[slf4j-dev] Release schedule?

[Ceki Gülcü]

Hello Jukka,

The core interface in SLF4J, namely the org.sfl4j.Logger,
org.slf4j.LoggerFactory class will never change in a disruptive
fashion, where a disruptive change is defined as a change adversely
affecting the end-user.

Considering the following three points, I think that the merge of the
MarkingLogger and the Logger interfaces is non-disruptive.

1) The MarkingLogger interface was not used by anyone. Thus, its
dissapperance will affect end-users nor bindings of the SLF4J API,
e.g. NLOG4J or x4juli.

2) Given their signatures, the newly *added* methods to the Logger
interface do not affect existing clients of that interface. However,
implementations of the Logger interface need to implement the new
methods. This has been done for all bindings shipping with SLF4J as
well as NLOG4J. By the way, the change could be applied on 5 different
implementations (nop, simple, jdk14, log4j, nlog4j) took about 60
seconds per binding.

In principle, it should be easy to apply on x4juli as well. However,
good developers such as Boris do not like changes pushed down their
throats and I apologize if the change was not sufficiently
explained. If there is still interest, I can provide an explanation
in a separate message.

3) Since each SLF4J binding contains its own copy of the SLF4J API, it
is impossible for an SLF4J binding to raise incompatibility issues. For
example, the user will be able to replace NLOG4J version 1.2.21 with a
later version of NLOG4J, say 1.2.22, even of 1.2.21 was compiled
against SLF4J 1.0RC5 and 1.2.22 against a later version of SLF4J.

Your concerns regarding stability of the SLF4J are very
understandable.  However, I would like to stress that there have not
been any disruptive changes in the SLF4J API since June 28th of last
year. Moreover, I do not expect any changes (except bug-fixes) in the
SLF4J API between the current SVN head and release 1.0. I plan to have
version 1.0final released sometime during this month.

Subject to Boris' consent, I hope to release 1.0RC6 in the next few
days.

Cheers,

At 11:27 AM 2/2/2006, Jukka Zitting wrote:
 > Hi,
 >
 > Are there any ideas about the expected timeframe for the SLF4J 1.0
 > release? I don't want to push for a premature release, but as Boris
 > Unckel already suspected, I got a little alarmed by the MarkingLogger
 > API change in RC5. With the API freeze and the sequence of release
 > candidates I had implicitly assumed that the official 1.0 release
 > would happen fairly soon, but RC5 broke this assumption. I'm
 > interested in the release schedule because I'd like to use SLF4J in
 > the upcoming 1.0 release of the Apache Jackrabbit project (see
 > http://issues.apache.org/jira/browse/JCR-215). BR, Jukka Zitting
 >
 > -- Yukatan - http://yukatan.fi/ - info at yukatan.fi Software
 > craftmanship, JCR consulting, and Java development


-- 
Ceki Gülcü