Raw Strings in Java - for regex in particular

2019-01-08 11:18发布

问题:

Is there any way to use raw strings in Java (without escape sequences)?

(I'm writing a fair amount of regex code and raw strings would make my code immensely more readable)

I understand that the language does not provide this directly, but is there any way to "simulate" them in any way whatsoever?

回答1:

No, there isn't.

Generally, you would put raw strings and regexes in a properties file, but those have some escape sequence requirements too.



回答2:

This is a work-around if you are using eclipse. You can automatically have long blocks of text correctly multilined and special characters automatically escaped when you paste text into a string literal

"-paste here-";

if you enable that option in window→preferences→java→Editor→Typing→"Escape text when pasting into a string literal"



回答3:

I use Pattern.quote. And it solves the problem of the question. Thusly:

Pattern pattern = Pattern.compile(Pattern.quote("\r\n?|\n"));

The quote method returns a string that would match the provided string argument, which the return string is the properly quoted string for our case.



回答4:

No (quite sadly).



回答5:

Have the raw text file in your class path and read it in with getResourceAsStream(....)



回答6:

( Properties files are common, but messy - I treat most regex as code, and keep it where I can refer to it, and you should too. As for the actual question: )

Yes, there are ways to get around the poor readability. You might try:

String s = "crazy escaped garbage"; //readable version//

though this requires care when updating. Eclipse has an option that lets you paste text in between quotes, and the escape sequences are applied for you. The tactic would be to edit the readable versions first, and then delete the garbage, and paste them in between the empty quotes "".


Idea time:

Hack your editor to convert them; release as a plugin. I checked around for plugins, but found none (try searching though). There's a one-to-one correspondence between escaped source strings and textbox text (discounting \n, \r\n). Perhaps highlighted text with two quotes on the ends could be used.

String s = "##########
#####";

where # is any character, which is highlighted - the break is treated as a newline. Text typed or pasted within the highlighted area are escaped in the 'real' source, and displayed as if they were not. (In the same way that Eclipse escapes pasted text, this would escape typed text, and also display it without the backslashes.) Delete one of the quotes to cause a syntax error if you want to edit normally. Hmm.



回答7:

Note : As of today, not available. Probably I'll edit this answer again whenever the feature release.

There is an ongoing proposal to introduce Raw Strings in Java. They actually much useful in the cases of regex.

Example 1: A regular expression string that was coded as

  System.out.println("this".matches("\\w\\w\\w\\w"));

may be alternately coded as

System.out.println("this".matches(`\w\w\w\w`));

since backslashes are not interpreted as having special meaning.

Example2 : A multi lines String literal with foreign language appends.

A multiple line string that was coded as 
    String html = "<html>\n" +
                "    <body>\n" +
                "         <p>Hello World.</p>\n" +
                "    </body>\n" +
                "</html>\n";

may be alternately coded as

 String html = `<html>
                       <body>
                           <p>Hello World.</p>
                       </body>
                   </html>
                  `;

which avoids the need for intermediate quotes, concatenation and explicit newlines.

Hopefully we can expect the release soon.



回答8:

String#getBytes() exposes a copy of the internal byte array contained in every single String object which actually contains the 16-bit UTF-16 encoded String - the byte array will contain the same string converted to match the platform's default charset. What I'm saying is that I think this is as close to "raw" string as you can ever get in Java.



回答9:

You could write your own, non-escaped property reader and put your strings in a resource file.



回答10:

I personally consider regex strings data and not code, so I don't like them in my code--but I realize that's impractical and unpopular (Yes, I realize it, you don't have to yell at me).

Given that there is no native way to do this, I can come up with two possibilities (well, three but the third is, umm, unnatural).

So my personal preference would be to just parse a file into strings. You could name each entry in the file and load them all into a hash table for easy access from your code.

Second choice, create a file that will be pre-processed into a java interface; it could escape the regex as it does so. Personally I hate code generation, but if the java file is 100% never human edited, it's not too bad (the real evil is generated files that you are expected to edit!)

Third (tricky and probably a bad idea): You might be able to create a custom doclet that will extract strings from your comments into a text file or a header file at compile time, then use one of the other two methods above. This keeps your strings in the same file in which they are being used. This could be really hard to do correctly, and the penalties of failure are extreme, so I wouldn't even consider it unless I had an overwhelming need and some pretty impressive talent.

I only suggest this because comments are free-form and things within a "pre" tag are pretty safe from formatters and other system uglies. The doclet could extract this before printing the javadocs, and could even add some of the generated javadocs indicating your use of regex strings.

Before downvoting and telling me this is a stupid idea--I KNOW, I just thought I'd suggest it because it's interesting, but my preference as I stated above is a simple text file...



回答11:

No. But there's an IntelliJ plug-in that makes this easier to deal with, called String Manipulation.

IntelliJ will also automatically escape a string pasted into it. (As @Dread points out, Eclipse has a plug-in to enable this.)