commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Luc Maisonobe <>
Subject Re: [Math] Toward 3.0 release: First deliverables
Date Sat, 03 Mar 2012 13:36:35 GMT
Hi Sebb,

Le 03/03/2012 00:41, sebb a écrit :
> On 2 March 2012 00:59, Gilles Sadowski <> wrote:
>> Hi.
>> I managed to complete part of the release process:
>> Tag on SVN:
>> Artefacts on Nexus:
>> I'm still stuck with the "staged" web site (cf. other post); but you can
>> already have a look at the above deliverables. Please let me know of
>> anything that requires correction.
> I think there are quite a few things that need checking before it is
> safe to do a release.
> This is the first release with the new package, so it's a chance to
> fix all the inherited poor APIs.

I think all the important problems have been addressed by now.

> We need to try and apply all the changes that will break binary
> compat. before this release, rather than after.

The API have been stabilized since a few weeks and reviewed. The latest
problems identified seems to me merely implementation details. such
problems can be fixed in a 3.1 version if needed. This would even force
us to really release more often, which is clearly our greatest problem
for [math] now.

> As such, any mutable public or protected variables should ideally be
> made private.

Some of these fields have been designed like this for years (the ODE
package was started in 2002 in another library and imported in [math] in
2006 for example). Of course we should go for better encapsulation,
better design. But we are late. Very late.

> Thread-safety is harder to achieve with mutable variables, so any
> classes with setter methods should be checked to see if the setters
> are really needed or not.

This cannot be adressed before 3.0. [math] is *not* thread safe and does
not intend to be, at least for now.

I strongly agree with Gilles desire to get 3.0 out soon. In fact, I know
two important operational projects that need it really really soon.
Delaying the release would be a very negative sign.

The current state of 3.0 is good. It has been used for months by
numerous users, but now some of them need an official release, they
cannot rely on a development snapshot any longer.

We must release 3.0 now.

best regards,

> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message