Mailing-List: contact user-help@commons.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Commons Users List" <user@commons.apache.org>
Received-SPF: pass (nike.apache.org: domain of nzipsi@gmail.com designates
 74.125.92.148 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:date:message-id:subject:from:to
         :content-type:content-transfer-encoding;
        b=KLM2YBHNpvr1WlSAxLsjsrCde0IWwG83jvehctcDrkYEJJ8jYVHSKFZ/EIDbedmuO9
         Qwm7PmcjrJxMXqKnkuSWfHtIZNNVtp3FxD8U7/oXUJWEYce/kd0oPp7xeOzF7jylTmyl
         fiRKMs+hH75PXsQ98EwmTEXECw7ExR+En3Tuw=
MIME-Version: 1.0
In-Reply-To: <fc59cc3a0909021818t25904059xb8d5f1062cc06f31@mail.gmail.com>
References: <fc59cc3a0909012213q1296894cj84d620d71d2771c5@mail.gmail.com>
	 <4A9EA5D8.5020104@oliver-heger.de>
	 <fc59cc3a0909021818t25904059xb8d5f1062cc06f31@mail.gmail.com>
Date: Thu, 3 Sep 2009 15:02:50 +1200
Message-ID: <fc59cc3a0909022002r529f3910me22169194ae01dc6@mail.gmail.com>
Subject: Re: [CONFIGURATION] Configuration Interpolates when Web Application
	is run directly under Glassfish, but not when debugging under Eclipse
From: Andrew Thorburn <nzipsi@gmail.com>
To: Commons Users List <user@commons.apache.org>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

Yeah, you know what? I was wrong. I had just loaded up a second copy
of my application into Glassfish without realising it. Go me?

- Andrew

On Thu, Sep 3, 2009 at 1:18 PM, Andrew Thorburn<nzipsi@gmail.com> wrote:
> Thanks Oliver.
>
> I have no idea why, but it's suddenly started working. I messed about
> with a few things, including dumping the combined configuration (but
> nothing was interpolated), and removing breakpoints which I had set on
> several different copies of my Configurator class.
>
> Absolutely bizarre. All I can do now is hope that it doesn't recur...
>
> Thanks,
>
> - Andrew Thorburn
>
> On Thu, Sep 3, 2009 at 5:05 AM, Oliver
> Heger<oliver.heger@oliver-heger.de> wrote:
>> Hi Andrew,
>>
>> what you describe seems really complicated. I am not sure whether I full=
y
>> understand the problem, but I hope at least to give you some hints. Comm=
ents
>> inline...
>>
>> Andrew Thorburn schrieb:
>>>
>>> This is something that doesn't make an awful lot of sense to me:
>>>
>>> I've got a web application that uses CommonsConfiguration to configure
>>> the logging (first it loads a configuration file as =C2=A0base, then lo=
ads
>>> a new configuration from the database), and for some reason it will
>>> not interpolate the various parameters when I try and debug it under
>>> Eclipse (using the Glassfish WTP plugin). It just doesn't happen.
>>>
>>> I'm using Java 1.5, Commons Configuration 1.6, Commons Lang 2.2,
>>> Eclipse 3.5 and Glassfish 2.1.
>>>
>>> I know that it works when I run it under Glassfish directly as I can
>>> see the log file being created in the right directory and such, and I
>>> can also see it printing out the parameters with everything
>>> successfully interpolated and expanded.
>>>
>>> I'm not sure exactly why it's not doing anything, but there we go.
>>>
>>> It doesn't even interpolate System parameters, including ones that
>>> I've set my self.
>>>
>>> I can't provide the code which is actually causing the problem, and
>>> I'm not entirely sure how to replicate it. The best I can do is show
>>> code which behaves in a similar fashion:
>>>
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0/**
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 * @param args
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 */
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0public static void main(String[] args)
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0{
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0System.setProper=
ty("ProcessID", "pid");
>>>
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0System.out.print=
ln("java.io.tmpdir [" +
>>> System.getProperty("java.io.tmpdir") + "]");
>>>
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0Configuration c =
=3D new BaseConfiguration();
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0c.setProperty("p=
rocess.default.logfile_directory",
>>> "${java.io.tmpdir}");
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0c.setProperty("l=
ogger.default.filename_prefix", "PREFIX");
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0c.setProperty("l=
ogger.default.filename_suffix", "log");
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0c.setProperty("l=
ogger.default.filename",
>>>
>>> "${logger.default.filename_prefix}${ProcessID}.${logger.default.filenam=
e_suffix}");
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0c.setProperty("l=
ogger.log4j.rootLogger", "ERROR, APP_ROD,
>>> APP_EMAIL");
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0c.setProperty("l=
ogger.log4j.appender.APP_ROD",
>>> "org.apache.log4j.DailyRollingFileAppender");
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0c.setProperty("l=
ogger.log4j.appender.APP_ROD.File",
>>>
>>> "${process.default.logfile_directory}/${ProcessID}/${logger.default.fil=
ename}");
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0c.setProperty("l=
ogger.log4j.appender.APP_ROD.DatePattern",
>>> "'.'yyyyMMdd");
>>>
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0CompositeConfigu=
ration cc =3D new CompositeConfiguration();
>>>
>>>
>>> =C2=A0cc.addConfiguration((AbstractConfiguration)c.subset("logger"));
>>>
>>>
>>> =C2=A0System.out.println(cc.getString("log4j.appender.APP_ROD.File"));
>>>
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0printProperties(=
ConfigurationConverter.getProperties(cc));
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0printProperties(=
ConfigurationConverter.getProperties(c));
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0}
>>>
>>>
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0private static void printProperties(Properti=
es props)
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0{
>>>
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0for(Entry<Object=
, Object> ent : props.entrySet())
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0{
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =
=C2=A0 =C2=A0System.out.println("[" + ent.getKey() + "] [" +
>>> ent.getValue() + "]");
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0}
>>> =C2=A0 =C2=A0 =C2=A0 =C2=A0}
>>>
>>> And the output from the above:
>>>
>>> java.io.tmpdir [/tmp]
>>>
>>> ${process.default.logfile_directory}/${ProcessID}/${logger.default.file=
name}
>>>
>>> [log4j.rootLogger] [ERROR,APP_ROD,APP_EMAIL]
>>> [log4j.appender.APP_ROD.DatePattern] ['.'yyyyMMdd]
>>> [log4j.appender.APP_ROD] [org.apache.log4j.DailyRollingFileAppender]
>>> [log4j.appender.APP_ROD.File]
>>>
>>> [${process.default.logfile_directory}/${ProcessID}/${logger.default.fil=
ename}]
>>> [default.filename_prefix] [PREFIX]
>>> [default.filename_suffix] [log]
>>> [default.filename]
>>>
>>> [${logger.default.filename_prefix}${ProcessID}.${logger.default.filenam=
e_suffix}]
>>>
>>> [process.default.logfile_directory] [${java.io.tmpdir}]
>>> [logger.default.filename_prefix] [PREFIX]
>>> [logger.default.filename_suffix] [log]
>>> [logger.log4j.appender.APP_ROD.File]
>>> [${java.io.tmpdir}/${ProcessID}/PREFIX${ProcessID}.log]
>>> [logger.default.filename] [PREFIX${ProcessID}.log]
>>> [logger.log4j.appender.APP_ROD.DatePattern] ['.'yyyyMMdd]
>>> [logger.log4j.appender.APP_ROD]
>>> [org.apache.log4j.DailyRollingFileAppender]
>>> [logger.log4j.rootLogger] [ERROR,APP_ROD,APP_EMAIL]
>>>
>>> Which seems a little... odd... since it left
>>> process.default.logfile_directory as ${java.io.tmpdir} despite the
>>> property definitely being set.
>>
>> If you work with a plain configuration like BaseConfiguration, system
>> properties are not interpolated out of the box. You need to prefix the
>> variable with the "sys:" prefix. There is an example in the user guide:
>> http://commons.apache.org/configuration/userguide/howto_basicfeatures.ht=
ml#Variable_Interpolation
>>
>> However, when using a CombinedConfiguration that includes a
>> SystemConfiguration the situation is different as the SystemConfiguratio=
n
>> should make all its properties available.
>>
>>>
>>> This differs from my actual problem in that *no* interpolation is
>>> performed when converting to a Properties Object. None at all. Also,
>>> the properties are loaded from an XML file, and then from a
>>> CombinedConfiguration constructed from the XMLFile, System Properties
>>> and a Database Table.
>>>
>>> I *know* that it works normally, as a temporary log winds up /tmp.
>>>
>>> Having said the above, if I stop Glassfish in Eclipse and then start
>>> it from the command line, I get the same issue (of no interpolation).
>>>
>>> There's a variety of issues that could be the problem here. It could
>>> be something funny about the Glassfish plugin. It could be something
>>> to do with how I've got the web project set up - prior to this, it was
>>> built via Ant and deployed directly. It could be something to do with
>>> Eclipse's compiler/classpath/whatever settings.
>>>
>>> However, I really don't have a clue where to begin here - has anyone
>>> seen this sort of problem before and been able to find a solution? Any
>>> hints would be most welcome, as I'm not even certain that this is
>>> actually a commons-configuration problem. It's just that
>>> commons-configuration is the visible symptom.
>>>
>>> Apologies if the above seems disjointed or unclear.
>>
>> I think we can assume that the interpolation features of Commons
>> Configuration work as expected - otherwise we would see lots of bug repo=
rts.
>> So if interpolation does not work in your case, the combined configurati=
on
>> probably does not contain the properties you expect.
>>
>> To debug the problem you can dump the complete content of the
>> CombinedConfiguration. An easy way to achieve this is to create an
>> XMLConfiguration from the CombinedConfiguration and store it somewhere, =
e.g.
>> if cc is the CombinedConfiguration:
>>
>> XMLConfiguration xmlConf =3D new XMLConfiguration(cc);
>> xmlConf.save(new File("dump.xml"));
>>
>> Maybe, depending on the environment you start the application system
>> properties are not propagated as expected?
>>
>> HTH
>> Oliver
>>
>>>
>>> Thanks,
>>>
>>> - Andrew Thorburn
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: user-unsubscribe@commons.apache.org
>>> For additional commands, e-mail: user-help@commons.apache.org
>>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscribe@commons.apache.org
>> For additional commands, e-mail: user-help@commons.apache.org
>>
>>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@commons.apache.org
For additional commands, e-mail: user-help@commons.apache.org