Re: [Evolution] More authentication headaches

(Sorry for the awkward quoting but I'm having to use Google webmail to post this, see the end of the post for why this is so).

On 15 September 2016 at 11:20, Milan Crha <mcrha redhat com> wrote:
On Wed, 2016-09-14 at 18:46 +0100, Patrick O'Callaghan wrote:
> I then fired up Evo again and created new mail accounts. I noticed
> that after creating my Gmail account (using the Evo wizard with the
> default settings) I was not asked for the password, but the account
> loaded anyway.

it found the token in the keyring from the earlier time. Seahorse helps
to find it and eventually delete it, but I do not know whether I would
do it, because your test as a new user does pretty much the same.

> However after a while Evo started saying:
> Failed to connect 'Gmail'
> HTTP Error: Unauthorized

The 'after a while' is an important detail for me. Does it also mean
that the email works for say an hour, then it starts failing?

This seems to be a moving target. When I reported it, the error appeared as soon as I started Evo. Now, after I've reset everything for a second time and followed Pete's advice about removing entries from the keyring, it's not appearing for mail, but still appears for calendar.
By the
way, is this error returned from the mail part or the calendar part?

 Mail, but right now it has gone.

You can try to log something from the mail with:
   $ CAMEL_DEBUG=imapx:io evolution
Search for the LOGIN command failure, which may eventually show also a
reason given by the Gmail server.

Can't see anything in the log (captured for a whole session which had the error). The text of the error isn't in the log and the IMAP authentication appears to be working, despite the error message.

> Furthermore, my Google calendar (which *did* ask for the password
> first time round) simply doesn't load. Sometimes the same error
> message as above, sometimes not, but no calendar entries in either
> case.

To debug the calendar part run:
   $ CALDAV_DEBUG=all /usr/libexec/evolution-calendar-factory -w
then wait a bit (few seconds), thus the service is known to D-Bus, and
then open the evolution/calendar. Apart of a raw communication between
the client and the server, it also shows how many items are in the
cache and other CalDAV related information (the lines are usually
prefixed or contain 'CalDAV' string).
 I get a variety of errors:

The task list backend servicing "Personal" has quit unexpectedly.
Some of your tasks may not be available until Evolution is restarted.
Error creating view for the memo list 'Personal'
Cannot invoke method; proxy is for a well-known name without an owner and proxy was constructed with the G_DBUS_PROXY_FLAGS_DO_NOT_AUTO_START flag
Usual procedure after load is that the CalDAV calendar checks for ctag
and if it doesn't match the one stored in
then it checks for changes, otherwise it does nothing. Could you try to
create an event in your main Google calendar, then right-click the
calendar and pick 'Refresh' from the context menu? The log will show a
PUT request, immediately followed by a GET request, then (as a response
for the Refresh) it will show the ctag check, which should change, due
to added event.
Got this:
Event cannot be edited, because the selected calendar is read only

I then deleted the calendar and created a new one pointing at my Google calendar. Same errors, calendar is not synced with the Google one, and I wasn't asked for authentication. The log shows:

> OPTIONS /caldav/v2/ HTTP/1.1
> Soup-Debug-Timestamp: 1473937214
> Soup-Debug: SoupSessionSync 1 (0x55b1e1e41140), SoupMessage 1 (0x7fa2440088e0), SoupSocket 1 (0x7fa244001b90)
> User-Agent: Evolution/3.20.5
> Connection: close
> Accept-Language: en-gb, en;q=0.9
< HTTP/1.1 403 Forbidden
< Soup-Debug-Timestamp: 1473937214
< Soup-Debug: SoupMessage 1 (0x7fa2440088e0)
< Vary: X-Origin
< Content-Type: text/html; charset=UTF-8
< Date: Thu, 15 Sep 2016 11:00:14 GMT
< Expires: Thu, 15 Sep 2016 11:00:14 GMT
< Cache-Control: private, max-age=0
< X-Content-Type-Options: nosniff
< X-Frame-Options: SAMEORIGIN
< X-XSS-Protection: 1; mode=block
< Server: GSE
< Alt-Svc: quic=":443"; ma=2592000; v="36,35,34,33,32"
< Accept-Ranges: none
< Vary: Origin,Accept-Encoding
< Connection: close
< Daily Limit for Unauthenticated Use Exceeded. Continued use requires signup.
Finally, why am I posting this via webmail? Because when I try to post using my normal account (*not* the one which has been giving me the problems mentioned above), the sending process simply hangs. Recall that I deleted all my Evo config and keyring info. I then recreated my personal account (on Gmail) and my work account (this one, which is on Google Apps). I was asked to authenticate the Gmail one, but not this one. Possibly because the address (poc usb ve) doesn't indicate Google directly so the setup wizard treats it differently, but that's speculation. However this had been working correctly before I tried to reset the whole shebang. I speculate that enough had been left over from before I switched from GOA to direct OAuth2 to keep it working, but now it has lost the tokens and doesn't know how to get them back.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]