This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author paulj
Recipients dalke, facundobatista, ijmorlan, jjlee, orsenthil, paulj, skip.montanaro, vila, vincentk
Date 2009-10-03.22:37:31
SpamBayes Score 5.551115e-16
Marked as misclassified No
Message-id <1254609454.97.0.969042025037.issue1462525@psf.upfronthosting.co.za>
In-reply-to
Content
Senthil wrote:
> > Senthil <orsenthil@gmail.com> added the comment:
> >
> > Hello Paul, 
> > Have you beeing keeping track of urlparse changes in Python2.6? 

No - do you have pointers to the particular changes you're
referring to?  I've done a bit of trying to catch up by searching
the mailing list, but want to make sure I know what you're
referring to in particular.

> > I 
> > anaylzed your patch and read through the RFC3986 and have the 
> > following comments:
> >
> > 1) The usage of this module is very diffirent from the current 
> > urlparse module's usage. It might be that this module was designed to 
> > co-exist with urlparse, providing certain additional functionalities. 
> > But inorder to replace urlparse, I find this module is "Backward 
> > Incompatible with the code base". 
> >
> > Some comments extra features provided /claims of this module.
> >   

Yes, there was no design goal of backward compatibility.

> > 2) The module provides URI handling framework that includes default 
> > URI Parsers for common URI Schemes.
> >     - RFC3986 specifies that scheme handling part is left to the 
> > separate RFC describing the schemes. 
> >     - uriparse library attempts that by providing default port and 
> > default hostname for certain schemes, but that can be made available 
> > as a patch to urlparse rather than new library. The need for such a 
> > change in urlparse needs to be analyzed, as there has not been any 
> > requirement raised as such for proving default port, default host for 
> > schemes whenever it is applicable.
> >   

Okay; It just seemed completist to provide said defaults.

> > 3) urlsplit, urlunsplit, spliting the authority into sub-components is
> > available in the current urlparse library itself and is RFC3986 
> > conformant.
> >   

Ah... it used to not do this for unknown schemes, which was my
original impetus for this.

> > 4) urljoin in the current urlparse ( patched with fixes) is currently 
> > RFC3986conformant.
> >
> > What urlparse further requires and this patch also lacks is ( as 
> > commented by John J Lee)
> > 1) Handling of IRIs.
> > 2) Python Unicode Strings.
> > 3) Percent- Encodings for IRIs and Python Unicode Strings.
> > ( There is a discussion going on on quote and unquote of unicode, and 
> > thatwould be basically be extended to above points as well)
> >
> > - If required, we can adopt the default host and port provision 
> > mechanisms as  mentioned in this patch to the current urlparse. 
> >
> > Other that that, I see that urlparse currently has all changes as 
> > mentioned inthis patch and makes the attached patch an obsolete one.
> >
> > Please let me know your comments/ thoughts.
> >   

It seems that urlparse now works for the case that caused me to rewrite
this (see the first comment on this bug for a link to the python-dev 
archives
where I posted about the 'itch' this code 'scratched'), so it's fine 
with me if
it just goes away now.
History
Date User Action Args
2009-10-03 22:37:35pauljsetrecipients: + paulj, skip.montanaro, facundobatista, jjlee, dalke, orsenthil, vila, ijmorlan, vincentk
2009-10-03 22:37:34pauljsetmessageid: <1254609454.97.0.969042025037.issue1462525@psf.upfronthosting.co.za>
2009-10-03 22:37:33pauljlinkissue1462525 messages
2009-10-03 22:37:31pauljcreate