84 lines · cpp
1//===----------------------------------------------------------------------===//2//3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.4// See https://llvm.org/LICENSE.txt for license information.5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception6//7//===----------------------------------------------------------------------===//8 9#include "llvm/ADT/STLExtras.h"10#include "llvm/ADT/StringExtras.h"11#include "llvm/Support/ConvertUTF.h"12#include "llvm/Support/MemoryBuffer.h"13#include "llvm/Support/raw_ostream.h"14 15using namespace llvm;16 17int main(int argc, char *argv[]) {18 auto ErrorOrBuffer = MemoryBuffer::getFile(argv[1], true);19 if (!ErrorOrBuffer)20 return 1;21 std::unique_ptr<MemoryBuffer> Buffer = std::move(ErrorOrBuffer.get());22 StringRef Content = Buffer->getBuffer();23 Content = Content.drop_until([](char C) { return C == '#'; });24 SmallVector<StringRef> Lines;25 SplitString(Content, Lines, "\r\n");26 27 std::vector<std::pair<llvm::UTF32, SmallVector<llvm::UTF32>>> Entries;28 SmallVector<StringRef> Values;29 for (const StringRef Line : Lines) {30 if (Line.starts_with("#"))31 continue;32 33 Values.clear();34 Line.split(Values, ';');35 if (Values.size() < 2) {36 errs() << "Failed to parse: " << Line << "\n";37 return 2;38 }39 40 const llvm::StringRef From = Values[0].trim();41 llvm::UTF32 CodePoint = 0;42 From.getAsInteger(16, CodePoint);43 44 SmallVector<llvm::UTF32> To;45 SmallVector<StringRef> ToN;46 Values[1].split(ToN, ' ', -1, false);47 for (const StringRef ToI : ToN) {48 llvm::UTF32 ToCodePoint = 0;49 ToI.trim().getAsInteger(16, ToCodePoint);50 To.push_back(ToCodePoint);51 }52 // Sentinel53 To.push_back(0);54 55 Entries.emplace_back(CodePoint, To);56 }57 llvm::sort(Entries);58 59 const unsigned LargestValue =60 llvm::max_element(Entries, [](const auto &Entry0, const auto &Entry1) {61 return Entry0.second.size() < Entry1.second.size();62 })->second.size();63 64 std::error_code Ec;65 llvm::raw_fd_ostream Os(argv[2], Ec);66 67 // FIXME: If memory consumption and/or lookup time becomes a constraint, it68 // maybe worth using a more elaborate data structure.69 Os << "struct {llvm::UTF32 codepoint; llvm::UTF32 values[" << LargestValue70 << "];} "71 "ConfusableEntries[] = {\n";72 for (const auto &Values : Entries) {73 Os << " { ";74 Os << Values.first;75 Os << ", {";76 for (auto CP : Values.second)77 Os << CP << ", ";78 79 Os << "}},\n";80 }81 Os << "};\n";82 return 0;83}84