From a92917f947fabef7e798c1ed733e5d57444c8958 Mon Sep 17 00:00:00 2001 From: Marc Durdin Date: Thu, 25 May 2023 12:54:27 +0700 Subject: [PATCH] chore(developer): replace cwrap wasm bindings This moves the remainder of the WASM interfaces in kmcmplib to using emscripten bind. It is a little bit of a step backwards at present for parseUnicodeSet, because I've changed the output buffer type to an int for the purposes of simplifying the binding just now. But that can be improved later, and at least we are consistent with the binding methods. Next step is to move the filesystem access out of kmcmplib. --- .../src/kmc-kmn/src/compiler/compiler.ts | 65 ++++--------------- developer/src/kmcmplib/include/kmcmplibapi.h | 15 ++--- .../src/kmcmplib/src/CompilerInterfaces.cpp | 34 +--------- developer/src/kmcmplib/src/meson.build | 2 +- developer/src/kmcmplib/src/uset-api.cpp | 13 ++-- .../src/kmcmplib/tests/uset-api-test.cpp | 30 +++++---- 6 files changed, 43 insertions(+), 116 deletions(-) diff --git a/developer/src/kmc-kmn/src/compiler/compiler.ts b/developer/src/kmc-kmn/src/compiler/compiler.ts index c46b894cb3..a88999e99c 100644 --- a/developer/src/kmc-kmn/src/compiler/compiler.ts +++ b/developer/src/kmc-kmn/src/compiler/compiler.ts @@ -42,50 +42,10 @@ const baseOptions: CompilerOptions = { */ let callbackProcIdentifier = 0; -/** - * Pointer in wasm-space - */ -type WasmPtr = number; - -/** - * The wrapped functions - */ -class WasmWrapper { - Module: any; - - compileKeyboardFile?: (pszInfile: string, pszOutfile: string, aSaveDebug: number, aCompilerWarningsAsErrors: number, aWarnDeprecatedCode: number, msgProc: string) => boolean; - parseUnicodeSet?: (pat: string, buf: WasmPtr, length: number) => number; - setCompilerOptions?: (shouldAddCompilerVersion: number) => boolean; - - constructor(wasmModule: any) { - this.Module = wasmModule; - if (!wasmModule) { - throw Error(`wasm host did not load`); - } - this.compileKeyboardFile = this.Module.cwrap('kmcmp_Wasm_CompileKeyboardFile', 'boolean', ['string', 'string', 'number', 'number', 'number', 'string']); - this.parseUnicodeSet = this.Module.cwrap('kmcmp_Wasm_ParseUnicodeSet', 'number', ['string', 'number', 'number']); - this.setCompilerOptions = this.Module.cwrap('kmcmp_Wasm_SetCompilerOptions', 'boolean', ['number']); - - if (this.parseUnicodeSet === undefined - || this.setCompilerOptions === undefined - || this.compileKeyboardFile === undefined) { - throw Error(`some wasm functions did not load properly.`); - } - } - - /** - * Entry point into Wasm functions - * @returns WasmWrapper - */ - public static async load() : Promise { - return new WasmWrapper(await loadWasmHost()); - } -}; - export class KmnCompiler { + private Module: any; callbackName: string; callbacks: CompilerCallbacks; - wasm: WasmWrapper; constructor() { this.callbackName = 'kmnCompilerCallback' + callbackProcIdentifier; @@ -94,9 +54,9 @@ export class KmnCompiler { public async init(callbacks: CompilerCallbacks): Promise { this.callbacks = callbacks; - if(!this.wasm) { + if(!this.Module) { try { - this.wasm = await WasmWrapper.load(); + this.Module = await loadWasmHost(); } catch(e: any) { this.callbacks.reportMessage(CompilerMessages.Fatal_MissingWasmModule({e})); return false; @@ -114,7 +74,7 @@ export class KmnCompiler { // Can't report a message here. throw Error('Must call Compiler.init(callbacks) before proceeding'); } - if(!this.wasm) { // fail if wasm not loaded or function not found + if(!this.Module) { // fail if wasm not loaded or function not found this.callbacks.reportMessage(CompilerMessages.Fatal_MissingWasmModule({})); return false; } @@ -153,7 +113,7 @@ export class KmnCompiler { private runCompiler(infile: string, outfile: string, options: CompilerOptions): CompilerResult { let result: CompilerResult = {}; - let wasm_interface = new this.wasm.Module.CompilerInterface(); + let wasm_interface = new this.Module.CompilerInterface(); let wasm_result = null; try { wasm_interface.saveDebug = options.saveDebug; @@ -161,7 +121,7 @@ export class KmnCompiler { wasm_interface.warnDeprecatedCode = options.warnDeprecatedCode; wasm_interface.messageCallback = this.callbackName; wasm_interface.loadFileCallback = this.callbackName; // TODO: this is wrong, needs to be a new callback; not yet used though - wasm_result = this.wasm.Module.kmcmp_compile(infile, wasm_interface); + wasm_result = this.Module.kmcmp_compile(infile, wasm_interface); if(!wasm_result.result) { return null; } @@ -175,7 +135,7 @@ export class KmnCompiler { result.kmx = { filename: outfile, - data: new Uint8Array(this.wasm.Module.HEAP8.buffer, wasm_result.kmx, wasm_result.kmxSize) + data: new Uint8Array(this.Module.HEAP8.buffer, wasm_result.kmx, wasm_result.kmxSize) }; return result; @@ -230,16 +190,15 @@ export class KmnCompiler { if (!bufferSize) { bufferSize = 100; // TODO-LDML: Preflight mode? Reuse buffer? } - const { Module } = this.wasm; - const buf = Module.asm.malloc(bufferSize * 2 * Module.HEAPU32.BYTES_PER_ELEMENT); + const buf = this.Module.asm.malloc(bufferSize * 2 * this.Module.HEAPU32.BYTES_PER_ELEMENT); // TODO-LDML: Catch OOM - const rc = this.wasm.parseUnicodeSet(pattern, buf, bufferSize); + const rc = this.Module.kmcmp_parseUnicodeSet(pattern, buf, bufferSize); if (rc >= 0) { const ranges = []; - const startu = (buf / Module.HEAPU32.BYTES_PER_ELEMENT); + const startu = (buf / this.Module.HEAPU32.BYTES_PER_ELEMENT); for (let i = 0; i < rc; i++) { - const low = Module.HEAPU32[startu + (i * 2) + 0]; - const high = Module.HEAPU32[startu + (i * 2) + 1]; + const low = this.Module.HEAPU32[startu + (i * 2) + 0]; + const high = this.Module.HEAPU32[startu + (i * 2) + 1]; ranges.push([low, high]); } // TODO-LDML: no free?? diff --git a/developer/src/kmcmplib/include/kmcmplibapi.h b/developer/src/kmcmplib/include/kmcmplibapi.h index 2fb35979b2..f4591612ec 100644 --- a/developer/src/kmcmplib/include/kmcmplibapi.h +++ b/developer/src/kmcmplib/include/kmcmplibapi.h @@ -62,7 +62,7 @@ EXTERN bool kmcmp_ValidateJsonFile( ); /** - * kmcmp_ParseUnicodeSet is successful if it returns >= USET_OK + * kmcmp_parseUnicodeSet is successful if it returns >= USET_OK */ static const int KMCMP_USET_OK = 0; @@ -83,24 +83,19 @@ static const int KMCMP_ERROR_UNSUPPORTED_PROPERTY = -3; */ static const int KMCMP_FATAL_OUT_OF_RANGE = -4; -/** - * Function pointer to kmcmp_ParseUnicodeSet - */ -typedef int (*kmcmp_ParseUnicodeSetProc)(const char* szText, uint32_t* output, uint32_t outputLength); - /** * Parse a UnicodeSet into 32-bit ranges. * For example, "[]" will return 0 (KMCMP_USET_OK) as a zero-length set. * "[" will return KMCMP_ERROR_SYNTAX_ERR, * and "[x A-C]" will return 2 and [0x41, 0x43, 0x78, 0x78] - * @param szText input txt, null terminated, in UTF-8 format + * @param text input txt, null terminated, in UTF-8 format * @param outputBuffer output buffer, owned by caller: Pairs of ranges in order * @param outputBufferSize length of output buffer. Needs to be twice the number of expected ranges * @return If >= KMCMP_USET_OK, number of ranges, otherwise one of the negative error values. */ -EXTERN int kmcmp_ParseUnicodeSet( - const char* szText, - uint32_t* outputBuffer, +EXTERN int kmcmp_parseUnicodeSet( + const std::string text, + uintptr_t outputBuffer_, uint32_t outputBufferSize ); diff --git a/developer/src/kmcmplib/src/CompilerInterfaces.cpp b/developer/src/kmcmplib/src/CompilerInterfaces.cpp index a1609d3544..56ecf4cdcf 100644 --- a/developer/src/kmcmplib/src/CompilerInterfaces.cpp +++ b/developer/src/kmcmplib/src/CompilerInterfaces.cpp @@ -46,39 +46,6 @@ int wasm_CompilerMessageProc(int line, uint32_t dwMsgCode, char* szText, void* c return wasm_msgproc(line, dwMsgCode, szText, msgProc); } -//DEPRECATED -EXTERN bool kmcmp_Wasm_SetCompilerOptions(int ShouldAddCompilerVersion) { - KMCMP_COMPILER_OPTIONS options; - options.dwSize = sizeof(KMCMP_COMPILER_OPTIONS); - options.ShouldAddCompilerVersion = ShouldAddCompilerVersion; - return kmcmp_SetCompilerOptions(&options); -} - -//DEPRECATED -EXTERN bool kmcmp_Wasm_CompileKeyboardFile(char* pszInfile, - char* pszOutfile, int ASaveDebug, int ACompilerWarningsAsErrors, - int AWarnDeprecatedCode, char* msgProc -) { - return kmcmp_CompileKeyboardFile( - pszInfile, - pszOutfile, - ASaveDebug, - ACompilerWarningsAsErrors, - AWarnDeprecatedCode, - wasm_CompilerMessageProc, - msgProc - ); -} - -EXTERN int kmcmp_Wasm_ParseUnicodeSet(char* pat, - uint32_t* buf, int length -) { - return kmcmp_ParseUnicodeSet( - pat, buf, length - ); -} - - struct COMPILER_INTERFACE { bool saveDebug; bool compilerWarningsAsErrors; @@ -162,6 +129,7 @@ EMSCRIPTEN_BINDINGS(compiler_interface) { ; emscripten::function("kmcmp_compile", &kmcmp_compile); + emscripten::function("kmcmp_parseUnicodeSet", &kmcmp_parseUnicodeSet); } #endif diff --git a/developer/src/kmcmplib/src/meson.build b/developer/src/kmcmplib/src/meson.build index 292b158ce2..3cd9abd6e1 100644 --- a/developer/src/kmcmplib/src/meson.build +++ b/developer/src/kmcmplib/src/meson.build @@ -24,7 +24,7 @@ endif name_suffix = [] if cpp_compiler.get_id() == 'emscripten' - links += ['-lnodefs.js', '-sMODULARIZE', '-sEXPORT_ES6', '--whole-archive', '--bind', '-sEXPORTED_RUNTIME_METHODS=[\'cwrap\', \'UTF8ToString\']'] + links += ['-lnodefs.js', '-sMODULARIZE', '-sEXPORT_ES6', '--whole-archive', '--bind', '-sEXPORTED_RUNTIME_METHODS=[\'UTF8ToString\']'] # tests are building as ES6 so we need to declare the file extension # note that meson currently struggles with the sanitycheckc_cross.exe # program, because it has a hard coded extension (.exe) which is not diff --git a/developer/src/kmcmplib/src/uset-api.cpp b/developer/src/kmcmplib/src/uset-api.cpp index 1aa22477b2..c2c72cab34 100644 --- a/developer/src/kmcmplib/src/uset-api.cpp +++ b/developer/src/kmcmplib/src/uset-api.cpp @@ -4,16 +4,13 @@ #include "unicode/uniset.h" #include "unicode/unistr.h" -EXTERN int kmcmp_ParseUnicodeSet( - const char* szText, - uint32_t* outputBuffer, +EXTERN int kmcmp_parseUnicodeSet( + const std::string text, + uintptr_t outputBuffer_, uint32_t outputBufferSize ) { - if (szText == nullptr) { - // null string coming in - return KMCMP_ERROR_SYNTAX_ERR; - } - const icu::UnicodeString str = icu::UnicodeString::fromUTF8(szText); + uint32_t* outputBuffer = reinterpret_cast(outputBuffer_); + const icu::UnicodeString str = icu::UnicodeString::fromUTF8(text.c_str()); if (str.isBogus() || str.isEmpty()) { // empty string return KMCMP_ERROR_SYNTAX_ERR; diff --git a/developer/src/kmcmplib/tests/uset-api-test.cpp b/developer/src/kmcmplib/tests/uset-api-test.cpp index 4fe599ca2e..61d263ad15 100644 --- a/developer/src/kmcmplib/tests/uset-api-test.cpp +++ b/developer/src/kmcmplib/tests/uset-api-test.cpp @@ -19,30 +19,32 @@ #include "../src/compfile.h" #include -void test_kmcmp_ParseUnicodeSetProc(); +void test_kmcmp_parseUnicodeSet(); // std::vector error_vec; int main(int argc, char *argv[]) { - test_kmcmp_ParseUnicodeSetProc(); + test_kmcmp_parseUnicodeSet(); return 0; } -void test_kmcmp_ParseUnicodeSetProc() { +void test_kmcmp_parseUnicodeSet() { { // null test const auto bufsiz = 128; uint32_t buf[bufsiz]; - int rc = kmcmp_ParseUnicodeSet(u8"[]", buf, bufsiz); + uintptr_t buf_ = reinterpret_cast(buf); + int rc = kmcmp_parseUnicodeSet(u8"[]", buf_, bufsiz); assert(rc == KMCMP_USET_OK); } { // basic test const auto bufsiz = 128; uint32_t buf[bufsiz]; - int rc = kmcmp_ParseUnicodeSet(u8"[x A-C]", buf, bufsiz); + uintptr_t buf_ = reinterpret_cast(buf); + int rc = kmcmp_parseUnicodeSet(u8"[x A-C]", buf_, bufsiz); assert(rc == 2); assert(buf[0] == 0x41); assert(buf[1] == 0x43); @@ -53,7 +55,8 @@ void test_kmcmp_ParseUnicodeSetProc() { // bigger test const auto bufsiz = 128; uint32_t buf[bufsiz]; - int rc = kmcmp_ParseUnicodeSet(u8"[[🙀A-C]-[CB]]", buf, bufsiz); + uintptr_t buf_ = reinterpret_cast(buf); + int rc = kmcmp_parseUnicodeSet(u8"[[🙀A-C]-[CB]]", buf_, bufsiz); assert(rc == 2); assert(buf[0] == 0x41); assert(buf[1] == 0x41); @@ -64,35 +67,40 @@ void test_kmcmp_ParseUnicodeSetProc() { // overflow test const auto bufsiz = 1; uint32_t buf[bufsiz]; - int rc = kmcmp_ParseUnicodeSet(u8"[x A-C]", buf, bufsiz); + uintptr_t buf_ = reinterpret_cast(buf); + int rc = kmcmp_parseUnicodeSet(u8"[x A-C]", buf_, bufsiz); assert(rc == KMCMP_FATAL_OUT_OF_RANGE); } { // err test const auto bufsiz = 128; uint32_t buf[bufsiz]; - int rc = kmcmp_ParseUnicodeSet(u8"[:Adlm:]", buf, bufsiz); + uintptr_t buf_ = reinterpret_cast(buf); + int rc = kmcmp_parseUnicodeSet(u8"[:Adlm:]", buf_, bufsiz); assert(rc == KMCMP_ERROR_UNSUPPORTED_PROPERTY); } { // err test const auto bufsiz = 128; uint32_t buf[bufsiz]; - int rc = kmcmp_ParseUnicodeSet(u8"[[\\p{Mn}]&[A-Z]]", buf, bufsiz); + uintptr_t buf_ = reinterpret_cast(buf); + int rc = kmcmp_parseUnicodeSet(u8"[[\\p{Mn}]&[A-Z]]", buf_, bufsiz); assert(rc == KMCMP_ERROR_UNSUPPORTED_PROPERTY); } { // err test const auto bufsiz = 128; uint32_t buf[bufsiz]; - int rc = kmcmp_ParseUnicodeSet(u8"[abc{def}]", buf, bufsiz); + uintptr_t buf_ = reinterpret_cast(buf); + int rc = kmcmp_parseUnicodeSet(u8"[abc{def}]", buf_, bufsiz); assert(rc == KMCMP_ERROR_HAS_STRINGS); } { // err test const auto bufsiz = 128; uint32_t buf[bufsiz]; - int rc = kmcmp_ParseUnicodeSet(u8"[[]", buf, bufsiz); + uintptr_t buf_ = reinterpret_cast(buf); + int rc = kmcmp_parseUnicodeSet(u8"[[]", buf_, bufsiz); assert(rc == KMCMP_ERROR_SYNTAX_ERR); } }