On Tue, 23 Sep 2014 18:46:12 +0200, Thomas Neidhart wrote:
On 09/23/2014 03:27 PM, Gilles wrote:
On Tue, 23 Sep 2014 14:58:56 +0200, luc wrote:
Hi Gilles,
Le 2014-09-23 12:51, Gilles a écrit :
Hi Luc.
The commit contains many more changes that just the added
annotation:
removed imports, removed variables, type changes, deleted code...
Most changes are probably harmless but it is not obvious for all.
Could it be that you unintentionally mixed too many commits?
Yes, I'm sorry about this.
I was fixing a lot of warnings (more than 2600!) that appeared
after
some updates
(Eclipse update to Eclipse Luna and probably also Java 8 being
installed when
I updated my Debian Jessie distribution).
So the overall goal is still to simply remove warnings, most of
which
are related to
deprecated classes, but not all of them.
Most of the changes were done in the test part. The few changes in
the library part concern
ValueServer, which delegated some code to deprecated
RandomDataImpl,
which itself delegated to
RandomDataGenerator, so I have streamlined this ti directly use
RandomDataGenerator and keep RandomDataImpl
only in a cosntructor that was already deprecated.
I'll try to refrain to mix things again. The huge commit for the
@Deprecated test classes was worth doing
in a single pass, but the ValueServer changes should have been in a
separate commit.
It seems that two private methods were deleted from
org.apache.commons.math3.ml.neuralnet.sofm.KohonenTrainingTaskTest
They are unused but could come in handy for debugging.[1]
Unless I miss something, that change at least should be reverted.
Best regards,
Gilles
[1] The more so that this test sometimes fails randomly.
can we disable the test when we know that it is failing randomly?
I don't think that it is a good idea to sweep things under the rug.
Actually, I've seen the failure reported in the mails sent from
continuous
build software. As it does not happen on my machine, it's fairly
difficult
to investigate...
I guess it could be pretty annoying if this happens to somebody
during a
release vote.
If there is a problem, better late than never (IMHO)...
Gilles
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org
