| 1 | #pragma once |
| 2 | |
| 3 | #include <set> |
| 4 | #include <memory> |
| 5 | |
| 6 | #include <Core/Defines.h> |
| 7 | #include <Core/Types.h> |
| 8 | #include <Parsers/IAST.h> |
| 9 | #include <Parsers/TokenIterator.h> |
| 10 | |
| 11 | |
| 12 | namespace DB |
| 13 | { |
| 14 | |
| 15 | namespace ErrorCodes |
| 16 | { |
| 17 | extern const int TOO_DEEP_RECURSION; |
| 18 | extern const int LOGICAL_ERROR; |
| 19 | } |
| 20 | |
| 21 | |
| 22 | /** Collects variants, how parser could proceed further at rightmost position. |
| 23 | */ |
| 24 | struct Expected |
| 25 | { |
| 26 | const char * max_parsed_pos = nullptr; |
| 27 | std::set<const char *> variants; |
| 28 | |
| 29 | /// 'description' should be statically allocated string. |
| 30 | void add(const char * current_pos, const char * description) |
| 31 | { |
| 32 | if (!max_parsed_pos || current_pos > max_parsed_pos) |
| 33 | { |
| 34 | variants.clear(); |
| 35 | max_parsed_pos = current_pos; |
| 36 | } |
| 37 | |
| 38 | if (!max_parsed_pos || current_pos >= max_parsed_pos) |
| 39 | variants.insert(description); |
| 40 | } |
| 41 | |
| 42 | void add(TokenIterator it, const char * description) |
| 43 | { |
| 44 | add(it->begin, description); |
| 45 | } |
| 46 | }; |
| 47 | |
| 48 | |
| 49 | /** Interface for parser classes |
| 50 | */ |
| 51 | class IParser |
| 52 | { |
| 53 | public: |
| 54 | /// Token iterator augmented with depth information. This allows to control recursion depth. |
| 55 | struct Pos : TokenIterator |
| 56 | { |
| 57 | using TokenIterator::TokenIterator; |
| 58 | |
| 59 | uint32_t depth = 0; |
| 60 | uint32_t max_depth = 1000; |
| 61 | |
| 62 | void increaseDepth() |
| 63 | { |
| 64 | ++depth; |
| 65 | if (depth > max_depth) |
| 66 | throw Exception("Maximum parse depth exceeded" , ErrorCodes::TOO_DEEP_RECURSION); |
| 67 | } |
| 68 | |
| 69 | void decreaseDepth() |
| 70 | { |
| 71 | if (depth == 0) |
| 72 | throw Exception("Logical error in parser: incorrect calculation of parse depth" , ErrorCodes::LOGICAL_ERROR); |
| 73 | --depth; |
| 74 | } |
| 75 | }; |
| 76 | |
| 77 | /** Get the text of this parser parses. */ |
| 78 | virtual const char * getName() const = 0; |
| 79 | |
| 80 | /** Parse piece of text from position `pos`, but not beyond end of line (`end` - position after end of line), |
| 81 | * move pointer `pos` to the maximum position to which it was possible to parse, |
| 82 | * in case of success return `true` and the result in `node` if it is needed, otherwise false, |
| 83 | * in `expected` write what was expected in the maximum position, |
| 84 | * to which it was possible to parse if parsing was unsuccessful, |
| 85 | * or what this parser parse if parsing was successful. |
| 86 | * The string to which the [begin, end) range is included may be not 0-terminated. |
| 87 | */ |
| 88 | virtual bool parse(Pos & pos, ASTPtr & node, Expected & expected) = 0; |
| 89 | |
| 90 | bool ignore(Pos & pos, Expected & expected) |
| 91 | { |
| 92 | ASTPtr ignore_node; |
| 93 | return parse(pos, ignore_node, expected); |
| 94 | } |
| 95 | |
| 96 | bool ignore(Pos & pos) |
| 97 | { |
| 98 | Expected expected; |
| 99 | return ignore(pos, expected); |
| 100 | } |
| 101 | |
| 102 | /** The same, but do not move the position and do not write the result to node. |
| 103 | */ |
| 104 | bool check(Pos & pos, Expected & expected) |
| 105 | { |
| 106 | Pos begin = pos; |
| 107 | ASTPtr node; |
| 108 | if (!parse(pos, node, expected)) |
| 109 | { |
| 110 | pos = begin; |
| 111 | return false; |
| 112 | } |
| 113 | else |
| 114 | return true; |
| 115 | } |
| 116 | |
| 117 | /** The same, but doesn't move the position even if parsing was successful. |
| 118 | */ |
| 119 | bool checkWithoutMoving(Pos pos, Expected & expected) |
| 120 | { |
| 121 | ASTPtr node; |
| 122 | return parse(pos, node, expected); |
| 123 | } |
| 124 | |
| 125 | virtual ~IParser() {} |
| 126 | }; |
| 127 | |
| 128 | using ParserPtr = std::unique_ptr<IParser>; |
| 129 | |
| 130 | } |
| 131 | |