• Michael Rynn (18/18) Aug 11 2009 I have made a little beginning attempt at understanding the phobos xml
• Saaa (4/8) Aug 12 2009 Ok, so I can use this one as well, thanks!
• Bill Baxter (13/31) Aug 12 2009 Great. Thanks for the contribution. I've checked your ports into svn n...
• Robert Fraser (16/27) Aug 12 2009 Well, different people feel differently about these things. IMO, the

Michael Rynn <michaelrynn optushome.com.au> writes:

I have made a little beginning attempt at understanding the phobos xml
modules.  For a warm up exercise, I took the std.xml and std.encoding
from phobos 2, and made them into std2.xml and std2.encoding in the
DSource project std2, so the dmd 1.0 compiler took a tour tripping up
all over the code.

They are attached as files on the http://www.dsource.org/projects/std2
page.  

I cannot believe that anybody actually uses the std2 project files.

I  took away all the safety features of const, immutable and any other
things that dmd 1.0 complains about, as "recommended" for the std2.

Inserted a few support functions not immediately avalable in Phobos
1.0,  notably some  suggested better code  for the encode (xml entity)
, and renamed it as encodeStdEntity, from a Tango fan in the D bugs
postings.

Ran unit tests and also successfully rab the books.xml example.

Re-enabled the disabled unittest in std.encoding, and fixed the little
bug that was spoiling it, also posting that to the D bugs.

..

"Saaa" <empty needmail.com> writes:

 I cannot believe that anybody actually uses the std2 project files.

Why do you believe this?
I do..

 Ran unit tests and also successfully rab the books.xml example.

Ok, so I can use this one as well, thanks!

 Re-enabled the disabled unittest in std.encoding, and fixed the little
 bug that was spoiling it, also posting that to the D bugs.

nice 

Bill Baxter <wbaxter gmail.com> writes:

On Tue, Aug 11, 2009 at 11:03 PM, Michael
Rynn<michaelrynn optushome.com.au> wrote:
 I have made a little beginning attempt at understanding the phobos xml
 modules. =A0For a warm up exercise, I took the std.xml and std.encoding
 from phobos 2, and made them into std2.xml and std2.encoding in the
 DSource project std2, so the dmd 1.0 compiler took a tour tripping up
 all over the code.

 They are attached as files on the http://www.dsource.org/projects/std2
 page.

Great.  Thanks for the contribution.  I've checked your ports into svn now.


 I cannot believe that anybody actually uses the std2 project files.

What do you mean by that?

 I =A0took away all the safety features of const, immutable and any other
 things that dmd 1.0 complains about, as "recommended" for the std2.

 Inserted a few support functions not immediately avalable in Phobos
 1.0, =A0notably some =A0suggested better code =A0for the encode (xml enti=

ty)
 , and renamed it as encodeStdEntity, from a Tango fan in the D bugs
 postings.

 Ran unit tests and also successfully rab the books.xml example.

 Re-enabled the disabled unittest in std.encoding, and fixed the little
 bug that was spoiling it, also posting that to the D bugs.

 ..

Very cool.  You should probably know, though, that std.xml is not very
popular.  I'm don't have much to do with XML -- by choice, horrid
stuff if you ask me -- but folks who have played with std.xml have
found it buggy and very slow.  And the original author has
disappeared.  There has been talk that it needs to be rewritten from
scratch.  Or perhaps replaced with a port of tango's very speedy xml
library.

--bb

Robert Fraser <fraserofthenight gmail.com> writes:

 On Tue, Aug 11, 2009 at 11:03 PM, Michael
 Rynn<michaelrynn optushome.com.au> wrote:
 I  took away all the safety features of const, immutable and any other
 things that dmd 1.0 complains about, as "recommended" for the std2.


Well, different people feel differently about these things. IMO, the 
complication of const/immutable and loss of productivity doesn't justify 
the small gains in safety. Whether it justifies the muiltithreading 
advantages is yet to be seen, though there's some promising stuff coming 
from the minds of A&W.

Bill Baxter wrote:
 Very cool.  You should probably know, though, that std.xml is not very
 popular.  I'm don't have much to do with XML -- by choice, horrid
 stuff if you ask me -- but folks who have played with std.xml have
 found it buggy and very slow.  And the original author has
 disappeared.  There has been talk that it needs to be rewritten from
 scratch.  Or perhaps replaced with a port of tango's very speedy xml
 library.

Yup... among other things, it does not correctly handle elements that 
are closed with /> (i.e. `<x/>` instead of `<x></x>`). This is basically 
a show-stopper bug, IMO. As far as Tango's XML goes, yeah, it's awesome; 
check out:

http://dotnot.org/blog/archives/2008/03/10/xml-benchmarks-updated-graphs-with-rapidxml/
http://dotnot.org/blog/archives/2008/03/10/xml-benchmarks-parsequerymutateserialize/
http://dotnot.org/blog/archives/2008/03/09/xml-benchmarks-pros-and-cons-of-each-library/
http://dotnot.org/blog/archives/2008/03/12/why-is-dtango-so-fast-at-parsing-xml/

That second link in particular is a "holy ****; look at those bars" type 
of experience.