Christopher Bonhage created COUCHDB-1639:
--------------------------------------------
Summary: doc_to_multi_part_stream does not include headers for attachments
Key: COUCHDB-1639
URL: https://issues.apache.org/jira/browse/COUCHDB-1639
Project: CouchDB
Issue Type: Bug
Components: Database Core, Replication
Reporter: Christopher Bonhage
I started a push replication job that contained a document with image attachments and noticed
that the replication job was taking far too long, so I dug into the logs and found the culprit:
[Tue, 08 Jan 2013 05:28:45 GMT] [error] [emulator] Error in process <0.22592.0> with
exit value: {{badmatch,{[137,80,78,71],[]}},[{couch_httpd,split_header,1},{couch_httpd,'-parse_part_header/1-fun-1-',2},{lists,foldl,3},{couch_httpd,parse_part_header,1},{couch_httpd,parse_multipart_request,3},{couch_doc,'-doc_from_multi_part_stream/2-fun-1-'...
I was able to deduce that the {[137,80,78,71],[]} that split_header was encountering while
trying to parse_part_header was actually the (partial) body of the attachment! After adding
some logging to confirm my suspicions I discovered that CouchDB does not appear to be sending
MIME multipart/related headers for attachments (only the application/json for the doc itself)
when converting a document to a multipart stream, which is causing the problem.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
|