Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Threaded View
Avro >> mail # user >> Best practice for versioning IDLs?


Copy link to this message
-
Re: Best practice for versioning IDLs?
Hi Scott,

Thanks for the suggestion. I like the metadata approach!

I just tried that property on one of my protocol IDL files and it didn't
work (parsing error). I discussed this is Doug Cutting as part of JIRA
ticket 886 (https://issues.apache.org/jira/browse/AVRO-886). However,
that ticket is now closed so I should probably open a new one unless I'm
doing something wrong. I was hoping to work on this a while back but got
busy with other tasks. Looks like now might be the right time:)

Thanks,
George

On 11/29/11 1:22 PM, Scott Carey wrote:
> I don't think there are yet best practices for what you are trying to do.
>
> However, I suggest you first consider embedding the version as metadata in
> the schema, rather than data.  If you put it in a Record, it will be data
> serialized with every record.  If you put it as schema metadata, it will
> only exist in the schemas and not the data.
>
> In raw JSON schema form, the metadata can be added to any named type:
> Record, Fixed, Enum, Protocol.   The "doc" field is a special named
> metadata field, you can use it or add your own:
>
> {
>    "namespace": "com.acme",
>    "protocol": "HelloWorld",
>    "doc": "Protocol Greetings",
>    "acme.version": "1.22.3",
>
>    "types": [
>      {"name": "Greeting", "type": "record", "fields": [
>        {"name": "message", "type": "string"}]},
>      {"name": "Curse", "type": "error", "fields": [
>        {"name": "message", "type": "string"}]}
>    ],
>
>    "messages": {
>      "hello": {
>        "doc": "Say hello.",
>        "request": [{"name": "greeting", "type": "Greeting" }],
>        "response": "Greeting",
>        "errors": ["Curse"]
>      }
>    }
> }
>
> http://avro.apache.org/docs/current/spec.html#Protocol+Declaration
>
> For IDL, it should be possible to add a property using the
> @propname("propval") annotation on the protocol.
> http://avro.apache.org/docs/current/idl.html#defining_protocol
>
> I have not tried this myself however.
>
> If I had the setup to test it now, I would try to see if the below AvroIDL
> creates an empty protocol with the "acme.version" property set:
>
> @acme.version("1.22.3")
> @namespace("com.acme")
> protocol HelloWorld {
>
> }
>
>
>
> On 11/29/11 9:20 AM, "George Fletcher"<[EMAIL PROTECTED]>  wrote:
>
>
>>
>>
>>
>>
>>
>>     Hi,
>>
>>       I'd like to incorporate a semver.org style versioning structure
>>       for the IDL files we are using. The IDLs represent interfaces of
>>       services (ala SOA).
>>
>>       We currently manage our IDL files separately from the
>>       implementation as multiple services might use the same IDL. This
>>       makes it critical to have the IDL's understand their version. I'd
>>       like to see our build process be able to inject into the IDL the
>>       version from the build environment (currently maven). Another
>>       option would be to define the version within the IDL. However, the
>>       only way I can think of to do this, is to create a "Version"
>>       Record within each IDL and then maybe have the Record contains 3
>>       string fields (major, minor, patch).
>>
>>       Just wondering if there are any best practices already established
>>       for this kind of IDL versioning.
>>
>>       Thanks,
>>       George
>>
>
>

--
Chief Architect                   AIM:  gffletch
Identity Services Engineering     Work: [EMAIL PROTECTED]
AOL Inc.                          Home: [EMAIL PROTECTED]
Mobile: +1-703-462-3494           Blog: http://practicalid.blogspot.com
Office: +1-703-265-2544           Twitter: http://twitter.com/gffletch

NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB