directory-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Emmanuel Lécharny <>
Subject Re: Ldap API 2.0 roadmap
Date Thu, 03 Aug 2017 13:26:47 GMT

Le 03/08/2017 à 15:05, Shawn McKinney a écrit :
>> On Aug 3, 2017, at 7:03 AM, Emmanuel Lécharny <> wrote:
>> Beside this effort, I started to work on a branch
>> (
>> which was a refactoring of the Value class, in order to simplify what we
>> had in 1.0. The rational was to get some major errors being fixed in
>> ApacheDS (mainly related to some special chars being mis-handled in
>> DNs). The consequences are huge in term of performances (20% faster),
>> but impacts the projects using this API.
>> At this point, I'd like to suggest we start with a 2.0 because of those
>> API changes. FTR, I ave carrefully ported all the changes made in 1.0 to
>> the branch, and I also have a branch for Apacheds which relies on the
>> API branch. What remains to be done is to switch to this branch for Studio.
>> So let's thing bigger : If we go for a 2.0, I also suggest we move to
>> Java 8 only for this version (I mean, Java 8 and higher). ApacheDS will
>> also switch to Java 8 and will use this API 2.0 in M25, and teh next
>> Studio release should also use the API 2.0 and ApacheDS with API 2.0.
>> I would also suggest we switch to git for the API, now that 1.0 is out.
>> SVN is outdated, and it's quite an anchor for us anyway (I have to use
>> svn *and* git daily, it makes things more complex...). Nor sure we
>> should'nt move to git for all teh projects, but startng wih teh API
>> sounds reasonable atm. In any case, I'll write another mail for this change.
>> I'd like to have your opinion about those proposed changes, before
>> starting an official vote.
> Seems a little strange to have a 2.0 this close to a 1.0 release but in the end the version
doesn’t matter all that much.  The main thing is we keep forward progress so…

Well, I think it conveys a message : it's alive ! Once upon a time, OSS
was very conservative, and projects were moving from X to X+1 with
extreme caution (httpd, etc). Then Chrome changed the game with its fast
incremental version releases, and Fireforx followd, and now, many other
projects do the same.

I do think it's easier for people to follow up when the major is
changing rather than going on with somethig that remains teh same for

But that may be just me.

Otherwise, considering teh API changes, we can't make it a minor (ie
1.1), so 2.0 is quite necesary.

Btw, 1.0 will be maintained, so we are likey to have a 1.0.1 and some
other versions later o (with Java 7 support, of course).

Emmanuel Lecharny

View raw message