[llvm] Create a CharSetConverter class with both iconv and icu support. (PR #138893)
Sergei Barannikov via llvm-commits
llvm-commits at lists.llvm.org
Fri May 9 13:34:20 PDT 2025
================
@@ -0,0 +1,351 @@
+//===-- CharSet.cpp - Characters sets conversion class ------------*- C++ -*-=//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// This file provides utility classes to convert between different character
+/// set encodings.
+///
+//===----------------------------------------------------------------------===//
+
+#include "llvm/Support/CharSet.h"
+#include "llvm/ADT/SmallString.h"
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/ADT/StringExtras.h"
+#include "llvm/Support/ConvertEBCDIC.h"
+#include "llvm/Support/raw_ostream.h"
+#include <algorithm>
+#include <limits>
+#include <system_error>
+
+#if HAVE_ICU
+#include <unicode/ucnv.h>
+#elif HAVE_ICONV
+#include <iconv.h>
+#endif
+
+using namespace llvm;
+
+// Normalize the charset name with the charset alias matching algorithm proposed
+// in https://www.unicode.org/reports/tr22/tr22-8.html#Charset_Alias_Matching.
+static void normalizeCharSetName(StringRef CSName,
+ SmallVectorImpl<char> &Normalized) {
+ bool PrevDigit = false;
+ for (auto Ch : CSName) {
+ if (isAlnum(Ch)) {
+ Ch = toLower(Ch);
+ if (Ch != '0' || PrevDigit) {
+ PrevDigit = isDigit(Ch);
+ Normalized.push_back(Ch);
+ }
+ }
+ }
+}
+
+// Maps the charset name to enum constant if possible.
+static std::optional<TextEncoding> getKnownEncoding(StringRef CSName) {
+ SmallString<16> Normalized;
+ normalizeCharSetName(CSName, Normalized);
+ if (Normalized.equals("utf8"))
+ return TextEncoding::UTF8;
+ if (Normalized.equals("ibm1047"))
+ return TextEncoding::IBM1047;
+ return std::nullopt;
+}
+
+LLVM_ATTRIBUTE_UNUSED static void
+HandleOverflow(size_t &Capacity, char *&Output, size_t &OutputLength,
+ SmallVectorImpl<char> &Result) {
+ // No space left in output buffer. Double the size of the underlying
+ // memory in the SmallVectorImpl, adjust pointer and length and continue
+ // the conversion.
+ Capacity = (Capacity < std::numeric_limits<size_t>::max() / 2)
+ ? 2 * Capacity
+ : std::numeric_limits<size_t>::max();
+ Result.resize(0);
+ Result.resize_for_overwrite(Capacity);
+ Output = static_cast<char *>(Result.data());
+ OutputLength = Capacity;
+}
+
+namespace {
+enum ConversionType {
+ UTF8ToIBM1047,
+ IBM1047ToUTF8,
+};
+
+// Support conversion between EBCDIC 1047 and UTF-8. This class uses
+// built-in translation tables that allow for translation between the
+// aforementioned character sets. The use of tables for conversion is only
+// possible because EBCDIC 1047 is a single-byte, stateless encoding; other
+// character sets are not supported.
+class EncodingConverterTable : public details::EncodingConverterImplBase {
+ const ConversionType ConvType;
+
+public:
+ EncodingConverterTable(ConversionType ConvType) : ConvType(ConvType) {}
+
+ std::error_code convertString(StringRef Source,
+ SmallVectorImpl<char> &Result) override;
+
+ void reset() override {}
+};
+
+std::error_code
+EncodingConverterTable::convertString(StringRef Source,
+ SmallVectorImpl<char> &Result) {
+ switch (ConvType) {
+ case IBM1047ToUTF8:
+ ConverterEBCDIC::convertToUTF8(Source, Result);
+ return std::error_code();
+ case UTF8ToIBM1047:
+ return ConverterEBCDIC::convertToEBCDIC(Source, Result);
+ }
+ llvm_unreachable("Invalid ConvType!");
+ return std::error_code();
+}
+
+#if HAVE_ICU
+struct UConverterDeleter {
+ void operator()(UConverter *Converter) const {
+ if (Converter)
+ ucnv_close(Converter);
+ }
+};
+using UConverterUniquePtr = std::unique_ptr<UConverter, UConverterDeleter>;
+
+class EncodingConverterICU : public details::EncodingConverterImplBase {
+ UConverterUniquePtr FromConvDesc;
+ UConverterUniquePtr ToConvDesc;
+
+public:
+ EncodingConverterICU(UConverterUniquePtr FromConverter,
+ UConverterUniquePtr ToConverter)
+ : FromConvDesc(std::move(FromConverter)),
+ ToConvDesc(std::move(ToConverter)) {}
+
+ std::error_code convertString(StringRef Source,
+ SmallVectorImpl<char> &Result) override;
+
+ void reset() override;
+};
+
+// TODO: The current implementation discards the partial result and restarts the
+// conversion from the beginning if there is a conversion error due to
+// insufficient buffer size. In the future, it would better to save the partial
+// result and redo the conversion for the remaining string.
+std::error_code
+EncodingConverterICU::convertString(StringRef Source,
+ SmallVectorImpl<char> &Result) {
+ // Setup the input in case it has no backing data.
+ size_t InputLength = Source.size();
+ const char *In = InputLength ? const_cast<char *>(Source.data()) : "";
+
+ // Setup the output. We directly write into the SmallVector.
+ size_t Capacity = Result.capacity();
+ size_t OutputLength = Capacity;
+ Result.resize_for_overwrite(Capacity);
+ char *Output;
+ UErrorCode EC = U_ZERO_ERROR;
+
+ ucnv_setToUCallBack(&*FromConvDesc, UCNV_TO_U_CALLBACK_STOP, NULL, NULL, NULL,
+ &EC);
+ ucnv_setFromUCallBack(&*ToConvDesc, UCNV_FROM_U_CALLBACK_STOP, NULL, NULL,
+ NULL, &EC);
+ assert(U_SUCCESS(EC));
+
+ do {
+ EC = U_ZERO_ERROR;
+ const char *Input = In;
+
+ Output = InputLength ? static_cast<char *>(Result.data()) : nullptr;
+ ucnv_convertEx(&*ToConvDesc, &*FromConvDesc, &Output, Result.end(), &Input,
+ In + InputLength, /*pivotStart=*/NULL,
+ /*pivotSource=*/NULL, /*pivotTarget=*/NULL,
+ /*pivotLimit=*/NULL, /*reset=*/true,
+ /*flush=*/true, &EC);
+ if (U_FAILURE(EC)) {
+ if (EC == U_BUFFER_OVERFLOW_ERROR &&
+ Capacity < std::numeric_limits<size_t>::max()) {
+ HandleOverflow(Capacity, Output, OutputLength, Result);
+ continue;
+ }
+ // Some other error occured.
+ Result.resize(Output - Result.data());
+ return std::error_code(EILSEQ, std::generic_category());
+ }
+ break;
+ } while (true);
+
+ Result.resize(Output - Result.data());
+ return std::error_code();
+}
+
+void EncodingConverterICU::reset() {
+ ucnv_reset(&*FromConvDesc);
+ ucnv_reset(&*ToConvDesc);
+}
+
+#elif HAVE_ICONV
+class EncodingConverterIconv : public details::EncodingConverterImplBase {
+ class UniqueIconvT {
+ iconv_t ConvDesc;
+
+ public:
+ operator iconv_t() const { return ConvDesc; }
+ UniqueIconvT(iconv_t CD) : ConvDesc(CD) {}
+ ~UniqueIconvT() {
+ if (ConvDesc != (iconv_t)-1) {
+ iconv_close(ConvDesc);
+ ConvDesc = (iconv_t)-1;
+ }
+ }
+ UniqueIconvT(UniqueIconvT &&Other) : ConvDesc(Other.ConvDesc) {
+ Other.ConvDesc = (iconv_t)-1;
+ }
+ UniqueIconvT &operator=(UniqueIconvT &&Other) {
+ if (&Other != this) {
+ ConvDesc = Other.ConvDesc;
+ Other.ConvDesc = (iconv_t)-1;
+ }
+ return *this;
+ }
+ };
+ UniqueIconvT ConvDesc;
+
+public:
+ EncodingConverterIconv(UniqueIconvT ConvDesc)
+ : ConvDesc(std::move(ConvDesc)) {}
+
+ std::error_code convertString(StringRef Source,
+ SmallVectorImpl<char> &Result) override;
+
+ void reset() override;
+};
+
+// TODO: The current implementation discards the partial result and restarts the
+// conversion from the beginning if there is a conversion error due to
+// insufficient buffer size. In the future, it would better to save the partial
+// result and redo the conversion for the remaining string.
+std::error_code
+EncodingConverterIconv::convertString(StringRef Source,
+ SmallVectorImpl<char> &Result) {
+ // Setup the output. We directly write into the SmallVector.
+ size_t Capacity = Result.capacity();
+ char *Output = static_cast<char *>(Result.data());
+ size_t OutputLength = Capacity;
+ Result.resize_for_overwrite(Capacity);
+
+ size_t Ret;
+ // Handle errors returned from iconv().
+ auto HandleError = [&Capacity, &Output, &OutputLength, &Result,
+ this](size_t Ret) {
+ if (Ret == static_cast<size_t>(-1)) {
+ // An error occured. Check if we can gracefully handle it.
+ if (errno == E2BIG && Capacity < std::numeric_limits<size_t>::max()) {
----------------
s-barannikov wrote:
Same here.
https://github.com/llvm/llvm-project/pull/138893
More information about the llvm-commits
mailing list