Mongrel for 1.9, for real?

Mongrel for 1.9, for real?

[Luis Lavena]

Hey Guys,

Heads up on this...

I want to pursue release an updated Mongrel gem that can compile and
run under Windows Ruby 1.8 and 1.9

This is due the preview1 versions of the installer just released:

http://groups.google.com/group/comp.lang.ruby/browse_thread/thread/c38255aacd68d6dc

For that purpose, 1.1.5 needs to be updated:

- consider MinGW as valid platform (it doesn't)
- change RSTRING macros

Now, I've struggled all day trying to figure out from where I should
get all the pieces to move forward, all without success.

Patched several Rakefiles to get extensions build, but tests fails.
Patched extension to get Ruby 1.9, and tests stall.

Cannot release a Gem for Windows since Echoe hook itself and don't let
rake-compiler tasks run well.

So: Should I consider Mongrel (and fauna/mongrel) repository dead and
move forward?

Mongrel is still being actively used by MANY users, even some moved to
Passenger, or Thin, Mongrel is still an excellent option.

I just want the "team" answer so I decide to invest time on getting
that fixed or invest time on getting "another project inspired by
Mongrel" build properly on Windows.

Please let me know as soon as possible in your agenda, since I'm
expecting lot of reports of failures when using the preview1
installers.

Thanks in advance for the feedback and your time.

Regards,
-- 
Luis Lavena
AREA 17
-
Perfection in design is achieved not when there is nothing more to add,
but rather when there is nothing more to take away.
Antoine de Saint-Exupéry

Re: Mongrel for 1.9, for real?

[Evan Weaver]

Ugh, sorry. Mongrel is not dead, but rather severely neglected by me.

We still use it at Twitter, for what it's worth. I will gather up the
1.9 bits I have and send you the details.

What cross-compilation strategy did you want to use for MinGW?

Evan

On Sun, Jun 28, 2009 at 9:46 PM, Luis Lavena<luislavena-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> Hey Guys,
>
> Heads up on this...
>
> I want to pursue release an updated Mongrel gem that can compile and
> run under Windows Ruby 1.8 and 1.9
>
> This is due the preview1 versions of the installer just released:
>
> http://groups.google.com/group/comp.lang.ruby/browse_thread/thread/c38255aacd68d6dc
>
> For that purpose, 1.1.5 needs to be updated:
>
> - consider MinGW as valid platform (it doesn't)
> - change RSTRING macros
>
> Now, I've struggled all day trying to figure out from where I should
> get all the pieces to move forward, all without success.
>
> Patched several Rakefiles to get extensions build, but tests fails.
> Patched extension to get Ruby 1.9, and tests stall.
>
> Cannot release a Gem for Windows since Echoe hook itself and don't let
> rake-compiler tasks run well.
>
> So: Should I consider Mongrel (and fauna/mongrel) repository dead and
> move forward?
>
> Mongrel is still being actively used by MANY users, even some moved to
> Passenger, or Thin, Mongrel is still an excellent option.
>
> I just want the "team" answer so I decide to invest time on getting
> that fixed or invest time on getting "another project inspired by
> Mongrel" build properly on Windows.
>
> Please let me know as soon as possible in your agenda, since I'm
> expecting lot of reports of failures when using the preview1
> installers.
>
> Thanks in advance for the feedback and your time.
>
> Regards,
> --
> Luis Lavena
> AREA 17
> -
> Perfection in design is achieved not when there is nothing more to add,
> but rather when there is nothing more to take away.
> Antoine de Saint-Exupéry
> _______________________________________________
> Mongrel-development mailing list
> Mongrel-development-GrnCvJ7WPxnNLxjTenLetw@public.gmane.org
> http://rubyforge.org/mailman/listinfo/mongrel-development
>



-- 
Evan Weaver

Re: Mongrel for 1.9, for real?

[Luis Lavena]

On Mon, Jun 29, 2009 at 2:41 AM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
> Ugh, sorry. Mongrel is not dead, but rather severely neglected by me.
>

And me. Been focusing a lot on RubyInstaller and the related tools.

> We still use it at Twitter, for what it's worth. I will gather up the
> 1.9 bits I have and send you the details.
>

Cool, you know my email already ;-)

> What cross-compilation strategy did you want to use for MinGW?
>

rake-compiler and fat binaries:

http://github.com/luislavena/rake-compiler/tree/

Cheers,
-- 
Luis Lavena
AREA 17
-
Perfection in design is achieved not when there is nothing more to add,
but rather when there is nothing more to take away.
Antoine de Saint-Exupéry

Re: Mongrel for 1.9, for real?

[Evan Weaver]

http://github.com/fauna/mongrel/tree/spinoff

Ragel 6.5 builds, and tests pass on 1.8 and 1.9, except for:

test_deflate(HandlersTest):
Zlib::DataError: incorrect header check
    /Volumes/Projects/fauna/mongrel/test/test_handlers.rb:109:in
`block in test_deflate'

Which I don't understand right now.

Go nuts with the packaging; I'm still not sure what you're trying to
accomplish exactly.

Evan

On Sun, Jun 28, 2009 at 11:26 PM, Luis Lavena<luislavena-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On Mon, Jun 29, 2009 at 2:41 AM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
>> Ugh, sorry. Mongrel is not dead, but rather severely neglected by me.
>>
>
> And me. Been focusing a lot on RubyInstaller and the related tools.
>
>> We still use it at Twitter, for what it's worth. I will gather up the
>> 1.9 bits I have and send you the details.
>>
>
> Cool, you know my email already ;-)
>
>> What cross-compilation strategy did you want to use for MinGW?
>>
>
> rake-compiler and fat binaries:
>
> http://github.com/luislavena/rake-compiler/tree/
>
> Cheers,
> --
> Luis Lavena
> AREA 17
> -
> Perfection in design is achieved not when there is nothing more to add,
> but rather when there is nothing more to take away.
> Antoine de Saint-Exupéry
> _______________________________________________
> Mongrel-development mailing list
> Mongrel-development-GrnCvJ7WPxnNLxjTenLetw@public.gmane.org
> http://rubyforge.org/mailman/listinfo/mongrel-development
>



-- 
Evan Weaver

Re: Mongrel for 1.9, for real?

[Evan Weaver]

I think we should call this version 1.2, but that's the least of our concerns.

I have a 1.2 branch in fauna but I don't know what's in it anymore...will audit.

Evan

On Tue, Jul 7, 2009 at 4:51 PM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
> http://github.com/fauna/mongrel/tree/spinoff
>
> Ragel 6.5 builds, and tests pass on 1.8 and 1.9, except for:
>
> test_deflate(HandlersTest):
> Zlib::DataError: incorrect header check
>    /Volumes/Projects/fauna/mongrel/test/test_handlers.rb:109:in
> `block in test_deflate'
>
> Which I don't understand right now.
>
> Go nuts with the packaging; I'm still not sure what you're trying to
> accomplish exactly.
>
> Evan
>
> On Sun, Jun 28, 2009 at 11:26 PM, Luis Lavena<luislavena-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>> On Mon, Jun 29, 2009 at 2:41 AM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
>>> Ugh, sorry. Mongrel is not dead, but rather severely neglected by me.
>>>
>>
>> And me. Been focusing a lot on RubyInstaller and the related tools.
>>
>>> We still use it at Twitter, for what it's worth. I will gather up the
>>> 1.9 bits I have and send you the details.
>>>
>>
>> Cool, you know my email already ;-)
>>
>>> What cross-compilation strategy did you want to use for MinGW?
>>>
>>
>> rake-compiler and fat binaries:
>>
>> http://github.com/luislavena/rake-compiler/tree/
>>
>> Cheers,
>> --
>> Luis Lavena
>> AREA 17
>> -
>> Perfection in design is achieved not when there is nothing more to add,
>> but rather when there is nothing more to take away.
>> Antoine de Saint-Exupéry
>> _______________________________________________
>> Mongrel-development mailing list
>> Mongrel-development-GrnCvJ7WPxnNLxjTenLetw@public.gmane.org
>> http://rubyforge.org/mailman/listinfo/mongrel-development
>>
>
>
>
> --
> Evan Weaver
>



-- 
Evan Weaver

Re: Mongrel for 1.9, for real?

[Luis Lavena]

On Tue, Jul 7, 2009 at 8:56 PM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
> I think we should call this version 1.2, but that's the least of our concerns.
>

Well, internally it didn't change a lot.

Only packaging and cross compilation.

It also disabled the dependencies on fastthread and cgi multipart,
which means only work for >= 1.8.6

But maybe you're right, and we can call it 1.2

> I have a 1.2 branch in fauna but I don't know what's in it anymore...will audit.
>

Bummer, auditing sucks.

-- 
Luis Lavena
AREA 17
-
Perfection in design is achieved not when there is nothing more to add,
but rather when there is nothing more to take away.
Antoine de Saint-Exupéry

Re: Mongrel for 1.9, for real?

[Luis Lavena]

On Tue, Jul 7, 2009 at 8:51 PM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
> http://github.com/fauna/mongrel/tree/spinoff
>
> Ragel 6.5 builds, and tests pass on 1.8 and 1.9, except for:
>
> test_deflate(HandlersTest):
> Zlib::DataError: incorrect header check
>    /Volumes/Projects/fauna/mongrel/test/test_handlers.rb:109:in
> `block in test_deflate'
>
> Which I don't understand right now.
>

Me neither, since tested manually Zlib at IRB and the deflating worked
without the header.

> Go nuts with the packaging; I'm still not sure what you're trying to
> accomplish exactly.
>

I'm trying to simplify the building and release of new gems. MinGW
didn't work with 1.1.5 and catching up and build new releases for
Mongrel got stuck in 2007/08.

What I'm trying: allow mongrel "run" on 1.9, next version (1.2 or 2.0)
can go nuts and do Rack and proper threading on 1.9...

Right now, we just need it to "work".

-- 
Luis Lavena
AREA 17
-
Perfection in design is achieved not when there is nothing more to add,
but rather when there is nothing more to take away.
Antoine de Saint-Exupéry

Re: Mongrel for 1.9, for real?

[Evan Weaver]

Cursory view looks like the stable-1.2 branch had mostly packaging
changes (many for Windows) and Rack support. Agree with ignoring
Rack/threading in the interests of having a quick release.

To be clear, I did not merge the stable-1.2 branch into fauna/spinoff.
Stable-1.2 branch is dead and fauna/spinoff is my new HEAD based on
your's and Matt Aimonett's.

Evan

On Tue, Jul 7, 2009 at 5:06 PM, Luis Lavena<luislavena-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On Tue, Jul 7, 2009 at 8:51 PM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
>> http://github.com/fauna/mongrel/tree/spinoff
>>
>> Ragel 6.5 builds, and tests pass on 1.8 and 1.9, except for:
>>
>> test_deflate(HandlersTest):
>> Zlib::DataError: incorrect header check
>>    /Volumes/Projects/fauna/mongrel/test/test_handlers.rb:109:in
>> `block in test_deflate'
>>
>> Which I don't understand right now.
>>
>
> Me neither, since tested manually Zlib at IRB and the deflating worked
> without the header.
>
>> Go nuts with the packaging; I'm still not sure what you're trying to
>> accomplish exactly.
>>
>
> I'm trying to simplify the building and release of new gems. MinGW
> didn't work with 1.1.5 and catching up and build new releases for
> Mongrel got stuck in 2007/08.
>
> What I'm trying: allow mongrel "run" on 1.9, next version (1.2 or 2.0)
> can go nuts and do Rack and proper threading on 1.9...
>
> Right now, we just need it to "work".
>
> --
> Luis Lavena
> AREA 17
> -
> Perfection in design is achieved not when there is nothing more to add,
> but rather when there is nothing more to take away.
> Antoine de Saint-Exupéry
> _______________________________________________
> Mongrel-development mailing list
> Mongrel-development-GrnCvJ7WPxnNLxjTenLetw@public.gmane.org
> http://rubyforge.org/mailman/listinfo/mongrel-development
>



-- 
Evan Weaver

Re: Mongrel for 1.9, for real?

[Luis Lavena]

On Tue, Jul 7, 2009 at 9:11 PM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
> Cursory view looks like the stable-1.2 branch had mostly packaging
> changes (many for Windows) and Rack support. Agree with ignoring
> Rack/threading in the interests of having a quick release.
>

So we agree :-)

> To be clear, I did not merge the stable-1.2 branch into fauna/spinoff.
> Stable-1.2 branch is dead and fauna/spinoff is my new HEAD based on
> your's and Matt Aimonett's.
>

I don't like this commit:

http://github.com/mattetti/mongrel/commit/7c9d988d4de2e08d67f95ca209196427fd89c9af

I see a lot of changes there to be implemented in a point release.

Need more time to actually check them, or maybe slice them.

-- 
Luis Lavena
AREA 17
-
Perfection in design is achieved not when there is nothing more to add,
but rather when there is nothing more to take away.
Antoine de Saint-Exupéry

Re: Mongrel for 1.9, for real?

[Evan Weaver]

It's not my favorite but there is some magic in there.

Should be possible to slice out and look at one-by-one.

Evan

On Tue, Jul 7, 2009 at 5:15 PM, Luis Lavena<luislavena-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On Tue, Jul 7, 2009 at 9:11 PM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
>> Cursory view looks like the stable-1.2 branch had mostly packaging
>> changes (many for Windows) and Rack support. Agree with ignoring
>> Rack/threading in the interests of having a quick release.
>>
>
> So we agree :-)
>
>> To be clear, I did not merge the stable-1.2 branch into fauna/spinoff.
>> Stable-1.2 branch is dead and fauna/spinoff is my new HEAD based on
>> your's and Matt Aimonett's.
>>
>
> I don't like this commit:
>
> http://github.com/mattetti/mongrel/commit/7c9d988d4de2e08d67f95ca209196427fd89c9af
>
> I see a lot of changes there to be implemented in a point release.
>
> Need more time to actually check them, or maybe slice them.
>
> --
> Luis Lavena
> AREA 17
> -
> Perfection in design is achieved not when there is nothing more to add,
> but rather when there is nothing more to take away.
> Antoine de Saint-Exupéry
> _______________________________________________
> Mongrel-development mailing list
> Mongrel-development-GrnCvJ7WPxnNLxjTenLetw@public.gmane.org
> http://rubyforge.org/mailman/listinfo/mongrel-development
>



-- 
Evan Weaver

Re: Mongrel for 1.9, for real?

[Evan Weaver]

Luis,

What else do you need me to do/look at?

Evan

On Tue, Jul 7, 2009 at 5:18 PM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
> It's not my favorite but there is some magic in there.
>
> Should be possible to slice out and look at one-by-one.
>
> Evan
>
> On Tue, Jul 7, 2009 at 5:15 PM, Luis Lavena<luislavena-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>> On Tue, Jul 7, 2009 at 9:11 PM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
>>> Cursory view looks like the stable-1.2 branch had mostly packaging
>>> changes (many for Windows) and Rack support. Agree with ignoring
>>> Rack/threading in the interests of having a quick release.
>>>
>>
>> So we agree :-)
>>
>>> To be clear, I did not merge the stable-1.2 branch into fauna/spinoff.
>>> Stable-1.2 branch is dead and fauna/spinoff is my new HEAD based on
>>> your's and Matt Aimonett's.
>>>
>>
>> I don't like this commit:
>>
>> http://github.com/mattetti/mongrel/commit/7c9d988d4de2e08d67f95ca209196427fd89c9af
>>
>> I see a lot of changes there to be implemented in a point release.
>>
>> Need more time to actually check them, or maybe slice them.
>>
>> --
>> Luis Lavena
>> AREA 17
>> -
>> Perfection in design is achieved not when there is nothing more to add,
>> but rather when there is nothing more to take away.
>> Antoine de Saint-Exupéry
>> _______________________________________________
>> Mongrel-development mailing list
>> Mongrel-development-GrnCvJ7WPxnNLxjTenLetw@public.gmane.org
>> http://rubyforge.org/mailman/listinfo/mongrel-development
>>
>
>
>
> --
> Evan Weaver
>



-- 
Evan Weaver

Re: Mongrel for 1.9, for real?

[Luis Lavena]

On Tue, Jul 7, 2009 at 9:51 PM, Evan Weaver<evan-72XWLPH10WVXUHR/Jj/Uug@public.gmane.org> wrote:
> Luis,
>
> What else do you need me to do/look at?
>

I think that from the spinoff branch we can do a 1.1.6 release.

Let me outline the changes and by tomorrow let's see a checklist if
something is missing.
-- 
Luis Lavena
AREA 17
-
Perfection in design is achieved not when there is nothing more to add,
but rather when there is nothing more to take away.
Antoine de Saint-Exupéry