incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dmitriy Setrakyan <>
Subject Re: [DISCUSSION] Graduate Ignite from the Apache Incubator
Date Thu, 23 Jul 2015 00:48:47 GMT
On Wed, Jul 22, 2015 at 2:07 PM, Daniel Gruno <> wrote:

> On 07/22/2015 09:23 PM, Julian Hyde wrote:
> > I reviewed Ignite’s commit log and email lists. I got the impression of
> a split personality: The dev list[1] is very open and clearly following the
> Apache Way. Meanwhile, the commit log[2] is (to my eyes at least) difficult
> to decipher.
> >
> > In the commit log, messages such as "Merge remote-tracking branch
> 'origin/master’” and "# master minor”, "GG-10559 - Improvements.” are
> typical. Very few descriptive comments or references to IGNITE-xxx cases
> that could provide further explanation. Those are indications to me that
> development is been driven by off-list meetings.
> >
> > If the Ignite committers want to build a diverse community of
> developers, I think they need to make their commits more transparent to
> match the excellent transparency they have achieved on their mailing list.
> >
> > Julian
> >
> > [1]
> > [2]
> >
> >
> The GG-10559 commit log (along with the 'oops' email sent to dev by
> mistake today) seems to indicate (this is just speculation, I would love
> for someone to answer whether this is the case) that commits are being
> based off an internal JIRA instance (or other bug tracking system) not
> hosted at the ASF.

Let me clarify this.

Denis is a GridGain employee and, apart from Ignite, he also works on
GridGain product fixing various issues. At the end of the day he sent an
email notifying other GridGain team members about the work he has done and
by accident sent it to the dev list. This has nothing to do with the Ignite
project and we do not maintain any separate Jiras for the Ignite project.

As far as Ignite project is concerned, all Ignite work is being done in the
open and all the issues/questions related to the Ignite are discussed on
the dev list.

> Furthermore, I have found that there are 207 JIRA tickets with NO
> description, discussion or useful information other than ownership
> changing hands and then suddenly the tickets have been resolved with no
> explanation, suggesting that either people have not been properly taught
> about how to share information and collaborate as a community, or the
> actual discussions and 'ping-pong' happen elsewhere, outside our view.

There are certainly tickets like that, but most of them were early in the
incubation process. As we started learning the Apache way, the Jira
communication has significantly improved and most tickets have relatively
good commentary and description. If you find anything to the contrary, it
is definitely not the norm.

For example, I have looked through the ticket flow for the past several
days and most of them have comments. Here are some examples:


Also, there are some tickets that are self explanatory and do not require
explanation (you can grasp the meaning from the title). Case in hand is
this ticket: . Granted
some commentary would not hurt and I will advise for that on the dev list,
but it is clear that the ticket has been reviewed without comments and

Also, keep in mind that every ticket is worked on in a separate branch, and
every branch is named consistently after the ticket. Some reviews happen in
the branch, and that is why not every ticket has a patch attached to it.
This is documented on the website and Wiki:

Additionally, all the discussions about any issue occurring throughout
working on the ticket usually take place on the dev list:


> I naturally hope there is no reason for concern, and I would love to get
> some insight into why these things pop into view when you examine the
> development process.
> With regards,
> Daniel.
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message