[Regex]Greta不支持“Named Groups”特性

微软研究院出品的 Greta ( http://research.microsoft.com/projects/greta/ )是一款非常优秀的 正则表达式模板类库 , GRETA 的匹配速度比 boost(http://www.boost.org) 正则表达式库大约快 7 倍,比 ATL7 的 CATLRegExp 快 10 倍。你可以通过 http://research.microsoft.com/projects/greta/regex_perf.html 了解他的超快速度。

但是他并不支持 “Named Groups” 特性,这在 dotNet Regex 引擎中是提供的。

在 Greta 花费了半天时间总是试验不出来这个特性后,只好去问了维护这个引擎的微软工程师。下面是来往信件:

** 发件人 ** ** : ** Chris Quirk [mailto:chrisq…(at)microsoft.com]
** 发送时间 ** ** : ** 2004 年 9 月 21 日 2:59
** 收件人 ** ** : ** Yun Zheng; Ashish Aggarwal
** 主题 ** ** : ** RE: greta gripe:for groupname syntax

Greta doesn’t support named groups; instead, it uses the more standard approach of numbering them.

So you should check your regex to

"^(\\w+):(\\w+)"

Then you access the first group as capture number 1 and the second as capture number 2, as in Perl, etc.

As far as I know, named captures are a peculiarity of the .NET Regex engine.


** From: ** Yun Zheng
Sent: Sunday, September 19, 2004 9:27 PM
To: Ashish Aggarwal; Chris Quirk
Subject: greta gripe:for groupname syntax

Hi!

I'm using greta now.It's very perfect,but I have a question:

In C#,I can do this:

[C#]
Regex r = new Regex("^(?\\w+):(?\\w+)");
Match m = r.Match("Section1:119900");

But, in VC7+greta-2[1].6.4,I run this codes:

std::string strDest("Section1:119900");
regex::rpattern pat("^(?\\w+):(?\\w+)",
regex::NOCASE|regex::GLOBAL|regex::EXTENDED);
regex::match_results::backref_type br
= pat.match( strDest, results );

when run to construct rpattern,I get bad_regexpr( "bad extension sequence" ) exception.

How does Greta support groupname syntax?

Thanks!

Yun Zheng

Published At
Categories with Web编程
Tagged with
comments powered by Disqus