Re: [Duplicity-talk] Does duplicity rely on a special namespace abbreviation in webdav respones?

[edgar . soldin]

On 27.07.2022 19:17, edgar.soldin@web.de wrote:
> On 27.07.2022 17:59, Kenneth Loafman via Duplicity-talk wrote:
> > I don't have the spec for WebDAV, but I'm guessing that your addition to duplicity 
> > will do no harm.  Perhaps you could submit a bug request with the fix to issues 
> > <https://gitlab.com/duplicity/duplicity/-/issues> so we can track it.
> >
> > ...Thanks,
> > ...Ken
> >
> >
> > On Wed, Jul 27, 2022 at 6:22 AM Felix Prüter via Duplicity-talk 
> > <duplicity-talk@nongnu.org <mailto:duplicity-talk@nongnu.org>> wrote:
> >
> >     Hi ede,
> >     thanks for your fast reply!
> >
> >     Am 2022-07-19 12:11, schrieb edgar.soldin--- via Duplicity-talk:
> > >     duplicity is outdated and  not maintained any more. please upgrade to 
> > > latest 0.8.23 and see if the error persists.
> >
> >     1) I know that the 0.7-series is EOL but my hoster still use it :-(
> >     2) The error persits even with the newest version
> >
> > > >     Does the webdav backend require a "d" as abbreviation for the xml namespace 
> > > > and does not parse the xml response otherwise?
> > >
> > >     can't tell, but feel free to have a look in
> > >
> > >     
> > > https://gitlab.com/duplicity/duplicity/-/blob/main/duplicity/backends/webdavbackend.py
> > >  
> > > <https://gitlab.com/duplicity/duplicity/-/blob/main/duplicity/backends/webdavbackend.py>
> >
> >     My assumption seems to be correct: in the webdavbackend.py in line 349
> >
> > >     for href in dom.getElementsByTagName(u'd:href') + 
> > > dom.getElementsByTagName(u'D:href'):
> >
> >     Adding an additional dom.getElementsByTagName(u'ns0:href') to that line 
> > solves the problem!
> >
> >     Question: is it conceivable that the namespace prefix gets configurable 
> > within duplicity?
> >     Or should that issue ("wrong" namespace prefix: ns0 instead of d/D) be 
> > solved on the backend's side?
>
> quick search
>   spec - http://www.webdav.org/specs/rfc4918.xml#METHOD_PROPFIND
>   example - 
> https://github.com/laurent22/joplin/commit/01614b5a136cab669e961ef4d1b95c3738afad6a
> suggests that the namespace(s) should be parsed from the response and used 
> accordingly.
>
> surely that'd be the cleaner solution.
>
> ..ede

this looks promising, just ignore the namespace.
 
https://stackoverflow.com/questions/2528852/how-do-i-get-minidom-to-ignore-namespaces

the backends "tastes" the hrefs anyway in 
https://gitlab.com/duplicity/duplicity/-/blob/main/duplicity/backends/webdavbackend.py#L350
 so it should be safe to simply check out all `href`s given in the response.

Felix, would you mind to try that solution? ..ede/duply.net