leaving mirah

155 views
Skip to first unread message

Michal Hantl

unread,
Jul 19, 2011, 5:56:41 AM7/19/11
to mirah
Hey guys. I'm sad to say that I'm giving up on Mirah.

I want something that works, makes me productive and happy. So I am
going back to Ruby.

I may return someday because I still think Mirah would be sweet
language. But right now the bugs, boxing and compilation time make it
a nightmare to program in.

--
S pozdravem, Regards
Michal Hantl

tel: +420 777 812 167
gtalk: michal...@gmail.com
skype: michal.hantl

Roger Pack

unread,
Jul 19, 2011, 9:34:51 AM7/19/11
to mi...@googlegroups.com
I may return someday because I still think Mirah would be sweet
language. But right now the bugs, boxing and compilation time make it
a nightmare to program in.

And lack of response to pull requests...maybe mirah should become "open commit" policy like rubinius has?
-roger-

Rui Neto

unread,
Jul 19, 2011, 9:38:29 AM7/19/11
to mi...@googlegroups.com
+1

2011/7/19 Roger Pack <roger...@gmail.com>:

Marko Kocić

unread,
Jul 19, 2011, 10:51:41 AM7/19/11
to mi...@googlegroups.com
Not only that, but the lacks of any visible development of Mirah core. I'm just looking at the commit log and see only 2 commits in ~1.5 months.
Also, the mailing list is pretty silent lately.

It's a pitty, since Mirah is a nice language, and IMHO still has much to offer, but the problems OP describes are killers.
I know there are too few commiters, but the lack of responsivness, documentation, roadmap of future goals and push doesn't help  to attract new ones.

Regards,
Marko Kocić

Roger Pack

unread,
Jul 19, 2011, 11:53:52 AM7/19/11
to mi...@googlegroups.com


Not only that, but the lacks of any visible development of Mirah core. I'm just looking at the commit log and see only 2 commits in ~1.5 months.
Also, the mailing list is pretty silent lately.

Yeah it's the stagnant aspect that is the most frustrating.  So it either needs to "recruit" more dedicated core developers, or alternatively allow more common folk to help with commits, so having a more open commit policy might help there, though only in part I suppose.
-roger-

Charles Oliver Nutter

unread,
Jul 19, 2011, 12:11:35 PM7/19/11
to mi...@googlegroups.com
I want to take this opportunity to apologize personally for the pace
of development of late.

I still believe Mirah has a great future, but the problems you
(Michal) list definitely need to be prioritized. My own time has been
split between JRuby, Ruboto, Mirah, conferences, and of course family.
Excuses excuses...I know.

I'm open to having an open commit policy, if that might help the pace
increase. Up to now very few folks have expressed interest in being a
committer, so I haven't just handed out the keys willy-nilly. Pull
requests are delayed, but they do get in (or they should be...it's
possible I've missed some recently due to travel). If there are folks
here interested in being a committer (and you've submitted a patch or
two), please say so...I will add you.

I promise that Mirah is not dead...it's just wanting for resources.
All these issues will be fixed, but I have no timeline.

I'm going to try to evaluate where I'd like to see Mirah go and
reprioritize what issues need to be addressed.

Michal: I hope it's possible to win you back to Mirah with some
improved resourcing and key fixes/features. You've been an excellent,
patient user, and I'd hate to lose you permanently.

- Charlie

Roger Pack

unread,
Jul 19, 2011, 12:36:13 PM7/19/11
to mi...@googlegroups.com

I'm open to having an open commit policy, if that might help the pace
increase. Up to now very few folks have expressed interest in being a
committer, so I haven't just handed out the keys willy-nilly. Pull
requests are delayed, but they do get in (or they should be...it's
possible I've missed some recently due to travel).

Hmm.  I think an assurance of "I really had planned on getting around to it in 6 months" isn't the best.
I don't even think anybody *looked* at my pull requests until now, except maybe barbequecat.
I'm all for the open commit policy.
Cheers!
-roger-

Charles Oliver Nutter

unread,
Jul 19, 2011, 12:52:33 PM7/19/11
to mi...@googlegroups.com
On Tue, Jul 19, 2011 at 4:36 PM, Roger Pack <roger...@gmail.com> wrote:
> Hmm.  I think an assurance of "I really had planned on getting around to it
> in 6 months" isn't the best.
> I don't even think anybody *looked* at my pull requests until now, except
> maybe barbequecat.
> I'm all for the open commit policy.
> Cheers!

Yeah, for some reason I thought I had already added baroquebobcat, but
I probably conflated their license acceptance as getting commit
access. That was a big error on my part. I also lost track of pull
requests at some point because I thought other folks would merge
them...but there's no other committers active lately.

baroquebobcat is now a committer. Anyone else who has submitted a
patch and agreed to Apache License 2.0 (with an email to this group)
is welcome to become a committer too. I have also dealt with all
outstanding pull requests.

And yes, promises to fix things that go unhandled for months are not
very good promises. It has been weighing on my mind heavily.

Here's brainstorming on a few tasks to get Mirah going again:

* Start an official language specification. This would go off what we
have now and just fill in blanks until we reach the edges of the
languages. Then we'd fix/add and document those edges and continue.
* Get another release out. It's been a while. (Some tests are not
passing on master at the moment, probably due to one of the pull
requests I just merged, but we'll clean that up)
* Prioritize key Java 5 features that have been holding people up.
Specifically, I think boxing should be fixed before anything else, and
then varargs dispatch should be added. Generics are a harder
problem...an important problem, but a hard one.
* Start filling out a page for what's missing. I need user help on
this to know where you all think Mirah should go. The issue tracker
might be a fine place for this, since it can track discussions.

I think we can turn around the dev process a lot in the next couple
weeks, and I'll make a point to prioritize replies on this group, pull
requests, and bug fixes.

- Charlie

Charles Oliver Nutter

unread,
Jul 19, 2011, 12:56:31 PM7/19/11
to mi...@googlegroups.com
@ashee and @rdp have now been added as committers (Roger: I assumed
you'd want commit access, since you suggested the open commit policy.
If you haven't officially consented to Apache License 2.0 on this
list, please do so now).

Roger Pack

unread,
Jul 19, 2011, 1:27:48 PM7/19/11
to mi...@googlegroups.com
* Get another release out. It's been a while. (Some tests are not
passing on master at the moment, probably due to one of the pull
requests I just merged, but we'll clean that up)

+1
and I consent :)
-roger-

Charles Oliver Nutter

unread,
Jul 19, 2011, 1:43:46 PM7/19/11
to mi...@googlegroups.com
On Tue, Jul 19, 2011 at 5:27 PM, Roger Pack <roger...@gmail.com> wrote:
> +1
> and I consent :)

I hate to be a legal asshole, but can you explicitly say something
like "I consent to my code being released under the terms of Apache
License 2.0"? I know it won't ever matter, but at least we'll have the
consent and what you're consenting to in the same email.

- Charlie

Roger Pack

unread,
Jul 19, 2011, 3:31:40 PM7/19/11
to mi...@googlegroups.com

Tom Lee

unread,
Jul 19, 2011, 4:29:47 PM7/19/11
to mi...@googlegroups.com
Can I please get commit access too? "thomaslee" on github.

I consent to release my code under the Apache License 2.0.

Cheers,
Tom
--

Tom Lee
Software Developer / Consultant
http://tomlee.co | @tglee | 0450 112 893

(or check out my code on github)

Charles Oliver Nutter

unread,
Jul 19, 2011, 7:16:43 PM7/19/11
to mi...@googlegroups.com
Thank you...I'll add you now.

consiliens

unread,
Jul 19, 2011, 7:46:00 PM7/19/11
to mi...@googlegroups.com

I created a wiki page that attempts to define the commit policy and
provides a standard license statement [1]. The statement is based on
the language used last year [2]. The benefits of having a standardized
statement include:

- Developers requesting commit access can copy and paste the statement
- Each developer avoids inventing yet another creative way to license
their code

Feel free to edit the page.

[1] https://github.com/mirah/mirah/wiki/Commit-Access

[2]
http://groups.google.com/group/mirah/browse_thread/thread/b13631a923f1f02/4a04fc7595161f18

Charles Oliver Nutter

unread,
Jul 19, 2011, 7:50:42 PM7/19/11
to mi...@googlegroups.com
On Tue, Jul 19, 2011 at 11:46 PM, consiliens <consi...@gmail.com> wrote:
> I created a wiki page that attempts to define the commit policy and provides
> a standard license statement [1].  The statement is based on the language
> used last year [2].  The benefits of having a standardized statement
> include:
>
> - Developers requesting commit access can copy and paste the statement
> - Each developer avoids inventing yet another creative way to license their
> code
>
> Feel free to edit the page.
>
> [1] https://github.com/mirah/mirah/wiki/Commit-Access

Looks good to me! I nuked the default empty wiki Home and added a new
welcome message (still very short) and a link to this one page.

I think we should move the wiki content out of mirah.org into the
Github wiki. Anyone want to start copy/pasting stuff over?

- Charlie

consiliens

unread,
Jul 19, 2011, 8:03:09 PM7/19/11
to mi...@googlegroups.com
On 07/19/2011 05:50 PM, Charles Oliver Nutter wrote:
> I think we should move the wiki content out of mirah.org into the
> Github wiki.

I agree.

I updated the page to include information about contributing.

consiliens

unread,
Jul 19, 2011, 8:15:39 PM7/19/11
to mi...@googlegroups.com
On 07/19/2011 05:50 PM, Charles Oliver Nutter wrote:
> Anyone want to start copy/pasting stuff over?

Where are the markdown source files? I think someone with AppEngine
access will have to download them.

Charles Oliver Nutter

unread,
Jul 19, 2011, 8:22:33 PM7/19/11
to mi...@googlegroups.com

It should be possible to edit the pages and get the content out (other
than the main page). The markdown is probably in a google data store
behind the GAE app that runs mirah.org, and Ryan would have to get it
out.

- Charlie

John Woodell

unread,
Jul 19, 2011, 8:25:24 PM7/19/11
to mi...@googlegroups.com
<center><img src="http://i.min.us/imc8fs.png"/></center>

The Mirah Programming Language
=============================

Mirah is a new way of looking at JVM languages. In attempting to build a replacement for Java, we have followed a few guiding principals:

* No runtime library

Mirah does not impose any jar files upon you. YOU decide what your application's dependencies should be.

* Clean, simple syntax

We have borrowed heavily from Ruby, but added static typing and minor syntax changes to support the JVM's type system. The result is pleasing to the eye, but as powerful as Java.

* Metaprogramming and macros

Mirah supports various mechanisms for compile-time metaprogramming and macros. Much of the "open class" feel of dynamic languages is possible in Mirah.

* No performance penalty

Because Mirah directly targets the JVM's type system and JVM bytecode, it performs exactly as well as Java.

More about => MirahFeatures

Getting Started
===============

There's a few ways to get your hands on Mirah.

* Download a Mirah distribution here: <https://github.com/mirah/mirah/downloads>

* If you're using JRuby, `gem install mirah`

* For maven users, see the sample project here: <https://github.com/mirah/mirah/tree/master/examples/maven>

Test it out by running

`mirah -e 'puts "Hello, Mirah!"'`

Then check out the examples directory in the distribution or on Github here: <https://github.com/mirah/mirah/tree/master/examples>

This Wiki
=========

The Mirah.org site and wiki are actually written in Mirah and deployed to Google AppEngine. Check out the wiki source here: <https://github.com/mirah/mirah/tree/master/examples/wiki>

About the Name
==============

The word "mirah" means "ruby" in Javanese, the language of the island of Java. Get it? "Ruby" in "Java"?

Rubyconf 2009 Presentation
==========================




</video>

Please join the discussion on the Mirah [mailing lists][2] or stop in to the #mirah IRC channel on \FreeNode.


consiliens

unread,
Jul 19, 2011, 8:29:28 PM7/19/11
to mi...@googlegroups.com
On 07/19/2011 06:22 PM, Charles Oliver Nutter wrote:
> It should be possible to edit the pages and get the content out (other
> than the main page).

You have to be signed in to see the edit option. I'll try moving the
wiki pages over now.

consiliens

unread,
Jul 19, 2011, 8:38:39 PM7/19/11
to mi...@googlegroups.com
Thank you John!

consiliens

unread,
Jul 19, 2011, 8:48:01 PM7/19/11
to mi...@googlegroups.com

I think I moved everything over. I couldn't find a sitemap on mirah.org
so I might of missed some.

consiliens

unread,
Jul 19, 2011, 9:48:41 PM7/19/11
to mi...@googlegroups.com
@fontana kindly provided feedback [1] on the License Grant text. I've
incorporated his suggestion [2].

If there are any questions or comments let me know. Note that this
isn't a new policy, merely a documentation of an existing practice
started last year.

[1] http://identi.ca/notice/79225440

[2]
https://github.com/mirah/mirah/wiki/_compare/0111b5bb81a7b1867a5ea0c468d66d667175e281...85570896a06728e044af454e04f7900bb06544cc

Charles Oliver Nutter

unread,
Jul 20, 2011, 12:57:10 PM7/20/11
to mi...@googlegroups.com
On Wed, Jul 20, 2011 at 12:48 AM, consiliens <consi...@gmail.com> wrote:
> I think I moved everything over.  I couldn't find a sitemap on mirah.org so
> I might of missed some.

Awesome, thank you!

If there's any pages missed they probably weren't linked properly or
prominently anyway.

- Charlie

Rib Rdb

unread,
Jul 22, 2011, 8:41:13 PM7/22/11
to mi...@googlegroups.com

Hi. I've been out camping this week and just got back into cellphone range. Looks like I've missed a bunch. I won't be able to catch up until I'm backnext week  but I wanted to give a quick update on what I've been working on. 

I've been working on the big project of moving the ast from ruby code to mirah. I got the new ast written a while ago but actually usin switching to it has been harder than I expected. Most of the type inference code needed to be updated, so I  ended up porting it to mirah and implementing a new algorithm I've been planning for a while. This week I've gotten the typer tests passing and started updating the I've backend.

I haven't got any of this checked in yet since it's so broken, but things are starting to work now so I should probably get this committed in another branch.

Ryan Brown

-- Sent from my phone

Charles Oliver Nutter

unread,
Jul 22, 2011, 10:42:04 PM7/22/11
to mi...@googlegroups.com
Exciting stuff, Ryan. I'm looking forward to helping out with it once
you have something to land.

Moving more of Mirah into Mirah will start to pay dividends when it
comes to startup time, plus it will make it easier for us to interface
with Mirah's compiler chain from other JVM languages. That's the right
way to go.

A rewrite might make it easier for us to incorporate newer features
like boxing and generics, as well.

Ryan: If there's anything I can do to help you get it all functional,
let me know. I'm ok, hacking around on a (broken) branch too.

- Charlie

John Woodell

unread,
Jul 23, 2011, 1:28:43 AM7/23/11
to mi...@googlegroups.com
The current version of Mirah is proof that the concepts behind Mirah
are sound.
Reply all
Reply to author
Forward
0 new messages