On Sun, Dec 16, 2012 at 04:51:25PM +0100, Thomas Neidhart wrote:
> On 12/16/2012 04:39 PM, Gilles Sadowski wrote:
> > On Sun, Dec 16, 2012 at 04:09:54PM +0100, Thomas Neidhart wrote:
> >> On 12/16/2012 04:03 PM, Gilles Sadowski wrote:
> >>> On Sun, Dec 16, 2012 at 02:34:51PM +0100, Thomas Neidhart wrote:
> >>>> On 12/16/2012 01:42 PM, Gilles Sadowski wrote:
> >>>>> Hi.
> >>>>>
> >>>>> Please have a look at the Commons Math 3.1 release candidate.
> >>>>>
> >>>>> Note about the web site: I've now found that the problem reported in
> >>>>> https://issues.apache.org/jira/browse/MATH-912
> >>>>> originates from not updating
> >>>>> src/site/site.xml
> >>>>> For this RC, I manually changed the "index.html" in the staging area (so
> >>>>> that the site will have the link to the new apidocs named correctly).
> >>>>> As this does not touch the artefacts, I hope that this small glitch[1]
> >>>>> will
> >>>>> not require a new RC.[2]
> >>>>>
> >>>>> ----------
> >>>>> Tag:
> >>>>>
> >>>>> https://svn.apache.org/repos/asf/commons/proper/math/tags/MATH_3_1_RC1/
> >>>>>
> >>>>> Site:
> >>>>> http://people.apache.org/builds/commons/math/3.1/RC1/
> >>>>>
> >>>>> Binaries:
> >>>>>
> >>>>> https://repository.apache.org/content/repositories/orgapachecommons-023/org/apache/commons/commons-math3/3.1/
> >>>>>
> >>>>> [ ] +1 Release it.
> >>>>> [ ] +0 Go ahead; I don't care.
> >>>>> [ ] -0 There are a few minor glitches: ...
> >>>>> [x] -1 No, do not release it because ...
> >>>>
> >>>> there are several clirr errors reported in:
> >>>>
> >>>> * DifferentiableUnivariateFunction
> >>>> * LogNormalDistribution
> >>>> * NormalDistribution
> >>>>
> >>>> which are not clear to me, and are also not mentioned in the release
> >>>> notes.
> >>>
> >>> I agree that in the first instances, no change should have been done (the
> >>> more so that they are modifications of deprecated methods).
> >>>
> >>> The last two are false positives, for some unknown reason.
> >>>
> >>>> Some findbugs warnings have been introduced due to the copy of classes
> >>>> to the new optim package, the exclude list should be adapted.
> >>>
> >>> Done.
> >>>
> >>>>
> >>>> the changes.xml does not specify a release date for 3.1
> >>>
> >>> Done. (Look at the date! ;-)
> >>>
> >>>> the tagged pom.xml does not specify the correct scm connection (to the
> >>>> final release tag) - just a minor glitch.
> >>>
> >>> I don't quite understand this one. [IIUC, I don't see why those should be
> >>> adapted to the tag (which is not supposed to be changed).]
> >>> If someone thinks that this is really a release blocker, please give an
> >>> explanation on the reason (and the right way to update it).
> >>
> >> It is just a minor thing, but I found it in one of the release guides. I
> >> think most of the releases have missed it so far. The scm links should
> >> be updated to the final release tag, e.g.:
> >>
> >> <scm>
> >>
> >> <connection>scm:svn:http://svn.apache.org/repos/asf/commons/proper/math/tags/MATH_3_1</connection>
> >>
> >> <developerConnection>scm:svn:https://svn.apache.org/repos/asf/commons/proper/math/tags/MATH_3_1</developerConnection>
> >>
> >> <url>http://svn.apache.org/viewvc/commons/proper/math/tags/MATH_3_1</url>
> >> </scm>
> >>
> >> I do not think it is a blocker, but when you prepare the next RC you
> >> could also adapt these links.
> >
> > It was not in my release recipe (and probably not in "UsingNexus" on which I
> > based my previous, and first, attempt to prepare a release). So I'll wait
> > for a rationale. [IMHO, and IIUC (which is not sure of course), in CM, there
> > is reason not to point people to a tag...]
>
> It is in the UsingNexus guide, section Create the SVN tags (Manual method):
>
> ...
> Edit the SCM entries in the parent POM. Note: use the final tag, without
> any RC suffix
> ...
>
> But again, I do not think it is a blocker in any way, I checked many
> other releases and most of them have the link to the latest trunk there.
I don't understand how that change would be useful in any way.
None of the CM releases has those links pointing to something other than
"trunk".
[The reason I wrote another "release recipe" was that because the
"UsingNexus" file contains so many outdated, unnecessarily complicated,
half-explained or incomplete (thus non-working) stuff...
As I said last time, it is useless for everyone: for the expert (who doesn't
need it) and for the newbie (because it simply does not provide a fail-safe
route to the release).]
Regards,
Gilles
P.S. I've updated my local version of the "release mini-howto".
Any objection to committing it to the CM's trunk?
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org
