff line change
@@ -3,10 +3,13 @@
33
#pragma once
44

55
#include <string>
6-
#include <scratchcpp/veque.h>
76

87
#include "value.h"
98
#include "entity.h"
9+
#include "string_pool.h"
10+
#include "string_functions.h"
11+
#include "stringptr.h"
12+
#include "veque.h"
1013

1114
namespace libscratchcpp
1215
{
@@ -143,21 +146,21 @@ class LIBSCRATCHCPP_EXPORT List : public Entity
143146
return m_dataPtr->operator[](index);
144147
}
145148

146-
/*! Joins the list items with spaces or without any separator if there are only digits and stores the result in dst. */
147-
inline void toString(std::string &dst) const
149+
/*! Joins the list items with spaces or without any separator if there are only digits and returns the result as StringPtr. */
150+
inline StringPtr *toStringPtr() const
148151
{
149-
dst.clear();
150-
veque::veque<std::string> strings;
152+
veque::veque<StringPtr *> strings;
153+
size_t size = 0;
151154
strings.reserve(m_size);
152155
bool digits = true;
153156
size_t i;
154157

155158
for (i = 0; i < m_size; i++) {
156159
const ValueData *item = &m_dataPtr->operator[](i);
157-
strings.push_back(std::string());
158-
value_toString(item, &strings.back());
160+
strings.push_back(value_toStringPtr(item));
161+
size += strings.back()->size;
159162

160-
if (value_isValidNumber(item) && !value_isBool(item) && !strings.back().empty()) {
163+
if (value_isValidNumber(item) && !value_isBool(item) && strings.back()->size > 0) {
161164
double doubleNum = value_toDouble(item);
162165
long num = value_toLong(item);
163166

@@ -176,42 +179,58 @@ class LIBSCRATCHCPP_EXPORT List : public Entity
176179
}
177180
}
178181

179-
std::string s;
182+
StringPtr *ret = string_pool_new();
183+
ret->size = 0;
180184

181185
if (digits) {
182-
for (i = 0; i < strings.size(); i++)
183-
dst.append(strings[i]);
186+
string_alloc(ret, size);
187+
188+
for (i = 0; i < strings.size(); i++) {
189+
memcpy(ret->data + ret->size, strings[i]->data, strings[i]->size * sizeof(char16_t));
190+
ret->size += strings[i]->size;
191+
string_pool_free(strings[i]);
192+
}
184193

185194
for (; i < m_size; i++) {
186-
value_toString(&m_dataPtr->operator[](i), &s);
187-
dst.append(s);
195+
StringPtr *item = value_toStringPtr(&m_dataPtr->operator[](i));
196+
size += item->size + 1;
197+
string_alloc(ret, size);
198+
memcpy(ret->data + ret->size, item->data, item->size * sizeof(char16_t));
199+
ret->size += item->size;
200+
string_pool_free(item);
188201
}
189202
} else {
203+
size += strings.size() - 1;
204+
string_alloc(ret, size);
205+
190206
for (i = 0; i < strings.size(); i++) {
191-
dst.append(strings[i]);
207+
memcpy(ret->data + ret->size, strings[i]->data, strings[i]->size * sizeof(char16_t));
208+
ret->size += strings[i]->size;
209+
string_pool_free(strings[i]);
192210

193211
if (i + 1 < m_size)
194-
dst.push_back(' ');
212+
ret->data[ret->size++] = u' ';
195213
}
196214

197215
for (; i < m_size; i++) {
198-
value_toString(&m_dataPtr->operator[](i), &s);
199-
dst.append(s);
216+
StringPtr *item = value_toStringPtr(&m_dataPtr->operator[](i));
217+
size += item->size + 1;
218+
string_alloc(ret, size);
219+
memcpy(ret->data + ret->size, item->data, item->size * sizeof(char16_t));
220+
ret->size += item->size;
221+
string_pool_free(item);
200222

201223
if (i + 1 < m_size)
202-
dst.push_back(' ');
224+
ret->data[ret->size++] = u' ';
203225
}
204226
}
205-
}
206227

207-
/*! Same as the other method, but returns the result directly. */
208-
inline std::string toString() const
209-
{
210-
std::string ret;
211-
toString(ret);
228+
ret->data[ret->size] = u'\0';
212229
return ret;
213230
}
214231

232+
std::string toString() const;
233+
215234
std::shared_ptr<List> clone();
216235

217236
private:

‎include/scratchcpp/string_functions.h

+34
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,34 @@
1+
// SPDX-License-Identifier: Apache-2.0
2+
3+
#pragma once
4+
5+
#include <cstddef>
6+
7+
#include "global.h"
8+
9+
namespace libscratchcpp
10+
{
11+
12+
struct StringPtr;
13+
14+
extern "C"
15+
{
16+
LIBSCRATCHCPP_EXPORT void string_alloc(StringPtr *str, size_t size);
17+
18+
LIBSCRATCHCPP_EXPORT void string_assign(StringPtr *str, const StringPtr *another);
19+
LIBSCRATCHCPP_EXPORT void string_assign_cstring(StringPtr *str, const char *another);
20+
21+
LIBSCRATCHCPP_EXPORT int string_compare_raw_case_sensitive(const char16_t *str1, size_t n1, const char16_t *str2, size_t n2);
22+
LIBSCRATCHCPP_EXPORT int string_compare_case_sensitive(const StringPtr *str1, const StringPtr *str2);
23+
24+
LIBSCRATCHCPP_EXPORT int string_compare_raw_case_insensitive(const char16_t *str1, size_t n1, const char16_t *str2, size_t n2);
25+
LIBSCRATCHCPP_EXPORT int string_compare_case_insensitive(const StringPtr *str1, const StringPtr *str2);
26+
27+
LIBSCRATCHCPP_EXPORT bool string_contains_raw_case_sensitive(const char16_t *str, const char16_t *substr);
28+
LIBSCRATCHCPP_EXPORT bool string_contains_case_sensitive(const StringPtr *str, const StringPtr *substr);
29+
30+
LIBSCRATCHCPP_EXPORT bool string_contains_raw_case_insensitive(const char16_t *str, const char16_t *substr);
31+
LIBSCRATCHCPP_EXPORT bool string_contains_case_insensitive(const StringPtr *str, const StringPtr *substr);
32+
}
33+
34+
} // namespace libscratchcpp

‎include/scratchcpp/string_pool.h

+18
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,18 @@
1+
// SPDX-License-Identifier: Apache-2.0
2+
3+
#pragma once
4+
5+
#include "global.h"
6+
7+
namespace libscratchcpp
8+
{
9+
10+
struct StringPtr;
11+
12+
extern "C"
13+
{
14+
LIBSCRATCHCPP_EXPORT StringPtr *string_pool_new(bool internal = false);
15+
LIBSCRATCHCPP_EXPORT void string_pool_free(StringPtr *str);
16+
}
17+
18+
} // namespace libscratchcpp

‎include/scratchcpp/stringptr.h

+34
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,34 @@
1+
// SPDX-License-Identifier: Apache-2.0
2+
3+
#pragma once
4+
5+
#include "global.h"
6+
#include "string_functions.h"
7+
8+
namespace libscratchcpp
9+
{
10+
11+
extern "C"
12+
{
13+
/*! \brief The StringPtr struct holds a string data pointer and string size */
14+
struct LIBSCRATCHCPP_EXPORT StringPtr
15+
{
16+
// NOTE: Constructors and destructors only work in C++ code and are not supposed to be used in LLVM IR
17+
StringPtr() = default;
18+
StringPtr(const std::string &str) { string_assign_cstring(this, str.c_str()); }
19+
StringPtr(const StringPtr &) = delete;
20+
21+
~StringPtr()
22+
{
23+
if (data && allocatedSize > 0)
24+
free(data);
25+
}
26+
27+
// NOTE: Any changes must also be done in the LLVM code builder!
28+
char16_t *data = nullptr;
29+
size_t size = 0;
30+
size_t allocatedSize = 0;
31+
};
32+
}
33+
34+
} // namespace libscratchcpp

‎include/scratchcpp/thread.h

+1
Original file line numberDiff line numberDiff line change
@@ -20,6 +20,7 @@ class LIBSCRATCHCPP_EXPORT Thread
2020
public:
2121
Thread(Target *target, IEngine *engine, Script *script);
2222
Thread(const Thread &) = delete;
23+
~Thread();
2324

2425
Target *target() const;
2526
IEngine *engine() const;

‎include/scratchcpp/value.h

+1-1
Original file line numberDiff line numberDiff line change
@@ -132,7 +132,7 @@ class LIBSCRATCHCPP_EXPORT Value
132132
};
133133

134134
/*! Returns the UTF-16 representation of the value. */
135-
std::u16string toUtf16()
135+
std::u16string toUtf16() const
136136
{
137137
std::u16string ret;
138138
value_toUtf16(&m_data, &ret);

‎include/scratchcpp/value_functions.h

+6-5
Original file line numberDiff line numberDiff line change
@@ -56,6 +56,7 @@ extern "C"
5656
LIBSCRATCHCPP_EXPORT void value_assign_bool(ValueData *v, bool boolValue);
5757
LIBSCRATCHCPP_EXPORT void value_assign_string(ValueData *v, const std::string &stringValue);
5858
LIBSCRATCHCPP_EXPORT void value_assign_cstring(ValueData *v, const char *stringValue);
59+
LIBSCRATCHCPP_EXPORT void value_assign_stringPtr(ValueData *v, const StringPtr *stringValue);
5960
LIBSCRATCHCPP_EXPORT void value_assign_copy(ValueData *v, const ValueData *another);
6061

6162
LIBSCRATCHCPP_EXPORT bool value_isInfinity(const ValueData *v);
@@ -72,16 +73,16 @@ extern "C"
7273
LIBSCRATCHCPP_EXPORT double value_toDouble(const ValueData *v);
7374
LIBSCRATCHCPP_EXPORT bool value_toBool(const ValueData *v);
7475
LIBSCRATCHCPP_EXPORT void value_toString(const ValueData *v, std::string *dst);
75-
LIBSCRATCHCPP_EXPORT char *value_toCString(const ValueData *v);
76+
LIBSCRATCHCPP_EXPORT StringPtr *value_toStringPtr(const ValueData *v);
7677
LIBSCRATCHCPP_EXPORT void value_toUtf16(const ValueData *v, std::u16string *dst);
7778
LIBSCRATCHCPP_EXPORT Rgb value_toRgba(const ValueData *v);
7879

7980
LIBSCRATCHCPP_EXPORT bool value_doubleIsInt(double v);
8081

81-
LIBSCRATCHCPP_EXPORT char *value_doubleToCString(double v);
82-
LIBSCRATCHCPP_EXPORT const char *value_boolToCString(bool v);
83-
LIBSCRATCHCPP_EXPORT double value_stringToDouble(const char *s);
84-
LIBSCRATCHCPP_EXPORT bool value_stringToBool(const char *s);
82+
LIBSCRATCHCPP_EXPORT StringPtr *value_doubleToStringPtr(double v);
83+
LIBSCRATCHCPP_EXPORT const StringPtr *value_boolToStringPtr(bool v);
84+
LIBSCRATCHCPP_EXPORT double value_stringToDouble(const StringPtr *s);
85+
LIBSCRATCHCPP_EXPORT bool value_stringToBool(const StringPtr *s);
8586

8687
LIBSCRATCHCPP_EXPORT void value_add(const ValueData *v1, const ValueData *v2, ValueData *dst);
8788
LIBSCRATCHCPP_EXPORT void value_subtract(const ValueData *v1, const ValueData *v2, ValueData *dst);

‎include/scratchcpp/valuedata.h

+3-2
Original file line numberDiff line numberDiff line change
@@ -9,6 +9,8 @@
99
namespace libscratchcpp
1010
{
1111

12+
struct StringPtr;
13+
1214
enum class LIBSCRATCHCPP_EXPORT ValueType
1315
{
1416
Number = 0,
@@ -26,11 +28,10 @@ extern "C"
2628
{
2729
double numberValue;
2830
bool boolValue;
29-
char *stringValue;
31+
StringPtr *stringValue;
3032
};
3133

3234
ValueType type;
33-
size_t stringSize; // allocated size, not length
3435
};
3536
}
3637

‎src/blocks/controlblocks.cpp

+8-3
Original file line numberDiff line numberDiff line change
@@ -11,6 +11,9 @@
1111
#include <scratchcpp/istacktimer.h>
1212
#include <scratchcpp/variable.h>
1313
#include <scratchcpp/sprite.h>
14+
#include <scratchcpp/stringptr.h>
15+
#include <scratchcpp/string_functions.h>
16+
#include <utf8.h>
1417

1518
#include "controlblocks.h"
1619

@@ -215,13 +218,15 @@ extern "C" void control_create_clone_by_index(ExecutionContext *ctx, double inde
215218
static_cast<Sprite *>(target)->clone();
216219
}
217220

218-
extern "C" void control_create_clone(ExecutionContext *ctx, const char *spriteName)
221+
extern "C" void control_create_clone(ExecutionContext *ctx, const StringPtr *spriteName)
219222
{
220-
if (strcmp(spriteName, "_myself_") == 0)
223+
static const StringPtr myself("_myself_");
224+
if (string_compare_case_sensitive(spriteName, &myself) == 0)
221225
control_create_clone_of_myself(ctx->thread()->target());
222226
else {
223227
IEngine *engine = ctx->engine();
224-
auto index = engine->findTarget(spriteName);
228+
// TODO: Use UTF-16 in engine
229+
auto index = engine->findTarget(utf8::utf16to8(std::u16string(spriteName->data)));
225230
Target *target = engine->targetAt(index);
226231

227232
if (!target->isStage())

‎src/blocks/eventblocks.cpp

+5-2
Original file line numberDiff line numberDiff line change
@@ -9,6 +9,8 @@
99
#include <scratchcpp/thread.h>
1010
#include <scratchcpp/compilerconstant.h>
1111
#include <scratchcpp/promise.h>
12+
#include <scratchcpp/stringptr.h>
13+
#include <utf8.h>
1214

1315
#include "eventblocks.h"
1416

@@ -125,11 +127,12 @@ CompilerValue *EventBlocks::compileWhenKeyPressed(Compiler *compiler)
125127
return nullptr;
126128
}
127129

128-
extern "C" void event_broadcast(ExecutionContext *ctx, const char *name, bool wait)
130+
extern "C" void event_broadcast(ExecutionContext *ctx, const StringPtr *name, bool wait)
129131
{
130132
Thread *thread = ctx->thread();
131133
IEngine *engine = thread->engine();
132-
std::vector<int> broadcasts = engine->findBroadcasts(name);
134+
// TODO: Use UTF-16 in engine
135+
std::vector<int> broadcasts = engine->findBroadcasts(utf8::utf16to8(std::u16string(name->data)));
133136

134137
for (int index : broadcasts)
135138
engine->broadcast(index, thread, wait);

0 commit comments

Comments
 (0)