From explosiv0SX at datafront.com  Fri Feb 17 17:03:40 2006
From: explosiv0SX at datafront.com (Explosiv0SX)
Date: Fri, 17 Feb 2006 17:03:40 -0500
Subject: [libxml-devel] building libxml on os x
Message-ID: <AABE549F-4626-400C-AA26-DC97D18BF215@datafront.com>

After banging at this all day, I finally realized why libxml.bundle  
won't compile on OS X. Simply remove the -fno-common flag from the  
CFLAGS variable after you generate the Makefile with ruby extconf.rb.  
Compiles, links, and brings me joy!

Keep up the great work!

Rick

From rosco at roscopeco.co.uk  Sun Feb 19 00:17:10 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 19 Feb 2006 05:17:10 -0000
Subject: [libxml-devel] Attribute fix?
Message-ID: <op.s47fmwlqkc9f7k@jukebox.roscopeco>

Hi,

The following seems to fix the problems I discussed before, doesn't break  
any tests. I'd like to get a sanity check on it though, rather than just  
committing it, because I've not had much time to spend getting used to  
libxml still...

Index: ruby_xml_attr.c
===================================================================
RCS file: /var/cvs/libxml/libxml/ruby_xml_attr.c,v
retrieving revision 1.2
diff -u -r1.2 ruby_xml_attr.c
--- ruby_xml_attr.c	2 Jan 2006 23:19:21 -0000	1.2
+++ ruby_xml_attr.c	19 Feb 2006 05:03:06 -0000
@@ -166,11 +166,11 @@
    ruby_xml_attr *rxa;

    rxa = ALLOC(ruby_xml_attr);
-  rxa->attr = attr;
+  rxa->attr = xmlCopyProp(attr->parent, attr);
    rxa->xd = xd;
-  rxa->is_ptr = 1;
+  rxa->is_ptr = 0;
    return(Data_Wrap_Struct(class, ruby_xml_attr_mark,
-			  ruby_xml_attr_free, rxa));
+			   ruby_xml_attr_free, rxa));
  }


Index: ruby_xml_node.c
===================================================================
RCS file: /var/cvs/libxml/libxml/ruby_xml_node.c,v
retrieving revision 1.3
diff -u -r1.3 ruby_xml_node.c
--- ruby_xml_node.c	2 Jan 2006 23:19:21 -0000	1.3
+++ ruby_xml_node.c	19 Feb 2006 05:03:09 -0000
@@ -1568,16 +1568,21 @@
  ruby_xml_node_property_get(VALUE self, VALUE prop) {
    ruby_xml_node *rxn;
    xmlChar *p;
-
+  VALUE r;
+
    Check_Type(prop, T_STRING);

    Data_Get_Struct(self, ruby_xml_node, rxn);
    p = xmlGetProp(rxn->node, (xmlChar*)StringValuePtr(prop));

    if (p == NULL)
-    return(Qnil);
-  else
-    return(rb_str_new2((const char*)p));
+    r = Qnil;
+  else {
+    r = rb_str_new2((const char*)p);
+    xmlFree(p);
+  }
+
+  return r;
  }


@@ -1626,7 +1631,7 @@

    if (node->node->type == XML_ELEMENT_NODE) {
      attr = node->node->properties;
-    return(ruby_xml_attr_new(cXMLAttr, node->xd, attr));
+    return(ruby_xml_attr_new2(cXMLAttr, node->xd, attr));
    } else {
      return(Qnil);
    }

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

From sean at gigave.com  Sun Feb 19 15:33:26 2006
From: sean at gigave.com (Sean Chittenden)
Date: Sun, 19 Feb 2006 12:33:26 -0800
Subject: [libxml-devel] Attribute fix?
In-Reply-To: <op.s47fmwlqkc9f7k@jukebox.roscopeco>
References: <op.s47fmwlqkc9f7k@jukebox.roscopeco>
Message-ID: <20060219203326.GB24308@mailhost.gigave.com>

> The following seems to fix the problems I discussed before, doesn't break  
> any tests. I'd like to get a sanity check on it though, rather than just  
> committing it, because I've not had much time to spend getting used to  
> libxml still...

Looks sane to me, you changed the object from a shadow object and are
passing around a copy that can be free(3)'ed when the object is
destroyed.  -sc

-- 
Sean Chittenden

From rosco at roscopeco.co.uk  Mon Feb 20 06:34:03 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Mon, 20 Feb 2006 11:34:03 -0000
Subject: [libxml-devel] Attribute fix?
In-Reply-To: <20060219203328.33294216E02@mailhost.gigave.com>
References: <op.s47fmwlqkc9f7k@jukebox.roscopeco>
	<20060219203328.33294216E02@mailhost.gigave.com>
Message-ID: <op.s49rq1hgkc9f7k@jukebox.roscopeco>

On Sun, 19 Feb 2006 20:33:26 -0000, Sean Chittenden <sean at gigave.com>  
wrote:

>> The following seems to fix the problems I discussed before, doesn't  
>> break
>> any tests. I'd like to get a sanity check on it though [...]
>
> Looks sane to me [...]

Okay thanks, I've committed it just now.

I think with that fixed, it might be worth thinking about getting that  
release out?

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

From rosco at roscopeco.co.uk  Mon Feb 20 07:09:18 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Mon, 20 Feb 2006 12:09:18 -0000
Subject: [libxml-devel] building libxml on os x
In-Reply-To: <AABE549F-4626-400C-AA26-DC97D18BF215@datafront.com>
References: <AABE549F-4626-400C-AA26-DC97D18BF215@datafront.com>
Message-ID: <op.s49tdszzkc9f7k@jukebox.roscopeco>

On Fri, 17 Feb 2006 22:03:40 -0000, Explosiv0SX  
<explosiv0SX at datafront.com> wrote:

> After banging at this all day, I finally realized why libxml.bundle
> won't compile on OS X. Simply remove the -fno-common flag from the
> CFLAGS variable after you generate the Makefile with ruby extconf.rb.
> Compiles, links, and brings me joy!
>

Thanks :)

I'm not sure why that flag is being used, presumably it's used to build  
Ruby itself on OSX...?

Accepting that we don't have much opportunity for experimenting on OSX  
right now, I wonder if we could just hack this for now, something like:

	if RUBY_PLATFORM == ??whatgoeshere??
	  mf = File.read('Makefile')
	  File.open('Makefile','w+') do |f|
	    f << mf.gsub(/^CFLAGS\s+=\s+(.*)/) { "CFLAGS =  
#{$1.gsub('-fno-common','')}" }
	  end
	end

at the end of extconf.rb?

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

From explosiv0SX at datafront.com  Tue Feb 21 00:16:45 2006
From: explosiv0SX at datafront.com (Explosiv0SX)
Date: Tue, 21 Feb 2006 00:16:45 -0500
Subject: [libxml-devel] building libxml on os x
In-Reply-To: <op.s49tdszzkc9f7k@jukebox.roscopeco>
References: <AABE549F-4626-400C-AA26-DC97D18BF215@datafront.com>
	<op.s49tdszzkc9f7k@jukebox.roscopeco>
Message-ID: <0D91CA94-0924-40E9-8632-63F773D572F1@datafront.com>

Ross,

This should take care of this OS X compilation issue for now. I'm  
still not sure why -fno-common is causing issues, but your extension  
now builds without problems on my Macs.

if RUBY_PLATFORM =~ /darwin/
   mf = File.read('Makefile')
   File.open('Makefile','w+') do |f|
     f << mf.gsub(/^CFLAGS\s+=\s+(.*)/) { "CFLAGS = #{$1.gsub('-fno- 
common','')}" }
   end
end

I also made a change to the ruby_xml_schema_init_from_uri function in  
ruby_xml_schema.c to get schema validation going. The patch is  
included below -- please check my code, as I have just started  
working with the Ruby extension mechanism. It's nice to now have some  
xml validation in Ruby...

Thanks!
Rick

--- libxml.1/libxml/ruby_xml_schema.c	2006-01-02 18:19:21.000000000  
-0500
+++ ruby_xml_schema.c	2006-02-19 20:49:58.000000000 -0500
@@ -26,23 +26,22 @@
   */
VALUE
ruby_xml_schema_init_from_uri(int argc, VALUE *argv, VALUE class) {
-  VALUE uri;
-  xmlSchemaParserCtxtPtr parser;
-  xmlSchemaPtr sptr;
-
-  switch (argc) {
-  case 1:
-    rb_scan_args(argc, argv, "10", &uri);
-
-    Check_Type(uri, T_STRING);
-
- 	parser = xmlSchemaNewParserCtxt(StringValuePtr(uri));
- 	sptr = xmlSchemaParse(parser);
- 	xmlSchemaFreeParserCtxt(parser);
-    break;
-  default:
-    rb_raise(rb_eArgError, "wrong number of arguments (need 1)");
-  }
+   VALUE uri;
+   xmlSchemaParserCtxtPtr parser;
+   ruby_xml_schema *rxschema;
+
+   switch (argc) {
+      case 1:
+         rb_scan_args(argc, argv, "10", &uri);
+         Check_Type(uri, T_STRING);
+         parser = xmlSchemaNewParserCtxt(StringValuePtr(uri));
+         rxschema = ALLOC(ruby_xml_schema);
+         rxschema->schema = xmlSchemaParse(parser);
+         xmlSchemaFreeParserCtxt(parser);
+         return(Data_Wrap_Struct(cXMLSchema, ruby_xml_schema_mark,  
ruby_xml_schema_free, rxschema));
+      default:
+         rb_raise(rb_eArgError, "wrong number of arguments (need 1)");
+   }
    return Qnil;
}



On Feb 20, 2006, at 7:09 AM, Ross Bamford wrote:

> On Fri, 17 Feb 2006 22:03:40 -0000, Explosiv0SX
> <explosiv0SX at datafront.com> wrote:
>
>> After banging at this all day, I finally realized why libxml.bundle
>> won't compile on OS X. Simply remove the -fno-common flag from the
>> CFLAGS variable after you generate the Makefile with ruby extconf.rb.
>> Compiles, links, and brings me joy!
>>
>
> Thanks :)
>
> I'm not sure why that flag is being used, presumably it's used to  
> build
> Ruby itself on OSX...?
>
> Accepting that we don't have much opportunity for experimenting on OSX
> right now, I wonder if we could just hack this for now, something  
> like:
>
> 	if RUBY_PLATFORM == ??whatgoeshere??
> 	  mf = File.read('Makefile')
> 	  File.open('Makefile','w+') do |f|
> 	    f << mf.gsub(/^CFLAGS\s+=\s+(.*)/) { "CFLAGS =
> #{$1.gsub('-fno-common','')}" }
> 	  end
> 	end
>
> at the end of extconf.rb?
>
> -- 
> Ross Bamford - rosco at roscopeco.co.uk
> _______________________________________________
> libxml-devel mailing list
> libxml-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/libxml-devel


From rosco at roscopeco.co.uk  Tue Feb 21 05:53:57 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Tue, 21 Feb 2006 10:53:57 -0000
Subject: [libxml-devel] building libxml on os x
In-Reply-To: <0D91CA94-0924-40E9-8632-63F773D572F1@datafront.com>
References: <AABE549F-4626-400C-AA26-DC97D18BF215@datafront.com>
	<op.s49tdszzkc9f7k@jukebox.roscopeco>
	<0D91CA94-0924-40E9-8632-63F773D572F1@datafront.com>
Message-ID: <op.s5bkj7rdkc9f7k@jukebox.roscopeco>

On Tue, 21 Feb 2006 05:16:45 -0000, Explosiv0SX  
<explosiv0SX at datafront.com> wrote:

>> [OSX build fix]
>
> This should take care of this OS X compilation issue for now. I'm
> still not sure why -fno-common is causing issues, but your extension
> now builds without problems on my Macs.
>

Cool, I've just committed that in as an interim measure. I guess in the  
end we'll need to look into this but for now it's nice to have some  
OSXability. Thanks again for that :)

>
> I also made a change to the ruby_xml_schema_init_from_uri function in
> ruby_xml_schema.c to get schema validation going. The patch is
> included below -- please check my code, as I have just started
> working with the Ruby extension mechanism. It's nice to now have some
> xml validation in Ruby...
>

Well, thanks again, again :)

I'll hold your patch along with another feature patch we have, I think the  
plan is to get a ground-zero release with fixes only, and then start  
working on new stuff (or finishing old stuff) once that's out the door.

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

From rosco at roscopeco.co.uk  Tue Feb 21 09:12:22 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Tue, 21 Feb 2006 14:12:22 -0000
Subject: [libxml-devel] Gem packaging
Message-ID: <op.s5btqwybkc9f7k@jukebox.roscopeco>

Hi,

I've been working on getting version number management and gem packaging  
into the Rakefile, which is working pretty well so far, except for one  
hiccup - it appears that we can't keep the 'xml/libxml' namespace under  
gems without rearranging the project so that all the code is in an 'xml'  
directory in the project root.

I think the options are:

	* Stick as it is, if you install by hand you need 'xml/libxml'
	  but under gems you need 'libxml'. Bad idea IMHO.

	* Drop the 'xml' prefix in the makefile, so that no matter how
	  you install you need 'libxml'.

	* Do the rearranging, keeping 'xml/libxml' everywhere but needing
	  more work creating and maintaining a top-level makefile setup
	  (or something) for those installing by hand.

I know the second might not seem like a good idea, but I think in this  
case, given that there aren't other libxml bindings for Ruby, and that  
xml/libxml has some redundancy, maybe it's worth considering. Perhaps  
though it'd need to be changed to 'libxml2' to avoid confusion.

Any thoughts?

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

From explosiv0SX at datafront.com  Tue Feb 21 09:48:55 2006
From: explosiv0SX at datafront.com (Explosiv0SX)
Date: Tue, 21 Feb 2006 09:48:55 -0500
Subject: [libxml-devel] Gem packaging
In-Reply-To: <op.s5btqwybkc9f7k@jukebox.roscopeco>
References: <op.s5btqwybkc9f7k@jukebox.roscopeco>
Message-ID: <256B643F-6AA8-4437-8EA5-0895F3043543@datafront.com>

I vote for second option with the libxml2 name.

Rick

On Feb 21, 2006, at 9:12 AM, Ross Bamford wrote:

> Hi,
>
> I've been working on getting version number management and gem  
> packaging
> into the Rakefile, which is working pretty well so far, except for one
> hiccup - it appears that we can't keep the 'xml/libxml' namespace  
> under
> gems without rearranging the project so that all the code is in an  
> 'xml'
> directory in the project root.
>
> I think the options are:
>
> 	* Stick as it is, if you install by hand you need 'xml/libxml'
> 	  but under gems you need 'libxml'. Bad idea IMHO.
>
> 	* Drop the 'xml' prefix in the makefile, so that no matter how
> 	  you install you need 'libxml'.
>
> 	* Do the rearranging, keeping 'xml/libxml' everywhere but needing
> 	  more work creating and maintaining a top-level makefile setup
> 	  (or something) for those installing by hand.
>
> I know the second might not seem like a good idea, but I think in this
> case, given that there aren't other libxml bindings for Ruby, and that
> xml/libxml has some redundancy, maybe it's worth considering. Perhaps
> though it'd need to be changed to 'libxml2' to avoid confusion.
>
> Any thoughts?
>
> Cheers,
> -- 
> Ross Bamford - rosco at roscopeco.co.uk
> _______________________________________________
> libxml-devel mailing list
> libxml-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/libxml-devel


From sean at gigave.com  Tue Feb 21 14:26:39 2006
From: sean at gigave.com (Sean Chittenden)
Date: Tue, 21 Feb 2006 11:26:39 -0800
Subject: [libxml-devel] Gem packaging
In-Reply-To: <op.s5btqwybkc9f7k@jukebox.roscopeco>
References: <op.s5btqwybkc9f7k@jukebox.roscopeco>
Message-ID: <20060221192639.GZ17911@mailhost.gigave.com>

> I've been working on getting version number management and gem
> packaging into the Rakefile, which is working pretty well so far,
> except for one hiccup - it appears that we can't keep the
> 'xml/libxml' namespace under gems without rearranging the project so
> that all the code is in an 'xml' directory in the project root.
> 
> I think the options are:
> 
> 	* Stick as it is, if you install by hand you need 'xml/libxml'
> 	but under gems you need 'libxml'. Bad idea IMHO.
> 
> 	* Drop the 'xml' prefix in the makefile, so that no matter how
> 	you install you need 'libxml'.
> 
> 	* Do the rearranging, keeping 'xml/libxml' everywhere but
> 	needing more work creating and maintaining a top-level
> 	makefile setup (or something) for those installing by hand.
> 
> I know the second might not seem like a good idea, but I think in
> this case, given that there aren't other libxml bindings for Ruby,
> and that xml/libxml has some redundancy, maybe it's worth
> considering. Perhaps though it'd need to be changed to 'libxml2' to
> avoid confusion.

Ugh, the ugliness of gems rears its head again.  Ruby's flat module
space is going to be a problem at some point.  The reason for
xml/libxml was to subdivide the module namespace into functional
groups.  For instance, on the xsl front, libxslt is accessed via
'xml/libxslt'.  I'm not married to this notion, but it'll be an ugly
barrier at some point for the language to cross and is one of the
niceties of Perl, by comparison.  rexml, should be xml/rexml or some
such nonsense... but "should" doesn't carry much weight.

I wouldn't object to stashing everything in a subdirectory of xml/ as
a hack around for the limitations of gems.  *shrug*

-- 
Sean Chittenden

From rosco at roscopeco.co.uk  Tue Feb 21 14:40:09 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Tue, 21 Feb 2006 19:40:09 -0000
Subject: [libxml-devel] Gem packaging
In-Reply-To: <20060221192640.6EB49216C21@mailhost.gigave.com>
References: <op.s5btqwybkc9f7k@jukebox.roscopeco>
	<20060221192640.6EB49216C21@mailhost.gigave.com>
Message-ID: <op.s5b8w7kokc9f7k@jukebox.roscopeco>

On Tue, 21 Feb 2006 19:26:39 -0000, Sean Chittenden <sean at gigave.com>  
wrote:

>> I've been working on getting version number management and gem
>> packaging into the Rakefile, which is working pretty well so far,
>> except for one hiccup - it appears that we can't keep the
>> 'xml/libxml' namespace under gems without rearranging the project so
>> that all the code is in an 'xml' directory in the project root.
>>
> I wouldn't object to stashing everything in a subdirectory of xml/ as
> a hack around for the limitations of gems.  *shrug*
>

Okay, I'm on it. It'll probably be something like ext/xml/*.c in the  
source, with the gemspec libpath set to ext, and make install working as  
now. That way we avoid pulling in 'tests' and what not into the load path  
when running under gems.

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

From rosco at roscopeco.co.uk  Tue Feb 21 15:59:12 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Tue, 21 Feb 2006 20:59:12 -0000
Subject: [libxml-devel] Gem packaging
In-Reply-To: <op.s5b8w7kokc9f7k@jukebox.roscopeco>
References: <op.s5btqwybkc9f7k@jukebox.roscopeco>
	<20060221192640.6EB49216C21@mailhost.gigave.com>
	<op.s5b8w7kokc9f7k@jukebox.roscopeco>
Message-ID: <op.s5cckyjckc9f7k@jukebox.roscopeco>

On Tue, 21 Feb 2006 19:40:09 -0000, Ross Bamford <rosco at roscopeco.co.uk>  
wrote:

> On Tue, 21 Feb 2006 19:26:39 -0000, Sean Chittenden <sean at gigave.com>
> wrote:
>
>> I wouldn't object to stashing everything in a subdirectory of xml/ as
>> a hack around for the limitations of gems.  *shrug*
>>
>
> Okay, I'm on it. It'll probably be something like ext/xml/*.c in the
> source, with the gemspec libpath set to ext, and make install working as
> now. That way we avoid pulling in 'tests' and what not into the load path
> when running under gems.
>

I've rearranged stuff and committed the changes, it now builds, tests,  
installs normally, packages and installs as a gem, with 'xml/libxml' as  
the require.

I've expanded the Rakefile, and it now can handle almost everything  
(calling out to make for the actual building of course). There are  
packaging tasks in there, and a release task that handles version number  
management. I've removed the Rubynet utility stuff now that's done.

I didn't hook the release task up to CVS or web upload yet, we'll still  
have to 'do' releases, there's only so much automation I can take in one  
day ;D

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

From mcix at gmx.net  Wed Feb 22 05:55:15 2006
From: mcix at gmx.net (mcix@gmx.net)
Date: Wed, 22 Feb 2006 11:55:15 +0100 (MET)
Subject: [libxml-devel] (no subject)
Message-ID: <19756.1140605715@www087.gmx.net>

Hi,

I am writing an article about XML validation using your libxml for ruby.

I have 2 questions:

1 - How does it relate with the libxml from http://xml-tools.rubyforge.org
2 - it seems that you are going to relase it as ruby gem.
    Is it true? If it is, by when?

Thanks for the nice work you are doing.

Mauro

--
mailto:mauro at cicio.org

-- 
Lust, ein paar Euro nebenbei zu verdienen? Ohne Kosten, ohne Risiko!
Satte Provisionen für GMX Partner: http://www.gmx.net/de/go/partner

From rosco at roscopeco.co.uk  Thu Feb 23 05:32:49 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Thu, 23 Feb 2006 10:32:49 -0000
Subject: [libxml-devel] (no subject)
In-Reply-To: <19756.1140605715@www087.gmx.net>
References: <19756.1140605715@www087.gmx.net>
Message-ID: <op.s5e8wzi1kc9f7k@jukebox.roscopeco>

Hi there,

On Wed, 22 Feb 2006 10:55:15 -0000, <mcix at gmx.net> wrote:

> I am writing an article about XML validation using your libxml for ruby.
>

Cool, look forward to that :)

> I have 2 questions:
>
> 1 - How does it relate with the libxml from  
> http://xml-tools.rubyforge.org

The code we're now working on in libxml.rubyforge.org is based off of the  
xml-tools code, which came from Sean's original codebase. When we started  
the current effort to bring libxml-ruby 'back from the dead' we took the  
xml-tools code and fixed up some code-rot / critical bugs, to become the  
libxml.rubyforge.org project.

(Maybe we should see about getting a notice of some kind posted on  
xml-tools website about this).

> 2 - it seems that you are going to relase it as ruby gem.
>     Is it true? If it is, by when?
>

It most certainly is, and as soon as possible. Earlier this week we  
squashed a critical bug we were holding back for, and I took some time to  
implement a fuller Rake build and gem packaging. Currently, in CVS head,  
gem packaging is in and working (type 'rake release') so with any luck  
(barring new critical bugs ;)) we should have a release with gem  
installation in the *very* near future (this week/next week).

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

From joe.rice at mercurymd.com  Thu Feb 23 18:27:41 2006
From: joe.rice at mercurymd.com (Joe Rice)
Date: Thu, 23 Feb 2006 18:27:41 -0500
Subject: [libxml-devel] Zlib on windows problem
Message-ID: <8A2F25E15E39D9409B45F57EBB2443AF6E1A8F@EXCHANGE1.MERCURYMD.RTP>

I am trying to move away from REXML because of performance problems on
larger files.  I've been trying to get the libxml-ruby project to work
but when I try the gem install libxml-ruby command it fails because I
need zlib.  I'm running on Windows with ruby 1.8.4.  Can anyone point me
in the direction of any information that would help me get zlib
installed/working on a windows box? 
 
Thanks in advance,
 
Joe Rice
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/libxml-devel/attachments/20060223/4a026c64/attachment.htm

From rosco at roscopeco.co.uk  Fri Feb 24 04:47:20 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Fri, 24 Feb 2006 09:47:20 -0000
Subject: [libxml-devel] Zlib on windows problem
In-Reply-To: <8A2F25E15E39D9409B45F57EBB2443AF6E1A8F@EXCHANGE1.MERCURYMD.RTP>
References: <8A2F25E15E39D9409B45F57EBB2443AF6E1A8F@EXCHANGE1.MERCURYMD.RTP>
Message-ID: <op.s5g1g6s0kc9f7k@jukebox.roscopeco>

On Thu, 23 Feb 2006 23:27:41 -0000, Joe Rice <joe.rice at mercurymd.com>  
wrote:

> I am trying to move away from REXML because of performance problems on
> larger files.  I've been trying to get the libxml-ruby project to work
> but when I try the gem install libxml-ruby command it fails because I
> need zlib.  I'm running on Windows with ruby 1.8.4.  Can anyone point me
> in the direction of any information that would help me get zlib
> installed/working on a windows box?

Looking around a bit, it seems like these might be useful to you:

	http://www.winimage.com/zLibDll/
	http://gnuwin32.sourceforge.net/packages/zlib.htm
	http://www.zlib.net/DLL_FAQ.txt

I guess you'll need to source distribution, since you're going to need  
zlib.h . I'm afraid I don't know how any of this works on Windows but  
there does seem to be a decent amount of documentation that may help.

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

From rosco at roscopeco.co.uk  Sat Feb 25 19:34:08 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 26 Feb 2006 00:34:08 -0000
Subject: [libxml-devel] Fwd: Re: [ANN] LibXML-Ruby 0.3.6 (from ruby-talk)
Message-ID: <op.s5j066l8kc9f7k@jukebox.roscopeco>

Hi,

Can anyone tell me, does this look right?

On Sun, 2006-02-26 at 00:31 +0000, Ross Bamford wrote:
> On Sun, 2006-02-26 at 08:04 +0900, Daniel Harple wrote:
> > On Feb 23, 2006, at 8:32 PM, Ross Bamford wrote:
> > > LibXML-Ruby 0.3.6 is now available from Rubyforge
> >> I am having some trouble with building.
> >> 1) I have two libiconvs installed, one supplied with my OS in
> >     /usr and another in /opt/local. Using
> >     --with-iconv-dir=/opt/local fails to locate libiconv, but
> >     --with-iconv-dir=/usr works. I made a change to extconf.rb:
> >> $ diff libxml-ruby-0.3.6.orig/ext/xml/extconf.rb libxml-ruby-0.3.6/>  
> ext/xml/extconf.rb
> > 45,46c45,49
> > < unless have_library('iconv','iconv_open') or have_library>  
> ('c','iconv_open') or
> > <   have_library('recode','iconv_open')
> > ---
> >  >
> >  > unless have_library('iconv') or
> >  >        have_library('iconv','iconv_open') or
> >  >        have_library('c','iconv_open') or
> >  >        have_library('recode','iconv_open')
> >>     and --with-iconv-dir=/opt/local worked.
> >I'm not sure about this - I'm still finding my way around some things
> here, but I'm pretty sure we do need iconv_open(3) in libiconv. Does it
> compile with this change?
>I'm going to forward this as a possible change to the mailing list at
> libxml-devel at rubyforge.org .

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

From rosco at roscopeco.co.uk  Sat Feb 25 21:55:55 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 26 Feb 2006 02:55:55 -0000
Subject: [libxml-devel] Fwd: Re: [ANN] LibXML-Ruby 0.3.6 (from ruby-talk)
In-Reply-To: <op.s5j066l8kc9f7k@jukebox.roscopeco>
References: <op.s5j066l8kc9f7k@jukebox.roscopeco>
Message-ID: <op.s5j7rhewkc9f7k@jukebox.roscopeco>

Hi again,

Further to my last message, I can't actually see why we're requiring  
iconv? I'm guessing it's because libxml2 has it as an optional dep? I  
couldn't find any references to iconv.h in the source.

I just tried commenting it in the extconf.rb and it doesn't appear to make  
any difference to the makefile or build, and all tests pass. Am I missing  
something, or can/should we remove it?

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

From sean at gigave.com  Sun Feb 26 01:39:40 2006
From: sean at gigave.com (Sean Chittenden)
Date: Sat, 25 Feb 2006 22:39:40 -0800
Subject: [libxml-devel] Fwd: Re: [ANN] LibXML-Ruby 0.3.6 (from ruby-talk)
In-Reply-To: <op.s5j7rhewkc9f7k@jukebox.roscopeco>
References: <op.s5j066l8kc9f7k@jukebox.roscopeco>
	<op.s5j7rhewkc9f7k@jukebox.roscopeco>
Message-ID: <20060226063940.GW2796@mailhost.gigave.com>

> Further to my last message, I can't actually see why we're requiring
> iconv? I'm guessing it's because libxml2 has it as an optional dep?
> I couldn't find any references to iconv.h in the source.
> 
> I just tried commenting it in the extconf.rb and it doesn't appear
> to make any difference to the makefile or build, and all tests
> pass. Am I missing something, or can/should we remove it?

UTF-8.  :) Ruby isn't there, but libxml2(3) will happily hand Ruby a
steaming pile of bits that Ruby can't handle well.  The biggy is
making sure the .so links.  I remember some systems having problems
linking the lib without iconv(3), but it's been a while and my memory
has about 5 days of space in its queue before bits start hitting
/dev/null.  Thank god for CRMs and databases.  -sc

-- 
Sean Chittenden

From rosco at roscopeco.co.uk  Sun Feb 26 06:13:25 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 26 Feb 2006 11:13:25 -0000
Subject: [libxml-devel] Fwd: Re: [ANN] LibXML-Ruby 0.3.6 (from ruby-talk)
In-Reply-To: <20060226063942.03B9F216E01@mailhost.gigave.com>
References: <op.s5j066l8kc9f7k@jukebox.roscopeco>
	<op.s5j7rhewkc9f7k@jukebox.roscopeco>
	<20060226063942.03B9F216E01@mailhost.gigave.com>
Message-ID: <op.s5kusns1kc9f7k@jukebox.roscopeco>

On Sun, 26 Feb 2006 06:39:40 -0000, Sean Chittenden <sean at gigave.com>  
wrote:

>> Further to my last message, I can't actually see why we're requiring
>> iconv? I'm guessing it's because libxml2 has it as an optional dep?
>> I couldn't find any references to iconv.h in the source.
>>
>> I just tried commenting it in the extconf.rb and it doesn't appear
>> to make any difference to the makefile or build, and all tests
>> pass. Am I missing something, or can/should we remove it?
>
> UTF-8.  :) Ruby isn't there, but libxml2(3) will happily hand Ruby a
> steaming pile of bits that Ruby can't handle well.  The biggy is
> making sure the .so links.  I remember some systems having problems
> linking the lib without iconv(3), but it's been a while and my memory
> has about 5 days of space in its queue before bits start hitting
> /dev/null.  Thank god for CRMs and databases.  -sc
>

5 days? Wow, I guess I need an upgrade ... ;)

So we should keep it in then? I suppose it's pretty ubiquitous... Apart  
 from that one report it doesn't seem to be causing problems, so I'm happy  
to consider that a case of broken version or install (until it comes up  
again, at least).

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

From rosco at roscopeco.co.uk  Sun Feb 26 06:13:39 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 26 Feb 2006 11:13:39 -0000
Subject: [libxml-devel] Fwd: Re: LibXML-Ruby 0.3.6 (from ruby-talk)
Message-ID: <op.s5kus1k7kc9f7k@jukebox.roscopeco>

On Sun, 2006-02-26 at 13:01 +0900, Daniel Harple wrote:
There is small  documentation error at
> http://libxml.rubyforge.org/doc/classes/XML/Dtd.html
>The class is named Dtd, but DTD is used. Also, the name of two classes, 
> Dtd and SaxParser, is inconsistent with the naming conventions of the 
> library -- it should be DTD and SAXParser. </pedantic>
>-- Daniel
>

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

From rosco at roscopeco.co.uk  Sun Feb 26 14:49:39 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Sun, 26 Feb 2006 19:49:39 -0000
Subject: [libxml-devel] fno-common issue: possible fix for OSX
Message-ID: <op.s5lio1iokc9f7k@jukebox.roscopeco>

Hopefully the attached patch will fix the -fno-common issues on OSX and  
get rid of some problems with the build. I wonder if anyone on OSX would  
be willing to give this a try before I commit it?

Cheers,
-- 
Ross Bamford - rosco at roscopeco.co.uk
-------------- next part --------------
A non-text attachment was scrubbed...
Name: libxml-nocommon.patch
Type: application/octet-stream
Size: 4016 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/libxml-devel/attachments/20060226/731b99ac/libxml-nocommon.obj

From dharple at generalconsumption.org  Mon Feb 27 07:34:42 2006
From: dharple at generalconsumption.org (Daniel Harple)
Date: Mon, 27 Feb 2006 13:34:42 +0100
Subject: [libxml-devel] fno-common issue: possible fix for OSX
Message-ID: <5E1EDB4A-7C28-47D6-845A-8E11521543DF@generalconsumption.org>

Works for me. Using gcc4.0 on Mac OS X 10.4.

-- Daniel

From rosco at roscopeco.co.uk  Mon Feb 27 08:10:16 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Mon, 27 Feb 2006 13:10:16 -0000
Subject: [libxml-devel] fno-common issue: possible fix for OSX
In-Reply-To: <5E1EDB4A-7C28-47D6-845A-8E11521543DF@generalconsumption.org>
References: <5E1EDB4A-7C28-47D6-845A-8E11521543DF@generalconsumption.org>
Message-ID: <op.s5muvecfkc9f7k@jukebox.roscopeco>

On Mon, 27 Feb 2006 12:34:42 -0000, Daniel Harple  
<dharple at generalconsumption.org> wrote:

> Works for me. Using gcc4.0 on Mac OS X 10.4.
>

Thanks, now committed.

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

From rosco at roscopeco.co.uk  Mon Feb 27 08:13:28 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Mon, 27 Feb 2006 13:13:28 -0000
Subject: [libxml-devel] Possible patch
Message-ID: <op.s5mu0qzpkc9f7k@jukebox.roscopeco>

Attached is a feature patch left over from when I was collecting what was  
out there. IMHO we should apply this.

-- 
Ross Bamford - rosco at roscopeco.co.uk
-------------- next part --------------
A non-text attachment was scrubbed...
Name: registerErrorHandler.patch
Type: application/octet-stream
Size: 2375 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/libxml-devel/attachments/20060227/3e2494c3/registerErrorHandler.obj

From rosco at roscopeco.co.uk  Mon Feb 27 08:20:08 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Mon, 27 Feb 2006 13:20:08 -0000
Subject: [libxml-devel] libxslt bindings
Message-ID: <op.s5mvbutbkc9f7k@jukebox.roscopeco>

Over the past couple of days I've started looking at bringing the libxslt  
bindings up to date, but am having some 'fun' with the compile/linking and  
trying to have it work as a Rubygem.

Mainly the problem is that it requires the libxml-ruby headers, but we  
can't guarantee that they'll be present, or in the right place, and if  
libxml-ruby is installed without gems it's even more difficult. Passing in  
options to extconf is what I currently have (with some basic searching,  
../libxml etc) but obviously this precludes building from a Gem.

So anyway, here's the question: Does anyone know any links pertaining to  
building and linking interdependent ruby extensions? Or have any tips for  
me?

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

From sean at gigave.com  Mon Feb 27 12:34:57 2006
From: sean at gigave.com (Sean Chittenden)
Date: Mon, 27 Feb 2006 09:34:57 -0800
Subject: [libxml-devel] libxslt bindings
In-Reply-To: <op.s5mvbutbkc9f7k@jukebox.roscopeco>
References: <op.s5mvbutbkc9f7k@jukebox.roscopeco>
Message-ID: <20060227173457.GX2796@mailhost.gigave.com>

> So anyway, here's the question: Does anyone know any links
> pertaining to building and linking interdependent ruby extensions?
> Or have any tips for me?

During the dist-bzip2 phase of the makefile, *copy* a version of
ruby-libxml's headers into the libxslt tree.  Then install a version
check at the load of the .so to confirm that ruby's libxml/libxslt
bindings are the same as current or within a micro release.

How's that sound?  -sc

-- 
Sean Chittenden

From rosco at roscopeco.co.uk  Mon Feb 27 18:20:03 2006
From: rosco at roscopeco.co.uk (Ross Bamford)
Date: Mon, 27 Feb 2006 23:20:03 -0000
Subject: [libxml-devel] libxslt bindings
In-Reply-To: <20060227173458.C1DDB216C21@mailhost.gigave.com>
References: <op.s5mvbutbkc9f7k@jukebox.roscopeco>
	<20060227173458.C1DDB216C21@mailhost.gigave.com>
Message-ID: <op.s5nm3pc6kc9f7k@jukebox.roscopeco>

On Mon, 27 Feb 2006 17:34:57 -0000, Sean Chittenden <sean at gigave.com>  
wrote:

>> So anyway, here's the question: Does anyone know any links
>> pertaining to building and linking interdependent ruby extensions?
>> Or have any tips for me?
>
> During the dist-bzip2 phase of the makefile, *copy* a version of
> ruby-libxml's headers into the libxslt tree.  Then install a version
> check at the load of the .so to confirm that ruby's libxml/libxslt
> bindings are the same as current or within a micro release.
>
> How's that sound?  -sc
>

Well, it works, which is good enough for me :) I've now got the  
libxslt-ruby source compiling without warnings and passing all the tests.  
I've not put the version check in yet - is there a better way than just  
comparing them with the header consts at the start of Init_libxslt ?

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

From sean at gigave.com  Mon Feb 27 19:47:23 2006
From: sean at gigave.com (Sean Chittenden)
Date: Mon, 27 Feb 2006 16:47:23 -0800
Subject: [libxml-devel] libxslt bindings
In-Reply-To: <op.s5nm3pc6kc9f7k@jukebox.roscopeco>
References: <op.s5mvbutbkc9f7k@jukebox.roscopeco>
	<20060227173458.C1DDB216C21@mailhost.gigave.com>
	<op.s5nm3pc6kc9f7k@jukebox.roscopeco>
Message-ID: <20060228004723.GJ2796@mailhost.gigave.com>

> >> So anyway, here's the question: Does anyone know any links
> >> pertaining to building and linking interdependent ruby
> >> extensions?  Or have any tips for me?
> >
> > During the dist-bzip2 phase of the makefile, *copy* a version of
> > ruby-libxml's headers into the libxslt tree.  Then install a
> > version check at the load of the .so to confirm that ruby's
> > libxml/libxslt bindings are the same as current or within a micro
> > release.
> >
> > How's that sound?  -sc
> 
> Well, it works, which is good enough for me :) I've now got the
> libxslt-ruby source compiling without warnings and passing all the
> tests.  I've not put the version check in yet - is there a better
> way than just comparing them with the header consts at the start of
> Init_libxslt ?

Nope, that's about it.  I'd look for major and minor versions.  Daniel
is pretty good about not breaking ABI with micro releases and think
I've done a decent job in the past with the libxml/xsl bindings.  -sc

-- 
Sean Chittenden