[
https://issues.apache.org/jira/browse/GEODE-9387?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17366805#comment-17366805
]
ASF GitHub Bot commented on GEODE-9387:
---------------------------------------
pdxcodemonkey opened a new pull request #822:
URL: https://github.com/apache/geode-native/pull/822
- makes the tool robust against weird versions of NC
- no longer any need to combine rolled log files to parse messages
Also fixed several issues I noticed in timestamps:
- chunked server responses didn't have timestamps
- some timestamps were strings, others datetime objects, so they printed out
differently ('/' vs '-')
- gnmsg wasn't parsing out the date, only the timestamp, so output had
today's date rather than the date in the log file
@todmorrison
--
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]
> gnmsg: remove version detection and switch regex based on found log entries
> ---------------------------------------------------------------------------
>
> Key: GEODE-9387
> URL: https://issues.apache.org/jira/browse/GEODE-9387
> Project: Geode
> Issue Type: Improvement
> Components: native client
> Reporter: Blake Bender
> Priority: Major
>
> gnmsg currently looks for the log line containing the version of native
> client being used, in order to detect which regex to use to parse message
> traces, among other things. This has several implications:
>
> i. gnmsg cannot parse messages in a "rolled" log file, because only the first
> file contains the version stamp
> ii. the parsing code is very brittle, and routinely breaks when coming across
> a log from a point release we haven't seen logs for, or a developer build
> with an uninitialized version stamp, etc.
>
> What should happen is, where there are possible variants for a given trace,
> gnmsg should check for all variants until it finds a match, then thereafter
> only check for that variant. This should still yield decent performance, but
> also no longer care about the version stamp, allowing us to parse rolled log
> files and be much more robust to the types of logs we can read.
>
>
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)