-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7780/
-----------------------------------------------------------

Review request for Flume.
Description
-------

The SyslogUtils class doesn't properly parse rfc 3164 style messages containing a null (hyphen) value. e.g.,
<10>Apr 1 13:14:04 ubuntu-11.cloudera.com - rest_of_message

It tries to parse it as a 5424 style message, skips over the date information, and interprets the first hyphen as a null timestamp. Part of the problem is the use of a Scanner and regex. This skips over a properly formatted 3164 style message until it finds anything that matches the 5424 regex, including a hyphen.
This addresses bug FLUME-1227.
    https://issues.apache.org/jira/browse/FLUME-1227
Diffs
-----

  flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java 4cded11
  flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java acfb29d

Diff: https://reviews.apache.org/r/7780/diff/
Testing
-------

Unit test added and units pass
Thanks,

Brock Noland

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7780/
-----------------------------------------------------------

(Updated Oct. 30, 2012, 7:10 p.m.)
Review request for Flume.
Changes
-------

Updated JIRA #
Description
-------

The SyslogUtils class doesn't properly parse rfc 3164 style messages containing a null (hyphen) value. e.g.,
<10>Apr 1 13:14:04 ubuntu-11.cloudera.com - rest_of_message

It tries to parse it as a 5424 style message, skips over the date information, and interprets the first hyphen as a null timestamp. Part of the problem is the use of a Scanner and regex. This skips over a properly formatted 3164 style message until it finds anything that matches the 5424 regex, including a hyphen.
This addresses bug FLUME-1277.
    https://issues.apache.org/jira/browse/FLUME-1277
Diffs
-----

  flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java 4cded11
  flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java acfb29d

Diff: https://reviews.apache.org/r/7780/diff/
Testing
-------

Unit test added and units pass
Thanks,

Brock Noland

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7780/#review13132
-----------------------------------------------------------

flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java
<https://reviews.apache.org/r/7780/#comment28291>

    Can we change cloudera.com into something else?
- Alexander Alten-Lorenz
On Oct. 30, 2012, 7:10 p.m., Brock Noland wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/7780/
> -----------------------------------------------------------
>
> (Updated Oct. 30, 2012, 7:10 p.m.)
>
>
> Review request for Flume.
>
>
> Description
> -------
>
> The SyslogUtils class doesn't properly parse rfc 3164 style messages containing a null (hyphen) value. e.g.,
> <10>Apr 1 13:14:04 ubuntu-11.cloudera.com - rest_of_message
>
> It tries to parse it as a 5424 style message, skips over the date information, and interprets the first hyphen as a null timestamp. Part of the problem is the use of a Scanner and regex. This skips over a properly formatted 3164 style message until it finds anything that matches the 5424 regex, including a hyphen.
>
>
> This addresses bug FLUME-1277.
>     https://issues.apache.org/jira/browse/FLUME-1277
>
>
> Diffs
> -----
>
>   flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java 4cded11
>   flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java acfb29d
>
> Diff: https://reviews.apache.org/r/7780/diff/
>
>
> Testing
> -------
>
> Unit test added and units pass
>
>
> Thanks,
>
> Brock Noland
>
>

> On Nov. 6, 2012, 8:42 a.m., Alexander Alten-Lorenz wrote:
> > flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java, line 162
> > <https://reviews.apache.org/r/7780/diff/1/?file=182870#file182870line162>
> >
> >     Can we change cloudera.com into something else?

Doesn't really matter does it? Just a place holder domain.
- Brock
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7780/#review13132
-----------------------------------------------------------
On Oct. 30, 2012, 7:10 p.m., Brock Noland wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/7780/
> -----------------------------------------------------------
>
> (Updated Oct. 30, 2012, 7:10 p.m.)
>
>
> Review request for Flume.
>
>
> Description
> -------
>
> The SyslogUtils class doesn't properly parse rfc 3164 style messages containing a null (hyphen) value. e.g.,
> <10>Apr 1 13:14:04 ubuntu-11.cloudera.com - rest_of_message
>
> It tries to parse it as a 5424 style message, skips over the date information, and interprets the first hyphen as a null timestamp. Part of the problem is the use of a Scanner and regex. This skips over a properly formatted 3164 style message until it finds anything that matches the 5424 regex, including a hyphen.
>
>
> This addresses bug FLUME-1277.
>     https://issues.apache.org/jira/browse/FLUME-1277
>
>
> Diffs
> -----
>
>   flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java 4cded11
>   flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java acfb29d
>
> Diff: https://reviews.apache.org/r/7780/diff/
>
>
> Testing
> -------
>
> Unit test added and units pass
>
>
> Thanks,
>
> Brock Noland
>
>

> On Nov. 6, 2012, 8:42 a.m., Alexander Alten-Lorenz wrote:
> > flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java, line 162
> > <https://reviews.apache.org/r/7780/diff/1/?file=182870#file182870line162>
> >
> >     Can we change cloudera.com into something else?
>
> Brock Noland wrote:
>     Doesn't really matter does it? Just a place holder domain.

not really, but some kind of example.com|apache.org would look better, I think. Just my impression. Ship it
- Alexander
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7780/#review13132
-----------------------------------------------------------
On Oct. 30, 2012, 7:10 p.m., Brock Noland wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/7780/
> -----------------------------------------------------------
>
> (Updated Oct. 30, 2012, 7:10 p.m.)
>
>
> Review request for Flume.
>
>
> Description
> -------
>
> The SyslogUtils class doesn't properly parse rfc 3164 style messages containing a null (hyphen) value. e.g.,
> <10>Apr 1 13:14:04 ubuntu-11.cloudera.com - rest_of_message
>
> It tries to parse it as a 5424 style message, skips over the date information, and interprets the first hyphen as a null timestamp. Part of the problem is the use of a Scanner and regex. This skips over a properly formatted 3164 style message until it finds anything that matches the 5424 regex, including a hyphen.
>
>
> This addresses bug FLUME-1277.
>     https://issues.apache.org/jira/browse/FLUME-1277
>
>
> Diffs
> -----
>
>   flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java 4cded11
>   flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java acfb29d
>
> Diff: https://reviews.apache.org/r/7780/diff/
>
>
> Testing
> -------
>
> Unit test added and units pass
>
>
> Thanks,
>
> Brock Noland
>
>

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7780/#review13150
-----------------------------------------------------------

Ship it!
Ship It!

- Alexander Alten-Lorenz
On Oct. 30, 2012, 7:10 p.m., Brock Noland wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/7780/
> -----------------------------------------------------------
>
> (Updated Oct. 30, 2012, 7:10 p.m.)
>
>
> Review request for Flume.
>
>
> Description
> -------
>
> The SyslogUtils class doesn't properly parse rfc 3164 style messages containing a null (hyphen) value. e.g.,
> <10>Apr 1 13:14:04 ubuntu-11.cloudera.com - rest_of_message
>
> It tries to parse it as a 5424 style message, skips over the date information, and interprets the first hyphen as a null timestamp. Part of the problem is the use of a Scanner and regex. This skips over a properly formatted 3164 style message until it finds anything that matches the 5424 regex, including a hyphen.
>
>
> This addresses bug FLUME-1277.
>     https://issues.apache.org/jira/browse/FLUME-1277
>
>
> Diffs
> -----
>
>   flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java 4cded11
>   flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java acfb29d
>
> Diff: https://reviews.apache.org/r/7780/diff/
>
>
> Testing
> -------
>
> Unit test added and units pass
>
>
> Thanks,
>
> Brock Noland
>
>

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7780/#review13315
-----------------------------------------------------------

flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java
<https://reviews.apache.org/r/7780/#comment28581>

    We should avoid calling baos.toString() inside the loop over the configured formats. Better to call it once at the top since this will likely have a performance impact.
- Mike Percy
On Oct. 30, 2012, 7:10 p.m., Brock Noland wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/7780/
> -----------------------------------------------------------
>
> (Updated Oct. 30, 2012, 7:10 p.m.)
>
>
> Review request for Flume.
>
>
> Description
> -------
>
> The SyslogUtils class doesn't properly parse rfc 3164 style messages containing a null (hyphen) value. e.g.,
> <10>Apr 1 13:14:04 ubuntu-11.cloudera.com - rest_of_message
>
> It tries to parse it as a 5424 style message, skips over the date information, and interprets the first hyphen as a null timestamp. Part of the problem is the use of a Scanner and regex. This skips over a properly formatted 3164 style message until it finds anything that matches the 5424 regex, including a hyphen.
>
>
> This addresses bug FLUME-1277.
>     https://issues.apache.org/jira/browse/FLUME-1277
>
>
> Diffs
> -----
>
>   flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java 4cded11
>   flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java acfb29d
>
> Diff: https://reviews.apache.org/r/7780/diff/
>
>
> Testing
> -------
>
> Unit test added and units pass
>
>
> Thanks,
>
> Brock Noland
>
>

> On Nov. 10, 2012, 2:05 a.m., Mike Percy wrote:
> > flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java, line 247
> > <https://reviews.apache.org/r/7780/diff/1/?file=182869#file182869line247>
> >
> >     We should avoid calling baos.toString() inside the loop over the configured formats. Better to call it once at the top since this will likely have a performance impact.

(Otherwise, this patch looks fine to me)
- Mike
-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7780/#review13315
-----------------------------------------------------------
On Oct. 30, 2012, 7:10 p.m., Brock Noland wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/7780/
> -----------------------------------------------------------
>
> (Updated Oct. 30, 2012, 7:10 p.m.)
>
>
> Review request for Flume.
>
>
> Description
> -------
>
> The SyslogUtils class doesn't properly parse rfc 3164 style messages containing a null (hyphen) value. e.g.,
> <10>Apr 1 13:14:04 ubuntu-11.cloudera.com - rest_of_message
>
> It tries to parse it as a 5424 style message, skips over the date information, and interprets the first hyphen as a null timestamp. Part of the problem is the use of a Scanner and regex. This skips over a properly formatted 3164 style message until it finds anything that matches the 5424 regex, including a hyphen.
>
>
> This addresses bug FLUME-1277.
>     https://issues.apache.org/jira/browse/FLUME-1277
>
>
> Diffs
> -----
>
>   flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java 4cded11
>   flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java acfb29d
>
> Diff: https://reviews.apache.org/r/7780/diff/
>
>
> Testing
> -------
>
> Unit test added and units pass
>
>
> Thanks,
>
> Brock Noland
>
>

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7780/
-----------------------------------------------------------

(Updated Nov. 10, 2012, 4:24 p.m.)
Review request for Flume.
Changes
-------

Updated patch based on review
Description
-------

The SyslogUtils class doesn't properly parse rfc 3164 style messages containing a null (hyphen) value. e.g.,
<10>Apr 1 13:14:04 ubuntu-11.cloudera.com - rest_of_message

It tries to parse it as a 5424 style message, skips over the date information, and interprets the first hyphen as a null timestamp. Part of the problem is the use of a Scanner and regex. This skips over a properly formatted 3164 style message until it finds anything that matches the 5424 regex, including a hyphen.
This addresses bug FLUME-1277.
    https://issues.apache.org/jira/browse/FLUME-1277
Diffs (updated)
-----

  flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java 4cded11
  flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java acfb29d

Diff: https://reviews.apache.org/r/7780/diff/
Testing
-------

Unit test added and units pass
Thanks,

Brock Noland

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7780/#review13331
-----------------------------------------------------------

Ship it!
Ship It!

- Mike Percy
On Nov. 10, 2012, 4:24 p.m., Brock Noland wrote:
>
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/7780/
> -----------------------------------------------------------
>
> (Updated Nov. 10, 2012, 4:24 p.m.)
>
>
> Review request for Flume.
>
>
> Description
> -------
>
> The SyslogUtils class doesn't properly parse rfc 3164 style messages containing a null (hyphen) value. e.g.,
> <10>Apr 1 13:14:04 ubuntu-11.cloudera.com - rest_of_message
>
> It tries to parse it as a 5424 style message, skips over the date information, and interprets the first hyphen as a null timestamp. Part of the problem is the use of a Scanner and regex. This skips over a properly formatted 3164 style message until it finds anything that matches the 5424 regex, including a hyphen.
>
>
> This addresses bug FLUME-1277.
>     https://issues.apache.org/jira/browse/FLUME-1277
>
>
> Diffs
> -----
>
>   flume-ng-core/src/main/java/org/apache/flume/source/SyslogUtils.java 4cded11
>   flume-ng-core/src/test/java/org/apache/flume/source/TestSyslogUtils.java acfb29d
>
> Diff: https://reviews.apache.org/r/7780/diff/
>
>
> Testing
> -------
>
> Unit test added and units pass
>
>
> Thanks,
>
> Brock Noland
>
>