You are viewing a plain text version of this content. The canonical link for it is here.
Posted to log4j-dev@logging.apache.org by mw...@apache.org on 2006/01/20 05:18:51 UTC
svn commit: r370726 - in /logging/log4j/trunk/tests/witness/watchdog:
watchdog.FileWatchdog.test1.txt watchdog.FileWatchdog.test2.txt
Author: mwomack
Date: Thu Jan 19 20:18:47 2006
New Revision: 370726
URL: http://svn.apache.org/viewcvs?rev=370726&view=rev
Log:
Restoring file watchdog witness files
Modified:
logging/log4j/trunk/tests/witness/watchdog/watchdog.FileWatchdog.test1.txt
logging/log4j/trunk/tests/witness/watchdog/watchdog.FileWatchdog.test2.txt
Modified: logging/log4j/trunk/tests/witness/watchdog/watchdog.FileWatchdog.test1.txt
URL: http://svn.apache.org/viewcvs/logging/log4j/trunk/tests/witness/watchdog/watchdog.FileWatchdog.test1.txt?rev=370726&r1=370725&r2=370726&view=diff
==============================================================================
--- logging/log4j/trunk/tests/witness/watchdog/watchdog.FileWatchdog.test1.txt (original)
+++ logging/log4j/trunk/tests/witness/watchdog/watchdog.FileWatchdog.test1.txt Thu Jan 19 20:18:47 2006
@@ -3,7 +3,6 @@
WARN - warn message
ERROR - error message
FATAL - fatal message
-DEBUG - debug message
INFO - info message
WARN - warn message
ERROR - error message
Modified: logging/log4j/trunk/tests/witness/watchdog/watchdog.FileWatchdog.test2.txt
URL: http://svn.apache.org/viewcvs/logging/log4j/trunk/tests/witness/watchdog/watchdog.FileWatchdog.test2.txt?rev=370726&r1=370725&r2=370726&view=diff
==============================================================================
--- logging/log4j/trunk/tests/witness/watchdog/watchdog.FileWatchdog.test2.txt (original)
+++ logging/log4j/trunk/tests/witness/watchdog/watchdog.FileWatchdog.test2.txt Thu Jan 19 20:18:47 2006
@@ -3,7 +3,6 @@
WARN - warn message
ERROR - error message
FATAL - fatal message
-DEBUG - debug message
INFO - info message
WARN - warn message
ERROR - error message
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: svn commit: r370726 - in /logging/log4j/trunk/tests/witness/watchdog: watchdog.FileWatchdog.test1.txt watchdog.FileWatchdog.test2.txt
Posted by Mark Womack <mw...@apache.org>.
That could be. I did tighten up the times. I'll change the times and add
the test back in.
-Mark
----- Original Message -----
From: "Curt Arnold" <ca...@apache.org>
To: "Log4J Developers List" <lo...@logging.apache.org>
Sent: Thursday, January 19, 2006 8:46 PM
Subject: Re: svn commit: r370726 - in
/logging/log4j/trunk/tests/witness/watchdog: watchdog.FileWatchdog.test1.txt
watchdog.FileWatchdog.test2.txt
>I don't think this fixes the problem since I was seeing that sometimes the
>generated files contained DEBUG and other times it didn't. My Mac build
>will now fails with the new witness files since the generated files
>contain the DEBUG line.
>
> The previous Gump failures show inconsistent generated files. In
> http://marc.theaimsgroup.com/?l=log4j-dev&m=113766882315669&w=2, DEBUG is
> missing in the generated files but present in the witness. In
> http://marc.theaimsgroup.com/?l=log4j-dev&m=113761865108336&w=2, DEBUG is
> present in the generated files but missing in the witness.
>
> My supposition is there is some delay in these tests that is sometime
> sufficient for a watchdog action to be completed, but sometime isn't.
>
> On Jan 19, 2006, at 10:18 PM, mwomack@apache.org wrote:
>
>> Author: mwomack
>> Date: Thu Jan 19 20:18:47 2006
>> New Revision: 370726
>>
>> URL: http://svn.apache.org/viewcvs?rev=370726&view=rev
>> Log:
>> Restoring file watchdog witness files
>>
>> Modified:
>> logging/log4j/trunk/tests/witness/watchdog/
>> watchdog.FileWatchdog.test1.txt
>> logging/log4j/trunk/tests/witness/watchdog/
>> watchdog.FileWatchdog.test2.txt
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
> For additional commands, e-mail: log4j-dev-help@logging.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: svn commit: r370726 - in /logging/log4j/trunk/tests/witness/watchdog: watchdog.FileWatchdog.test1.txt watchdog.FileWatchdog.test2.txt
Posted by Mark Womack <mw...@apache.org>.
Actually, looking at it closer than I did last night, the actual
recommendations are still good. I was confusing it with the recent
change in the relationship between Priority/Level. But the actual
suggestions are good. Sorry.
I still think this page or a page like it should be part of the actual
documentation, not just an external link. But then, documentation in
general needs work, as noted before.
(glad to see you are still out there, Ceki!)
-Mark
On 1/25/06, Ceki Gülcü <ce...@qos.ch> wrote:
> At 05:46 AM 1/25/2006, Mark Womack wrote:
>
> >>Do we want to add a log4j-all-${version}.jar that would contain all the
> >>class files which could be used as a replacement for log4j-1.2.x.jar?
> >
> >+1
> >
> >We also need to update the HISTORY.txt. Looking at the state of the
> >documentation for 1.3, there are links to Ceki's "Preparing for 1.3"
> >document on his website. I don't think it is as valid as it used to be at
> >this point. We need to create our own that is part of our docs. I'm not
> >suggesting we delay the build, but we need to address it soon.
>
> Just curious, what part of that document is no longer valid?
>
> >-Mark
>
> --
> Ceki Gülcü
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
> For additional commands, e-mail: log4j-dev-help@logging.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: svn commit: r370726 - in
/logging/log4j/trunk/tests/witness/watchdog:
watchdog.FileWatchdog.test1.txt watchdog.FileWatchdog.test2.txt
Posted by Ceki Gülcü <ce...@qos.ch>.
At 05:46 AM 1/25/2006, Mark Womack wrote:
>>Do we want to add a log4j-all-${version}.jar that would contain all the
>>class files which could be used as a replacement for log4j-1.2.x.jar?
>
>+1
>
>We also need to update the HISTORY.txt. Looking at the state of the
>documentation for 1.3, there are links to Ceki's "Preparing for 1.3"
>document on his website. I don't think it is as valid as it used to be at
>this point. We need to create our own that is part of our docs. I'm not
>suggesting we delay the build, but we need to address it soon.
Just curious, what part of that document is no longer valid?
>-Mark
--
Ceki Gülcü
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: svn commit: r370726 - in /logging/log4j/trunk/tests/witness/watchdog: watchdog.FileWatchdog.test1.txt watchdog.FileWatchdog.test2.txt
Posted by Mark Womack <mw...@apache.org>.
> On Jan 24, 2006, at 10:02 PM, Mark Womack wrote:
>
>> Curt, before I re-enable the FileWatchdog test case, can you try it on
>> your machine? I cleaned it up some, but looking at the code I'm not
>> sure how it was still printing the debug message. It explicitly checks
>> the level on the logger is INFO before printing any messages. But try
>> it now.
>
> Test had been passing with my earlier modifications, now fails. The test
> did not previously check for the level change, I added that a few days
> ago when I addressed the issue which got both my local machine and Gump
> working. See http://marc.theaimsgroup.com/?
> l=log4j-dev&m=113805291616535&w=2.
>
> I had introduced adding an intentional delay before rewriting the new
> configuration file. The resolution in the last modification time varies
> between platforms and file systems. Without the delay, it is possible
> that the last modification time does not "appear" to change since the
> last modification times are identical. The documentation for
> File.setLastModified() indicates that you can not assume resolution less
> than a second on the filesystem. I think it is more reliable just to
> wait since writing and setting the last modified time could result in the
> new configuration file being processed twice (once on the write and once
> on the change of modification time).
Erg. So, it was working fine and now I have broken it? Guess I'll revert
it back then.
>
>
>> I am still planning to do the next 1.3 build (alpha 8) tomorrow evening.
>> If you have any changes you want to make it in, please have them checked
>> into svn by 7pm US pacific time.
>
>
> Unless I get an objection, I'll put the NTEventLogAppender back in
> tomorrow. The build will be like the MinGW build for the
> NTEventLogAppender.dll in the log4j 1.2 branch. To build the
> NTEventLogAppender.dll, you will need:
>
> a) Have MinGW (http://mingw.sourceforge.net) installed and on the path.
> Must be ahead of Cygwin, use gcc --dumpversion to make sure.
> b) Have the Microsoft Message Compiler from the Microsoft Platform SDK on
> the path (see http://msdn.microsoft.com/platformsdk)
I'll have to get all that set up on my machine.
> Do we want to add a log4j-all-${version}.jar that would contain all the
> class files which could be used as a replacement for log4j-1.2.x.jar?
+1
We also need to update the HISTORY.txt. Looking at the state of the
documentation for 1.3, there are links to Ceki's "Preparing for 1.3"
document on his website. I don't think it is as valid as it used to be at
this point. We need to create our own that is part of our docs. I'm not
suggesting we delay the build, but we need to address it soon.
-Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: svn commit: r370726 - in /logging/log4j/trunk/tests/witness/watchdog: watchdog.FileWatchdog.test1.txt watchdog.FileWatchdog.test2.txt
Posted by Curt Arnold <ca...@apache.org>.
On Jan 24, 2006, at 10:02 PM, Mark Womack wrote:
> Curt, before I re-enable the FileWatchdog test case, can you try it
> on your machine? I cleaned it up some, but looking at the code I'm
> not sure how it was still printing the debug message. It
> explicitly checks the level on the logger is INFO before printing
> any messages. But try it now.
Test had been passing with my earlier modifications, now fails. The
test did not previously check for the level change, I added that a
few days ago when I addressed the issue which got both my local
machine and Gump working. See http://marc.theaimsgroup.com/?
l=log4j-dev&m=113805291616535&w=2.
I had introduced adding an intentional delay before rewriting the new
configuration file. The resolution in the last modification time
varies between platforms and file systems. Without the delay, it is
possible that the last modification time does not "appear" to change
since the last modification times are identical. The documentation
for File.setLastModified() indicates that you can not assume
resolution less than a second on the filesystem. I think it is more
reliable just to wait since writing and setting the last modified
time could result in the new configuration file being processed twice
(once on the write and once on the change of modification time).
> I am still planning to do the next 1.3 build (alpha 8) tomorrow
> evening. If you have any changes you want to make it in, please
> have them checked into svn by 7pm US pacific time.
Unless I get an objection, I'll put the NTEventLogAppender back in
tomorrow. The build will be like the MinGW build for the
NTEventLogAppender.dll in the log4j 1.2 branch. To build the
NTEventLogAppender.dll, you will need:
a) Have MinGW (http://mingw.sourceforge.net) installed and on the
path. Must be ahead of Cygwin, use gcc --dumpversion to make sure.
b) Have the Microsoft Message Compiler from the Microsoft Platform
SDK on the path (see http://msdn.microsoft.com/platformsdk)
Do we want to add a log4j-all-${version}.jar that would contain all
the class files which could be used as a replacement for
log4j-1.2.x.jar?
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: svn commit: r370726 - in /logging/log4j/trunk/tests/witness/watchdog: watchdog.FileWatchdog.test1.txt watchdog.FileWatchdog.test2.txt
Posted by Mark Womack <mw...@apache.org>.
Curt, before I re-enable the FileWatchdog test case, can you try it on your
machine? I cleaned it up some, but looking at the code I'm not sure how it
was still printing the debug message. It explicitly checks the level on the
logger is INFO before printing any messages. But try it now.
thanks,
-Mark
----- Original Message -----
From: "Curt Arnold" <ca...@apache.org>
To: "Log4J Developers List" <lo...@logging.apache.org>
Sent: Thursday, January 19, 2006 8:46 PM
Subject: Re: svn commit: r370726 - in
/logging/log4j/trunk/tests/witness/watchdog: watchdog.FileWatchdog.test1.txt
watchdog.FileWatchdog.test2.txt
>I don't think this fixes the problem since I was seeing that sometimes the
>generated files contained DEBUG and other times it didn't. My Mac build
>will now fails with the new witness files since the generated files
>contain the DEBUG line.
>
> The previous Gump failures show inconsistent generated files. In
> http://marc.theaimsgroup.com/?l=log4j-dev&m=113766882315669&w=2, DEBUG is
> missing in the generated files but present in the witness. In
> http://marc.theaimsgroup.com/?l=log4j-dev&m=113761865108336&w=2, DEBUG is
> present in the generated files but missing in the witness.
>
> My supposition is there is some delay in these tests that is sometime
> sufficient for a watchdog action to be completed, but sometime isn't.
>
> On Jan 19, 2006, at 10:18 PM, mwomack@apache.org wrote:
>
>> Author: mwomack
>> Date: Thu Jan 19 20:18:47 2006
>> New Revision: 370726
>>
>> URL: http://svn.apache.org/viewcvs?rev=370726&view=rev
>> Log:
>> Restoring file watchdog witness files
>>
>> Modified:
>> logging/log4j/trunk/tests/witness/watchdog/
>> watchdog.FileWatchdog.test1.txt
>> logging/log4j/trunk/tests/witness/watchdog/
>> watchdog.FileWatchdog.test2.txt
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
> For additional commands, e-mail: log4j-dev-help@logging.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: svn commit: r370726 - in /logging/log4j/trunk/tests/witness/watchdog: watchdog.FileWatchdog.test1.txt watchdog.FileWatchdog.test2.txt
Posted by Curt Arnold <ca...@apache.org>.
I don't think this fixes the problem since I was seeing that
sometimes the generated files contained DEBUG and other times it
didn't. My Mac build will now fails with the new witness files since
the generated files contain the DEBUG line.
The previous Gump failures show inconsistent generated files. In
http://marc.theaimsgroup.com/?l=log4j-dev&m=113766882315669&w=2,
DEBUG is missing in the generated files but present in the witness.
In http://marc.theaimsgroup.com/?l=log4j-dev&m=113761865108336&w=2,
DEBUG is present in the generated files but missing in the witness.
My supposition is there is some delay in these tests that is sometime
sufficient for a watchdog action to be completed, but sometime isn't.
On Jan 19, 2006, at 10:18 PM, mwomack@apache.org wrote:
> Author: mwomack
> Date: Thu Jan 19 20:18:47 2006
> New Revision: 370726
>
> URL: http://svn.apache.org/viewcvs?rev=370726&view=rev
> Log:
> Restoring file watchdog witness files
>
> Modified:
> logging/log4j/trunk/tests/witness/watchdog/
> watchdog.FileWatchdog.test1.txt
> logging/log4j/trunk/tests/witness/watchdog/
> watchdog.FileWatchdog.test2.txt
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org