[ 
https://issues.apache.org/jira/browse/AVRO-219?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Jeff Hammerbacher updated AVRO-219:
-----------------------------------

    Attachment: AVRO-219.patch.schema

Okay, I've got most schemas parsing correctly in my new implementation. See the 
attached patch or http://github.com/hammer/avro.

I'd really appreciate a review of the direction for new_schema.py; I'm moving 
on tonight to io.py and genericio.py and should be able to work out any issues 
with the implementation through that process, but I'd appreciate an extra set 
of eyes.

One note: I didn't want to maintain state when serializing a schema object to a 
string or when comparing to another schema object, so I kept around private 
variables for the schema objects which could have children to keep track of 
whether or not a child's name was resolved from the names cache. It's a bit 
hacky but I can easily switch to the alternative method if people don't like it.

Also, there are a few things not implemented: I don't check for correct default 
values, I don't handle the error schema, and I don't actually parse arbitrary 
properties (though they should not call a parse failure). On the other hand, I 
handle a far wider variety of schemas correctly than the existing Python 
implementation, and the variety of schemas used to test my implementation is 
wider.

> Break testio.py into testschema.py, testio.py, testgenericio.py, and 
> testdatafile.py
> ------------------------------------------------------------------------------------
>
>                 Key: AVRO-219
>                 URL: https://issues.apache.org/jira/browse/AVRO-219
>             Project: Avro
>          Issue Type: Improvement
>          Components: python
>            Reporter: Jeff Hammerbacher
>            Assignee: Jeff Hammerbacher
>         Attachments: AVRO-219.patch.schema
>
>
> Currently, the unit tests for schema.py, genericio.py, and datafile.py are 
> grouped in with the unit tests for io.py in testio.py. We should break the 
> tests into individual files so that we have better modularization of tests.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to