url_canon_internal.h | Explore in Territory

// Copyright 2013 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

#ifdef UNSAFE_BUFFERS_BUILD
// TODO(crbug.com/40285824): Remove this and convert code to safer constructs.
#pragma allow_unsafe_buffers
#endif

// This is a copy of url/url_canon_internal.h circa 2023. It should be used only
// by components/feedback/redaction_tool/. We need a copy because the
// components/feedback/redaction_tool source code is shared into ChromeOS and
// needs to have no dependencies outside of base/.

#ifndef COMPONENTS_FEEDBACK_REDACTION_TOOL_URL_CANON_INTERNAL_H_
#define COMPONENTS_FEEDBACK_REDACTION_TOOL_URL_CANON_INTERNAL_H_

// This file is intended to be included in another C++ file where the character
// types are defined. This allows us to write mostly generic code, but not have
// template bloat because everything is inlined when anybody calls any of our
// functions.

#include <stddef.h>
#include <stdlib.h>

#include "components/feedback/redaction_tool/url_canon.h"

namespace redaction_internal {

// Character type handling -----------------------------------------------------

// Bits that identify different character types. These types identify different
// bits that are set for each 8-bit character in the kSharedCharTypeTable.
enum SharedCharTypes { … };

// This table contains the flags in SharedCharTypes for each 8-bit character.
// Some canonicalization functions have their own specialized lookup table.
// For those with simple requirements, we have collected the flags in one
// place so there are fewer lookup tables to load into the CPU cache.
//
// Using an unsigned char type has a small but measurable performance benefit
// over using a 32-bit number.
extern const unsigned char kSharedCharTypeTable[0x100];

// More readable wrappers around the character type lookup table.
inline bool IsCharOfType(unsigned char c, SharedCharTypes type) { … }
inline bool IsQueryChar(unsigned char c) { … }
inline bool IsIPv4Char(unsigned char c) { … }
inline bool IsHexChar(unsigned char c) { … }
inline bool IsComponentChar(unsigned char c) { … }

#ifndef WIN32

// Implementations of Windows' int-to-string conversions
int _itoa_s(int value, char* buffer, size_t size_in_chars, int radix);

// Secure template overloads for these functions
template <size_t N>
inline int _itoa_s(int value, char (&buffer)[N], int radix) { … }

// _strtoui64 and strtoull behave the same
inline uint64_t _strtoui64(const char* nptr, char** endptr, int base) { … }

#endif  // WIN32

}  // namespace redaction_internal

#endif  // COMPONENTS_FEEDBACK_REDACTION_TOOL_URL_CANON_INTERNAL_H_
chromium/components/feedback/redaction_tool/url_canon_internal.h
