Can you check which version of Couchbase Mobile for Android is in use? We found/resolved an issue with the replicator in December.
http://www.couchbase.com/issues/browse/CBMA-22 If I remember correctly, the "bad_utf8_character_code" is a red herring caused by us logging some binary attachment data. The fix for this bug was included in the latest release. marty On Mon, Mar 19, 2012 at 5:29 AM, Karoly Szanto <[email protected]> wrote: > Hi Paul, > Thanks for the fast reply. > > I am replicating from couchDB instances running on Android phones towards > the server (and the other way around). The couchDB for Android is > https://github.com/couchbase/Android-Couchbase which is probably based on > a different (older) version of couchDB. > > The sever is running *version:"1.1.1"* > * > * > Karoly > > On Mon, Mar 19, 2012 at 10:21 AM, Paul Davis > <[email protected]>wrote: > >> Are you replicating between different versions of CouchDB? My first >> guess is that you've replicating older to newer and its hitting a >> stricter enforcement of Unicode semantics in surrogate pairs. >> >> On Mon, Mar 19, 2012 at 3:04 AM, Karoly Szanto <[email protected]> >> wrote: >> > Hi, >> > During the replication proces I get the following error in the couchdb >> > server log: >> > >> > {<0.13261.0>,crash_report, >> > [[{initial_call, >> > {mochiweb_acceptor,init, >> > ['Argument__1','Argument__2','Argument__3']}}, >> > {pid,<0.13261.0>}, >> > {registered_name,[]}, >> > {error_info, >> > {exit, >> > {ucs,{bad_utf8_character_code}}, >> > [{xmerl_ucs,from_utf8,1}, >> > {mochijson2,json_encode_string,2}, >> > >> > {mochijson2,'-json_encode_proplist/2-fun-0-',3}, >> > {lists,foldl,3}, >> > {mochijson2,json_encode_proplist,2}, >> > {couch_httpd,send_json,4}, >> > {couch_httpd,handle_request_int,5}, >> > {mochiweb_http,headers,5}]}}, >> > {ancestors, >> > >> [https,couch_secondary_services,couch_server_sup, >> > <0.35.0>]}, >> > {messages, >> > [{#Ref<0.0.8.97038>, >> > {ok, >> > {proc,<0.7774.0>,<<"javascript">>, >> > [{<<"_design/monarcaserver">>, >> > >> > <<"469-a002d8a3c537c48875b315164d8eadb6">>}], >> > {couch_os_process,prompt}, >> > {couch_os_process,set_timeout}, >> > {couch_os_process,stop}}, >> > {[{<<"reduce_limit">>,true}, >> > {<<"timeout">>,30000}]}}}]}, >> > {links,[<0.136.0>]}, >> > {dictionary, >> > [{mochiweb_request_body,<<>>}, >> > {mochiweb_request_qs, >> > [{"filter","monarcaserver/userfilter"}, >> > {"userid","70"}, >> > {"feed","continuous"}, >> > {"heartbeat","10000"}, >> > {"style","all_docs"}, >> > {"since","0"}]}, >> > {jsonp,no_jsonp}, >> > {mochiweb_request_cookie,[]}]}, >> > {trap_exit,false}, >> > {status,running}, >> > {heap_size,4108475}, >> > {stack_size,24}, >> > {reductions,5319640}], >> > []]}} >> > >> > My first guess is that this might be solved by specifying the charset in >> > the replication header but I'd like to know your opinion as well. >> > >> > Thank you, >> > Karoly >>
