Re: [Jython-dev] xml.etree.ElementTree?

[Sébastien B. <Seb...@en...>]

Philip Jenvey wrote:
>
> On Nov 14, 2008, at 2:30 PM, Sébastien Boisgérault wrote:
>
>> Sébastien Boisgérault wrote:
>>>
>>> I plan to totally REMOVE from the subversion repository any ElementTree
>>> 1.3 file in the next few days, to replace them with the standard 2.5
>>> package.
>>>
>> That much is done.
>>
>>   Homepage: http://code.google.com/p/jython-elementtree/
>>   Download: 
>> http://jython-elementtree.googlecode.com/files/jython-elementtree-0.2.0.tar.gz 
>>
>>
>> Any feedback would be appreciated at this point.
>>
>> If you wish too look and comment the code, you may use the simple 
>> code review support of google code hosting (see 
>> http://code.google.com/p/support/wiki/CodeReviews).
>
> Interesting that you have partial support for the expat module. I 
> wonder how it would fair against something like Genshi.
>
I'll have a look.
> Does this require Java 6 (or do only some parts?). Not that that would 
> bar it from being incorporated into Jython 2.5.
Arf, I don't know. I wrote "Java 6" out of habit in the source code but 
I don't know if I specifically use Java 6 features ...
Until now, I have never tried the code with Java 5.
>
> As for tests, CPython 2.5 also has a test_xml_etree.py. I'd like to 
> see how it fairs against that.
Not too bad actually: there is 1 test that fails out of 96, because the 
"mac-roman" encoding is not supported in Jython.
More detailled tests are available in ElementTree downloads from 
effbot's source distributions (selftest.py file). So far:
  * ET 1.2.6: 3 tests out of 296 fail
  * ET 1.3: 6 tests out of 337 fail.
There are three categories:
  - expat version detection (minor issue),
  - undefined entities error reporting, something I need to have a 
closer look at. This is related to the DefaultHandlerExpand expat 
callback that I don't support for now. Not that all 3 ET 1.2.6 errors 
fall in this category.
  - error locations (line, column number): expat and java sax seem to 
have a different behavior here, I am not sure that something can be done.
>
> The FormEncode tests use elementtree for a few simple things, I'll see 
> if I can try it against jython-elementtree:
>
> http://pylonshq.com:8014/builders/Jython%202.5-trunk%20on%201.6.0_03%20%28server%29%20FreeBSD/builds/509/steps/shell_26/logs/stdio 
>
Good ! I'll have a look too.

> -- 
> Philip Jenvey
>
>
>