[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [DISCUSS] Towards Calcite 1.17.0

Thanks Julian + Sergey!

Can someone please confirm that the following PRs can be closed?

If so, I'll open a case with INFRA to close them.


On 30/06/2018 7:14 AM, Julian Hyde wrote:
Thanks for bringing this up, Francis. Note that sometimes there are
comments on the JIRA case (and so sometimes the ball might be in the
contributor's court). But yes, let's either review all of these or
close them as stale.

As it happens, I picked up a half-finished patch yesterday, finished
it, and merged it. This is never a small amount of work. Contributors
need to realize that if they do not provide clean, running code with a
test case, then their PR is unlikely to make it. I could have watched
a full-length movie with my kids in the time it took to get that PR
into shape.


On Fri, Jun 29, 2018 at 8:29 AM, Sergey Nuyanzin <snuyanzin@xxxxxxxxx> wrote:
adapted existing changes to current master + added tests

On Fri, Jun 29, 2018 at 2:14 AM, Francis Chuang <francischuang@xxxxxxxxxx>

I would love to see if we can close out/merge some of the really stale PRs
on Github.

Here are a few that I think we should resolve before the release:

[CALCITE-1025] Add support for HTTP Basic auth (for proxies) in Avatica: (JIRA is marked as resolved
and this PR is probably irrelevant since Avatica is now in a different repo)

Add support for converting a MongoDB date to a SQL timestamp type: (Could not find relevant JIRA.
No tests. Original contributor probably lost context, since it's almost 4
years old)

CALCITE-1681 Provide a way to copy RelNode trees between clusters: (No review, despite the
contributor asking for one. Do you think it would be possible to finish
this? PR is more than a year old, so contributor might have lost context)

CALCITE-1748: make method getSchema to be overridable: (The proposed solution
appeared to be unnecessary according to the comments on JIRA, so this is
probably safe to close. Can the JIRA issue be marked as resolved or do we
still need to work on it?)

[CALCITE-1866] dateTime FLOOR to HOUR cause MySQL connector throw
SQLException: (Original
contribution missing a test case. Seems to be abandoned by the contributor.
I think it shouldn't be too onerous if someone could carry the PR and add a

[CALCITE-1882] Can't obtain the user defined aggregate function such as
sum,avg by calcite: (Was dead
for a while, but contributor addressed comments 2 months ago. Can someone
please review so we can get this in?)

These are just some PRs I found from a quick look through Github. I think
a few of them should be easy to carry so that we can get them into the
release. For PRs #180 and #422 can someone confirm if they are safe to
close? If so, I'll ask INFRA to close them.

There are also a bunch of PRs from a year ago that are going stale. If
possible we should review, carry,  close or merge them. Having a lot of
stale PRs without any activity or review may deter contributors new to the
community from contributing.


On 29/06/2018 3:54 AM, Michael Mior wrote:

I'll add CALCITE-2331 as at least a nice-to-have. I believe Andrei is
working on a PR and this would be a good bug to have fixed.

* evaluation of
predicate (A or B) and C is failing for some adapters

Michael Mior

Le jeu. 28 juin 2018 à 13:34, Julian Hyde <jhyde@xxxxxxxxxx> a écrit :

To answer Michael's question: I don't think we should delay the
report, nor should we hurry the release. It's fair to say that 1.17 is
in its final stages. There was a lot of activity on avatica-1.12 from
a lot of individuals, and the community is in great shape.

Here are the cases listed by Volodymyr, converted into links:

* CVSS - ?
* Performance issue
in getPulledUpPredicates - Zoltan
* Upgrade Avatica -
* EXTRACT - Julian

* Arrays in MongoDB
- Volodymyr
* Java8 syntax -
* Babel - Julian
(blocked by 2259)
* JDBC adapter
timestamps - Julian to review


On Thu, Jun 28, 2018 at 10:15 AM, Volodymyr Vysotskyi
<volodymyr@xxxxxxxxxx> wrote:

Hi Michael,

There are several Jiras, which I think are blockers for the
release: CALCITE-2379, CALCITE-2384 and CALCITE-2365 + CALCITE-2303.
Also, there is a list of the Jiras, which would be good to include to
1.17: CALCITE-194, CALCITE-2259, CALCITE-2280, CALCITE-2339.

So I think we need at least couple of weeks to resolve these issues and
release 1.17.

Also, please let me know if there are other issues-blockers for the
release, or if some of these issues may be omitted.

Kind regards,
Volodymyr Vysotskyi

чт, 28 черв. 2018 о 20:13 Enrico Olivelli <eolivelli@xxxxxxxxx> пише:

It also would be great to have at least a BETA version of Babel parser

Il gio 28 giu 2018, 18:48 Michael Mior <mmior@xxxxxxxxxx> ha scritto:

Just wanted to check in with how we're doing with progress towards a
release. It's not a rush at all, but I'm preparing the board report

July and wondering if I should wait to include the 1.17.0 release. It
sounds like there are a few other things that still need to be wrapped


so I'm fine if we don't release for another couple weeks.

Thanks again Volodymyr for taking this on!

Michael Mior

Le mer. 30 mai 2018 à 07:44, Volodymyr Vysotskyi <

écrit :

Calcite 1.16.0 was released on March 19 (more than two months ago).
We have solved over 48 issues[1] since then, therefore we should

discussing about releasing Calcite 1.17.0.
I have created [2] to track the release.

We will start a release process after the Avatica 1.12 is released.

I went through the list of opened PRs during the last month, and

the list of Jiras which possibly be merged before the release:
CALCITE-2303: Support DECADE time unit in EXTRACT function - changes

done, depends on the Avatica release
CALCITE-2329: Enhance SubQueryRemoveRule to rewrite IN operator with

constant from the left side more optimally - I will make required
at the beginning of next week
* CALCITE-2321: Support ragged fixed length value union be variable

In review (Julian)*
* CALCITE-2291: Add rule to push Project past Correlate - I will

this next week*
List of other pull requests which require review or additional

CALCITE-2327: In 3 valued logic mode (b and not b) may not be

CALCITE-2302 / CALCITE-2325: Implicit type cast support - Julian

comment into the Jira
CALCITE-2331: evaluation of predicate (A or B) and C is failing for

adapters - created PR only with a test for this Jira
CALCITE-2209: Support loading JSON model file through URL - Shuyi

added a comment into the Jira
CALCITE-2324: Extract seconds, minutes from date works not correct

CALCITE-2319: Druid Expressions - Output Type of Boolean expressions


be set to FLOAT.
CALCITE-500: Ensure EnumerableJoin hashes the smallest input - PR is
CALCITE-2301: Remove the 10-second-timeout restriction in
ResultSetEnumerable - needs additional rework.

Please let me know, if there are any Jiras should be included into


Kind regards,
Volodymyr Vysotskyi


-- Enrico Olivelli

Best regards,