You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@avro.apache.org by "Daniel Schierbeck (JIRA)" <ji...@apache.org> on 2015/03/04 14:10:04 UTC
[jira] [Updated] (AVRO-1645) Improved handling of missing named
types
[ https://issues.apache.org/jira/browse/AVRO-1645?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Daniel Schierbeck updated AVRO-1645:
------------------------------------
Description:
I've written a Ruby library that wraps the Avro Ruby library and provides a "registry" of schemas. The idea is to allow users to avoid repeating themselves by allowing inter-schema references using the type names. At runtime any missing type is resolved by loading the relevant schema from disk.
This works great, but depends on parsing the error message from the Avro::SchemaParseError exception, extracting the name of the missing type. If there was a more specific subclass, e.g. Avro::MissingSchemaError, that provided an attribute with the name of the schema, my library would be a lot more robust. I'd rather not have everything break if you change the error message.
I'd be happy to code this myself (it's a rather trivial and backwards compatible change), but I'd prefer to get a +1 here first.
Here's the project: https://github.com/dasch/avro_turf (still in development)
was:
I've written a Ruby library that wraps the Avro Ruby library and provides a "registry" of schemas. The idea is to allow users to avoid repeating themselves by allowing inter-schema references using the type names. At runtime any missing type is resolved by loading the relevant schema from disk.
This works great, but depends on parsing the error message from the Avro::SchemaParseError exception, extracting the name of the missing type. If there was a more specific subclass, e.g. Avro::MissingSchemaError, that provided an attribute with the name of the schema, my library would be a lot more robust. I'd rather not have everything break if you change the error message.
I'd be happy to code this myself (it's a rather trivial and backwards compatible change), but I'd prefer to get a +1 here first.
> Improved handling of missing named types
> ----------------------------------------
>
> Key: AVRO-1645
> URL: https://issues.apache.org/jira/browse/AVRO-1645
> Project: Avro
> Issue Type: Improvement
> Reporter: Daniel Schierbeck
>
> I've written a Ruby library that wraps the Avro Ruby library and provides a "registry" of schemas. The idea is to allow users to avoid repeating themselves by allowing inter-schema references using the type names. At runtime any missing type is resolved by loading the relevant schema from disk.
> This works great, but depends on parsing the error message from the Avro::SchemaParseError exception, extracting the name of the missing type. If there was a more specific subclass, e.g. Avro::MissingSchemaError, that provided an attribute with the name of the schema, my library would be a lot more robust. I'd rather not have everything break if you change the error message.
> I'd be happy to code this myself (it's a rather trivial and backwards compatible change), but I'd prefer to get a +1 here first.
> Here's the project: https://github.com/dasch/avro_turf (still in development)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)