cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <>
Subject Re: [proposal] Cocoon documentation system
Date Sat, 15 Jan 2005 19:30:55 GMT
Geoff Howard wrote:
> On Sat, 15 Jan 2005 09:42:17 +0100, Reinhard Poetz <> wrote:
> ...
>>>>My first goal is
>>>>reusing existing functionality (SVN, Forrest, static web pages). Writing
>>>>the missing pieces (a small web application to edit docs online) shouldn't
>>>>be too difficult - especially using CocoonForms (HTMLArea with HTMLCleaning
>>>>convertor are shining here!) and Flowscript.
>>>>At the end of the document you find a roadmap. It also contains some parts
>>>>where I *really* need some help by others:
>>>>- component (Java) that reads from and commits to SVN - should work over
>>>>  WebDAV, shouldn't it?
>>>>  I know there are some specialist in this community who could help out
>>>>  here (Unico, Gianugo, Guido)
>>>We have a huge team of specialists on the dev list.
>>Yep. I hope that I can convince at least one of them to help me. Actually I need
>>some Java code ;-) So here I ask officially: Who volunteers to write the Java
>>code that communicates with the SVN repository (read and commit)?
> You may find this more problematic than you'd expect.  I follow the
> subclipse list a little and have the impression that pure-java library
> development against SVN lags behind because it's difficult to keep up
> with the evolving standard.  I should have thought that the rate of
> evolution has significantly slowed, but that is not the impression I
> get.  Subclipse for instance uses jni for the low level svn operations
> so it can go against the original native libraries.
> On the other hand, there are some attempts to provide pure java svn
> libraries (one is called javaSVN I think?) so if these are usable in
> practice, the work could be trivial.

People, again, let's be brave and get this of this silly "pure java" 
nonsense. the JNI connector works. Today! there is no need to play 
catchup with the SVN protocol (which is a proprietary extension/subset 
of WebDAV/DeltaV and very painful, I'll tell you)

Besides, it's not that we have to distribute such a system.

Let's do the simplest thing that can possibly work, that SWIG-based API 
is very very bad in the eyes of a java programmer, but well, it's way 
better than rewriting it from scratch.

one day, hopefully, somebody will create a JSR170 wrapper for SVN... but 
until then, well, let's do what we can.


View raw message