Support multiline

[okkez]

No description provided.

[okkez]

There are some problems in this implementation.

• %{GREEDYDATA} does not much multiline
  • so test is failed, for now.
• In parse(text), text does not include line break.
  • Maybe this line break is deleted by fluentd automatically.
  • Should we append line break?
• This plugin cannot process last record forever
  • How to detect the end of last record?

Request for comments, please.
Thanks.

[okkez]

This is for #6

[visualphoenix]

maybe it can buffer until an end regex is matched or until the start regex matches again? Obviously this would not work for multiple workers (same problem as multiline in logstash)

[visualphoenix]

could also possibly flush out after a timeout

[visualphoenix]

for parser problem maybe check how they did multilinecsv? https://github.com/fluent/fluentd/pull/543/files

[okkez]

Thank you for your comment!

maybe it can buffer until an end regex is matched or until the start regex matches again?

Yes, I think so, too.

could also possibly flush out after a timeout

Maybe, this is fluentd's in_tail plugin's issue.
I will check fluentd's in_tail plugin codes.

[okkez]

@visualphoenix How about 9f3b25f and db49d34?

[visualphoenix]

Looks good to me. We can invent new refinements as we use it more.

On Monday, November 16, 2015, okkez notifications@github.com wrote:

@visualphoenix https://github.com/visualphoenix How about 9f3b25f
9f3b25f
and db49d34
db49d34
?

—
Reply to this email directly or view it on GitHub
#10 (comment)
.

Raymond "VisualPhoenix" Barbiero

[okkez]

Thanks merged.