CFEclipse Still Being Developed?

[Tiggrstaar]

It's been nearly a year since there were any new releases and checking
out trac it seems that there haven't been any commits in almost a year
and nearly as long for tickets. So I have to wonder is anything still
being developed for cfeclipse?

It's pretty good how it is but there was some talk about advanced
parsing eventually that would be really beneficial since it would
probably yield object insight and then as eclipse changes eventually
this will need to be updated. I do appreciate cfeclipse and use it
every day but it would be great to get some insight into how
development is going.

[Andrew Myers]

I've been wondering the same thing.

In fact I've been thinkinh for some time about digging into the code myself and just today l figured out how to add in the completion of function args when you type:

arguments.

I am not certain I did it the 'right' way, but I have it working and I'll submit a pull request for this tonight.  

There's also a tiny fix I submitted to a dictionary awaiting a pull request that hasn't been actioned yet as well.

If anyone from the dev team is willing to 'mentor' me I'm very happy to get more actively involved.  

Andrew

> --
> For more information on CFEclipse visit: cfeclipse.org
> For support, FAQ and tips and tricks visit: trac.cfeclipse.org
>
> You are subscribed to the Google Groups "CFEclipse Users" group.
> To post send email to: cfeclip...@googlegroups.com
> To unsubscribe send email to: cfeclipse-use...@googlegroups.com
> For more options, visit this group online: http://groups.google.com/group/cfeclipse-users?hl=en

--
Sent from Gmail Mobile

[Mark Drew]

I have been working on it yesterday if that helps? :) adding the railo 3.3.1 dictionary. 

Btw, most of the source  is now over in GitHub. 

MD

Sent from one of my many iDevices

[Jim Priest]

Hopefully Denny will chime in.... but this is traditionally how it's
been with CFEclipse :)

Someone works on it for a bit, then it's quiet then someone works on
it a bit, rinse repeat.

Andrew - I'm sure both Mark and Denny can give you pointers if you ask...

And I'd be all up for moving everything to GIT and getting rid of Trac
altogether...

If anyone wants to help in a 'non-coding' way - moving the wiki info
over to git (along with some housecleaning) would be a tremendous
help.

Jim

[Andrew Myers]

Thanks guys. Really appreciate your work on the project btw...

Sent from my mobile

[Mark Drew]

Some way of getting all the tickets into GitHub would be awesome. I think the issue is that comments do not have an api in GitHub so we would have to add them somehow to the body of the ticket. (remember I copied them from Tigris first time!)

MD
Sent from one of my many iDevices

[Mark Drew]

I should mention I copied the tickets via selenium btw.

MD
Sent from one of my many iDevices

[Jim Priest]

I'm wondering if we really need all the tickets? I haven't even
looked lately to see how many there are.

Though it looks like people have already done this :)
https://github.com/adamcik/github-trac-ticket-import
https://gist.github.com/589126

It'd be tedious but might be worth a going through manually?

Or maybe import and then go through?

Jim

[Mark Drew]

I twas thinking of just importing them. They are stored in mysql if I remember right anyway!

MD

[Jim Priest]

Yeah - I believe so... I could get a dump of that if you want it...

Jim

[Mark Drew]

That would rock sir!!!

MD
Sent from one of my many iDevices

[Peter Boughton]

Jim wrote:
> If anyone wants to help in a 'non-coding' way - moving the wiki info
> over to git (along with some housecleaning) would be a tremendous
> help.

What's involved in moving it? Is it basically copy and paste pages,
and update links accordingly?

And what sort of housecleaning did you have in mind?

I'd probably be willing to help with either half of that, if someone
else wanted to take the other half.

[Jim Priest]

Cut and paste, and since no two wikis share the same formatting - I
imagine some things will need to be tweaked in regards to links,
formatting,etc.

Housecleaning would basically be... "is the info still relevant" and
could we improve it or just remove it.

Jim

[Peter Boughton]

Formatting issues are avoided by passing the Trac HTML pages into a
Markdown converter. :)

Which I think leaves fixing links, importing images, and noting down
which pages could do with an update/etc.

So, if you want to turn on the GitHub wiki and/or give me relevant
permissions, I can probably get it done tonight or tomorrow. :)

[Andrew Myers]

Sounds awesome guys.

[Stan Winchester]

Is anyone working on the OpenBD dictionary?

[Mark Drew]

I am not sure, I am sure the OpenBD guys would either send it or if you are into it, you could create one and submit it to the project?

Regards

Mark Drew

[Peter Boughton]

Stan wrote:
> Is anyone working on the OpenBD dictionary?

If they are, they haven't mentioned it on this list nor the dev list.

So, if you're thinking of doing so yourself, I'd say go ahead. :)

[Stan Winchester]

I just submitted to the OpenBD list "if a CFEclipse dictionary xml file could be auto-generated with the nightly build like the documentation is."

The reply is they believe it is possible.

[Matthew Woodward]

On Mon, Jan 16, 2012 at 8:54 AM, Stan Winchester <stanwin...@gmail.com> wrote:

I just submitted to the OpenBD list "if a CFEclipse dictionary xml file could be auto-generated with the nightly build like the documentation is."

The reply is they believe it is possible.

Thanks Stan--this has been on my very long to-do list for a while (i.e. getting a dictionary file put together for CFEclipse), so thanks for helping push it forward.

--
Matthew Woodward
ma...@mattwoodward.com
http://blog.mattwoodward.com
identi.ca / Twitter: @mpwoodward

Please do not send me proprietary file formats such as Word, PowerPoint, etc. as attachments.
http://www.gnu.org/philosophy/no-word-attachments.html

[Mark Drew]

Matt, I can send you my script if you like, it uses the railo functions getTagList and getFunctionList as well as getFunctionData and getTagData so I am not sure if they are available in OpenBD, but well, you have most of the ways of creating the XML there at least! 

[Matthew Woodward]

Sure--can't hurt! Those functions are available in OpenBD, and also our manual (http://www.openbd.org/manual) is generated straight from the engine source code so between the two we should be able to create something self-sustaining. Thanks!

[Jim Priest]

I've turned on the wiki.  I'm not sure how prevalent wiki spam is so I've set it to only allow editing by team members for now.

If you have a github account - let me know you username and I can add you.

https://github.com/cfeclipse/cfeclipse/wiki 

I've imported the initial homepage content just so there is something there...

Jim

[Peter Boughton]

Thanks Jim, username is boughtonp.

I'm off out shortly, but back in a few hours so can crack on with it then. :)

[Mark Drew]

Nice one Jim! 

You are indeed a star! 

MD 

[Will Coleda]

The parrot project just imported all their tickets from a trac
instance to github.

See: http://lists.parrot.org/pipermail/parrot-dev/2012-January/006525.html

--
Will "Coke" Coleda

[Andrew Myers]

Please add me Jim - username am2605

Sent from my mobile

[Will Coleda]

Ah, Andy wrote an article about it:

http://perlbuzz.com/2012/01/parrot-tickets-now-converted-to-github.html

--
Will "Coke" Coleda

[Jim Priest]

Just added Andrew and Peter.  If anyone else wants access, please let me know.

Reading the Parrot article it looks like you may be able to use this ""net-trac" to do things remotely.

Tomorrow I will still look into getting a dump of the SQL - if for nothing else a good backup :)

Jim

[MichaelEvangelista]

>> Hopefully Denny will chime in....

where'd he go anyway?
I think it was at last year's openCFsummit when we heard he was
joining Railo... and they shuttled him off to an undisclosed location.

[Mark Drew]

I think it's called Albuquerque... 

[Nathan Strutz]

On 17 Jan 2012, at 02:20, MichaelEvangelista wrote:

they shuttled him off to an undisclosed location.

On Mon, Jan 16, 2012 at 7:31 PM, Mark Drew <mark...@gmail.com> wrote:

I think it's called Albuquerque... 

Yeah I wouldn't want to disclose that location either. 

*ba-dum ching*

nathan strutz
[www.dopefly.com] [hi.im/nathanstrutz] [about.me/nathanstrutz]

[Peter Boughton]

A quick update...

The initial version of the wiki import is now on GitHub. :)

There's still some tidying up to be done (I've already spotted several
links I need to fix, and there's no doubt other stuff), so I haven't
removed the "under construction" text yet. I'll be going through all
the pages again checking/fixing stuff later this evening (I'm off out
now).

Also, I've made a mental note of various housekeeping tasks that need
doing (e.g. updating images, better section organisation/structure,
improve wording, etc), and when I go through again I'll turn this into
an actual list, which I guess can then be added to the "Current Items
We Need Help With" section.

[Jim Priest]

Awesome!! Please let me know when you update the todo list and I can
help fix some of that stuff...

Jim

[Peter Boughton]

Ok, I've just done my second run through, and hopefully fixed all
links and formatting issues and so on.

I've also added the various housekeeping tasks to this page:
https://github.com/cfeclipse/cfeclipse/wiki/How-You-Can-Contribute-To-CFEclipse

So anyone that wants to can pick on some of those issues to attack, or
of course work on anything else there that looks like it could do with
some polish. :)

A quick thing to remember: the wiki is itself powered by a Git repository.

So, for example, to add/modify images you do a `git clone
g...@github.com:cfeclipse/cfeclipse.wiki.git` and then you can commit
updated images as you would in a normal git repository, and once done
you `git push origin master` to get your changes live.

This also makes it easier to bulk edit files, or just for editing
larger files (which can be sluggish in the browser).

You can read more about it here:
https://github.com/blog/699-making-github-more-open-git-backed-wikis

[Tiggrstaar]

Glad to see people are indeed still active on cfeclipse! I really
don't want to switch over to cf builder. I'd even like to help with
working on cfcelipse myself but I don't have a ton of java knowledge
so far and also right now I have no extra time to help but in a few
months I could.

The big things really seem to be about improving how the parsing is
done so that old bigger pages don't take forever to parse and then
also so that code insight for user generated code could be done.
What's involved with changing and improving the parsing?

[JPotts]

I believe there was talk about moving to a new parsing engine, but I
don't think the project ever got off the ground. It is a pretty major
endeavor, and I'm not sure how much work would be involved in
converting the dictionaries, if at all necessary. I agree though, the
biggest weak point is the parsing performance. I've also noticed some
syntax highlighting bugs in cfscript as well, but it never got any
attention. I suppose I could always post a bug on the GitHub project.

[Andrew Myers]

Is there anything in railo or openbd that would be of use?

Do they need to "parse" cfml internally?

[Mark Drew]

This is something I was discussing with Gert today, the problem is that Railo will parse COMPLETE and correct pages, that if you think about it, the code you are writing (and parsing as you write) is almost ALWAYS incorrect. i.e. 

<cfset var elvis = 

Which is incorrect, so Railo will immediately throw an error rather than giving you a good option. 

A language parser for editors can be generated with Antler, which will parse incorrect files as the break point ids closer to the error. 

Does this make sense?

Regards

Mark Drew

--
Mark Drew

[Andrew Myers]

Of course - totally makes sense now that you explain it Mark.

[Peter Boughton]

Tiggrstaar wrote:
> I'd even like to help with working on cfcelipse myself but I don't
> have a ton of java knowledge so far and also right now I have
> no extra time to help but in a few months I could.

Then bookmark this page and check it in a few months:
https://github.com/cfeclipse/cfeclipse/wiki/How-You-Can-Contribute-To-CFEclipse
:)

[Peter Boughton]

JPotts wrote:
> I believe there was talk about moving to a new parsing engine,
> but I don't think the project ever got off the ground.

I think a fairer description would be that efforts floundered, due to
the inability/difficulty of expressing the complexity of CFML syntax
using ANTLR v3, which needed all these wacky "predicate" things to
determine what characters are special and when. (Most languages don't
have that problem; they are syntactically very simple.)

With ANTLR v4 things are done in an entirely different way - modal
grammars are supported (which makes predicates mostly/entirely
unnecessary). Without saying too much too soon, it basically means the
only roadblock is time. ;)

[Randy Merrill]

Isn't OpenBD now using an ANTLR grammer to parse CFML?

I thought I remembered a conversation between Denny and OpenBD to see about using their grammar to power CFEclipse (I believe that the license of OpenBD would normally not allow it but they were thinking of allowing for CFEclipse to use it under a different license?!? I think...). I don't remember when it was or what became of the conversation though...

[Peter J. Farrell]

Randy Merrill said the following on 01/24/2012 05:44 PM:

> Isn't OpenBD now using an ANTLR grammer to parse CFML?
>
> I thought I remembered a conversation between Denny and OpenBD to see
> about using their grammar to power CFEclipse (I believe that the
> license of OpenBD would normally not allow it but they were thinking
> of allowing for CFEclipse to use it under a different license?!? I
> think...). I don't remember when it was or what became of the
> conversation though...
>

That conversation happened during OpenCF Summit 2011 -- last year. I
don't know what was decided but IIRC the OpenBD guys were open to the idea.

.pjf

--
Peter J. Farrell
pe...@mach-ii.com
p...@maestropublishing.com
http://blog.maestropublishing.com
Identi.ca / Twitter: @maestrofjp

[Mark Drew]

The problem is me too, for some reason, part of my knowledge with
regards to grammar (I am talking english, spanish or programming
languages) is missing. the same with Trigonometry and RegEx.

Mark Drew

[Randy Merrill]

Here is the conversation I was thinking of with the ANTLR from OpenBD:

https://groups.google.com/d/msg/cfeclipse-users/45z7gY2kZ4U/ifYkBLDJ3AkJ
--
Randy Merrill

[Mark Drew]

Well, as long as we get an AST out of it, I am happy and can work with the rest!

Mark Drew

[Peter Boughton]

Randy wrote:
> Isn't OpenBD now using an ANTLR grammer to parse CFML?

That was only for cfscript, which of course is far closer to a
standard language.

Dunno what the status of that is though - I think it might be
somewhere on GitHub.

[Randy Merrill]

https://github.com/denuno/cfml.parsing

I think that he may already have the ANTLR cfscript parsing in there.

Now just to get one for working with the tags and put them together?

If I remember correctly there are some tests in there that we can build on for unit testing everything too.
--
Randy Merrill

[Andrew Myers]

Denny has a cfml.parsing project on github but I couldn't get it to
compile unfortunately. I haven't looked into it that deeply, but I
suspect this is the stuff in question.

I'd love to be able to download a dump of his brain database and
upload it into mine ;-)

[Randy Merrill]

Be careful what you ask for... he probably already has an ANT script for it... just needs documentation...
--
Randy Merrill

[Tiggrstaar]

Glad to see good discussion happening. Does anyone have any ideas how
cfbuilder does their parsing? I only used it briefly but it did parse
and return code insight.

One thing I was wondering about is there any way to use existing cf
server compiliation or java abilities to improve parsing or at least
code insight?

[Peter Boughton]

Tiggrstaar wrote:
> One thing I was wondering about is there any way to use existing cf
> server compiliation or java abilities to improve parsing or at least
> code insight?

Yes and no.

See what Mark said about parsing - there's no use/benefit in even
asking the server to compile a page being edited - you pretty much get
a works/broken answer out of it, and an editor needs more than that.

For code insight, that is another thing Denny was investigating -
being able to connect to a running instance of a project and
automatically obtain extra information (e.g. mappings and custom tag
paths, variables which currently exist in scopes, and so on).

[Andrew Myers]

On 25 January 2012 15:27, Peter Boughton <boug...@gmail.com> wrote:

> For code insight, that is another thing Denny was investigating -
> being able to connect to a running instance of a project and
> automatically obtain extra information (e.g. mappings and custom tag
> paths, variables which currently exist in scopes, and so on).

I believe this is how ColdFusion Builder works, using the RDS server.
Of course that is specific to Adobe CF, however I did recently notice
a railo rds project on riaforge: http://railords.riaforge.org/. The
other thought I had is whether something like Colddoc could be used
for code insight. In a server agnostic way. It seems to allow
outputting in alternative formats and presumably you could use it
(either in real time or probably more likely as a "compile task") to
produce some kind of output format which could then be "queried" to
allow function and function argument completion from CFCs other than
the one you were working on.

The only problem is that once we start bringing in stuff like this, we
then need to have people set up servers stuff on their and configure
cfeclipse to point to it. So I do like the current approach where it
reads files from your project without further setup.

I did get a "proof of concept" working a couple of weeks ago which a
CFC and I think it's something that could be built upon...but it's
very much a work in progress at this stage and getting it working
properly really is going to depend on the parsing of the "currently
open file" to be modified as well. I submitted a pull request but I'm
not sure if it's made it into the dev builds yet. If anyone wants a
look let me know and I can provide a zip of the update site.

Andrew.

[Andrew Myers]

Sorry I just re-read that and it's terrible. I think I was so excited
my fingers couldn't keep up with my thoughts. Hopefully it makes a
little it of sense!

[Andrew Myers]

Looks like I may be wrong that RDS is used for code insight:

http://web-rat.com/blog/post.cfm/using-cfbuilder-with-railo

But from my experience it does require to have a server running, and
configured correctly to get it to work, even if it's not the RDS
component it's using.

[Randy Merrill]

I'm pretty sure that they would have to have something in Eclipse doing the parsing on the raw file.

Not sure how CFBuilder handle mappings and the such for autocomplete. Seems like that would be a good thing to have in the project specific settings.

Other than having the requirement of having a CF server running I think that it would be a Bad Idea to have an external source parsing since it will probably completely die or slow things down on large documents. (Not to mention the added complexity)

I think that we are on the right track with the ANTLR parsing to get the AST that Mark was talking about. That basically gives you all of the information that you need to get the parsing and also can be used to do auto-complete (I believe).

Just need to do some more work on the cfml.parser project to get more tests. Then we know what things don't work that should and can tell quickly when we break something that was working before. (Standard TDD argument).

So if denny can chime in with some instructions on getting it to compile and run the tests that would be a good step that people can help with: writing tests, even if we don't know how to change the grammar to make them work.

</cfdump>

--
Randy Merrill

[Tiggrstaar]

I noticed recently that the JSEclipse plugin(formerly available from
adobe though not sure if was written by them) creates a xml file with
doc type info for javascript functions and classes in a project as you
use them. It then provides some code insight and lets people jump to
declarations(with F3) based on this metadata. If cfeclipse could get
metadata(like with getMetadata() in CF) about a CFC for instance then
that insight info could be generated as a xml file.

Once a xml file of metadata exists then the editor would just need to
search it and make suggestions but I suspect that this would be easy
since the dictionary stuff already works that way.

Thought, would be possible for individual developers now to glean
metadata from their CFCs and append it to the CF dictionary file? At
least as a quick solution but otherwise for the broader solution it
would be some other xml files for metadata.

Thoughts?

[denstar]

After wandering the high desert for a while (yes, Albuquerque, one
awesome town to hail from, fer reals! =]), growing a beard, pondering my
navel, etc., I'm happy to announce that with the contributions of
Andrew, combined with the luck I had with ANTLR, things are about to get
a LOT sweeter in CFE land.

There's a highly untested dev build out, sans a list of what's new. If
you can't wait to see things like outline/method view of cfscript-based
cfcs, and component assist (yes, we're not trapped in the currently open
CFC/dictionary anymore, thanks to Andrew and Mark!), feel free to give
it a spin!

The current implementation is geared towards Railo ATM, so it won't flag
missing semi-colons, and allows

for(var item in collection) { AndSomeOtherStuff() }

(the var'ed item would throw an error prior, and as such I don't think
it works in OpenBD, which is where the script parsing originated).

I'm hoping that what I did to the grammar wasn't /so/ horrible, and that
maybe OpenBD can use it, or get ideas from it, since I added bits to
parse "component { }" and ternary "if ? then : else" and a few other
things in there.

:DeN

--
Always remember you're unique, just like everyone else.

[Andrew Myers]

Awesome Denny.

Can we get this from the update site, or do we need to build it ourselves?

I haven't seen any updates in github - is your intention to put them
into master or develop?

So happy to see you back and contributing to my favourite CF ide. I'm
using CFEclipse it as I type (well took a break from using it to type
this!). Actually just noted down a list of stuff to look at
surrounding my latest changes. In particular, function method
completion only works in .cfm files at the moment, and not cfc's,
which is where I need it most! I might try and figure that one out
this weekend.

Andrew.

> --
> For more information on CFEclipse visit: cfeclipse.org

> For support, FAQ and tips and tricks visit: https://github.com/cfeclipse/cfeclipse/wiki

[Randy Merrill]

Indeed, that sounds pretty amazing. Which update site did you push to?

I am using the http://cfeclipse.org/update-dev and don't see any updates..?
--
Randy Merrill

[denstar]

Pushed another build to the "dev" update site.

As for git, I majorly fuxored up the merge some time ago, so, lemme see
if I can clean up git a bit, or maybe just push a new branch. Apologies
if the result messes up pulls!

Couple changes to this push:

Compromise on engines: Semicolons are required again, as two of the 3
CFML engines require them.

Support for Railo specific stuff like:
savecontent variable="wee" { writeOutput("woohoo"); }

is there though, so, fair's fair. :)

Eventually we'll split out the parsers or some such...

In general, I'd rather miss errors than have false errors.

Anywho, function method completion /should/ work in cfc's... maybe try
this latest build and see. I tried to hook up everything that had been
done for the tags, for the script equivs, but we really need a single
model (maybe just copy java's, with "compilationUnits" and whatnot) vs.
what we're rocking now (checking context, etc.).

Lemme see if I can get this merge sorted... I did push the parser
changes (https://github.com/denuno/cfml.parsing).

:den

--
Guys: No Shirt, No Service - Gals: No Shirt, No Charge

[Chris Phillips]

Den et al.,

I updated to the dev build (extremely excited about the strides you guys are making). And I am looking at an "all-script" component and not seeing any outline of methods view. I tried closing eclipse and starting -clean. I also closed the views and made sure I had the correct ones open.

Here is the version from help > about eclipse:
  CFEclipse (including MXUnit and Frameworks)    1.4.5.201202261042    org.cfeclipse.cfml.feature.feature.group

Is it working for anyone else that has updated?

Chris Phillips
www.cfchris.com

[Stephen Moretti]

On 28 February 2012 22:12, Chris Phillips <cfc...@gmail.com> wrote:

Den et al.,

I updated to the dev build (extremely excited about the strides you guys are making). And I am looking at an "all-script" component and not seeing any outline of methods view. I tried closing eclipse and starting -clean. I also closed the views and made sure I had the correct ones open.

Here is the version from help > about eclipse:
  CFEclipse (including MXUnit and Frameworks)    1.4.5.201202261042    org.cfeclipse.cfml.feature.feature.group

Is it working for anyone else that has updated?

No. I noticed that the method view (Probably the my most used functionality in CFEclipse)  wasn't working.

Bizarrely, at one point the method view did populate, but clicking on any of the methods threw an error.  Unfortunately I've been running from one meeting to the next this afternoon, so never got to grab the error and post it.

Stephen

[Chris Phillips]

Funny thing is. The methods did just show up today for an all-script CFC. And when I clicked one, I got the following error:

"An error has occurred. See error log for more details.
org.cfeclipse.cfml.views.cfcmethods.CFCMethodViewScriptItem cannot be cast to org.cfeclipse.cfml.views.cfcmethods.CFCMethodViewItem"

Also, I'm getting error notes next to perfectly valid lines.

missing SEMICOLON at 'function' 0 cfml.parsing.cfscript.CFParseException    JiraAPI.cfc    /local DealerPeak/com/jira    line 11

Anyway, I'm really excited to see progress. I'm in CFEclipse every day. Improvements are really appreciated you guys!

Chris Phillips
www.cfchris.com

--

[Tiggrstaar]

It's great to see this progress happening!! I'm really looking forward
to having cfc introspection in the near future.

Would this also make it possible to be also to highlight a method on a
cfc object and then have a shortcut key that could be used to jump to
that method?

[Alan Holden]

When working in the Flex perspective, you can select a method/function
name and hold down Ctrl. This turns the selection into a link - works
like you describe below.

Yes, +1 for that idea.

Al

[Andrew Myers]

Do you mean like the ctrl-click that you can do in Java? Where it opens
the file that method is defined in?

That'd definitely be sweet...

[Tiggrstaar]

Any updates on the new awesome development? It's pretty exciting that
we might have new capabilities with the new parser.