[Duplicity-talk] Fwd: Re: Duplicity Backup bug of GoogleDocs backend

[edgar . soldin]

can please somebody using GDocs backend check if Carlos's suggestion below 
works as expected? it's ugly but a bit more generic then the previous patch on 
pastebin.

if so, i would prepare launchpad branch for inclusion into trunk so the error 
can be fixed.

..ede/duply.net


-------- Original Message --------
Subject:        Re: Duplicity Backup bug of GoogleDocs backend
Date:   Tue, 19 Nov 2013 17:34:10 +0100
From:   Carlos Abalde <address@hidden>
To:     address@hidden



Hi Edgar,

I'm sorry but I missed the e-mail about the problem with the GDocs backend. I 
have not been using the GDocs backend lately and I unsubscribed the mailing 
list some time ago. I've reviewing the code today and I think the best approach 
is replacing the check in __fetch_entries() from,

if (not type) or (entry.get_resource_type() == type):

to,

resource_type = entry.get_resource_type()
if (not type) or
   (type == 'folder' and resource_type == 'folder') or
   (type == GDocsBackend.BACKUP_DOCUMENT_TYPE and resource_type != 'folder'):

In general the extra filtering in this function when filtering by entry title 
is ugly (and I was not able to find an alternative). This change simply adds a 
little bit more ugliness to this part of the code. Checking against the exact 
mime type I think is overkill. This change should be good enough for all cases.

Cheers,

--
Carlos Abalde.



On Tue, Nov 19, 2013 at 11:35 AM, <address@hidden <mailto:address@hidden>> 
wrote:

    On 19.11.2013 01:09, Kostas Papadopoulos wrote:
    > Hi,
    >
    > Please note that when trying to use duplicity 0.6.22 and duply 1.5.11 
with GoogleDocs backend (under Debian Wheezy), I experienced the same errors as 
described here:
    >
    > http://lists.nongnu.org/archive/html/duplicity-talk/2013-07/msg00007.html
    >
    > I just applied his patch, and duplicity+duply now seem to work fine.
    >

    did anybody else experience this issue and can confirm the patch works?

    looking at the patch i feel a bit unsure if it is safe to just compare left 
values of mime-types. wouldn't it make more sense to not compare mime-types at 
all here?

    Carlos: what do you think?

    ..ede/duply.net <http://duply.net>




-- 
Carlos.