From rosco at roscopeco.co.uk  Sun May  7 14:08:30 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 07 May 2006 19:08:30 +0100
Subject: [libxml-devel] Adding 'previous' node, adding doctype to doc.
In-Reply-To: <868dee660604280130kae91572t79d762e065e8f00b@mail.gmail.com>
References: <868dee660604280130kae91572t79d762e065e8f00b@mail.gmail.com>
Message-ID: <op.s860ogelkc9f7k@jukebox.roscopeco>

On Fri, 28 Apr 2006 09:30:44 +0100, Rik Hemsley <rikkus at gmail.com> wrote:

> Hi,
>
> I have a couple of small questions:
>
> I have XML like this:
>
> <a><b/></a>
> I would like to add a node before <b/>, i.e.:
> <a><c/><b/></a>
>
> With the C library of libxml2, I could use xmlAddPrevSibling()
> Is there an equivalent for ruby-libxml?
>

Currently, I think we only have support for adding a sibling after all  
existing siblings. I'm proposing the attached patch to add support for  
both prev= and next= on XML::Node.

> I would also like to know how it is possible to add a Dtd to a
> Document. There seems to be no dtd= or equivalent.
>

I'm afraid we still have some work to do in this area. I don't know of a  
way to achieve this at present.

Cheers,
-- 
Ross Bamford - rosco at roscopeco.co.uk
-------------- next part --------------
A non-text attachment was scrubbed...
Name: set_next_prev.patch
Type: application/octet-stream
Size: 5159 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/libxml-devel/attachments/20060507/2cc5db2a/attachment-0001.obj 

From rikkus at gmail.com  Sun May  7 17:17:41 2006
From: rikkus at gmail.com (Rik Hemsley)
Date: Sun, 7 May 2006 22:17:41 +0100
Subject: [libxml-devel] Adding 'previous' node, adding doctype to doc.
In-Reply-To: <op.s860ogelkc9f7k@jukebox.roscopeco>
References: <868dee660604280130kae91572t79d762e065e8f00b@mail.gmail.com>
	<op.s860ogelkc9f7k@jukebox.roscopeco>
Message-ID: <868dee660605071417w7e9f1676m3bcc0666a85f4fa2@mail.gmail.com>

Thanks for the reply. I've checked out the source from CVS and will
have a look to see if I can figure out how to add support for setting
a DTD.

Cheers,
Rik


From rosco at roscopeco.co.uk  Thu May 11 12:18:56 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Thu, 11 May 2006 17:18:56 +0100
Subject: [libxml-devel] XPath problems with validation
Message-ID: <op.s9d99uglkc9f7k@jukebox.roscopeco>

Hi,

Has anyone noticed strange behaviour when executing XPath against XML with  
a DOCTYPE ? Doesn't seem to work, but I wonder if I'm missing something  
before I call it a bug. I attached a testcase (which, incidentally, REXML  
does pass so I guess this probably is a bug).

Any heads-up on what might be wrong would be appreciated.

Cheers,
-- 
Ross Bamford - rosco at roscopeco.co.uk

From rosco at roscopeco.co.uk  Thu May 11 12:21:46 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Thu, 11 May 2006 17:21:46 +0100
Subject: [libxml-devel] XPath problems with validation
In-Reply-To: <op.s9d99uglkc9f7k@jukebox.roscopeco>
References: <op.s9d99uglkc9f7k@jukebox.roscopeco>
Message-ID: <op.s9eaek0gkc9f7k@jukebox.roscopeco>

On Thu, 11 May 2006 17:18:56 +0100, Ross Bamford <rosco at roscopeco.co.uk>  
wrote:

> I attached a testcase (which, incidentally, REXML
> does pass so I guess this probably is a bug).
>

Apologies for noise, forgot the attachment. Why do I *always* do that ?  
}:-|

-- 
Ross Bamford - rosco at roscopeco.co.uk
-------------- next part --------------
A non-text attachment was scrubbed...
Name: xpath-validation-bug.rb
Type: text/x-ruby-script
Size: 769 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/libxml-devel/attachments/20060511/307f7531/attachment.bin 

From mvette13 at gmail.com  Sun May 14 01:56:47 2006
From: mvette13 at gmail.com (Mark Van Holstyn)
Date: Sun, 14 May 2006 01:56:47 -0400
Subject: [libxml-devel] A couple suggested updates...
Message-ID: <e45a289c0605132256k2afa1822o866da073103763ab@mail.gmail.com>

Hey everyone,

I had a little time today and made a couple of more changes to XML::Node. I
changed [], []=, and new to all accept symbols OR string, rather than just
strings. Does this seems like good functionality to anyone else? Attatched
is the patch should you find this functionality useful.

Thanks,

Mark Van Holstyn

-- 
Mark Van Holstyn
mvette13 at gmail.com
http://lotswholetime.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060514/7410f6b3/attachment.htm 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ruby_xml_node.c.PATCH
Type: text/x-patch
Size: 711 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/libxml-devel/attachments/20060514/7410f6b3/attachment.bin 

From bendavis78 at gmail.com  Thu May 18 14:57:13 2006
From: bendavis78 at gmail.com (Ben Davis)
Date: Thu, 18 May 2006 13:57:13 -0500
Subject: [libxml-devel] How to loop through element attributes?
Message-ID: <67da7c6a0605181157v20eb804ah729d091640e30d1d@mail.gmail.com>

Hi,  I'm using libxml-ruby in my app, and there's a point where I need to
loop through all the attributes in a particular element node.  I've tried
the following (assuming I have the correct node object called "node"):

node.properties.each do |attr|
  puts "#{attr.name} = {attr.value}"
end

For some reason, this only outputs the first attribute.  The node definately
has more attribtues -- I can access them using   node["theattribute"] .

What is the correct way to loop through the attributes?  Is there something
wrong w/ my installation?  A bug?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060518/bd04d046/attachment.htm 

From rosco at roscopeco.co.uk  Fri May 19 03:42:15 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Fri, 19 May 2006 08:42:15 +0100
Subject: [libxml-devel] A couple suggested updates...
In-Reply-To: <e45a289c0605132256k2afa1822o866da073103763ab@mail.gmail.com>
References: <e45a289c0605132256k2afa1822o866da073103763ab@mail.gmail.com>
Message-ID: <op.s9sfopw0kc9f7k@jukebox.roscopeco>

On Sun, 14 May 2006 06:56:47 +0100, Mark Van Holstyn <mvette13 at gmail.com>  
wrote:

> Hey everyone,
>
> I had a little time today and made a couple of more changes to  
> XML::Node. I
> changed [], []=, and new to all accept symbols OR string, rather than  
> just
> strings. Does this seems like good functionality to anyone else?  
> Attatched
> is the patch should you find this functionality useful.
>

Thanks Mark, I like the idea of this change and I'll look at integrating  
it.

Cheers,
-- 
Ross Bamford - rosco at roscopeco.co.uk

From rosco at roscopeco.co.uk  Fri May 19 03:44:18 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Fri, 19 May 2006 08:44:18 +0100
Subject: [libxml-devel] How to loop through element attributes?
In-Reply-To: <67da7c6a0605181157v20eb804ah729d091640e30d1d@mail.gmail.com>
References: <67da7c6a0605181157v20eb804ah729d091640e30d1d@mail.gmail.com>
Message-ID: <op.s9sfr4d3kc9f7k@jukebox.roscopeco>

On Thu, 18 May 2006 19:57:13 +0100, Ben Davis <bendavis78 at gmail.com> wrote:

> Hi,  I'm using libxml-ruby in my app, and there's a point where I need to
> loop through all the attributes in a particular element node.  I've tried
> the following (assuming I have the correct node object called "node"):
>
> node.properties.each do |attr|
>   puts "#{attr.name} = {attr.value}"
> end
>
> For some reason, this only outputs the first attribute.  The node  
> definately
> has more attribtues -- I can access them using   node["theattribute"] .
>
> What is the correct way to loop through the attributes?  Is there  
> something
> wrong w/ my installation?  A bug?
>

I've just spent a frustrating few minutes looking at this, and right now I  
think it looks like a bug. I'm afraid I don't even have a workaround to  
suggest right now, but I'll work on this today and see what I come up with.

Thanks,
-- 
Ross Bamford - rosco at roscopeco.co.uk

From strass at strasslab.net  Wed May 24 11:39:27 2006
From: strass at strasslab.net (Yann Klis)
Date: Wed, 24 May 2006 17:39:27 +0200
Subject: [libxml-devel] Thank you for reviving libxml
Message-ID: <44747E2F.2080104@strasslab.net>

Hi,

Just a quick message to let you know that I'm very pleased with the Ruby 
bindings for libxml2. The performance are far greater than with REXML 
for example.
The API is not always very "nice", but I think you will make a good job 
improving it.

BTW, which version of libxml this page http://libxml.rubyforge.org/doc/ 
refers to ?

++

yk

From zdennis at mktec.com  Wed May 24 17:53:57 2006
From: zdennis at mktec.com (zdennis)
Date: Wed, 24 May 2006 17:53:57 -0400
Subject: [libxml-devel] Thank you for reviving libxml
In-Reply-To: <44747E2F.2080104@strasslab.net>
References: <44747E2F.2080104@strasslab.net>
Message-ID: <4474D5F5.10403@mktec.com>

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Yann Klis wrote:
> Hi,
> 
> Just a quick message to let you know that I'm very pleased with the Ruby 
> bindings for libxml2. The performance are far greater than with REXML 
> for example.
> The API is not always very "nice", but I think you will make a good job 
> improving it.

Can you suggest where you find the API not nice?


Zach
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFEdNX1Myx0fW1d8G0RAqt1AJ9sb5w7/QVPYIr+omyZxMpz9UrfcACfd26+
N9RcD3SP3zJS1SlL5hMOFVc=
=L0TX
-----END PGP SIGNATURE-----

From bryant.doug at gmail.com  Thu May 25 13:09:13 2006
From: bryant.doug at gmail.com (Doug Bryant)
Date: Thu, 25 May 2006 13:09:13 -0400
Subject: [libxml-devel] create a document from xml text in memory
Message-ID: <78cf1ade0605251009h75b9d0e6r90b2a28ae62c131e@mail.gmail.com>

I'm in the process of moving to libxml-ruby from rexml.  The speed was
killing me.

Is there a way to create an XML::Document from a string rather than a file?
I store all the documents in the database, so there is no file to parse
from.

I have looked through the api docs, but don't see how to accomplish this.

Thanks, Doug
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060525/6af9a674/attachment.htm 

From mvette13 at gmail.com  Thu May 25 14:09:14 2006
From: mvette13 at gmail.com (Mark Van Holstyn)
Date: Thu, 25 May 2006 14:09:14 -0400
Subject: [libxml-devel] create a document from xml text in memory
In-Reply-To: <78cf1ade0605251009h75b9d0e6r90b2a28ae62c131e@mail.gmail.com>
References: <78cf1ade0605251009h75b9d0e6r90b2a28ae62c131e@mail.gmail.com>
Message-ID: <e45a289c0605251109ld6d3d9cx8ef078bf33edebfc@mail.gmail.com>

parser=XML::Parser.new
parser.string= '<root><another_node>hey</another_node></root>'
doc = parser.parse

Anyone else have an easier way?

Mark

On 5/25/06, Doug Bryant <bryant.doug at gmail.com> wrote:
>
> I'm in the process of moving to libxml-ruby from rexml.  The speed was
> killing me.
>
> Is there a way to create an XML::Document from a string rather than a
> file? I store all the documents in the database, so there is no file to
> parse from.
>
> I have looked through the api docs, but don't see how to accomplish this.
>
> Thanks, Doug
>
> _______________________________________________
> libxml-devel mailing list
> libxml-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/libxml-devel
>
>


-- 
Mark Van Holstyn
mvette13 at gmail.com
http://lotswholetime.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060525/a18ee703/attachment.htm 

From strass at strasslab.net  Thu May 25 14:15:00 2006
From: strass at strasslab.net (Yann Klis)
Date: Thu, 25 May 2006 20:15:00 +0200
Subject: [libxml-devel] Thank you for reviving libxml
In-Reply-To: <4474D5F5.10403@mktec.com>
References: <44747E2F.2080104@strasslab.net> <4474D5F5.10403@mktec.com>
Message-ID: <4475F424.2050604@strasslab.net>

zdennis a ?crit :
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Yann Klis wrote:
>   
>> Hi,
>>
>> Just a quick message to let you know that I'm very pleased with the Ruby 
>> bindings for libxml2. The performance are far greater than with REXML 
>> for example.
>> The API is not always very "nice", but I think you will make a good job 
>> improving it.
>>     
>
> Can you suggest where you find the API not nice?
>   
Typically, loading an XML chunk from a file or from a string may have a 
similar API, IMO (funny that this issue has just been raised in another 
thread).
And it would be really cool if I could write something like 
bar.children['Foo'] to get the first child node of the node "bar" whose 
name is "Foo", or, at least, an easy way to iterate through the children 
of a particular node.

++

yk

From mvette13 at gmail.com  Thu May 25 14:20:49 2006
From: mvette13 at gmail.com (Mark Van Holstyn)
Date: Thu, 25 May 2006 14:20:49 -0400
Subject: [libxml-devel] Thank you for reviving libxml
In-Reply-To: <4475F424.2050604@strasslab.net>
References: <44747E2F.2080104@strasslab.net> <4474D5F5.10403@mktec.com>
	<4475F424.2050604@strasslab.net>
Message-ID: <e45a289c0605251120l231d05caxf7085c588098badf@mail.gmail.com>

you do something like this... bar.find_first('Foo') to get the first child
node of bar whose name is Foo. Also, to get all the children, you can do
bar.find('*').each { |child| do something }

mark

On 5/25/06, Yann Klis <strass at strasslab.net> wrote:
>
> zdennis a ?crit :
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> >
> > Yann Klis wrote:
> >
> >> Hi,
> >>
> >> Just a quick message to let you know that I'm very pleased with the
> Ruby
> >> bindings for libxml2. The performance are far greater than with REXML
> >> for example.
> >> The API is not always very "nice", but I think you will make a good job
> >> improving it.
> >>
> >
> > Can you suggest where you find the API not nice?
> >
> Typically, loading an XML chunk from a file or from a string may have a
> similar API, IMO (funny that this issue has just been raised in another
> thread).
> And it would be really cool if I could write something like
> bar.children['Foo'] to get the first child node of the node "bar" whose
> name is "Foo", or, at least, an easy way to iterate through the children
> of a particular node.
>
> ++
>
> yk
> _______________________________________________
> libxml-devel mailing list
> libxml-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/libxml-devel
>



-- 
Mark Van Holstyn
mvette13 at gmail.com
http://lotswholetime.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060525/52410c31/attachment.htm 

From bryant.doug at gmail.com  Thu May 25 14:24:46 2006
From: bryant.doug at gmail.com (Doug Bryant)
Date: Thu, 25 May 2006 14:24:46 -0400
Subject: [libxml-devel] create a document from xml text in memory
In-Reply-To: <e45a289c0605251109ld6d3d9cx8ef078bf33edebfc@mail.gmail.com>
References: <78cf1ade0605251009h75b9d0e6r90b2a28ae62c131e@mail.gmail.com>
	<e45a289c0605251109ld6d3d9cx8ef078bf33edebfc@mail.gmail.com>
Message-ID: <78cf1ade0605251124j4182fafej21a7371deb1d6644@mail.gmail.com>

Thanks Mark,

That will get me by just fine for now.

It would be nice if XML::Document.new took a string.  Then you could use
files & strings with the same method.

doc = XML::Document.new("<foo><bar>doug</bar></foo>")

or

doc = XML::Document.new(File.open("my_file.xml"))

Doug



On 5/25/06, Mark Van Holstyn <mvette13 at gmail.com> wrote:
>
> parser=XML::Parser.new
> parser.string= '<root><another_node>hey</another_node></root>'
> doc = parser.parse
>
> Anyone else have an easier way?
>
> Mark
>
> On 5/25/06, Doug Bryant <bryant.doug at gmail.com> wrote:
>
> > I'm in the process of moving to libxml-ruby from rexml.  The speed was
> > killing me.
> >
> > Is there a way to create an XML::Document from a string rather than a
> > file? I store all the documents in the database, so there is no file to
> > parse from.
> >
> > I have looked through the api docs, but don't see how to accomplish
> > this.
> >
> > Thanks, Doug
> >
> > _______________________________________________
> > libxml-devel mailing list
> > libxml-devel at rubyforge.org
> > http://rubyforge.org/mailman/listinfo/libxml-devel
> >
> >
>
>
> --
> Mark Van Holstyn
> mvette13 at gmail.com
> http://lotswholetime.com
>
> _______________________________________________
> libxml-devel mailing list
> libxml-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/libxml-devel
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060525/59c56f9d/attachment.htm 

From strass at strasslab.net  Thu May 25 14:30:30 2006
From: strass at strasslab.net (Yann Klis)
Date: Thu, 25 May 2006 20:30:30 +0200
Subject: [libxml-devel] Thank you for reviving libxml
In-Reply-To: <e45a289c0605251120l231d05caxf7085c588098badf@mail.gmail.com>
References: <44747E2F.2080104@strasslab.net>
	<4474D5F5.10403@mktec.com>	<4475F424.2050604@strasslab.net>
	<e45a289c0605251120l231d05caxf7085c588098badf@mail.gmail.com>
Message-ID: <4475F7C6.8060808@strasslab.net>

Great, thanks.
Maybe the API could reflect this a little bit better.
Or there could be a small "cookbook" on the website for this kind of 
"standard" case.

++

yk

Mark Van Holstyn a ?crit :
> you do something like this... bar.find_first('Foo') to get the first 
> child node of bar whose name is Foo. Also, to get all the children, 
> you can do bar.find('*').each { |child| do something }
>
> mark
>
> On 5/25/06, *Yann Klis* <strass at strasslab.net 
> <mailto:strass at strasslab.net>> wrote:
>
>     zdennis a ?crit :
>     > -----BEGIN PGP SIGNED MESSAGE-----
>     > Hash: SHA1
>     >
>     > Yann Klis wrote:
>     >
>     >> Hi,
>     >>
>     >> Just a quick message to let you know that I'm very pleased with
>     the Ruby
>     >> bindings for libxml2. The performance are far greater than with
>     REXML
>     >> for example.
>     >> The API is not always very "nice", but I think you will make a
>     good job
>     >> improving it.
>     >>
>     >
>     > Can you suggest where you find the API not nice?
>     >
>     Typically, loading an XML chunk from a file or from a string may
>     have a
>     similar API, IMO (funny that this issue has just been raised in
>     another
>     thread).
>     And it would be really cool if I could write something like
>     bar.children['Foo'] to get the first child node of the node "bar"
>     whose
>     name is "Foo", or, at least, an easy way to iterate through the
>     children
>     of a particular node.
>
>     ++
>
>     yk
>     _______________________________________________
>     libxml-devel mailing list
>     libxml-devel at rubyforge.org <mailto:libxml-devel at rubyforge.org>
>     http://rubyforge.org/mailman/listinfo/libxml-devel
>
>
>
>
> -- 
> Mark Van Holstyn
> mvette13 at gmail.com <mailto:mvette13 at gmail.com>
> http://lotswholetime.com
> ------------------------------------------------------------------------
>
> _______________________________________________
> libxml-devel mailing list
> libxml-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/libxml-devel


From bryant.doug at gmail.com  Thu May 25 14:47:53 2006
From: bryant.doug at gmail.com (Doug Bryant)
Date: Thu, 25 May 2006 14:47:53 -0400
Subject: [libxml-devel] how to replace a node with another node?
Message-ID: <78cf1ade0605251147t48acb778r68076a3cb905b035@mail.gmail.com>

The one other thing (hopefully this is the last) that I can not figure out
how to do is replace a node with another node.

say I have

<foo>
  <bar>
    <baz>
      <woot/>
   </baz>
  </bar>
</foo>

I want to replace baz with <new_data/> so the result would look like this

<foo>
  <bar>
    <new_data/>
  </bar>
</foo>

with rexml, I had a helper method which would do it that consisted of

def replace_element(target_element, replacement_element)
   target_element.replace_with(replacement_element)
end

With rexml, I could also get an elements position relative to it's parent
and replace by index.

Any suggestions?  Thanks for helping the noob.

Doug
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060525/0261b632/attachment.htm 

From rosco at roscopeco.co.uk  Thu May 25 14:48:18 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Thu, 25 May 2006 19:48:18 +0100
Subject: [libxml-devel] create a document from xml text in memory
In-Reply-To: <e45a289c0605251109ld6d3d9cx8ef078bf33edebfc@mail.gmail.com>
References: <78cf1ade0605251009h75b9d0e6r90b2a28ae62c131e@mail.gmail.com>
	<e45a289c0605251109ld6d3d9cx8ef078bf33edebfc@mail.gmail.com>
Message-ID: <op.s94eisz2kc9f7k@jukebox.roscopeco>

On Thu, 25 May 2006 19:09:14 +0100, Mark Van Holstyn <mvette13 at gmail.com>  
wrote:

> parser=XML::Parser.new
> parser.string= '<root><another_node>hey</another_node></root>'
> doc = parser.parse
>
> Anyone else have an easier way?
>

I tend to use:

	doc = XML::Parser.string(str).parse

(unless the XML is inlined) but obviously that's just a matter of style. A  
few times I've thought we should support a Document.string method, guess  
it's about time I looked into that...

-- 
Ross Bamford - rosco at roscopeco.co.uk

From rosco at roscopeco.co.uk  Thu May 25 15:28:51 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Thu, 25 May 2006 20:28:51 +0100
Subject: [libxml-devel] how to replace a node with another node?
In-Reply-To: <78cf1ade0605251147t48acb778r68076a3cb905b035@mail.gmail.com>
References: <78cf1ade0605251147t48acb778r68076a3cb905b035@mail.gmail.com>
Message-ID: <op.s94gedjxkc9f7k@jukebox.roscopeco>

On Thu, 25 May 2006 19:47:53 +0100, Doug Bryant <bryant.doug at gmail.com>  
wrote:

> The one other thing (hopefully this is the last) that I can not figure  
> out
> how to do is replace a node with another node.
> [...]
> Any suggestions?  Thanks for helping the noob.
>

We should probably have a replace_with method on XML::Node. It can be done  
easily enough from ruby:

class XML::Node
   def replace_with(other)
     self.next = other
     remove!
   end
end

You'll have to give me some time on the replace by index functionality  
 from REXML. In the meantime, anyone have comments on adding this in?

-- 
Ross Bamford - rosco at roscopeco.co.uk

From bryant.doug at gmail.com  Thu May 25 15:50:24 2006
From: bryant.doug at gmail.com (Doug Bryant)
Date: Thu, 25 May 2006 15:50:24 -0400
Subject: [libxml-devel] how to replace a node with another node?
In-Reply-To: <op.s94gedjxkc9f7k@jukebox.roscopeco>
References: <78cf1ade0605251147t48acb778r68076a3cb905b035@mail.gmail.com>
	<op.s94gedjxkc9f7k@jukebox.roscopeco>
Message-ID: <78cf1ade0605251250p71833be9hcccef34244854c0d@mail.gmail.com>

Thanks Ross.  That will work well for now.

Are you responsible for reviving libxml-ruby?  If so,  thank you very much!

Doug

On 5/25/06, Ross Bamford <rosco at roscopeco.co.uk> wrote:
>
> On Thu, 25 May 2006 19:47:53 +0100, Doug Bryant <bryant.doug at gmail.com>
> wrote:
>
> > The one other thing (hopefully this is the last) that I can not figure
> > out
> > how to do is replace a node with another node.
> > [...]
> > Any suggestions?  Thanks for helping the noob.
> >
>
> We should probably have a replace_with method on XML::Node. It can be done
> easily enough from ruby:
>
> class XML::Node
>    def replace_with(other)
>      self.next = other
>      remove!
>    end
> end
>
> You'll have to give me some time on the replace by index functionality
> from REXML. In the meantime, anyone have comments on adding this in?
>
> --
> Ross Bamford - rosco at roscopeco.co.uk
> _______________________________________________
> libxml-devel mailing list
> libxml-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/libxml-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060525/6ee4abb4/attachment.htm 

From bryant.doug at gmail.com  Thu May 25 16:01:48 2006
From: bryant.doug at gmail.com (Doug Bryant)
Date: Thu, 25 May 2006 16:01:48 -0400
Subject: [libxml-devel] how to replace a node with another node?
In-Reply-To: <78cf1ade0605251250p71833be9hcccef34244854c0d@mail.gmail.com>
References: <78cf1ade0605251147t48acb778r68076a3cb905b035@mail.gmail.com>
	<op.s94gedjxkc9f7k@jukebox.roscopeco>
	<78cf1ade0605251250p71833be9hcccef34244854c0d@mail.gmail.com>
Message-ID: <78cf1ade0605251301w5a63718cg9f98c3b09c793a80@mail.gmail.com>

crud.  that's not going to work.  next is a read-only property.  trying to
assign to it gives this error

NoMethodError: undefined method `next='

Any other ideas?

Doug

On 5/25/06, Doug Bryant <bryant.doug at gmail.com> wrote:
>
> Thanks Ross.  That will work well for now.
>
> Are you responsible for reviving libxml-ruby?  If so,  thank you very
> much!
>
> Doug
>
>
> On 5/25/06, Ross Bamford <rosco at roscopeco.co.uk> wrote:
> >
> > On Thu, 25 May 2006 19:47:53 +0100, Doug Bryant < bryant.doug at gmail.com>
> > wrote:
> >
> > > The one other thing (hopefully this is the last) that I can not figure
> > > out
> > > how to do is replace a node with another node.
> > > [...]
> > > Any suggestions?  Thanks for helping the noob.
> > >
> >
> > We should probably have a replace_with method on XML::Node. It can be
> > done
> > easily enough from ruby:
> >
> > class XML::Node
> >    def replace_with(other)
> >      self.next = other
> >      remove!
> >    end
> > end
> >
> > You'll have to give me some time on the replace by index functionality
> > from REXML. In the meantime, anyone have comments on adding this in?
> >
> > --
> > Ross Bamford - rosco at roscopeco.co.uk
> > _______________________________________________
> > libxml-devel mailing list
> > libxml-devel at rubyforge.org
> > http://rubyforge.org/mailman/listinfo/libxml-devel
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060525/cde84ea1/attachment-0001.htm 

From bryant.doug at gmail.com  Thu May 25 16:11:34 2006
From: bryant.doug at gmail.com (Doug Bryant)
Date: Thu, 25 May 2006 16:11:34 -0400
Subject: [libxml-devel] how to replace a node with another node?
In-Reply-To: <78cf1ade0605251301w5a63718cg9f98c3b09c793a80@mail.gmail.com>
References: <78cf1ade0605251147t48acb778r68076a3cb905b035@mail.gmail.com>
	<op.s94gedjxkc9f7k@jukebox.roscopeco>
	<78cf1ade0605251250p71833be9hcccef34244854c0d@mail.gmail.com>
	<78cf1ade0605251301w5a63718cg9f98c3b09c793a80@mail.gmail.com>
Message-ID: <78cf1ade0605251311v632ca82brf7b73dcaacf92f5b@mail.gmail.com>

I just saw that next= & previous= was implemented a couple of weeks ago in
the trunk.

I'll give that a try.

Doug


On 5/25/06, Doug Bryant <bryant.doug at gmail.com> wrote:
>
> crud.  that's not going to work.  next is a read-only property.  trying to
> assign to it gives this error
>
> NoMethodError: undefined method `next='
>
> Any other ideas?
>
> Doug
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060525/0b7b6fb4/attachment.htm 

From bryant.doug at gmail.com  Thu May 25 22:03:40 2006
From: bryant.doug at gmail.com (Doug Bryant)
Date: Thu, 25 May 2006 22:03:40 -0400
Subject: [libxml-devel] error occurring after program using libxml-ruby runs
Message-ID: <78cf1ade0605251903y4d1099beh4140e23d44afb841@mail.gmail.com>

I just finished changing a portion of my code from rexml to
libxml-ruby.  Processing time went from 40 minutes to 3 seconds.
How's that for a speed increase!

Anyhow, when the program exits, ruby crashes with the below output.  I
thought I remembered seeing something about GCC4 and libxml-ruby, but
can't seem to find it now.

The program runs completely to the end.  It looks like the error
occurs when ruby is trying to clean up after itself.

[doug at arctic captiveware]$ gcc --version
gcc (GCC) 4.1.0 20060304 (Red Hat 4.1.0-3)

Any ideas?

Thanks, Doug


FORM GENERATED
Trying to write to file
After trying to write to file
nothing left to do but print this message
*** glibc detected *** ruby: munmap_chunk(): invalid pointer: 0x071f1804 ***
======= Backtrace: =========
/lib/libc.so.6(__libc_free+0x179)[0xbb04f0]
ruby(ruby_xfree+0x26)[0x806d936]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x16e)[0x713deae]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeDoc+0xcb)[0x713dbfb]
/usr/local/lib/ruby/site_ruby/1.8/i686-linux/xml/libxml_so.so(ruby_xml_document_free+0x95)[0x89e735]
ruby(rb_gc_call_finalizer_at_exit+0x87)[0x806db77]
ruby[0x80573e9]
ruby(ruby_cleanup+0xcc)[0x806327c]
ruby(ruby_stop+0x11)[0x8063351]
ruby[0x8068831]
ruby[0x80521d4]
/lib/libc.so.6(__libc_start_main+0xdc)[0xb5e724]
ruby[0x8052111]
======= Memory map: ========
00111000-00115000 r-xp 00000000 08:03 721638
/usr/local/lib/ruby/1.8/i686-linux/strscan.so
00115000-00116000 rwxp 00003000 08:03 721638
/usr/local/lib/ruby/1.8/i686-linux/strscan.so
00116000-0011a000 r-xp 00000000 08:03 721616
/usr/local/lib/ruby/1.8/i686-linux/stringio.so
0011a000-0011b000 rwxp 00003000 08:03 721616
/usr/local/lib/ruby/1.8/i686-linux/stringio.so
0011b000-00135000 r-xp 00000000 08:03 721634
/usr/local/lib/ruby/1.8/i686-linux/syck.so
00135000-00136000 rwxp 00019000 08:03 721634
/usr/local/lib/ruby/1.8/i686-linux/syck.so
00136000-0013f000 r-xp 00000000 08:03 721649
/usr/local/lib/ruby/1.8/i686-linux/socket.so
0013f000-00140000 rwxp 00008000 08:03 721649
/usr/local/lib/ruby/1.8/i686-linux/socket.so
00140000-00142000 r-xp 00000000 08:03 721605
/usr/local/lib/ruby/1.8/i686-linux/etc.so
00142000-00143000 rwxp 00001000 08:03 721605
/usr/local/lib/ruby/1.8/i686-linux/etc.so
00143000-00146000 r-xp 00000000 08:03 721607
/usr/local/lib/ruby/1.8/i686-linux/iconv.so
00146000-00147000 rwxp 00002000 08:03 721607
/usr/local/lib/ruby/1.8/i686-linux/iconv.so
00147000-0014f000 r-xp 00000000 08:03 721641
/usr/local/lib/ruby/1.8/i686-linux/zlib.so
0014f000-00150000 rwxp 00007000 08:03 721641
/usr/local/lib/ruby/1.8/i686-linux/zlib.so
00151000-00186000 r-xp 00000000 08:03 721609
/usr/local/lib/ruby/1.8/i686-linux/openssl.so
00186000-00188000 rwxp 00034000 08:03 721609
/usr/local/lib/ruby/1.8/i686-linux/openssl.so
00188000-00197000 r-xp 00000000 08:03 721608
/usr/local/lib/ruby/1.8/i686-linux/nkf.so
00197000-001a9000 rwxp 0000f000 08:03 721608
/usr/local/lib/ruby/1.8/i686-linux/nkf.so
001a9000-001aa000 rwxp 001a9000 00:00 0
001aa000-001b1000 r-xp 00000000 08:03 721349
/usr/local/lib/ruby/site_ruby/1.8/i686-linux/postgres.so
001b1000-001b2000 rwxp 00006000 08:03 721349
/usr/local/lib/ruby/site_ruby/1.8/i686-linux/postgres.so
001b2000-001bb000 r-xp 00000000 08:01 192232     /lib/libnss_files-2.4.so
001bb000-001bc000 r-xp 00008000 08:01 192232     /lib/libnss_files-2.4.so
001bc000-001bd000 rwxp 00009000 08:01 192232     /lib/libnss_files-2.4.so
001bd000-001bf000 r-xp 00000000 08:03 721633
/usr/local/lib/ruby/1.8/i686-linux/enumerator.so
001bf000-001c0000 rwxp 00001000 08:03 721633
/usr/local/lib/ruby/1.8/i686-linux/enumerator.so
00242000-0024d000 r-xp 00000000 08:01 192231
/lib/libgcc_s-4.1.0-20060304.so.1
0024d000-0024e000 rwxp 0000a000 08:01 192231
/lib/libgcc_s-4.1.0-20060304.so.1
002bb000-002bc000 r-xp 002bb000 00:00 0          [vdso]
002bc000-002d5000 r-xp 00000000 08:01 192210     /lib/ld-2.4.so
002d5000-002d6000 r-xp 00018000 08:01 192210     /lib/ld-2.4.so
002d6000-002d7000 rwxp 00019000 08:01 192210     /lib/ld-2.4.so
00310000-00312000 r-xp 00000000 08:01 192331     /lib/libcom_err.so.2.1
00312000-00313000 rwxp 00001000 08:01 192331     /lib/libcom_err.so.2.1
00315000-00318000 r-xp 00000000 08:03 525647     /usr/lib/libkrb5support.so.0.0
00318000-00319000 rwxp 00002000 08:03 525647     /usr/lib/libkrb5support.so.0.0
0031b000-0033f000 r-xp 00000000 08:03 527969     /usr/lib/libk5crypto.so.3.0
0033f000-00340000 rwxp 00024000 08:03 527969     /usr/lib/libk5crypto.so.3.0
00342000-0035a000 r-xp 00000000 08:03 527971     /usr/lib/libgssapi_krb5.so.2.2
0035a000-0035b000 rwxp 00017000 08:03 527971     /usr/lib/libgssapi_krb5.so.2.2
00389000-003a8000 r-xp 00000000 08:03 526444     /usr/lib/libpq.so.4.1
003a8000-003a9000 rwxp 0001f000 08:03 526444     /usr/lib/libpq.so.4.1
003c5000-003d7000 r-xp 00000000 08:01 192243     /lib/libnsl-2.4.so
003d7000-003d8000 r-xp 00011000 08:01 192243     /lib/libnsl-2.4.so
003d8000-003d9000 rwxp 00012000 08:01 192243     /lib/libnsl-2.4.so
003d9000-003db000 rwxp 003d9000 00:00 0
003dd000-004fc000 r-xp 00000000 08:01 192333     /lib/libcrypto.so.0.9.8a
004fc000-0050f000 rwxp 0011e000 08:01 192333     /lib/libcrypto.so.0.9.8a
0050f000-00512000 rwxp 0050f000 00:00 0
00514000-00587000 r-xp 00000000 08:03 527970     /usr/lib/libkrb5.so.3.2
00587000-00589000 rwxp 00073000 08:03 527970     /usr/lib/libkrb5.so.3.2
0058b000-005cc000 r-xp 00000000 08:01 192430     /lib/libssl.so.0.9.8a
005cc000-005d0000 rwxp 00040000 08:01 192430     /lib/libssl.so.0.9.8a
0061a000-0061b000 r-xp 00000000 08:03 721619
/usr/local/lib/ruby/1.8/i686-linux/digest/md5.so
0061b000-0061c000 rwxp 00000000 08:03 721619
/usr/local/lib/ruby/1.8/i686-linux/digest/md5.so
006c7000-006d6000 r-xp 00000000 08:01 192255     /lib/libresolv-2.4.so
006d6000-006d7000 r-xp 0000e000 08:01 192255     /lib/libresolv-2.4.so
006d7000-006d8000 rwxp 0000f000 08:01 192255     /lib/libresolv-2.4.so
006d8000-006da000 rwxp 006d8000 00:00 0
0072b000-0072d000 r-xp 00000000 08:03 721611
/usr/local/lib/ruby/1.8/i686-linux/digest.so
0072d000-0072e000 rwxp 00001000 08:03 721611
/usr/local/lib/ruby/1.8/i686-linux/digest.so
00888000-008a1000 r-xp 00000000 08:03 895239
/usr/local/lib/ruby/site_ruby/1.8/i686-linux/xml/libxml_so.so
008a1000-008a2000 rwxp 00019000 08:03 895239
/usr/local/lib/ruby/site_ruby/1.8/i686-linux/xml/libxml_so.so
00b49000-00c76000 r-xp 00000000 08:01 192217     /lib/libc-2.4.so
00c76000-00c78000 r-xp 0012d000 08:01 192217     /lib/libc-2.4.so
00c78000-00c79000 rwxp 0012f000 08:01 192217     /lib/libc-2.4.so
00c79000-00c7c000 rwxp 00c79000 00:00 0
00c7e000-00ca1000 r-xp 00000000 08:01 192221     /lib/libm-2.4.so
00ca1000-00ca2000 r-xp 00022000 08:01 192221     /lib/libm-2.4.so
00ca2000-00ca3000 rwxp 00023000 08:01 192221     /lib/libm-2.4.so
00ca5000-00ca7000 r-xp 00000000 08:01 192223     /lib/libdl-2.4.so
00ca7000-00ca8000 r-xp 00001000 08:01 192223     /lib/libdl-2.4.so
00ca8000-00ca9000 rwxp 00002000 08:01 192223     /lib/libdl-2.4.so
00cab000-00cbd000 r-xp 00000000 08:03 527663     /usr/lib/libz.so.1.2.3
00cbd000-00cbe000 rwxp 00011000 08:03 527663     /usr/lib/libz.so.1.2.3
00dde000-00dee000 r-xp 00000000 08:01 192227     /lib/libpthread-2.4.so
00dee000-00def000 r-xp 0000f000 08:01 192227     /lib/libpthread-2.4.so
00def000-00df0000 rwxp 00010000 08:01 192227     /lib/libpthread-2.4.so
00df0000-00df2000 rwxp 00df0000 00:00 0
00fa4000-00fa7000 r-xp 00000000 08:03 721626
/usr/local/lib/ruby/1.8/i686-linux/racc/cparse.so
00fa7000-00fa8000 rwxp 00002000 08:03 721626
/usr/local/lib/ruby/1.8/i686-linux/racc/cparse.so
00fdb000-00fdc000 r-xp 00000000 08:03 721629
/usr/local/lib/ruby/1.8/i686-linux/fcntl.so
00fdc000-00fdd000 rwxp 00000000 08:03 721629
/usr/local/lib/ruby/1.8/i686-linux/fcntl.so
070fe000-07222000 r-xp 00000000 08:03 527809     /usr/lib/libxml2.so.2.6.23
07222000-0722a000 rwxp 00124000 08:03 527809     /usr/lib/libxml2.so.2.6.23
0722a000-0722b000 rwxp 0722a000 00:00 0
0722d000-07232000 r-xp 00000000 08:01 192245     /lib/libcrypt-2.4.so
07232000-07233000 r-xp 00004000 08:01 192245     /lib/libcrypt-2.4.so
07233000-07234000 rwxp 00005000 08:01 192245     /lib/libcrypt-2.4.so
07234000-0725b000 rwxp 07234000 00:00 0
08048000-080ec000 r-xp 00000000 08:03 556491     /usr/local/bin/ruby
080ec000-080ed000 rw-p 000a4000 08:03 556491     /usr/local/bin/ruby
080ed000-080fd000 rw-p 080ed000 00:00 0
09519000-134b6000 rw-p 09519000 00:00 0          [heap]
b758f000-b76f6000 rw-p b758f000 00:00 0
b7737000-b776d000 rw-p b7737000 00:00 0
b77e4000-b7d80000 rw-p b77e4000 00:00 0
b7d81000-b7fc8000 rw-p b7d81000 00:00 0
b7fe3000-b7fe5000 rw-p b7fe3000 00:00 0
bfbc8000-bfbe4000 rw-p bfbc8000 00:00 0          [stack]
Aborted


From rosco at roscopeco.co.uk  Sun May 28 09:39:51 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 28 May 2006 14:39:51 +0100
Subject: [libxml-devel] exposing _private data
Message-ID: <op.s99j8pjqkc9f7k@jukebox.roscopeco>

Feature request 4565 [1] concerns exposing the libxml2 _private member in  
libxml-ruby. This looks doable, but I wonder if anyone who knows more  
about libxml2 than I do has any thoughts? One potential problem I can  
think of is that libxslt apparently makes use of _private, so it might  
open up a can of worms when they're used together...?

Thanks,
Ross

[1]:  
http://rubyforge.org/tracker/index.php?func=detail&aid=4565&group_id=494&atid=1974
-- 
Ross Bamford - rosco at roscopeco.co.uk

From rosco at roscopeco.co.uk  Sun May 28 09:44:55 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 28 May 2006 14:44:55 +0100
Subject: [libxml-devel] how to replace a node with another node?
In-Reply-To: <78cf1ade0605251250p71833be9hcccef34244854c0d@mail.gmail.com>
References: <78cf1ade0605251147t48acb778r68076a3cb905b035@mail.gmail.com>
	<op.s94gedjxkc9f7k@jukebox.roscopeco>
	<78cf1ade0605251250p71833be9hcccef34244854c0d@mail.gmail.com>
Message-ID: <op.s99kg5jmkc9f7k@jukebox.roscopeco>

On Thu, 25 May 2006 20:50:24 +0100, Doug Bryant <bryant.doug at gmail.com>  
wrote:

> Thanks Ross.  That will work well for now.
>
> Are you responsible for reviving libxml-ruby?  If so,  thank you very  
> much!
>

Cool, I assume you got it working with the CVS head - I forgot to mention  
that, sorry...

I wouldn't say I was responsible for reviving the project, more that I've  
been the one applying some great patches from the community. I've done  
some coding but a lot of the fixes and updates have come from people on  
this list. But in any case, thanks for the kind words, and you're welcome  
:)

-- 
Ross Bamford - rosco at roscopeco.co.uk

From rosco at roscopeco.co.uk  Sun May 28 09:50:28 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 28 May 2006 14:50:28 +0100
Subject: [libxml-devel] Thank you for reviving libxml
In-Reply-To: <4475F7C6.8060808@strasslab.net>
References: <44747E2F.2080104@strasslab.net> <4474D5F5.10403@mktec.com>
	<4475F424.2050604@strasslab.net>
	<e45a289c0605251120l231d05caxf7085c588098badf@mail.gmail.com>
	<4475F7C6.8060808@strasslab.net>
Message-ID: <op.s99kqei3kc9f7k@jukebox.roscopeco>

On Thu, 25 May 2006 19:30:30 +0100, Yann Klis <strass at strasslab.net> wrote:

> Mark Van Holstyn a ?crit :
>> you do something like this... bar.find_first('Foo') to get the first
>> child node of bar whose name is Foo. Also, to get all the children,
>> you can do bar.find('*').each { |child| do something }
>>
> Great, thanks.
> Maybe the API could reflect this a little bit better.
> Or there could be a small "cookbook" on the website for this kind of
> "standard" case.
>

I have been planning to write some 'cookbook' style documentation for the  
website, but at the minute I'm quite pressed for spare time. I agree it's  
definitely something we should have, though, and will try to get something  
together asap.

Thanks,
-- 
Ross Bamford - rosco at roscopeco.co.uk

From rosco at roscopeco.co.uk  Sun May 28 10:05:35 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 28 May 2006 15:05:35 +0100
Subject: [libxml-devel] A couple suggested updates...
In-Reply-To: <e45a289c0605132256k2afa1822o866da073103763ab@mail.gmail.com>
References: <e45a289c0605132256k2afa1822o866da073103763ab@mail.gmail.com>
Message-ID: <op.s99lfly2kc9f7k@jukebox.roscopeco>

On Sun, 14 May 2006 06:56:47 +0100, Mark Van Holstyn <mvette13 at gmail.com>  
wrote:

> I changed [], []=, and new to all accept symbols OR string, rather than  
> just strings. Does this seems like good functionality to anyone else?  
> Attatched is the patch should you find this functionality useful.

Thanks again Mark, I've applied this in cvs HEAD.

Cheers,
-- 
Ross Bamford - rosco at roscopeco.co.uk

From rosco at roscopeco.co.uk  Wed May 31 14:49:23 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Wed, 31 May 2006 19:49:23 +0100
Subject: [libxml-devel] Sax Api - reprise?
Message-ID: <op.tafiklz1kc9f7k@jukebox.roscopeco>

Hi,

Since implementing that first-cut Sax API, I've been having a few regrets  
about doing it with Procs - it's pretty slow (duh!). I've been working on  
an alternative implementation based on method calls to a strategy object  
held in the C struct for the Sax parser. The downside is, this breaks the  
API (more in a moment). The upside is it's a fair bit faster.

Breaking the API is, I think, not as big a thing as it might be - we're  
coming up for 0.4.0 next, and though I don't like this kind of breakage on  
a minor release, the sax parser API is also new in 0.3.8 and still  
experimental, so IMHO it's better to change it now rather than wait.

I've attached a patch against CVS head (update first, there've been  
changes today) that implements the new functionality if anyone wants to  
try it and let me know what they think. I've also run a few benchmarks.  
First up, this is using the sax parser, parse 1 being with no callbacks,  
parse 2 being with callbacks applied (empty procs / methods). Really it's  
just testing the overhead of calling the callbacks:

  ########### OLD (proc-based) ###############
       user     system      total        real
XML::SAXParser#parse 1  2.200000   0.210000   2.410000 (  5.066253)
       user     system      total        real
XML::SAXParser#parse 2 14.450000   0.370000  14.820000 ( 15.340662)
  ########### NEW (class-based) ##############
       user     system      total        real
XML::SAXParser#parse 1  1.850000   0.140000   1.990000 (  2.000680)
       user     system      total        real
XML::SAXParser#parse 2  6.570000   0.190000   6.760000 (  6.906034)

You can see that the overhead of the class-based approach is much lower. I  
also did a slightly more real-world benchmark using my digestr library.  
This compares the REXML-based XMLDigester, DigestR based on the proc-based  
sax parser, and DigestR based on the new sax parser. The difference is  
less marked, but still worthwhile I believe (these tests run 500  
iterations over an small addressbook XML file, and build a custom tree in  
the callbacks).

###### XMLDIGESTER ######
       user     system      total        real
  22.120000   0.920000  23.040000 ( 23.336247)

###### DIGESTR ######
       user     system      total        real
   8.400000   0.200000   8.600000 (  8.731552)

###### DIGESTR-NEWSAX ######
       user     system      total        real
   6.360000   0.120000   6.480000 (  6.591864)

Anyway, I'd really appreciate any thoughts or ideas anyone has about this.

Cheers,
-- 
Ross Bamford - rosco at roscopeco.co.uk
-------------- next part --------------
A non-text attachment was scrubbed...
Name: newsax.patch
Type: text/x-patch
Size: 29611 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/libxml-devel/attachments/20060531/99a76d6a/attachment-0001.bin 

From bryant.doug at gmail.com  Wed May 31 16:50:01 2006
From: bryant.doug at gmail.com (Doug Bryant)
Date: Wed, 31 May 2006 16:50:01 -0400
Subject: [libxml-devel] error when application finished running
Message-ID: <78cf1ade0605311350m725b0112oa8ac466b9588d58b@mail.gmail.com>

I'm getting the below error when the program I have written using
libxml-ruby finishes running (when the program terminates).

Is this something I should be concerned about?  It does not do this until
the program terminates.

The standalone app mentioned above is just a command line test of a larger
piece of functionality.  It will ultimately be part of a webapplication.

Thanks, Doug

*** glibc detected *** ruby: double free or corruption (!prev): 0x12a9cb48
***
======= Backtrace: =========
/lib/libc.so.6[0xbacf18]
/lib/libc.so.6(__libc_free+0x78)[0xbb03ef]
ruby(ruby_xfree+0x26)[0x806d936]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x11d)[0x713de5d]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeNodeList+0x96)[0x713ddd6]
/usr/lib/libxml2.so.2(xmlFreeDoc+0xcb)[0x713dbfb]
/usr/local/lib/ruby/site_ruby/1.8/i686-linux/xml/libxml_so.so(ruby_xml_document_free+0x95)[0x1707e5]
ruby(rb_gc_call_finalizer_at_exit+0x87)[0x806db77]
ruby[0x80573e9]
ruby(ruby_cleanup+0xcc)[0x806327c]
ruby(ruby_stop+0x11)[0x8063351]
ruby[0x8068831]
ruby[0x80521d4]
/lib/libc.so.6(__libc_start_main+0xdc)[0xb5e724]
ruby[0x8052111]
======= Memory map: ========
00111000-00115000 r-xp 00000000 08:03 721616
/usr/local/lib/ruby/1.8/i686-linux/stringio.so
00115000-00116000 rwxp 00003000 08:03 721616
/usr/local/lib/ruby/1.8/i686-linux/stringio.so
00116000-00130000 r-xp 00000000 08:03 721634
/usr/local/lib/ruby/1.8/i686-linux/syck.so
00130000-00131000 rwxp 00019000 08:03 721634
/usr/local/lib/ruby/1.8/i686-linux/syck.so
00131000-00133000 r-xp 00000000 08:03 721611
/usr/local/lib/ruby/1.8/i686-linux/digest.so
00133000-00134000 rwxp 00001000 08:03 721611
/usr/local/lib/ruby/1.8/i686-linux/digest.so
00134000-00143000 r-xp 00000000 08:03 721608
/usr/local/lib/ruby/1.8/i686-linux/nkf.so
00143000-00155000 rwxp 0000f000 08:03 721608
/usr/local/lib/ruby/1.8/i686-linux/nkf.so
00155000-00156000 rwxp 00155000 00:00 0
00156000-00159000 r-xp 00000000 08:03 721607
/usr/local/lib/ruby/1.8/i686-linux/iconv.so
00159000-0015a000 rwxp 00002000 08:03 721607
/usr/local/lib/ruby/1.8/i686-linux/iconv.so
0015a000-00173000 r-xp 00000000 08:03 895239
/usr/local/lib/ruby/site_ruby/1.8/i686-linux/xml/libxml_so.so
00173000-00174000 rwxp 00019000 08:03 895239
/usr/local/lib/ruby/site_ruby/1.8/i686-linux/xml/libxml_so.so
00174000-0017b000 r-xp 00000000 08:03 721349
/usr/local/lib/ruby/site_ruby/1.8/i686-linux/postgres.so
0017b000-0017c000 rwxp 00006000 08:03 721349
/usr/local/lib/ruby/site_ruby/1.8/i686-linux/postgres.so
0017c000-0018c000 r-xp 00000000 08:01 192227     /lib/libpthread-2.4.so
0018c000-0018d000 r-xp 0000f000 08:01 192227     /lib/libpthread-2.4.so
0018d000-0018e000 rwxp 00010000 08:01 192227     /lib/libpthread-2.4.so
0018e000-00190000 rwxp 0018e000 00:00 0
00190000-00192000 r-xp 00000000 08:03 721633
/usr/local/lib/ruby/1.8/i686-linux/enumerator.so
00192000-00193000 rwxp 00001000 08:03 721633
/usr/local/lib/ruby/1.8/i686-linux/enumerator.so
00242000-0024d000 r-xp 00000000 08:01 192231     /lib/libgcc_s-
4.1.0-20060304.so.1
0024d000-0024e000 rwxp 0000a000 08:01 192231     /lib/libgcc_s-
4.1.0-20060304.so.1
002bb000-002bc000 r-xp 002bb000 00:00 0          [vdso]
002bc000-002d5000 r-xp 00000000 08:01 192210     /lib/ld-2.4.so
002d5000-002d6000 r-xp 00018000 08:01 192210     /lib/ld-2.4.so
002d6000-002d7000 rwxp 00019000 08:01 192210     /lib/ld-2.4.so
00310000-00312000 r-xp 00000000 08:01 192331     /lib/libcom_err.so.2.1
00312000-00313000 rwxp 00001000 08:01 192331     /lib/libcom_err.so.2.1
00315000-00318000 r-xp 00000000 08:03 525647
/usr/lib/libkrb5support.so.0.0
00318000-00319000 rwxp 00002000 08:03 525647
/usr/lib/libkrb5support.so.0.0
0031b000-0033f000 r-xp 00000000 08:03 527969     /usr/lib/libk5crypto.so.3.0
0033f000-00340000 rwxp 00024000 08:03 527969     /usr/lib/libk5crypto.so.3.0
00342000-0035a000 r-xp 00000000 08:03 527971
/usr/lib/libgssapi_krb5.so.2.2
0035a000-0035b000 rwxp 00017000 08:03 527971
/usr/lib/libgssapi_krb5.so.2.2
00389000-003a8000 r-xp 00000000 08:03 526444     /usr/lib/libpq.so.4.1
003a8000-003a9000 rwxp 0001f000 08:03 526444     /usr/lib/libpq.so.4.1
003c5000-003d7000 r-xp 00000000 08:01 192243     /lib/libnsl-2.4.so
003d7000-003d8000 r-xp 00011000 08:01 192243     /lib/libnsl-2.4.so
003d8000-003d9000 rwxp 00012000 08:01 192243     /lib/libnsl-2.4.so
003d9000-003db000 rwxp 003d9000 00:00 0
003dd000-004fc000 r-xp 00000000 08:01 192333     /lib/libcrypto.so.0.9.8a
004fc000-0050f000 rwxp 0011e000 08:01 192333     /lib/libcrypto.so.0.9.8a
0050f000-00512000 rwxp 0050f000 00:00 0
00514000-00587000 r-xp 00000000 08:03 527970     /usr/lib/libkrb5.so.3.2
00587000-00589000 rwxp 00073000 08:03 527970     /usr/lib/libkrb5.so.3.2
0058b000-005cc000 r-xp 00000000 08:01 192430     /lib/libssl.so.0.9.8a
005cc000-005d0000 rwxp 00040000 08:01 192430     /lib/libssl.so.0.9.8a
005ed000-005f1000 r-xp 00000000 08:03 721638
/usr/local/lib/ruby/1.8/i686-linux/strscan.so
005f1000-005f2000 rwxp 00003000 08:03 721638
/usr/local/lib/ruby/1.8/i686-linux/strscan.so
0069a000-006a3000 r-xp 00000000 08:01 192232     /lib/libnss_files-2.4.so
006a3000-006a4000 r-xp 00008000 08:01 192232     /lib/libnss_files-2.4.so
006a4000-006a5000 rwxp 00009000 08:01 192232     /lib/libnss_files-2.4.so
006c7000-006d6000 r-xp 00000000 08:01 192255     /lib/libresolv-2.4.so
006d6000-006d7000 r-xp 0000e000 08:01 192255     /lib/libresolv-2.4.so
006d7000-006d8000 rwxp 0000f000 08:01 192255     /lib/libresolv-2.4.so
006d8000-006da000 rwxp 006d8000 00:00 0
008b8000-008b9000 r-xp 00000000 08:03 721629
/usr/local/lib/ruby/1.8/i686-linux/fcntl.so
008b9000-008ba000 rwxp 00000000 08:03 721629
/usr/local/lib/ruby/1.8/i686-linux/fcntl.so
008f0000-008f3000 r-xp 00000000 08:03 721626
/usr/local/lib/ruby/1.8/i686-linux/racc/cparse.so
008f3000-008f4000 rwxp 00002000 08:03 721626
/usr/local/lib/ruby/1.8/i686-linux/racc/cparse.so
00977000-00980000 r-xp 00000000 08:03 721649
/usr/local/lib/ruby/1.8/i686-linux/socket.so
00980000-00981000 rwxp 00008000 08:03 721649
/usr/local/lib/ruby/1.8/i686-linux/socket.so
00a97000-00a99000 r-xp 00000000 08:03 721605     /usr/local/liAborted
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060531/b3cf596d/attachment.htm 

From rosco at roscopeco.co.uk  Wed May 31 19:29:47 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Thu, 01 Jun 2006 00:29:47 +0100
Subject: [libxml-devel] error when application finished running
In-Reply-To: <78cf1ade0605311350m725b0112oa8ac466b9588d58b@mail.gmail.com>
References: <78cf1ade0605311350m725b0112oa8ac466b9588d58b@mail.gmail.com>
Message-ID: <op.tafvjx2pkc9f7k@jukebox.roscopeco>

On Wed, 31 May 2006 21:50:01 +0100, Doug Bryant <bryant.doug at gmail.com>  
wrote:

> I'm getting the below error when the program I have written using
> libxml-ruby finishes running (when the program terminates).
>

Can you post your code?

-- 
Ross Bamford - rosco at roscopeco.co.uk

From bryant.doug at gmail.com  Wed May 31 20:42:00 2006
From: bryant.doug at gmail.com (Doug Bryant)
Date: Wed, 31 May 2006 20:42:00 -0400
Subject: [libxml-devel] error when application finished running
In-Reply-To: <op.tafvjx2pkc9f7k@jukebox.roscopeco>
References: <78cf1ade0605311350m725b0112oa8ac466b9588d58b@mail.gmail.com>
	<op.tafvjx2pkc9f7k@jukebox.roscopeco>
Message-ID: <78cf1ade0605311742r5d290f0dhcf83819004969013@mail.gmail.com>

I was afraid you were going to ask that :)

I'll have try to duplicate it with some similar code.  The concrete
class which drives the program is tied very tightly to my application
and would require too much supporting information.

The behavior occurrs when doing lots of manipulation of existing xml.
It does not occur in other sections of my app which use libxml for
read-only access.

I'll try to post something tomorrow.

Thanks, Doug


On 5/31/06, Ross Bamford <rosco at roscopeco.co.uk> wrote:
> On Wed, 31 May 2006 21:50:01 +0100, Doug Bryant <bryant.doug at gmail.com>
> wrote:
>
> > I'm getting the below error when the program I have written using
> > libxml-ruby finishes running (when the program terminates).
> >
>
> Can you post your code?
>
> --
> Ross Bamford - rosco at roscopeco.co.uk
> _______________________________________________
> libxml-devel mailing list
> libxml-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/libxml-devel
>