Github user vanzin commented on the pull request:
https://github.com/apache/spark/pull/1222#issuecomment-66821777
The log file contains a header section that contains `key=value` pairs
preceeded by a 4-byte length field encoded in binary form. That makes it easy
to parse the header without reading past it, which would be tricky with
`BufferedReader.readLine()`. With that approach, you'd probably have to re-open
the stream and skip the header somehow before wrapping it (in the case of a
compressed log).
The tests for old logs do exist. They're now in FsHistoryProviderSuite,
which is where the code to handle the legacy format lives. In my view it
doesn't make sense to keep code to handle legacy stuff in EventLoggingListener.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
