我有一个这样的令牌：
<delim2=((?{ $MATCH{delim} }))>
我想要发生的是 delim2 捕获并设置为 delim 的值。当我运行此命令时，设置了 delim2，但捕获从未完成。我认为这是我的推理中的错误：我试图将这种形式：
<ALIAS= ( PATTERN )> Match pattern, save match in $MATCH{ALIAS}
和这种形式： (?{ MATCH{delim} }) 链接到类似的东西，
<ALIAS= ( (?{MATCH{delim}) )> Matches the value of $MATCH{delim} save to $MATCH{delim2}
但这似乎并不有效的。我可以验证我的原始令牌是否有效
I've got a token like such:
<delim2=((?{ $MATCH{delim} }))>
and what I want to happen is for delim2 to capture and be set to the value of delim. When I run this, delim2 is set, but the capture is never done. I think this is an error in my reasoning: I'm trying to chain this form:
<ALIAS= ( PATTERN )> Match pattern, save match in $MATCH{ALIAS}
and this form: (?{ MATCH{delim} }) into something like this
<ALIAS= ( (?{MATCH{delim}) )> Matches the value of $MATCH{delim} save to $MATCH{delim2}
but this simply doesn't seem valid. I can verify my original token works <delim2=((?{ die $MATCH{delim} }))> will die with the value, and, if I hard code it, I get the right capture and everything works <delim2=(')>? So how do I go about achieving sane results, while having a dynamic pattern?
发布评论
评论（1）
(?{ $MATCH{delim} })不会断言$MATCH{delim}出现在输入中；只是它是一个真正的价值。 Regexp::Grammars 应该有一个像 perl\k这样的“命名反向引用”结构，但它没有（并且你不能使用\k> 因为 Regexp::Grammars 将其结果存储在完全不同的地方）。你可以做一些类似的事情
，虽然很可怕，但在测试中似乎有效。或者你可以放弃并转到 Parse::RecDescent ，它对此类事情有更好的支持。或者你可以开始破解 R::G。
(?{ $MATCH{delim} })doesn't assert that$MATCH{delim}appears here in the input; only that it's a true value. Regexp::Grammars should have a "named-backreference" construct like perl\k<NAME>but it doesn't (and you can't use\k<NAME>because Regexp::Grammars stores its results somewhere entirely different).You could do something like
which is horrible but seems to work in testing. Or you could give in and go to Parse::RecDescent which has better support for this kind of thing. Or you could start hacking on R::G.