db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Myrna van Lunteren" <m.v.lunte...@gmail.com>
Subject Re: [VOTE] 10.3.1.3 release
Date Wed, 25 Jul 2007 16:47:52 GMT
On 7/24/07, Kathey Marsden <kmarsdenderby@sbcglobal.net> wrote:
> -1.   For three reasons:
>
> 1)
>  I think that collation is not quite ready but is close.  I finished the
> testing for DERBY-2656 and filed a lot of issues which Mamta has been
> fixing very rapidly.  I think if we wait a week or two we can have a
> much more solid picture for collation.   I would especially like to see
> these three fixed:
>
>
>    DERBY-2966  rs.moveToInsertRow() causes loss of connection with
> TERRITORY_BASED collation.
>    DERBY-2973 With collation TERRITORY_BASED, insert into table after
> changing type of column causes assert failure and loss of connection
>    DERBY-2972 Update or select with function in the where clause causes
> with TERRITORY_BASED collation fails with ERROR 42818: Comparisons
> between 'VARCHAR' and 'VARCHAR' are not supported.
>
> 2)  DERBY-2963
>  AccessControlException: Access denied java.net.SocketPermission <client
> ip> accept,resolve
>
> 3) I am also quite concerned that we are letting the release out with
> known regressions, DERBY-2905, DERBY-2892, but since I am not in a
> position to fix these in short order,  I would change my vote just on
> resolution of the other issues.
>
> Thanks to everyone for their hard work on the release.  I really think
> it is coming into form but is just not quite there yet.
>
> Kathey
>
>
>
>
Hi,

Thx for your vote, Kathey...

As DERBY-2963 is marked blocker, (and I think we all agree, we should
at least be able to start network server providing an ip address...)
we should really roll another release, when it is fixed (fix has been
applied to trunk, backport is pending test results).

In the mean time, 10.3.1.4 has now fixes in for DERBY-2966, DERBY-2955
and DERBY-2960.
Neither DERBY-2973 nor DERBY-2972 are marked as blockers, so they
should not hold up a release; if you think they should, please mark
them as blockers...
As there are no timeframes, or no takers for the other issues, and
they are not marked as blockers, the other issues should not be able
to hold up a release...

I have a question, though, in earlier votes/comments, you mentioned
you thought DERBY-2437/2925 should get fixed in 10.3.1. Is this no
longer the case (fix is currently in trunk)? Again, if you think it is
a blocker, can you mark it so?


More in general, note that I've not rushed into another release
because, well, there was no fix yet for the blocker (:-) ), and
secondly, I cannot conclude the work on a new release myself. I'll
have to let the baton go to someone else, or pick things up when I
can.

Rick has said he'd be willing to take over, but that was before he
went on vacation. So, at this point, I can either:
- Create one more set of release artifacts as soon as DERBY-2963 is
fixed, and call a vote on that hoping that Rick (or someone else) can
finish the release management tasks, or leave it for me in September
- Cancel the current vote for 10.3.1.3 and do nothing, and leave it to
someone else to finish before, or for me in September.

Opinions?

Myrna

Mime
View raw message