[logback-user] 1.0 release date?

Ari Meyer ari.meyer at gmail.com
Mon Sep 6 11:31:23 CEST 2010


Thanks Ralph -- good to see Ceki's response to the same question, albeit 
2 yrs earlier.  Makes me think I should have voiced my justifications 
for a 1.0 label years ago, as doubtless many people are likewise waiting 
for 1.0 to try out logback, but get by with log4j and don't press the 
issue.  Will wait to hear me directly from Ceki on this.  Thanks also 
for pointing out the issues lists.

On this note, here's something I dug up from over 3 yrs ago, including 
the "1.0 release" question: http://www.infoq.com/news/2007/08/logback.  
Again this brings me back to my related question: /Are there some 
must-have features that have yet to be implemented?/

Best,
Ari

Ralph Goers wrote:
> I'm not disagreeing with you that the version number should be 1.0 or 
> greater. I asked the question 2 years ago. 
> See http://marc.info/?t=122124087900007&r=1&w=2 
> <http://marc.info/?t=122124087900007&r=1&w=2>. I imagine the answer is 
> still the same. The bottom line is that since the project is run by a 
> single individual you are going to have to take it or leave it on his 
> terms.
>
> As for Log4j 1.2, Logback actually had some of the same problems.  The 
> way syncronization is done logging can become a serious bottleneck or 
> cause a deadlock. The caller data in Log4j is sort of a hack. Since 
> Log4j supports pre Java 1.4 applications it can't directly use 
> StackTraceElements. In recent versions they are used by detecting the 
> JVM version at runtime and then using reflection to invoke the needed 
> APIs.  Another way of assessing this is to 
> a) look at the Jira issues for Log4j 2.0 - essentially a wish list for 
> the next version 
> - https://issues.apache.org/jira/secure/IssueNavigator.jspa?reset=true&pid=12310790&status=1 
> <https://issues.apache.org/jira/secure/IssueNavigator.jspa?reset=true&pid=12310790&status=1>
> b) look at all the open issues for Log4j 1.2 
> - https://issues.apache.org/bugzilla/buglist.cgi?query_format=specific&order=relevance+desc&bug_status=__open__&product=Log4j&content= 
> <https://issues.apache.org/bugzilla/buglist.cgi?query_format=specific&order=relevance+desc&bug_status=__open__&product=Log4j&content=>
>
> Ralph
>
> On Sep 5, 2010, at 5:27 PM, Ari Meyer wrote:
>
>> Thanks Greg.  Ralph: again, it has nothing to do with anyone's 
>> personal opinion as to what is "production ready".  Literally 
>> THOUSANDS of organizations have established policies in place that do 
>> not permit the use of explicitly alpha/beta or pre-1.0 releases.  
>> Anyone who has worked for a Fortune 500, defense contractor, 
>> government/energy/military, etc. knows what I'm talking about.  New 
>> software components often have to be approved by a review board.  I 
>> can't tell you how much time I've spent pushing such organizations to 
>> accept FOSS, but they typically draw the line at components that 
>> aren't EXPLICITLY production releases (> 1.0 and not betas).  And 
>> (for at least larger components/app servers/etc.) they usually 
>> require indemnification and SLAs.  I do see that apparently Ceki 
>> offers support: http://www.qos.ch/shop/products/professionalSupport.
>>  
>> That all said, I'm interested to hear what are the "severe problems" 
>> with Log4j 1.2.  If that is documented, that might be enough to 
>> persuade some organizations to move to logback pre-1.0.  I've seen 
>> http://logback.qos.ch/reasonsToSwitch.html, but I have to assume if 
>> there were actually /severe/ problems with Log4j that people would 
>> have been obliged to ditch it long ago.  Countless FOSS, closed 
>> source, and internally created projects still depend in some way or 
>> another on Log4j.
>> Bottom line: if the current "production ready" logback release were 
>> simply labeled "1.0", logback would then be considered an acceptable 
>> alternative by thousands of groups that currently cannot do so due to 
>> internal policies.  You can say that's absurd, but that's the way 
>> things work.
>> Ari
>> On Sun, Sep 5, 2010 at 8:29 AM, Grzegorz Borkowski 
>> <grzegorzborkowski at o2.pl <mailto:grzegorzborkowski at o2.pl>> wrote:
>>
>>     Ralph, it's all true. But I haven't said Logback is not
>>     production ready or so. I use it in production environment too.
>>     The only question is: if it is stable and production ready, why
>>     it has number suggesting something opposite? If the common sense
>>     is that versions <1.0 are used for pre-final builds, then why
>>     Logback uses such number? (You can release some library, give it
>>     version number "0.0.0.012.pre-alpha" and say "this is final,
>>     production ready version, the number is of no importance" - but
>>     this would be really strange, right?)
>>     So if version number should not be picked up by accident, then
>>     we're just curious why this strange version 0.9.x number is used
>>     in case of Logback.
>>     Greg
>>
>>
>>     W dniu 2010-09-05 16:52, Ralph Goers pisze:
>>>     I think you are using the wrong criteria. SLF4J and Logback are certainly production ready. I have projects using them in production systems.  The version number is not a great way to determine that. IMO, what is more important is the health of the community and availability of support. In some ways this community is very healthy as there is a fair amount of activity on the mailing lists.  However, with only a single person with commit access you do need to be aware that there is a serious supportability problem with SLF4J and Logback.  However, I am sure Ceki is willing to accept contracts to fix any bugs you might encounter. You can also ask the rest of us for help, but the best we can do  is to fix the bug in a fork somewhere and try to get Ceki to incorporate it.
>>>
>>>     Having said that, the situation with Log4j isn't a lot better at the moment. Log4j 1.2 is ancient and has severe problems and Log4j 2.0 is still in the stages of experimental code.
>>>
>>>     Ralph
>>>
>>>     On Sep 5, 2010, at 12:33 AM, Ari Meyer wrote:
>>>
>>>             
>>>>     Jeff,
>>>>
>>>>     I am not trying to argue with you.  Please re-read what I wrote.  I am simpy saying that many organizations will not allow, except for very special circumstances, software components that are in alpha/beta stages of development (or appear to be so, as is the case with sub-1.0 releases) to be deployed to their servers.  If some mission-critical software has dependencies on pre-1.0 releases, that cannot be avoided.  But I cannot justify to my management the use of a pre-1.0 logback when log4j/SLF4J is still sufficient for most purposes.  This has nothing to do with marketing nor what I personally value.  And yes, of course I know who Ceki is and respect all that he has done, and I have used (and made small contributions to) log4j and SLF4J since their early releases.
>>>>
>>>>     Again, thanks to Ceki and all contributors,
>>>>     Ari
>>>>
>>>>     Jeff Jensen wrote:
>>>>               
>>>>>     I may be right?  Just look at the Logback home page and see a few projects using it, e.g. SpringSource’s dm Server.
>>>>>      
>>>>>     You place too much value in a marketing thing – a release number!
>>>>>      
>>>>>     You are also confusing authors of work – I am a fan and user of SLF4J and Logback.  All credit goes to Ceki and associates.
>>>>>      
>>>>>     For stability concerns, you should review the Logback history with its genesis from Log4j.  You do know Ceki, the Logback founder, is also the founder of Log4j?
>>>>>      
>>>>>     ‘nuff said…
>>>>>      
>>>>>      
>>>>>      
>>>>>     From: logback-user-bounces at qos.ch <mailto:logback-user-bounces at qos.ch> [mailto:logback-user-bounces at qos.ch] On Behalf Of Ari Meyer
>>>>>     Sent: Saturday, September 04, 2010 8:26 PM
>>>>>     To: logback users list
>>>>>     Subject: Re: [logback-user] 1.0 release date?
>>>>>      
>>>>>     Hi Jeff,
>>>>>      
>>>>>     You may be right about that (though I haven't seen it pulled down as a dependency for my Maven builds, yet...), but large organizations often don't see things the way you and I do.  They won't ditch an acceptable, stable log4j for what *appears* to be a beta.  After over 4 years of active development, though, it seems reasonable to expect multiple full releases of something as relatively small in scope as a logging framework.  The fact that we don't see a 1.0 yet perhaps indicates over-perfectionism.  This often happens with other OSS projects (JDOM being a notable case of this, as I remember).  For our sake, please just get to a reasonably stable build and label it "1.0".  We'll expect bug fixes and minor API changes, of course, as that's natural, and new features can be released incrementally.
>>>>>      
>>>>>     Are there some must-have features that have yet to be implemented?  It would be nice if the FAQ were updated for this, along with a high-level road map.
>>>>>      
>>>>>     Thanks for all for the hard work!
>>>>>     Ari
>>>>>     On Sat, Sep 4, 2010 at 8:39 AM, Jeff Jensen <jjensen at apache.org> <mailto:jjensen at apache.org> wrote:
>>>>>     Just because the release number begins with a value < 1 doesn't mean it is
>>>>>     in beta.  Lots and lots of products around the world use SLF4J and Logback
>>>>>     in production operations, and I bet including some of the FOS frameworks you
>>>>>     are using!
>>>>>
>>>>>
>>>>>     -----Original Message-----
>>>>>     From: logback-user-bounces at qos.ch <mailto:logback-user-bounces at qos.ch> [mailto:logback-user-bounces at qos.ch] On
>>>>>     Behalf Of Ari Meyer
>>>>>     Sent: Saturday, September 04, 2010 8:09 AM
>>>>>     To: logback-user at qos.ch <mailto:logback-user at qos.ch>
>>>>>     Subject: [logback-user] 1.0 release date?
>>>>>
>>>>>     Hi,
>>>>>
>>>>>     We'd like to switch from log4j to logback, but can't use beta releases.
>>>>>     logback has been in development for over 4 years now -- any idea when
>>>>>     we'll see a 1.0 release?
>>>>>
>>>>>     Thanks,
>>>>>     Ari
>>>>>     _______________________________________________
>>>>>     Logback-user mailing list
>>>>>     Logback-user at qos.ch <mailto:Logback-user at qos.ch>
>>>>>     http://qos.ch/mailman/listinfo/logback-user
>>>>>
>>>>>     _______________________________________________
>>>>>     Logback-user mailing list
>>>>>     Logback-user at qos.ch <mailto:Logback-user at qos.ch>
>>>>>     http://qos.ch/mailman/listinfo/logback-user
>>>>>      
>>>>>
>>>>>     _______________________________________________
>>>>>     Logback-user mailing list
>>>>>     Logback-user at qos.ch <mailto:Logback-user at qos.ch>
>>>>>     http://qos.ch/mailman/listinfo/logback-user
>>>>>       
>>>>>                 
>>>>     _______________________________________________
>>>>     Logback-user mailing list
>>>>     Logback-user at qos.ch <mailto:Logback-user at qos.ch>
>>>>     http://qos.ch/mailman/listinfo/logback-user
>>>>               
>>>
>>>
>>>     _______________________________________________
>>>     Logback-user mailing list
>>>     Logback-user at qos.ch <mailto:Logback-user at qos.ch>
>>>     http://qos.ch/mailman/listinfo/logback-user
>>>             
>>
>>
>>     _______________________________________________
>>     Logback-user mailing list
>>     Logback-user at qos.ch <mailto:Logback-user at qos.ch>
>>     http://qos.ch/mailman/listinfo/logback-user
>>
>>
>> _______________________________________________
>> Logback-user mailing list
>> Logback-user at qos.ch <mailto:Logback-user at qos.ch>
>> http://qos.ch/mailman/listinfo/logback-user
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Logback-user mailing list
> Logback-user at qos.ch
> http://qos.ch/mailman/listinfo/logback-user
>   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://qos.ch/pipermail/logback-user/attachments/20100906/cf80bda6/attachment-0001.html>


More information about the Logback-user mailing list