PyParsing解析Solidity方法签名时出现无限递归错误求助

问题描述 #

我编写了用于解析Solidity函数方法签名的PyParsing语法，代码如下：

from pyparsing import Word, alphas, alphanums, oneOf, Group, Forward, ZeroOrMore, Optional, delimitedList

solidity_type = Forward()
primitive = oneOf("address uint uint8 uint16 uint32 uint64 uint128 uint256 int int8 int16 int32 int64 int128 int256 bool bytes bytes1 bytes4 bytes32")
array = solidity_type + "[]"
tuple_type = "(" + delimitedList(solidity_type) + ")"
solidity_type <<= (primitive | array | tuple_type)
function_name = Word(alphas, alphanums + "_")
arguments = Optional(delimitedList(solidity_type))
function_signature = function_name + "(" + arguments + ")"

sig = "diamondCut((address,uint8,bytes4[])[],address,bytes)"
parsed_sig = function_signature.parseString(sig)
print(parsed_sig.dump())

运行时触发{RecursionError}maximum recursion depth exceeded错误，测试array.parseString("(bytes4)[]")也会出现同样问题。尝试过用Group包裹表达式、调整选项顺序、将array设为独立Forward，都无法解决，需要修改代码解析目标签名。

解决方法 #

递归超限的原因是array和tuple_type都直接依赖solidity_type，而solidity_type又包含这两者，导致解析器在尝试匹配时陷入无限递归。解决思路是调整语法定义的优先级，让解析器优先匹配最底层的类型，再逐步向上匹配复合类型，同时用Group明确结构避免歧义：

from pyparsing import Word, alphas, alphanums, oneOf, Group, Forward, Optional, delimitedList

solidity_type = Forward()
# 先定义最基础的原始类型
primitive = oneOf("address uint uint8 uint16 uint32 uint64 uint128 uint256 int int8 int16 int32 int64 int128 int256 bool bytes bytes1 bytes4 bytes32")
# 定义元组：用Group包裹，确保内部结构被正确识别
tuple_type = Group("(" + delimitedList(solidity_type) + ")")
# 定义数组：优先匹配原始类型或元组后再匹配数组后缀，避免递归循环
array = Group((primitive | tuple_type) + "[]")
# 最后把所有类型选项按优先级加入solidity_type
solidity_type <<= (primitive | tuple_type | array)

function_name = Word(alphas, alphanums + "_")
arguments = Optional(delimitedList(Group(solidity_type)))
function_signature = Group(function_name + "(" + arguments + ")")

sig = "diamondCut((address,uint8,bytes4[])[],address,bytes)"
parsed_sig = function_signature.parseString(sig)
print(parsed_sig.dump())

关键修改点 #

• 调整类型匹配顺序：让解析器先匹配primitive，再匹配tuple_type，最后匹配array，避免一开始就进入复合类型的递归匹配
• 用Group包裹复合类型：明确元组和数组的结构，让解析结果更清晰，同时帮助解析器识别完整的复合类型单元
• 数组定义优化：数组不再直接依赖solidity_type，而是基于已匹配的primitive或tuple_type，打破无限递归的循环

运行修改后的代码，就能成功解析目标签名，输出的结构化结果也能清晰展示函数名和每个参数的类型结构。

内容的提问来源于stack exchange，提问作者aram10