[
https://issues.apache.org/jira/browse/COUCHDB-2764?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14650476#comment-14650476
]
Alexander Shorin commented on COUCHDB-2764:
-------------------------------------------
Don't think that this is an issue because:
1. In regular CouchDB workflow, new_edits=false is only used by replicator
which known what it doing and actually replicated documents are valid for their
revisions.
2. If user updated documents with new_edits=false it also provides some
unspoken guarantee that document`s revisions are really belongs to them and
reflects their content uniquely and correctly. If they don't - user creates
problem for himself and there is nothing much to do if all documents in batch
with the different content has similar revisions.
> Etag can be made to not reflect content
> ---------------------------------------
>
> Key: COUCHDB-2764
> URL: https://issues.apache.org/jira/browse/COUCHDB-2764
> Project: CouchDB
> Issue Type: Bug
> Security Level: public(Regular issues)
> Reporter: Dale Harvey
>
> https://gist.github.com/daleharvey/2010c3d3f873bc03fc5e
> Is a curl example, basically if you new_edits: false a document you can use
> whatever data you want and the revision is not determined by its content,
> which makes can cause bugs if stuff depends on the etag being correct.
> We are hit this a lot in the pouchdb test suite since we delete databases and
> replicate into them often
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
