GNOME Bugzilla – Bug 637291
Can't send to both internal and external recipients
Last modified: 2012-06-25 18:14:51 UTC
Hello, I can use evolution + mapi to receive email in my organisation, and also to send emails and reply to emails. However for small subset of emailadresses in our organisation this does not work. Emails to these addresses, or replies to emails originating from such an address never make it out of the outbox. They also block other emails in the outbox until I remove them. When trying to force sending the messages in the outbox I get the following error message: "Error while Sending message. ModifyRecipients: MAPI error MAPI_E_CALL_FAILED (0x80004005) occurred" I have no idea what causes this message. I don't know how to get more information about what is happening. I have no clue about what the addresses in the subset that doesn't work have in common. So the first thing I need is information on how to investigate this further. (there is unfortunately very little I could find on the internet...)
Thanks for a bug report. What is your exact openchange version you are using with evolution-mapi 0.32, please? Is there anything special about those "other" addresses which are not working? I'm thinking of being only usernames, not email addresses, or are pointing to a different domain on the server (bug #621064), or something like that? It will be good if you could test a patch from bug #621064 too. You can also run evolution like this: $ MAPI_DEBUG=10 evolution &>evo.log but be ready that this log is pretty long and hard to read, thus rather search for "ecInvalidRecips" and see whether it's same/similar as in the above mentioned bug report. Thanks in advance.
I'm not using OpenChange. I'm using evolution + mapi in an attempt to survive in an organisation that is pathetically MS centric. So I have to work with an Exchange server. I've started to collect those email addresses that fail. I'll run evolution with more debugging and post the results here.
openchange provides libmapi, which is used in evolution-mapi. You have installed it, and I'm asking for openchange version you have installed. I just faced similar downstream bug report: https://bugzilla.redhat.com/show_bug.cgi?id=667167
As far as I can find out I've got openchagen 0.9
I also keep getting this error with the same problem: Could not send message: ModifyRecipients: MAPI error MAPI_E_CALL_FAILED (0x80004005) occurred I am using evolution-ews-3.4.2-2.fc17.i686 evolution-exchange-3.4.1-1.fc17.i686 (In reply to comment #0) > Hello, > > I can use evolution + mapi to receive email in my organisation, and also to > send emails and reply to emails. However for small subset of emailadresses in > our organisation this does not work. Emails to these addresses, or replies to > emails originating from such an address never make it out of the outbox. They > also block other emails in the outbox until I remove them. > > When trying to force sending the messages in the outbox I get the following > error message: > > > "Error while Sending message. > > ModifyRecipients: MAPI error MAPI_E_CALL_FAILED (0x80004005) occurred" > > I have no idea what causes this message. I don't know how to get more > information about what is happening. I have no clue about what the addresses in > the subset that doesn't work have in common. > > So the first thing I need is information on how to investigate this further. > (there is unfortunately very little I could find on the internet...)
(In reply to comment #5) > I also keep getting this error with the same problem: Could not send message: > ModifyRecipients: MAPI error MAPI_E_CALL_FAILED (0x80004005) occurred > > I am using > > evolution-ews-3.4.2-2.fc17.i686 > evolution-exchange-3.4.1-1.fc17.i686 It should be evolution-mapi here. I received a question from other reporter in a downstream bug where he got a suspicion whether this can be related to address length, like he has an issue with addresses longer than 30-37 letters. Do you think it can be it?
(In reply to comment #6) > (In reply to comment #5) > > I also keep getting this error with the same problem: Could not send message: > > ModifyRecipients: MAPI error MAPI_E_CALL_FAILED (0x80004005) occurred > > > > I am using > > > > evolution-ews-3.4.2-2.fc17.i686 > > evolution-exchange-3.4.1-1.fc17.i686 > > It should be evolution-mapi here. I received a question from other reporter in > a downstream bug where he got a suspicion whether this can be related to > address length, like he has an issue with addresses longer than 30-37 letters. > Do you think it can be it? Sorry, forgot to include evolution-mapi, evolution-mapi-3.4.2-1.fc17.i686 what I did find is that I had a an old rpm from fc12 still hanging around that checked duplicate addresses in the contact book. As soon as I deleted it, I was able to start sending email. Now it evolution crashes a lot in my kde env. However, I am not convinced that evolution is the problem, but something else
(In reply to comment #7) > Now it evolution crashes a lot in my kde env. > However, I am not convinced that evolution is the problem, but something else It would be good to open a _new_ bug report and attach backtrace of crashed evolution (with debug info packages installed for evolution-data-server, evolution, evolution-mapi), thus it could be checked where it crashes, and maybe why too. (Please CC me there, or write here a reference to it.)
With help from Gianluca from downstream bug [1] I investigated further and was finally able to reproduce this too, even in the latest development version of evolution-mapi. It seems to be the same problem as in OpenChange's [2], which is the way how I reproduced it, thus I'll try to do similar fix in evolution-mapi like Brad did in OpenChange. [1] https://bugzilla.redhat.com/show_bug.cgi?id=827371 [2] http://tracker.openchange.org/issues/343
Created attachment 217231 [details] [review] ema patch for evolution-mapi; After tweaking with slightly complicated patch I cut this down to basically one-liner. The PidTag7BitDisplayName property is causing trouble here. It is resolved for internal addresses, but is missing for external addresses. Rather than trying to provide some 7bit version of the display name I decided to drop the property from the list of required properties, which made it work.
Created commit eb0cfe3 in ema master (3.5.4+) Created commit 03006a1 in ema gnome-3-4 (3.4.4+)