[slf4j-dev] slf4j on the google android platform
Darrin Edelman
darrin501 at gmail.com
Tue Aug 24 17:16:02 CEST 2010
Thank you -
Comments below for those who follow...
Anyway - I'm back up and going on the official version - thanks much for your help!!
-Darrin
On Aug 24, 2010, at 8:43 AM, Thorsten Möller wrote:
> On Tuesday, August 24, 2010 3:05 PM [GMT+1=CET],
> Darrin Edelman <darrin501 at gmail.com> wrote (with possible deletions):
>
>> What of the other two points below :
>> 2. Where is this sourcecode? Am I missing this in Git somehow?
> [1], folder slf4j-android
>
Yes - I was able to git this with:
Also - I had to make a change to the pom.xml because for whatever reason my android sdk had a different structure:
--- a/slf4j-android/pom.xml
+++ b/slf4j-android/pom.xml
@@ -83,7 +83,7 @@
<artifactId>android</artifactId>
<version>1.5</version>
<scope>system</scope>
- <systemPath>${env.ANDROID_HOME}/platforms/android-1.5/android.jar</systemPath>
+ <systemPath>${env.ANDROID_HOME}/platforms/android-3/android.jar</systemPath>
</dependency>
</dependencies>
FYI - I believe you can now use a publicly available android jar instead of including the jar within git using something like this:
<dependency>
<groupId>com.google.android</groupId>
<artifactId>android</artifactId>
<version>1.5_r4</version>
<scope>system</scope>
</dependency>
Note that this would require a change to the imports from 'android.utils.log' to 'com.google.android.utils.log' which is how it should be...
>
>> 3. Are there plans to get slf4j-android into a maven repo?
> Not of high priority. However, since the project is already managed
> using Maven, it shouldn't be that complicated to check out the code and
> build and install it in your local repo. Note that I also refuse to
> upload it unless #173 is fixed. Together with this we intend to upgrade
> the code basis to SLF4J 1.6.1 as well.
>
True - but the main issue here is that I don't want members of my team to have to go through this process (understand and download slf4j-android, build, install) in order to build our application - also it makes it a bit more painful for continuous integration. Making it available to the community means that it'll be that much easier to use slf4j as the logger of choice for an android app (for those who are using maven with android anyway).
That said - what I'm doing as a workaround to make builds work for my team and our CI server until this happens is what you did above for the android.jar (which also wasn't originally available). I'm putting it the jar into source code control under my project's lib directory and added it as a 'system' maven dependency:
<dependency>
<groupId>org.slf4j</groupId>
<artifactId>slf4j-android</artifactId>
<version>1.5.8</version>
<scope>system</scope>
<systemPath>${basedir}/lib/slf4j-android-1.5.8.jar</systemPath>
</dependency>
When you have it publicly available I can remove this from my repo and reference the officially available jar.
> @Ceki: What would be required in order to publish the artifacts to Maven
> central - I never did this.
>
> Cheers,
> Thorsten
>
> [1] http://github.com/twwwt/slf4j
>
>
>
>> Thanks,
>> -Darrin
>>
>> I didn't see this in the android code base - have I missed something?
>> What On Aug 24, 2010, at 4:52 AM, Thorsten Möller wrote:
>>
>>> On Monday, August 23, 2010 10:51 PM [GMT+1=CET],
>>> Darrin Edelman <darrin501 at gmail.com> wrote (with possible deletions):
>>>
>>>> I did get slf4j-android working after going through the list here by
>>>> grabbing the source code and patching it myself but I'm wondering if
>>>> progress has been made on the issues below so that I can adopt a
>>>> more official release?
>>> The repository [1] contains a fix which simply truncates names longer
>>> than 23 characters at the right-hand side. This fix is considered a
>>> temporary solution and it is intended to eventually come up either
>>> with the solution proposed in comment #10 [2], or with the following
>>> solution (inspired by Eclipse IDE which has a similar function):
>>>
>>> tag = a.package.name.longer.than.twentythree.characters.MyClass
>>>
>>> would be shortened to
>>>
>>> tag = a.p*.n*.l*.t*.t*.c*.MyClass
>>>
>>> If tag would still be longer than 23 characters (e.g. long class
>>> name, deep hierarchy, or no dots at all), then one could further
>>> shorten it to
>>>
>>> tag = *TrailOfLongClassName
>>>
>>>
>>> Feedback welcome as well as patches. I suggest to continue the
>>> discussion via Bugzilla directly on the issue.
>>>
>>>
>>> Cheers,
>>> Thorsten
>>>
>>>
>>> [1] http://github.com/twwwt/slf4j
>>> [2] http://bugzilla.slf4j.org/show_bug.cgi?id=173#c10
>>>
>>>
>>>
>>>
>>>
>>>
>>>> Thanks in advance,
>>>> -Darrin
>>>>
>>>>>>
>>>>>> I've run into a couple frustrating issues with slf4j-android
>>>>>> that you're no doubt aware of as it seems that some work has
>>>>>> occurred
>>>>>> here:
>>>>>>
>>>>>> 1) http://bugzilla.slf4j.org/show_bug.cgi?id=173
>>>>>>
>>>>>> Given that I'm dealing with a cross-platform library with lots of
>>>>>> existing code the '23 character' android limit is a showstopper
>>>>>> for me. I cannot change all the library code to use 'simplename'
>>>>>> for example. Granted I see this as really an issue with Android
>>>>>> - but still we should be able to handle this in the android port.
>>>>>> This means that I cannot just handle it in the android specific
>>>>>> portions of the code - the already written libraries have to
>>>>>> properly deal with this issue by cropping the name of the logger
>>>>>> - so that the library functions don't throw exceptions. This to
>>>>>> me means that I need to change the StaticLoggerBinder to
>>>>>> instantiate a version of code that knows how to deal with this
>>>>>> issue properly even when the client code isn't aware of it.
>>>>>> Wendel's solution in the above thread
>>>>>> seems a reasonable solution.
>>>>>>
>>>>>> 2) Availability of android source-code from repo.
>>>>>>
>>>>>> Given that this is an for me issue I'd like to put the fix in that
>>>>>> wendell suggest... use as much of the useful name as possible and
>>>>>> hopefully
>>>>>> Android comes around and fixes this limitation someday. I
>>>>>> grabbed the codebase from git and I couldn't find references to
>>>>>> the android branch. Specifically it seems like wendell's fix is
>>>>>> a decent work-around for the android limitation. Unfortunately I
>>>>>> cannot easily do this and contribute any branches back to the
>>>>>> open source community because I cannot seem to find android code
>>>>>> in this branch. Now
>>>>>> this said - I did find the slf4j-android code base on the website
>>>>>> so I can move forward... but why isn't this in Git??
>>>>>>
>>>>>> 3) Really just a picky thing... slf4j-android does not appear to
>>>>>> be available from any publicly accessible maven repository.
>>>>>>
>>>>>> I've resolved this by adding the jar to my project and
>>>>>> automatically deploying it to the local repo - but this really
>>>>>> isn't a very elegant solution. Any chance the android version is
>>>>>> going to be made available anytime soon? I'd rather not package
>>>>>> the jar as a build
>>>>>> artifact.
>>>>
>>>>>
>>>>
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> slf4j-dev mailing list
>>>> slf4j-dev at qos.ch
>>>> http://qos.ch/mailman/listinfo/slf4j-dev
>>>
>>> _______________________________________________
>>> slf4j-dev mailing list
>>> slf4j-dev at qos.ch
>>> http://qos.ch/mailman/listinfo/slf4j-dev
>>
>> _______________________________________________
>> slf4j-dev mailing list
>> slf4j-dev at qos.ch
>> http://qos.ch/mailman/listinfo/slf4j-dev
>
> _______________________________________________
> slf4j-dev mailing list
> slf4j-dev at qos.ch
> http://qos.ch/mailman/listinfo/slf4j-dev
More information about the slf4j-dev
mailing list