Hi guys,
I've been thinking about this today and I think the best is for me to
continue writing the XWiki parser/Sinks inside of the XWiki code base
since this will allow me the best speed of development. In practice
the infra classes I need are minimal: a Sink interface, a Parser
interface and that's all so doing my dev inside or outside of Doxia
itself doesn't matter much.
Also I'd really like to have a wrapper in XWiki so that xwiki core
code doesn't directly depend on Doxia but instead depends on some
XWiki interfaces. Thus I need to do that wrapper work anyway.
What I'll do thus is move the code I started in the maven sandbox in
XWiki's code base and refactor the sandbox implementation I have done
to use the xwiki sources as a binary dependency. It should be quite
easy to make a bridge from the direction xwiki --> doxia since the
xwiki implementation will be richer than the doxia one.
In parallel I'll wait for the 1.0 release and then start contributing
the changes I have done to Doxia.
I'll start doing a POC right now and see if this can work fine.
Thanks
-Vincent
On Jan 23, 2008, at 8:44 PM, Dennis Lundberg wrote:
Here's my view on doxia right now. I understand Vincent M's need for
somewhere to commit his stuff to, so let's try to solve that as
quickly as possible. In order to do that I propose yet another
alternative
D) Create an xwiki-branch from the current trunk, where Vincent can
add him xwiki implementation. Including changes to the core.
My reasoning behind this is that I feel that we are in risk of
loosing control of doxia, if we don't create a release plan soon.
However that is the topic of another e-mail (which I will start
soon). And I don't want that discussion to stop Vincent M's xwiki
momentum.
Would you be OK with that solution Vincent M ?
If it turns out that nothing much happens on trunk after that, it
should be easy enough to merge Vincent's changes back to trunk if we
want to.
Lukas Theussl wrote:
I am for B) because that's what (I thought) we had agreed upon in
the past. I always understood that doxia is still alpha, that I was
allowed to make changes on trunk, and that we would stabilize the
API with the first beta release. Also, as you say, we have already
made some breaking changes to current trunk (eg DOXIA-137,
DOXIA-155), and there are a number of open bugs whose fix will
entail some breakig changes (since in practice people use hacks to
workaround those bugs). However, I haven't tested recently if
beta-1 works with the site plugin (it didn't when I tried a while
ago).
Anyway, in order to push things forward, I would agree to do a 1.0
release from the current alpha branch. I would vote with a belly-
ache though because I feel that for a 1.0-final label there are
some important things missing (DOXIA-123, DOXIA-138, DOXIA-145).
OTOH some of the things that are fixed in beta-1 could well go into
a 1.0 release, eg all the work Dave Syer has done on the confluence
module.
I'd first like to hear Dennis' opinion though, he should know best
how things work with various Maven plugins. I haven't had much time
for doxia recently and even though I still plan to do things, I
don't think I can contribute much in the near future...
Cheers,
-Lukas
Vincent Massol wrote:
On Jan 22, 2008, at 11:31 PM, Jason van Zyl wrote:
I think you should cut the 1.0 and then move on. You will
completely hose the site plugin will you not?
I'm all for changes. I want them but this can't happen before a
1.0.
I'm fine with either of these options:
A) Create a 1.0 branch so that work can be finished for 1.0 and
make trunk 1.1/2.0. I'll commit on trunk then.
B) Commit on trunk right now since Maven uses alphas of doxia and
there's an alpha branch for maven. And since I believe trunk is
already not working with maven.
C) Leave trunk for 1.0 and create a 1.1/2.0 branch so that I can
commit there while waiting for the 1.0 version to be released and
then merged back everything to trunk
My preference goes to A) but I'm ok also with C). I know Lukas
prefers B).
Just let me know what I should do. Only thing that's sure is that
I don't have close to enough knowledge/bandwidth to help with the
1.0 release. The other sure thing is that I need a place to
commit ASAP since I'm stuck right now and I'll loose my momentum
if I don't have a place to commit. I really need to make quick
progress to be aligned with xwiki's needs.
Thanks
-Vincent
We have to make something for the site plugin to use.
On 22-Jan-08, at 5:02 AM, Vincent Massol wrote:
Hi,
I'm just starting development on Doxia and I need to understand
the strategy since I need to make lots of breaking changes to
accomodate the XWiki use cases (I'm planning to replace XWiki's
rendering engine from Radeox to Doxia).
I've just talked to VincentS and Lukas and the consensus I got
was that I could just commit away on trunk and make breaking
changes since:
a) 1.0 has not been released yet and the API is not final and
thus it's the right time to do this
b) there an alpha branch that the current Doxia users (like
Maven itself) can use id they don't want to upgrade to the
changes made.
So I'd like agreement that I can make my changes on trunk.
Namely I'd like to implement these (and more to come):
* http://jira.codehaus.org/browse/DOXIA-203: Add support for
level 6 sections and generalize Sink API for sections
* http://jira.codehaus.org/browse/DOXIA-202: Add an API for
getting a tree of syntax blocks
* http://jira.codehaus.org/browse/DOXIA-165: Add support for
macros
* http://jira.codehaus.org/browse/DOXIA-164: Add support for
strikethroughs
* http://jira.codehaus.org/browse/DOXIA-163: Add support for
underscores
* http://jira.codehaus.org/browse/DOXIA-204: Add generic
parameters support to Figure and Link events
* http://jira.codehaus.org/browse/DOXIA-205: Add new standard
parameters to figure sink API
* http://jira.codehaus.org/browse/DOXIA-206: Add new standard
parameters to link sink API
* http://jira.codehaus.org/browse/DOXIA-207: Add events for
recognizing words in the Sink API
WDYT?
Thanks
-Vincent
Thanks,
Jason
----------------------------------------------------------
Jason van Zyl
Founder, Apache Maven
jason at sonatype dot com
----------------------------------------------------------
A man enjoys his work when he understands the whole and when he
is responsible for the quality of the whole
-- Christopher Alexander, A Pattern Language
--
Dennis Lundberg