Thanks for a bug report. I tried to reproduce this, but no luck, my server doesn't suffer of this issue. I waited for more than 5 minutes while doing other stuff in the Evolution, only not touching EWS mail.

Typical!  :D

First, thanks for your time in assisting me. Is there anything I can do for you to help you diagnose further?  It has done it with me this morning again.

I have an IMAPX and EWS account set up.  I can see when it fails on the two bidirectional arrows next to the account names.  The IMAP account's remains normal, but the EWS account has a little red circle next to the arrows.

Similarly, I opened a meeting request that had been mailed to me, and wanted to send a "Decline", however, I get the error:
Found the appointment in the calendar 'Calendar'
Unable to send item to calendar 'Calendar'. Cannot receive calendar objects: Authentication failed

The bidirectinoal arrow now has a small cross (x) at the bottom right of the arrows.  I clicked open Calendar, which it did.  Then I closed the Calendar and the invite message I was declining.  Now the bidirectional arrows have a small red circle at the bottom right -- like a "no entry" traffic sign.

Now, in this state, it seems strange, but I can open and read any message, even with the red "no entry" sign present.  However, I cannot SEND any new message, or reply to an existing one.  I also cannot save the message I have spent time composing to Draft because I get the error:
An error occurred while saving to your Drafts folder.
The reported error was "Invalid password for myemailaddress@mydomain".  The message has most likely not been saved.

Fortunately, I have just discovered that I can save the message to my Outbox on my local PC.  I then quit Evolution and reopen it, go to my Outbox and tell it to Flush.  Then the message sends.

But this is a painful way of working!

So please let me know anything I can do to help fix this!

I think (and hope) the communication log between the server and the client will provide the information what happened to the connection. As you see the differences in the folder tree, I'd guess that there was done some connection changes. That you can read messages while being offline can be possible only if the messages were already downloaded locally. A truly new messages might not be read from the server when in offline mode (it claims here an error when I try to view a message which is not downloaded locally while the EWS account is in offline mode).

You can capture the communication log with this command:
   $ EWS_DEBUG=2 evolution &>log.txt

The most interesting part will be when the connection icon in UI changes and when you'll try to send a message while the EWS account is in the offline state (it may not necessarily print anything in the log on the send, especially when the code will just realize that the account is in the offline mode.

Created attachment 997119 [details]
EWS_DEBUG=2 evolution &>log.txt

BTW, I *did* include debug output when I created the ticket.

I have run the process to log.txt.  I sent a test email, and then left it viewing the IMAP folder for a few minutes.  The arrows stayed "normal".  But when I clicked on EWS Inbox, the arrows got the X next to them.  After refreshing again the red no-entry sign was there.

Thanks!

Created attachment 997198 [details]
log.txt, generated by EWS_DEBUG=2; evolution & >/tmp/log.txt

Hello Gentlemen!
New version, same distinguished company. And a very similar problem. Hopefully, my attachment, generated as Milan has described few posts above, will help.

(In reply to Louis van Dyk from comment #4)
> BTW, I *did* include debug output when I created the ticket.

My bad, I'm sorry for that, I should look more carefully around, not mixing different bug reports all together (in my mind).

The log from comment #4 shows that the last working request was only slightly more than 3 minutes before the first failing request. The log from comment #0 shows exactly 3 minutes difference between the last working and the first failing request. Both seem to be related to some folder update, one with a very long SyncState.

The log from comment #5 has about 4 and half minute delay between the last working request and the first failed. The failed request also uses longer sync state, but not as that long as Louis has.

Dan, I vaguely recall a similar bug, there was some issue with a long sync state too, but we were not able to reproduce it, if I'm not mistaken. Do you know the bug and eventually more details than me, please?

I can reproduce at will. Please let me know if I can help in any way shape or form. Evolution is not very useful right now, I am using thunderbird with DavMail on a Windows 8.1 VM right now.

(In reply to Mladen Gogala from comment #8)
> I can reproduce at will. Please let me know if I can help in any way shape
> or form. Evolution is not very useful right now, I am using thunderbird with
> DavMail on a Windows 8.1 VM right now.

What do you do exactly to reproduce the issue, please? Is it similar with Louis, send a message through EWS, then wait for about 5 minutes, and another attempt to send a message fails with the authentication error?

I'm wondering whether the error is not fake, maybe the server tries to tell that there are too many opened connections and rejects to use another one. Though it doesn't explain why libsoup didn't ask for a password when it should.

What are your mail account settings, including authentication method? Did you enter your password to evolution, and saved it into keyring?

I may have slightly different setup, which can explain why I cannot reproduce it.

My setup is as follows:
1.  I have an IMAPX account to my Linux server at home.  This is the second account in my account order on Evolution.  (The first and Default is "On This Computer".)
2.  The EWS is the third account.  On my work notebook, I use the Gnome Online Accounts.  On my home PC I chose to add the account directly in Evolution's accounts.  I experience the same problem on both installations.
3.  I use search folders - I have quite a few - and they are based on source email addresses and scan "all local and remote accounts".

All my passwords are SAVED so I don't have to enter them.  Each account checks for new messages every 5 minutes.  So it would seem that when the next check happens, that that is when the failure is reported to me.

Does that help?

Partly yes, thanks. What is checked (and eventually what values are) i, Edit->Preferences->Mail Accounts->Evolution EWS account->Edit, basically in each tab? I'll just setup as close to you and will see. I suspect that the server has short connection close interval, which is not seen in the logs, despite using "Connection: Keep-Alive" in evolution-ews (which still can be related).

Milan, after 5 min or so, server will ask me for authentication, regardless of the operation. Even if I try just retrieving an already read message.

Milan, my connection timeout is set to 600 seconds (10 min). What particular settings are you interested in?

(In reply to Louis van Dyk from comment #0)
> On F20 there was a similar problem which Dan fixed in libsoup. Bug 1088458.

doh... it looks like I never committed that fix to F21.

Can you confirm that the packages at http://koji.fedoraproject.org/koji/taskinfo?taskID=9138330 fix things?

I just installed your version of libsoup:

 Installing : libsoup-2.48.1-1.danw1.fc21.x86_64                           1/1 
  Verifying  : libsoup-2.48.1-1.danw1.fc21.x86_64                           1/1 

Installed:
  libsoup.x86_64 0:2.48.1-1.danw1.fc21          

Will let you know shortly.

It looks good. Ever since I installed the new version of libsoup, there are no more requests for re-authentication. Thanks a lot. Will that be pushed to the regular F21 updates or do I have to keep it in versionlock?
Once again, thank you very much! Evolution-ews is so much faster than DavMail based solution.

This is definitely good. I've had no problems with my email for the whole afternoon. I consider the bug fixed. I don't know about Louis. Despite the fact that we seem to encounter the same problems, we don't know each other.

Dan, I am so pleased that you remembered!  The new libsoup has fixed the problem for me ... again.

Could you please roll out the update for Fedora 21 (and Rawhide)?

Again, Milan and Dan, thanks SO much for your assistance.

libsoup-2.48.1-2.fc21 has been submitted as an update for Fedora 21.
https://admin.fedoraproject.org/updates/libsoup-2.48.1-2.fc21

Package libsoup-2.48.1-2.fc21:
* should fix your issue,
* was pushed to the Fedora 21 testing repository,
* should be available at your local mirror within two days.
Update it with:
# su -c 'yum update --enablerepo=updates-testing libsoup-2.48.1-2.fc21'
as soon as you are able to.
Please go to the following url:
https://admin.fedoraproject.org/updates/FEDORA-2015-3303/libsoup-2.48.1-2.fc21
then log in and leave karma (feedback).

libsoup-2.48.1-2.fc21 has been pushed to the Fedora 21 stable repository.  If problems still persist, please make note of it in this bug report.