Forum: Redcloth Now what?

> $ rake test

On Sat, Mar 15, 2008 at 08:53:51AM -0400, Jason Garber wrote:
> > 392 tests, 392 assertions, 0 failures, 0 errors

Oh wow splendid!  Well done, Jason.  Crazy great work.

What Ruby version and Ragel version are you using?  I'm getting a
bunch of failed tests with:

  NoMethodError: private method `gsub' called for nil:NilClass
    ./test/../lib/superredcloth/html.rb:141:in `link'

Once I can get this working, I'll put up some release gems and
we can get this out to people.  Seriously, way to go!

_why

sweet.
i seem to remember there were 8 passing

On Sat, Mar 15, 2008 at 1:53 PM, Jason Garber <jg@jasongarber.com> 
wrote:
>  > 392 tests, 392 assertions, 0 failures, 0 errors

!!!

Awesome work Jason, kudos!

This still requires the Ragel trunk to generate the c files?

JS

Yes, t requires Ragel trunk.  I had to use a leaving action at the
end of a scanner pattern in a few cases, and that's only supported in
Ragel 6.1, which hasn't been released yet.  Wish it didn't have to be
that way, but it did.  I've asked Adrian when we can expect 6.1.

Let's talk a bit about interface.  I'm thinking we shouldn't break
Rails...

>             textilized.to_html
>           end
>         end


so maybe we need a dummy second parameter to RedCloth.new?  Ugh.

We had talked earlier about releasing this as RedCloth 4.0.  I'm all
for that, but how shall we go about warning people that it's a major
change?  We've got some people using RedCloth 3.0.4 with hard breaks
off, some people with it on and broken, some with it on and hacked to
work.  Maybe we should emit a warning if the second parameter isn't
passed [:hard_breaks] or #hard_breaks= isn't set to true?  At least
people using Rails' #textilize would never see it (thankfully, Rails
always had :hard_breaks on).

Should we switch the class name to RedCloth now?  Should we copy the
branch to trunk?

Jason

On Mon, Mar 17, 2008 at 12:26:14PM -0400, Jason Garber wrote:
> Yes, t requires Ragel trunk.

Perfeck, that did it.

> Let's talk a bit about interface.  I'm thinking we shouldn't break Rails...
> 
> so maybe we need a dummy second parameter to RedCloth.new?  Ugh.

Oh, that sounds good.  I doubt anyone is using hard breaks off any
more.  Maybe we get RedCloth 4 out and then whoever's got problems
will come out in the open and I think it'll be clear what to do.

> We had talked earlier about releasing this as RedCloth 4.0.  I'm all for 
> that, but how shall we go about warning people that it's a major change?  

Well, we could release it as a gem with the lowercase `redcloth`
name.  Then no one will auto-upgrade.  And perhaps there is a way to
throw a warning in extconf.rb that will require that RedCloth 3 be
uninstalled.

Either that or we could just pre-release on my gem server and just
put out 4.0 once we feel good enough about the upgrade.  Whatever
you like, really.

> Should we switch the class name to RedCloth now?  Should we copy the branch 
> to trunk?

Just finished that.  Seriously, this is tremendous, Jason!!

_why

Congrats Jason - great work!!!

Just one thing.... can we change the thread title - all I can think of
are boghogs! ;)

Tony

I also feel concerned about the ragel dependency. Please include
generated C files to avoid every redcloth user to install/compile
ragel.

The upgrade path for 3.x users should be as smooth as possible, but if
we cannot avoid some glitches, people upgrading there gems are usually
warned that they will have to do some small changes to their apps.

2008/3/17, Johan Sørensen <johan@johansorensen.com>:

With the gem, Ragel is not required.  Just

   gem install RedCloth --source http://code.whytheluckystiff.net

and you're done.