rnewson commented on issue #2015: parse_revid - Badarg error in HTTP request URL: https://github.com/apache/couchdb/issues/2015#issuecomment-487321849 @wohali agree that this has been the behaviour since forever (since 1.0 itself) and so this is definitely not a regression or should be a surprise to anyone making bespoke rev values. for 4.0, I have a mild preference for saying that only couchdb gets to generate these values but the reality is that pouchdb generates rev values using a different algorithm and we (couchdb) should also be free to change (i.e, improve) this in future. I note, for full disclosure, that prior to the "deterministic rev" work that _rev used to be a random uuid. Replication works so long as the rev stays the same if the doc does, and changes if the doc changes. Anything beyond that is enhancement. So... in 4.0 I'd be fine with us defining what the valid format for a _rev string is and explicitly supporting any string that matches that (and dropping the integer conversion when length is 32 chars). 3.0 could maybe drop the '== 32' clauses as a stepping stone? Other voices should chime in, I'm not comfortable defining this here. (cc @davisp @kocolosk @chewbranca @daleharvey, a response from all of you would help, of course other voices are equally welcome).
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
