From nicolas.delsaux at gmail.com  Mon Nov 13 14:53:31 2006
From: nicolas.delsaux at gmail.com (Nicolas Delsaux)
Date: Mon, 13 Nov 2006 20:53:31 +0100
Subject: [Pimki-users] register as developper for Pimki
In-Reply-To: <3ff47be90611091436p2552f639if25f068651f823c6@mail.gmail.com>
References: <20061107083038.C47305240F08@rubyforge.org>
	<3ff47be90611071523m314ede97m6ad3bdd95d3ec859@mail.gmail.com>
	<418361960611090901t55ae8eeeke0b51b9710d66868@mail.gmail.com>
	<3ff47be90611091436p2552f639if25f068651f823c6@mail.gmail.com>
Message-ID: <418361960611131153u599adf0gaabdf0b1ba155a7e@mail.gmail.com>

On 11/9/06, Assaph Mehr <assaph at gmail.com> wrote:
> Hi Nicolas,
>
>
> * Prepare the database
>
> Run 'rake migrate' from the project root. This should create and
> initialise the DB as db/dev.db3.
> (I'm betting this is the step you missed ;)

Obviously, you were right.
Well, in fact, there were also issues regarding dependancies, issues
that I solved using Locomotive (since I'm quite a Mac newbie, I had
weird issues regardin the install of sqlite-ruby package and the lack
of a valid ruby.h file).
>
> * Start the application
>
> Hopefully it should work. if not, send me the development.log.

Oh my god, it's alive, it's ALIVE :-)
>
> Looking forward to your feedback.

Nice new stylesheet.
So, now, i must do a svn update to sync with latest news, and start
coding my own features.
Besides, I''ve noticed in the instiki doc that there already exists a
footnote system, but not working like mine : the main difference is
that I add the footnote directly where its calling link appears,
allowing me not to handle those crazy numbers.
In fact, it's a port of docbook system making my footnotes look like :
My text needs a footnote<footnote>And here is the footnote</footnote>.

Do you think it could be added, since I find this system really cleaner ?
So, i must now check my suggested features and choose the one to work
on first ...
Well, the look was quite long, since dyndns seems to have borken my
nicolas-delsaux.homelinux.net website, which runs a pimki 1.8.2
containing all feature descriptions and details.

So .. I think i could start by taking a look at code syntax highlight.
Candidates highlighters are

http://rubyforge.org/projects/syntax
http://rd.cycnus.de/coderay/

I have a little preference for coderay, since it's more adapted to my
needs, which include Java, Javascript, XML, ... and Ruby of course.

So ... from what I can imagine (never forget I'm a complete Ruby
newbie), the first step would be to download my candidate (say
coderay) as a gem (like says the doc)

For the rendering, I have choose to use the method already used in my
good lod wordpress plugin :
a square bracketed tag at the begining, and the closing one at the end
and inside the code, like in
[RUBY]
 require 'coderay'

 tokens = CodeRay.scan "puts 'Hello, world!'", :ruby
 page = tokens.html :line_numbers => :inline, :wrap => :page
 puts page
[/RUBY]

>From what I've seen, i believe it should only be a rendering modification.
I guess it could be a custom chunk element ...
Sidenote : I believe this chunk system could also be used for my SF library

So, I create one more chunk with simple beta code :

# Chunk for including syntax highlighted computer code.
# This chunk uses the coderay library
class SyntacticHighlightChunk < Chunk
	SYNTAX_HIGHLIGHT_PATTERN = Regexp.new('RUBY', Regexp::MULTILINE)
	def self.pattern() SYNTAX_HIGHLIGHT_PATTERN end
	
	def resolved_text(view) 'toto' end
end

More due to my lack of skill on regexp than anything else
Besides, one should implement here a kind of proxy system like the one
mentionned in one of the ruby blogs I read ... that's it : eigenvalue
: http://eigenclass.org/hiki.rb?cmd=view&p=ruby+plugins&key=plugin

So, back to my code ...
I added my chunk to the POST_ENGINE ones and reloaded my page and
nothing happens.
Well, that may be called a bad karma.

To be continued ...

> Cheers,
> Assaph
>


-- 
Nicolas Delsaux
N'imprimez ce mail que si vous ne savez pas le lire sur l'?cran : les
?lectrons se recyclent bien, le papier, beaucoup moins bien.

From nicolas.delsaux at gmail.com  Tue Nov 14 15:03:23 2006
From: nicolas.delsaux at gmail.com (Nicolas Delsaux)
Date: Tue, 14 Nov 2006 21:03:23 +0100
Subject: [Pimki-users] register as developper for Pimki
In-Reply-To: <3ff47be90611132005i51cdd77av2b38e34737911b6f@mail.gmail.com>
References: <20061107083038.C47305240F08@rubyforge.org>
	<3ff47be90611071523m314ede97m6ad3bdd95d3ec859@mail.gmail.com>
	<418361960611090901t55ae8eeeke0b51b9710d66868@mail.gmail.com>
	<3ff47be90611091436p2552f639if25f068651f823c6@mail.gmail.com>
	<418361960611131153u599adf0gaabdf0b1ba155a7e@mail.gmail.com>
	<3ff47be90611132005i51cdd77av2b38e34737911b6f@mail.gmail.com>
Message-ID: <418361960611141203q35abb7cejd14249c9b824651b@mail.gmail.com>

On 11/14/06, Assaph Mehr <assaph at gmail.com> wrote:
> > Oh my god, it's alive, it's ALIVE :-)
>
> bwahahahaha! :-D
>
Once again, i put the list in copy, for the sake of memory.
>
> Thanks. It stretches my graphic design skills to the max....
>
I may provide here a little knowledge, since I made myself (with
obvious display bugs) my blog stylesheet.
>
> Yep. When you're ready to submit a patch just update your copy,
> resolve conflicts if any and re-run the tests. Use svn diff to create
> a patch and send it to me.

OK
>
> > Do you think it could be added, since I find this system really cleaner ?
>
> The current footnote mechanics are using textile (RedCloth) - there is
> no specific support for them. If you want to add a special chunk...

Let me guess ... If I want that chunk, I can just ... code it :-))
>
> Tried the site and got a 500 error. Let me know if you want me to take a look.

Should be corrected now.
Take a look at http://nicolas-delsaux.homelinux.net/leblog/show/Pimki+2
There is a bottom link for features proposals.
>
> To be more inline with the rest of the chunks syntax it should
> probably be something like:
>   [code[ruby]]
> at the start and a simple
>   [[code]]
> to end the section. This easily allows using other languages, like
> [code[java]]. I use the double brackets for links, and the
> [keyword[parameters]] syntax for "macros". There are already some
> special such constructs, like:
>   [bliki[thread]] - links to bliki threads
>   [>other web[page name]] - links to pages on other webs
>   [!my-macro[params...]] - macro expansion (planned :)

Suggestion included in feature !
>
> So i want to keep the same syntax. Note that I was still playing with
> the advanced syntax in the trunk, so it may change a little. Once I
> have something I like I'll make sure everything conforms.

And you'll make sure I can merge my changes with ease ?
>
> Not sure what you're after, but if you want to put plain text
> "objects" and later reference them then definitely yes.

Look :
[sf[isbn]]227363847495 for retrieving book info
[sf[title][sf[] for fetching nearby titles
[sf[person]] for a person name
All managed with chunks and page dynamic creation.
>
> You also need to be a bit careful when adding the chunk into the
> processing order. The "engine" part is the markup engine
> (textile/markdown etc). So for the syntax-highlight chunk it should be
> before the engine, and it should mask it's content from further
> processing.

I believe i can get inspired reading existing code, like, say, ...
GlossaryDefinition.
I've already notice i only have to insert my class name in the
PRE_ENGINE_ACTIONS array (in the revision class).
>
>     %r<\[\[code[([^\]]+)\]\](.*?)\[\[code\]\]>mi
>
>
> This should obviously be replaced with a call to CodeRay ;-)

In the final iteration, yes ;-)

> You can easily cache the results, as they will not change between page
> rendering (unlike links, which need to be resolved each time as the
> page may have been created/deleted etc).
>
Let's use cache in v. 2
>
> We all started there :)
>
Never forget I also begin in Rails ...

>
> I thought about chunks as plugins, but because the processing order
> for chunks is VERY important I am not sure this is the proper way to
> go. If someone already hacks the source, they can easily just add the
> extra call at the right place to add the new chunk, which is a lot
> simple and less error prone (at least until I have hundreds of people
> writing customisations....)

I'm OK with that, provided you never get that developper crowd.
>
> Check your regexp, check where you added the chunk etc. Start with
> testing the chunk as stand-alone, then as part of page processing.

How can I test correctly my chunk ?
>
> Looking forward to it! :)
>
You shouldn't hope so. Anyway ... to be continued ...

-- 
Nicolas Delsaux
N'imprimez ce mail que si vous ne savez pas le lire sur l'?cran : les
?lectrons se recyclent bien, le papier, beaucoup moins bien.

From assaph at gmail.com  Wed Nov 15 23:05:07 2006
From: assaph at gmail.com (Assaph Mehr)
Date: Thu, 16 Nov 2006 15:05:07 +1100
Subject: [Pimki-users] register as developper for Pimki
In-Reply-To: <418361960611141203q35abb7cejd14249c9b824651b@mail.gmail.com>
References: <20061107083038.C47305240F08@rubyforge.org>
	<3ff47be90611071523m314ede97m6ad3bdd95d3ec859@mail.gmail.com>
	<418361960611090901t55ae8eeeke0b51b9710d66868@mail.gmail.com>
	<3ff47be90611091436p2552f639if25f068651f823c6@mail.gmail.com>
	<418361960611131153u599adf0gaabdf0b1ba155a7e@mail.gmail.com>
	<3ff47be90611132005i51cdd77av2b38e34737911b6f@mail.gmail.com>
	<418361960611141203q35abb7cejd14249c9b824651b@mail.gmail.com>
Message-ID: <3ff47be90611152005i3ab8d1e4j806afd52e7843f75@mail.gmail.com>

Hi Nicolas,

Thanks for CC'ing the list. For those joining in, Nicolas has offered
some help with Pimki development.

If any one is interested in a sneak peek of the (very incomplete)
Pimki2, the RubyForge guys are graciously hosting an demo instance at
http://pimki.rubyforge.org/demo. I still haven't put the access
controls, so I realise that by posting the address it'll be spammed in
no time - but feel free to play with it. It also has a section on
implementing the Pimki2 features. Feel free to comment on those there
or on the list.

> > Thanks. It stretches my graphic design skills to the max....
> >
> I may provide here a little knowledge, since I made myself (with
> obvious display bugs) my blog stylesheet.

I try to break the CSS into 3 sections/stylesheets:
* positions (layout) - where each bit goes
* general style - like borders, heading styles etc
* colours - purely the colours of text and heading

> > The current footnote mechanics are using textile (RedCloth) - there is
> > no specific support for them. If you want to add a special chunk...
>
> Let me guess ... If I want that chunk, I can just ... code it :-))

:-)

> Take a look at http://nicolas-delsaux.homelinux.net/leblog/show/Pimki+2
> There is a bottom link for features proposals.

Cool. Will comment there.

> > So i want to keep the same syntax. Note that I was still playing with
> > the advanced syntax in the trunk, so it may change a little. Once I
> > have something I like I'll make sure everything conforms.
>
> And you'll make sure I can merge my changes with ease ?

Yep. I'll cleanup any existing code and any patch before merging.


> Look :
> [sf[isbn]]227363847495 for retrieving book info
> [sf[title][sf[] for fetching nearby titles
> [sf[person]] for a person name
> All managed with chunks and page dynamic creation.

I'd probably go with something like:

 [!sf-isbn[0380011220]]
 [!sf-title[Creatures of Light and Darkness]]
 [!sf-person[Roger Zelazny]]

Note how it starts with a ! mark, denoting a macro and that the
parameters are inside the inner brackets. If you want to create a
custom chunk then maybe use a different mark, say &. For macros (!) I
plan to make them be able to either expand to static XHTML or - if the
site admin is crazy enough - to ruby/rhtml code.


> I believe i can get inspired reading existing code, like, say, ...
> GlossaryDefinition.
> I've already notice i only have to insert my class name in the
> PRE_ENGINE_ACTIONS array (in the revision class).

Yep, that the way to do it.

> > You can easily cache the results, as they will not change between page
> > rendering (unlike links, which need to be resolved each time as the
> > page may have been created/deleted etc).
> >
> Let's use cache in v. 2

By "cache" I mean the CHUNKS table 'cache' column. I do not plan on
using Rails page caching. For the amount of users/hits a typical Pimki
install should get it's far better to always serve fresh data.

> > I thought about chunks as plugins, but because the processing order
> > for chunks is VERY important I am not sure this is the proper way to
> > go. If someone already hacks the source, they can easily just add the
> > extra call at the right place to add the new chunk, which is a lot
> > simple and less error prone (at least until I have hundreds of people
> > writing customisations....)
>
> I'm OK with that, provided you never get that developper crowd.

Well, there's Peter Fitzgibbons and now you. I'm on way to take over
the world, bwahahaha :-)


> > Check your regexp, check where you added the chunk etc. Start with
> > testing the chunk as stand-alone, then as part of page processing.
>
> How can I test correctly my chunk ?

First create a standalone test for it. Take a look at
./test/unit/glossary_chunks_test.rb or todo_chunk_test.rb for
examples. There are a few assert_chunk_* helper methods in
test_helper.rb.
You can then add a few more test methods about how the chunk behaves
together with other chunks. Look at the revision_test.rb for how to
use the Page.revise mechanics.

As always, if you run into any issue just ask.

Cheers,
Assaph

From nicolas.delsaux at gmail.com  Sat Nov 18 16:18:04 2006
From: nicolas.delsaux at gmail.com (Nicolas Delsaux)
Date: Sat, 18 Nov 2006 22:18:04 +0100
Subject: [Pimki-users] register as developper for Pimki
In-Reply-To: <3ff47be90611152005i3ab8d1e4j806afd52e7843f75@mail.gmail.com>
References: <20061107083038.C47305240F08@rubyforge.org>
	<3ff47be90611071523m314ede97m6ad3bdd95d3ec859@mail.gmail.com>
	<418361960611090901t55ae8eeeke0b51b9710d66868@mail.gmail.com>
	<3ff47be90611091436p2552f639if25f068651f823c6@mail.gmail.com>
	<418361960611131153u599adf0gaabdf0b1ba155a7e@mail.gmail.com>
	<3ff47be90611132005i51cdd77av2b38e34737911b6f@mail.gmail.com>
	<418361960611141203q35abb7cejd14249c9b824651b@mail.gmail.com>
	<3ff47be90611152005i3ab8d1e4j806afd52e7843f75@mail.gmail.com>
Message-ID: <418361960611181318y561ff4fbtf0c25472710814e7@mail.gmail.com>

On 11/16/06, Assaph Mehr <assaph at gmail.com> wrote:
> Hi Nicolas,
>
> Yep, that the way to do it.

That's now the way it's done.
>
> Well, there's Peter Fitzgibbons and now you. I'm on way to take over
> the world, bwahahaha :-)
>
Oh my god, I'm feeling like you're a purple tentacle :-)
>
> First create a standalone test for it. Take a look at
> ./test/unit/glossary_chunks_test.rb or todo_chunk_test.rb for
> examples. There are a few assert_chunk_* helper methods in
> test_helper.rb.

Is there a URL I could read for knowing how to test and how to launch my test.

>
> As always, if you run into any issue just ask.

Well, I now have a pure coding issue.
I've written the following chunk :

# Chunk for including syntax highlighted computer code.
# This chunk uses the coderay library
class SyntacticHighlightChunk < Chunk
	SYNTAX_HIGHLIGHT_PATTERN = %r<\[code\[([\w]+?)\]\](.*?)\[\[code\]\]>im
	def self.pattern() SYNTAX_HIGHLIGHT_PATTERN end

	def resolved_text(view)
		'todo: syntax highlight'
	end
end

My issue is that, whatever I try, I cannot put any dynamic content in
the resolve_content without getting a rendering error.
I've tried params[0], @match_data.captures[0], and even other things,
but none of them worked.

It seems to me it's due to my poor understanding of regexp. From what
I known, elements surrounded by "()" are memorized as regexp params
and can be reused latter. but it seems that Ruby does not works this
way ...

Like you said, I added ? in the places i want my content to be
memorized, but it neither worked.

So, without coding yourself my work, can you give me a hint on the way
to make it work.

In fact, looking at pimki from my J2EE background, it seems to me I've
a real poor understanding of the application flow. I now page
rendering is done in the rhtml by invoking
page.current_revision.display_content
But the revision code is too rubyish to me.
Furthermore, i don't really know when chunks are initialized, making
my use of the regexp a little too voodoo for me.
Are there any tools for easier ruby appls understanding (in J2EE
world, i would invoke classical UML diagrams (sequence and the such),
knowing there does not exists any tool for generating them).

Cheers

-- 
Nicolas Delsaux
N'imprimez ce mail que si vous ne savez pas le lire sur l'?cran : les
?lectrons se recyclent bien, le papier, beaucoup moins bien.

From nicolas.delsaux at gmail.com  Tue Nov 21 04:37:12 2006
From: nicolas.delsaux at gmail.com (Nicolas Delsaux)
Date: Tue, 21 Nov 2006 10:37:12 +0100
Subject: [Pimki-users] [doc]model and controllers
Message-ID: <418361960611210137q5c54b34ek182a6a6a460b4ab1@mail.gmail.com>

Using rav, I generated this diagram of my pimki version :
http://nicolas-delsaux.is-a-geek.net/wordpress/index.php/archives/2006/cest-quand-meme-bien-foutu-rails/
Maybe it could be of any help, although I regret the lack of web workflow

-- 
Nicolas Delsaux
N'imprimez ce mail que si vous ne savez pas le lire sur l'?cran : les
?lectrons se recyclent bien, le papier, beaucoup moins bien.

From assaph at gmail.com  Tue Nov 21 22:34:00 2006
From: assaph at gmail.com (Assaph Mehr)
Date: Wed, 22 Nov 2006 14:34:00 +1100
Subject: [Pimki-users] [doc]model and controllers
In-Reply-To: <418361960611210137q5c54b34ek182a6a6a460b4ab1@mail.gmail.com>
References: <418361960611210137q5c54b34ek182a6a6a460b4ab1@mail.gmail.com>
Message-ID: <3ff47be90611211934v6eb1f1eenb5a4f8156da311ef@mail.gmail.com>

On 11/21/06, Nicolas Delsaux <nicolas.delsaux at gmail.com> wrote:
> Using rav, I generated this diagram of my pimki version :
> http://nicolas-delsaux.is-a-geek.net/wordpress/index.php/archives/2006/cest-quand-meme-bien-foutu-rails/
> Maybe it could be of any help, although I regret the lack of web workflow

looks nice. I never actually bothered with an ERD though. I'll post a
bit of an explanation on the demo site shortly:
http://pimki.rubyforge.org/demo/show/InformationForDevelopers.

Cheers,
Assaph

From nicolas.delsaux at gmail.com  Fri Nov 24 05:17:49 2006
From: nicolas.delsaux at gmail.com (Nicolas Delsaux)
Date: Fri, 24 Nov 2006 11:17:49 +0100
Subject: [Pimki-users] register as developper for Pimki
In-Reply-To: <3ff47be90611212107u308842bco6c26dec116a99f1@mail.gmail.com>
References: <20061107083038.C47305240F08@rubyforge.org>
	<3ff47be90611071523m314ede97m6ad3bdd95d3ec859@mail.gmail.com>
	<418361960611090901t55ae8eeeke0b51b9710d66868@mail.gmail.com>
	<3ff47be90611091436p2552f639if25f068651f823c6@mail.gmail.com>
	<418361960611131153u599adf0gaabdf0b1ba155a7e@mail.gmail.com>
	<3ff47be90611132005i51cdd77av2b38e34737911b6f@mail.gmail.com>
	<418361960611141203q35abb7cejd14249c9b824651b@mail.gmail.com>
	<3ff47be90611152005i3ab8d1e4j806afd52e7843f75@mail.gmail.com>
	<418361960611181318y561ff4fbtf0c25472710814e7@mail.gmail.com>
	<3ff47be90611212107u308842bco6c26dec116a99f1@mail.gmail.com>
Message-ID: <418361960611240217j47dd5fd9x16a74ffa2b9ca2f0@mail.gmail.com>

On 11/22/06, Assaph Mehr <assaph at gmail.com> wrote:
>
> You can start with a general guide to testing rails:
> http://manuals.rubyonrails.com/read/book/5
>
>
> What's the error exactly?

I'm a morron. THAT is the error.
I was trying to do
def resolved_text()
  params[0]
end

which fails. But after many tests, I tried params.inspect which
returned me an array, and then I concluded with params[0].to_s which
works awfully well.

I will use cache latter, when all current issues will be solved.
>
>
> It should...

In fact, it is.
>
> I'll try to explain how it all works on
> http://pimki.rubyforge.org/demo/show/Pimki2DataModel (.... and also
> possibly better documenting my code ;).

I've taken a look at it, and, without further reading, i can already
say it will be helpful.
>
> BTW, As someone who has done UMLs professionally for a few years
> before running away screaming,

Normal reaction, I guess ;-)

> I believe Ruby code - once the penny
> drops - is actually more readable and compact than diagrams.

yup.

See ya latter for issue of the day :
coderay requiring failing and, maybe, locomotive restart

-- 
Nicolas Delsaux
N'imprimez ce mail que si vous ne savez pas le lire sur l'?cran : les
?lectrons se recyclent bien, le papier, beaucoup moins bien.

From nicolas.delsaux at gmail.com  Sat Nov 25 12:18:06 2006
From: nicolas.delsaux at gmail.com (Nicolas Delsaux)
Date: Sat, 25 Nov 2006 18:18:06 +0100
Subject: [Pimki-users] [chunk[syntax highlight]] alpha version
Message-ID: <418361960611250918t7ec2ad72yfcfe9f2d00702444@mail.gmail.com>

So ...
After just a few seconds for coding it :-)) I wrote a chunk for
managing syntax highlight. This chunk is described in the diff joined.

It requires coderay, so don't forget to gem install coderay
Furthermore, i didn't diffed the revision.rb wherre this chunk is
added as a pre-rendering one.

Finally, here are the following issues :
1 - regexp may not handle more than one highlighted section (i must
find out why)
2 - contrary to coderay's author description, i find coderay to be
quite slow (I guess it's more du to my poor ruby, pimki and coderay
knowledge, here is why coderay's author will receive a copy of this
mail, with all my thanks for his library), requiring caching (which i
was not able to install, due to a cache size issue - ah, the joices of
DB-driven development), so it mlay slow down rendering ... a solution
should be latter found.

Besides, it's a good alpha version and a good entry point in pimki's
marvelous world.

-- 
Nicolas Delsaux
N'imprimez ce mail que si vous ne savez pas le lire sur l'?cran : les
?lectrons se recyclent bien, le papier, beaucoup moins bien.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: chunk_diff_for_syntax_highlight.diff
Type: application/octet-stream
Size: 1608 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/pimki-users/attachments/20061125/740abed3/attachment.obj