如何在TextX语法中解析并保留C风格单行注释至模型？

我正尝试构建支持C/C++风格单行注释（例如//this is a comment）的TextX语法。测试阶段编写了如下语法规则：

Ledger:entries+=Entry;

Entry: Comment;

Comment:/\/\/.*$/;

但解析包含两行注释的文件：

//comment 1
//comment 2

时，出现错误：

textx.exceptions.TextXSyntaxError: file.txt:3:1: Expected Comment => 'omment 2 *'

解析器调试输出如下：

*** PARSING MODEL ***
>> Matching rule Model=Sequence at position 0 => *//comment
   >> Matching rule Ledger=Sequence in Model at position 0 => *//comment
      >> Matching rule __asgn_oneormore=OneOrMore[entries] in Ledger at position 0 => *//comment
         ?? Try match rule Comment=RegExMatch(\/\/.*$) in __asgn_oneormore at position 0 => *//comment
         ?? Try match rule Comment=RegExMatch(\/\/.*$) in __asgn_oneormore at position 0 => *//comment
' at 0 => '*//comment 1*'ent 1
         ?? Try match rule Comment=RegExMatch(\/\/.*$) in __asgn_oneormore at position 13 => omment 1 *//comment
' at 13 => 'omment 1 *//comment 2*'
         ?? Try match rule Comment=RegExMatch(\/\/.*$) in __asgn_oneormore at position 26 => omment 2 *
         -- NoMatch at 26
         -- NoMatch at 26
      <<- Not matched rule __asgn_oneormore=OneOrMore[entries] in __asgn_oneormore at position 0 => *//comment
   <<- Not matched rule Ledger=Sequence in Ledger at position 0 => *//comment
<<- Not matched rule Model=Sequence in Model at position 0 => *//comment

当为Entry规则添加FLOAT选项：

Ledger:entries+=Entry;

Entry: Comment | FLOAT;

Comment:/\/\/.*$/;

并在文件中添加数值：

//comment 1
//comment 2
7

此时不再抛出异常，但注释未被解析进模型，解析器调试输出如下：

>> Matching rule Model=Sequence at position 0 => *//comment
   >> Matching rule Ledger=Sequence in Model at position 0 => *//comment
      >> Matching rule __asgn_oneormore=OneOrMore[entries] in Ledger at position 0 => *//comment
         >> Matching rule Entry=OrderedChoice in __asgn_oneormore at position 0 => *//comment
            ?? Try match rule Comment=RegExMatch(\/\/.*$) in Entry at position 0 => *//comment
            ?? Try match rule Comment=RegExMatch(\/\/.*$) in Entry at position 0 => *//comment
' at 0 => '*//comment 1*'omment 1
            ?? Try match rule Comment=RegExMatch(\/\/.*$) in Entry at position 13 => omment 1 *//comment
' at 13 => 'omment 1 *//comment 2*'
            ?? Try match rule Comment=RegExMatch(\/\/.*$) in Entry at position 26 => omment 2 *7
            -- NoMatch at 26
            -- NoMatch at 26
            ?? Try match rule FLOAT=RegExMatch([+-]?(\d+(\.\d*)?|\.\d+)([eE][+-]?\d+)?(?<=[\w\.])(?![\w\.])) in Entry at position 0 => *//comment
            ++ Match '7' at 26 => 'omment 2 *7*'
         <<+ Matched rule Entry=OrderedChoice in Entry at position 27 => mment 2 7*
         >> Matching rule Entry=OrderedChoice in __asgn_oneormore at position 27 => mment 2 7*
            ?? Try match rule Comment=RegExMatch(\/\/.*$) in Entry at position 27 => mment 2 7*
            ?? Try match rule Comment=RegExMatch(\/\/.*$) in Entry at position 27 => mment 2 7*
            -- NoMatch at 27
            -- NoMatch at 27
            ?? Try match rule FLOAT=RegExMatch([+-]?(\d+(\.\d*)?|\.\d+)([eE][+-]?\d+)?(?<=[\w\.])(?![\w\.])) in Entry at position 27 => mment 2 7*
            -- NoMatch at 27
         <<- Not matched rule Entry=OrderedChoice in Entry at position 27 => mment 2 7*
      <<+ Matched rule __asgn_oneormore=OneOrMore[entries] in __asgn_oneormore at position 27 => mment 2 7*
   <<+ Matched rule Ledger=Sequence in Ledger at position 27 => mment 2 7*
   ?? Try match rule EOF in Model at position 27 => mment 2 7*
<<+ Matched rule Model=Sequence in Model at position 27 => mment 2 7*

问题原因 #

核心问题在于正则表达式/\/\/.*$/的行为：

• $默认匹配整个输入的结尾，而非单行结尾。TextX默认不启用多行模式，导致第一个注释的正则会匹配从//开始到整个文件末尾的所有内容，包括后续的换行和第二行注释。
• 当解析器尝试匹配第二个Comment时，指针已经跳到文件末尾附近，无法找到新的//开头，因此报错。
• 添加FLOAT后，解析器跳过了无法匹配的注释内容，直接匹配到最后的数值，所以不报错但注释被忽略。

解决方案 #

修改Comment的正则表达式，明确匹配到单行末尾而非整个输入末尾，有两种可行方案：

方案1：启用多行模式 #

使用(?m)标记开启多行模式，让$匹配每行的结束位置：

Ledger: entries+=Entry;

Entry: Comment | FLOAT;

Comment: /(?m)\/\/.*$/;

方案2：直接匹配非换行字符 #

用[^\r\n]匹配//之后所有非换行的字符，直观限制在单行内：

Ledger: entries+=Entry;

Entry: Comment | FLOAT;

Comment: /\/\/[^\r\n]*/;

补充说明 #

• 两种方案都能让每个注释仅匹配当前行的内容，解析器会自动处理空白字符（包括换行），匹配完单行注释后，会自动跳过换行符继续匹配下一个Entry。
• 测试时，两行注释会被识别为两个独立的Comment实例，存入entries列表；包含注释和数值的文件，两者都会被正确解析并存入模型。

内容的提问来源于stack exchange，提问作者kabla002