cmds/incident_helper/src/ih_util.h

/*
 * Copyright (C) 2017 The Android Open Source Project
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *      http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

#ifndef INCIDENT_HELPER_UTIL_H
#define INCIDENT_HELPER_UTIL_H

#include <map>
#include <stack>
#include <string>
#include <vector>

#include <android/util/ProtoOutputStream.h>

using namespace android::util;

typedef std::vector<std::string> header_t;
typedef std::vector<std::string> record_t;
typedef std::string (*trans_func) (const std::string&);

const std::string DEFAULT_WHITESPACE = " \t";
const std::string DEFAULT_NEWLINE = "\r\n";
const std::string TAB_DELIMITER = "\t";
const std::string COMMA_DELIMITER = ",";
const std::string PIPE_DELIMITER = "|";
const std::string PARENTHESES_DELIMITER = "()";

// returns true if c is a-zA-Z0-9 or underscore
bool isValidChar(char c);

// trim the string with the given charset
std::string trim(const std::string& s, const std::string& charset);

/**
 * When a text has a table format like this
 * line 1: HeadA HeadB HeadC
 * line 2: v1    v2    v3
 * line 3: v11   v12   v13
 *
 * We want to parse the line in structure given the delimiter.
 * parseHeader is used to parse the firse line of the table and returns a list of strings in lower case
 * parseRecord is used to parse other lines and returns a list of strings
 * empty strings are skipped
 */
header_t parseHeader(const std::string& line, const std::string& delimiters = DEFAULT_WHITESPACE);
record_t parseRecord(const std::string& line, const std::string& delimiters = DEFAULT_WHITESPACE);

/**
 * Gets the list of end indices of each word in the line and places it in the given vector,
 * clearing out the vector beforehand. These indices can be used with parseRecordByColumns.
 * Will return false if there was a problem getting the indices. headerNames
 * must be NULL terminated.
 */
bool getColumnIndices(std::vector<int>& indices, const char* headerNames[], const std::string& line);

/**
 * When a text-format table aligns by its vertical position, it is not possible to split them by purely delimiters.
 * This function allows to parse record by its header's column position' indices, must in ascending order.
 * At the same time, it still looks at the char at index, if it doesn't belong to delimiters, moves forward to find the delimiters.
 */
record_t parseRecordByColumns(const std::string& line, const std::vector<int>& indices, const std::string& delimiters = DEFAULT_WHITESPACE);

/** Prints record_t to stderr */
void printRecord(const record_t& record);

/**
 * When the line starts/ends with the given key, the function returns true
 * as well as the line argument is changed to the rest trimmed part of the original.
 * e.g. "ZRAM: 6828K physical used for 31076K in swap (524284K total swap)" becomes
 * "6828K physical used for 31076K in swap (524284K total swap)" when given key "ZRAM:",
 * otherwise the line is not changed.
 *
 * In order to prevent two values have same prefix which cause entering to incorrect conditions,
 * stripPrefix and stripSuffix can turn on a flag that requires the ending char in the line must not be a valid
 * character or digits, this feature is off by default.
 * i.e. ABC%some value, ABCD%other value
 */
bool stripPrefix(std::string* line, const char* key, bool endAtDelimiter = false);
bool stripSuffix(std::string* line, const char* key, bool endAtDelimiter = false);

/**
 * behead the given line by the cut, return the head and reassign the line to be the rest.
 */
std::string behead(std::string* line, const char cut);

/**
 * Converts string to the desired type
 */
int toInt(const std::string& s);
long long toLongLong(const std::string& s);
double toDouble(const std::string& s);

/**
 * Reader class reads data from given fd in streaming fashion.
 * The buffer size is controlled by capacity parameter.
 */
class Reader
{
public:
    explicit Reader(const int fd);
    ~Reader();

    bool readLine(std::string* line);
    bool ok(std::string* error);

private:
    FILE* mFile;
    char* mBuffer;
    std::string mStatus;
};

/**
 * The Table class is constructed from two arrays generated by the given message with
 * option (stream_proto.stream_msg).enable_fields_mapping = true.
 * The names are each field's names in the message and must corresponding to the header/name of
 * the text to be parsed, and the ids are the streaming proto encoded field ids.
 *
 * This class then allows users to insert the table values to proto based on its header.
 *
 * Advance feature: if some fields in the message are enums, user must explicitly add the
 * mapping from enum name string to its enum values.
 */
class Message;
class Table
{
friend class Message;
public:
    Table(const char* names[], const uint64_t ids[], const int count);
    ~Table();

    // Add enum names to values for parsing purpose.
    void addEnumTypeMap(const char* field, const char* enumNames[], const int enumValues[], const int enumSize);

    // Manually add enum names to values mapping, useful when an Enum type is used by
    // a number of fields, there must not be any enum name conflicts.
    void addEnumNameToValue(const char* enumName, const int enumValue);

    // Based on given name, find the right field id, parse the text value and insert to proto.
    // Return false if the given name can't be found.
    bool insertField(ProtoOutputStream* proto, const std::string& name, const std::string& value);
private:
    std::map<std::string, uint64_t> mFields;
    std::map<std::string, std::map<std::string, int>> mEnums;
    std::map<std::string, int> mEnumValuesByName;
};

/**
 * Reconstructs a typical proto message given its message Table, adds submessage fields explicitly.
 * It allows user to insert nested proto values purely by the names. See insertField for detail.
 */
class Message
{
public:
    explicit Message(Table* table);
    ~Message();

    // Reconstructs the typical proto message by adding its message fields.
    void addSubMessage(uint64_t fieldId, Message* fieldMsg);

    // Inserts value if the given name has the corresponding field in its message and return true.
    // It will recursively search the name in submessages and find the correct field to insert.
    // For example, when the name is dalvik_vm_heapsize, and the message's corresponding proto is:
    //     message Properties {
    //         message DalvikVm {
    //             int32 heapsize = 1;
    //             bool  usejit = 2;
    //         }
    //         DalvikVm dalvik_vm = 1;
    //         string hack_in = 2;
    //     }
    // The value will be inserted into field heapsize in dalvik_vm submessage.
    //
    // Also value belongs to same submessage MUST be inserted contiguously.
    // For example, dalvik_vm_usejit must be inserted directly after dalvik_vm_heapsize, otherwise
    // if hack_in attempts to be inserted before dalvik_vm_usejit, value of usejit isn't added as expected.
    bool insertField(ProtoOutputStream* proto, const std::string& name, const std::string& value);

    // Starts a new message field proto session.
    void startSession(ProtoOutputStream* proto, const std::string& name);

    // Ends the previous message field proto session.
    void endSession(ProtoOutputStream* proto);
private:
    Table* mTable;
    std::string mPreviousField;
    std::stack<uint64_t> mTokens;
    std::map<std::string, Message*> mSubMessages;
};

#endif  // INCIDENT_HELPER_UTIL_H