duplicity-team team mailing list archive

[edso <edgar.soldin@xxxxxx>]

On 15.06.2014 22:57, Michael Terry wrote:
>> i would bet that now every request needs to be re-authenticated now. not sure that's a good idea.
> 
> Well...  We already did the connection close in many cases.  I just made it consistent.

nope, just when failure occured, of which the error cause might have been the connection, hence the reset.
 
> I feel like the connection close to clear out the bad data/status was a bit of hack in the first place -- seems like really odd behavior of the library, but the workaround worked well.  I'm just extending the hack.

actually the workaround was to read the response, which seemed to be managed by the connection and not properly cleaned on response.close()
 
> If we could come up with a better way that doesn't mean we close the connection, I'm all ears.  

yeah.. just read out the response every time and close/reset only on errors.

>But I'm not very familiar with http python libraries.

pfffh.. don't ask me. i did this stuff in php & perl once upon a time, but i assume the mechanics shouldn't be to different.
 
>> PS: i wonder why the backend suddenly needs this other way to write the request data. any clue?
> 
> It was an extra check added during the backends re-organization to enforce correctness.  When writing to a Path object, you ideally call setdata() on it after.
> 

ohh, i see, so your added check was that actually failed. not the old code?

..ede

-- 
https://code.launchpad.net/~mterry/duplicity/webdav-fixes/+merge/223183
Your team duplicity-team is requested to review the proposed merge of lp:~mterry/duplicity/webdav-fixes into lp:duplicity.