From: Mike Dalessio on
[Note: parts of this message were removed to make it a legal post.]

On Sun, Jan 24, 2010 at 7:15 AM, Charles Oliver Nutter
<headius(a)headius.com>wrote:

> On Sat, Jan 23, 2010 at 11:49 PM, Roger Pack <rogerpack2005(a)gmail.com>
> wrote:
> > Looks like pure java Nokogiri is something popular--the bounty on it has
> > already risen to $225
>
> It's probably the most oft-encountered stumbling block for folks using
> JRuby (these days), since Nokogiri itself has become very popular and
> is now depended on by many other libraries. A pure-Java version would
> never need special handling on any platform, would work on any
> platform where JRuby works, and would not require native library
> support at all.
>
> I implore gem authors: think about who you might hurt with hard gem
> dependencies on native extensions. At least provide an alternative
> path.
>

I'm in for $200 for a Java Nokogiri implementation. So is my partner in
crime, Aaron.

Consider it "conscience money" for all the puppies who died as a result of
us writing and maintaining the JRuby FFI port of Nokogiri. :)



>
> - Charlie
>
>


--
mike dalessio
mike(a)csa.net

From: Phillip Gawlowski on
On 25.01.2010 06:29, Aaron Patterson wrote:

> I implore Ruby implementors to support the MRI C api, as it too is part
> of Ruby's api.

Either provide mswin32 (compiled with VC6), or mingw32 binaries (yes,
binaries, not the source code), or use Java.

--
Phillip Gawlowski

From: Phillip Gawlowski on
On 25.01.2010 08:29, Phillip Gawlowski wrote:
> On 25.01.2010 06:29, Aaron Patterson wrote:
>
>> I implore Ruby implementors to support the MRI C api, as it too is part
>> of Ruby's api.
>
> Either provide mswin32 (compiled with VC6), or mingw32 binaries (yes,
> binaries, not the source code), or use Java.

"mswin32 *and* mingw32" is what I actually wanted to write. My apologies.

--
Phillip Gawlowski

From: Tony Arcieri on
[Note: parts of this message were removed to make it a legal post.]

On Sun, Jan 24, 2010 at 10:29 PM, Aaron Patterson <
aaron(a)tenderlovemaking.com> wrote:

> I thought FFI was supposed to solve this problem. Is it not?
>

It "solves" it to a certain degree, but when deploying in enterprise
environments it can be quite a hassle. Think: ancient versions of RHEL with
ancient versions of libxml/libxslt. That's not to mention problems with
32-bit vs 64-bit environments: for some reason the 64-bit version of RedHat
didn't package a 32-bit libxslt, but they did provide a 64-bit one.
However, we originally installed a 32-bit JRE. Eep! 64-bit JRE is an
"unofficially supported" configuration for our environment (they typically
run 32-bit JRE on 64-bit RedHat for some reason).

Nokogiri's great and we have everything working now (we had to do some
symlink hackery though). However, it would definitely streamline the
deployment process a lot if we didn't have native code dependencies, and
right now Nokogiri is the only one we have.

--
Tony Arcieri
Medioh! A Kudelski Brand

From: Eleanor McHugh on
On 25 Jan 2010, at 05:29, Aaron Patterson wrote:
> On Sun, Jan 24, 2010 at 09:15:56PM +0900, Charles Oliver Nutter wrote:
>> I implore gem authors: think about who you might hurt with hard gem
>> dependencies on native extensions. At least provide an alternative
>> path.
>
> I implore Ruby implementors to support the MRI C api, as it too is part
> of Ruby's api. Think about who you hurt by not letting people reuse
> valuable libraries written in C. :-)

I implore Ruby developers to write in Pure Ruby and demand all these Ruby Implementors solve their "performance" problems ;p


Ellie

Eleanor McHugh
Games With Brains
http://slides.games-with-brains.net
----
raise ArgumentError unless @reality.responds_to? :reason