Can actual Perl Regular Expressions be implemented in Java via existing library?

Question

Is there an existing robust Java library which implements a fairly substantial subset of Perl regular expression syntax?

Background:

I wish to implement a file renamer where renaming is done using Perl regular expressions.

The trick is that the project containing said renamer as a component is, currently, 100% in Java in Windows.

NOTES:

I am setting aside the obvious workaround of "install Strawberry Perl; write the whole renamer in Perl (or teach Perl to the developer doing Java coding), pass a list of files to rename to the Perl renamer script via a system call from Java" as too crafty and too obvious :)

Also, please don't offer comparisons of how Java's latest and greatest RegEx engine is already good enough to do most of what Perl RegEx does - I'm fairly aware of its functionality (and can google); thus I already know that said statement might even be true; it is, however, irrelevant to my interest in seeing real Perl RegEx syntax implemented as a Java library.

How much is "fairly substantial"? The rest of your question reads like that would be in the 95% and above ballpark. — innaM, Dec 15 '09 at 20:27
At the very least, you won't be able to get `(?{ code })` support without a Perl interpreter. — Anon., Dec 15 '09 at 20:28
Can you access PCRE (perl compatible regex) through JNI (Java Native Interface)? Not much of a Java hacker, so I don't know what's involved. — daotoad, Dec 15 '09 at 20:35
@Manni - since, as the questions makes obvious, there are fairly doable workarounds, it is more of a "searching for a theoretically cool thing"; and thus I don't have a firm threshold. The more coverage the better. — DVK, Dec 15 '09 at 20:46
@Kevin - not being familiar with JNI at all, could you please explain why it is the worst? Thx — DVK, Dec 15 '09 at 20:49
@DVK You would be writing C code that interfaces with both the JVM and the Perl engine. So you would need convert Java's String object into a char pointer, then convert that into a Perl SV structure, call the regexp function, get another SV back, then convert that back to a char pointer, and finally back into a String object. It could be done, but I do not think it would be worth the trouble. — Kevin Panko, Dec 15 '09 at 20:58
Looks kind of like this... http://cpansearch.perl.org/src/PATL/Inline-Java-0.52/Java/JNI.xs Guess someone already did the dirty work for you. — Kevin Panko, Dec 15 '09 at 21:12

score 7 · Accepted Answer · answered Dec 15 '09 at 20:35

7

Have you tried JRegex? It boasts Perl 5.6 compatibility and doesn't have the worries of linking PCRE in directly.

answered Dec 15 '09 at 20:35

pestilence669

5,698
1
23
35

1

Thanks! Would you happen to know how (a) robust and (b) fast is it? – DVK Dec 15 '09 at 20:48
1

~10% faster than java.util.regex, much slower than native Perl. – pestilence669 Dec 15 '09 at 20:53
Should be fairly robust. Haven't tested it for named grouping or anything. – pestilence669 Dec 15 '09 at 20:56
Sux re: performance. I'm probably just going to batch off the list of strings and regexes and call a Perl script on the batch then... +1 nevertheless for excellent suggestion! – DVK Dec 16 '09 at 16:04

score 5 · Answer 2 · edited May 23 '17 at 12:30

5

For maximum Perl compatibility you would need to actually use Perl. You can do that using Inline::Java::Callback, which is distributed as part of the Inline::Java module.

See also: How can I call Perl from Java?

edited May 23 '17 at 12:30

Community

1
1

answered Dec 15 '09 at 21:02

Michael Carman

30,628
10
74
122

score 3 · Answer 3 · answered Dec 15 '09 at 21:37

3

The library you are looking for is the Apache ORO library, it is specifically implemented to handle Perl5 regular expressions.

answered Dec 15 '09 at 21:37

rsp

23,135
6
55
69

Just as a note, the stable version only supports Perl5.003 as per their docs. But a good option nonetheless! – DVK Dec 16 '09 at 16:02

Can actual Perl Regular Expressions be implemented in Java via existing library?

3 Answers3