[ragel-users] problems with java generation in ragel 6.x trunk

[Adrian Thurston]

For java code set eof to -1 to pacify ragel. Since java doesn't have pointers I use array offsets and -1 is used to represent null.

-Adrian
-----Original Message-----
From: Stephen Bannasch <stephen.bannasch at deanbrook.org>

Date: Mon, 24 Mar 2008 13:47:19 
To:ragel-users at googlegroups.com
Subject: [ragel-users] problems with java generation in ragel 6.x trunk



At 12:05 PM -0400 3/24/08, Adrian Thurston wrote:
>Stephen Bannasch wrote:
>> I'd like to be able to create a ragel file that will compile
>> in ragel 5.x and 6.x but didn't see any macro capability in Ragel.
>
>I'm not sure how much of Hpricot is affected by the 6.0 changes, but my
>feeling is that a compatibility layer isn't worth the effort. The
>changes to the operators > % :> :>> and <: could require some minor
>refactoring of ragel definitions.

In the C ragel source I only needed to:

  convert tokstart, tokend => ts, te

and failing tests went from 2 to 0 ;-)

In the Java ragel source I did the above and added:

  int eof=0;

There are a couple of failing tests but I don't think it has anything to do with ragel.

Regarding: changes to operators > % :> :>>

I'll take a look at where those are used and if needed try and make sure the test coverage includes problems that might occur.

>People normally side-step ragel version issues by distributing the files
>generated by ragel.

That makes sense -- distribute both *.c and *.java files (actually the Java class fles can be bundled into a Jar and the only dependency with be on JVM version).

Right now I'm working on the hpricot source release and on the redcloth source release -- fixing bugs, seeing about

Redcloth needs ragel trunk.

I think the changes I made to (Hpricot which I'm testing with ragel trunk) will also work fine with ragel v6.0 and 6.trunk.

I could create dependency on ragel >= 6.0 for compiling the ragel source for hpricot.

Now that 6.0 seems to be distributed widely that might be fine.