You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tinkerpop.apache.org by Stephen Mallette <sp...@gmail.com> on 2016/10/04 12:28:54 UTC
Re: [DISCUSS] breaking gremlin-server.sh just a little
The more I think about this, the more I think I would just prefer to merge
this on 3.3.0. We open that branch this weekend, so it's not as though
we're pushing PR out too far. Not sure if that changes anyone's silence on
this one, but that's what I'm thinking at this point.
On Fri, Sep 30, 2016 at 12:30 PM, Stephen Mallette <sp...@gmail.com>
wrote:
> This is a very neat change I think - our gremlin-server.sh looks pretty
> legit now. I'd just add one clarification that the proposal here is to push
> this breaking change into the next release of 3.2.3. the question is
> whether we wait for 3.3.0 which we would presumably start on pretty soon
> (couple of weeks) to avoid pushing a breaking change into the 3.2.x line.
> any thoughts on the matter?
>
> On Fri, Sep 30, 2016 at 12:22 PM, Robert Dale <ro...@gmail.com> wrote:
>
>> In relation to PR https://github.com/apache/tinkerpop/pull/439 , I'm
>> proposing that gremlin-server.sh will be an init script. As an init
>> script, the expectation of not providing any parameters is to display
>> usage/help:
>>
>> Usage: gremlin-server.sh {start|stop|restart|status|console|install
>> <group> <artifact> <version>|<conf file>}
>>
>> This breaks the current usage of starting the server in the foreground
>> with the default yaml file. Instead, a user would provide the command
>> `console`.
>>
>> I have tried to keep other backwards compatibility by accepting `-i`
>> (aka `install`) and a yaml file. If the yaml file is the only
>> parameter, the server will continue to start in the foreground.
>>
>> If there are no objections, then @spmallette will assume lazy
>> consensus after 72 hours and do his thing. I'm guessing. ;-)
>>
>> --
>> Robert Dale
>>
>
>
Re: [DISCUSS] breaking gremlin-server.sh just a little
Posted by Stephen Mallette <sp...@gmail.com>.
The breaking in change in 3.2.2 was a mistake that wasn't noticed in
review. Typically, that shouldn't happen.
On Tue, Oct 4, 2016 at 11:13 AM, Dylan Millikin <dy...@gmail.com>
wrote:
> I would wait for 3.3.0. I'm a little confused about the versions at the
> moment as I thought our minor version 3.2.x should not be breaking (but
> that doesn't seem to be the case since 3.2.2 was breaking compared to
> 3.2.0, though maybe that was a mishap).
>
> Other than that the PR is a really nice one.
>
> On Tue, Oct 4, 2016 at 8:28 AM, Stephen Mallette <sp...@gmail.com>
> wrote:
>
> > The more I think about this, the more I think I would just prefer to
> merge
> > this on 3.3.0. We open that branch this weekend, so it's not as though
> > we're pushing PR out too far. Not sure if that changes anyone's silence
> on
> > this one, but that's what I'm thinking at this point.
> >
> > On Fri, Sep 30, 2016 at 12:30 PM, Stephen Mallette <spmallette@gmail.com
> >
> > wrote:
> >
> > > This is a very neat change I think - our gremlin-server.sh looks pretty
> > > legit now. I'd just add one clarification that the proposal here is to
> > push
> > > this breaking change into the next release of 3.2.3. the question is
> > > whether we wait for 3.3.0 which we would presumably start on pretty
> soon
> > > (couple of weeks) to avoid pushing a breaking change into the 3.2.x
> line.
> > > any thoughts on the matter?
> > >
> > > On Fri, Sep 30, 2016 at 12:22 PM, Robert Dale <ro...@gmail.com>
> wrote:
> > >
> > >> In relation to PR https://github.com/apache/tinkerpop/pull/439 , I'm
> > >> proposing that gremlin-server.sh will be an init script. As an init
> > >> script, the expectation of not providing any parameters is to display
> > >> usage/help:
> > >>
> > >> Usage: gremlin-server.sh {start|stop|restart|status|console|install
> > >> <group> <artifact> <version>|<conf file>}
> > >>
> > >> This breaks the current usage of starting the server in the foreground
> > >> with the default yaml file. Instead, a user would provide the command
> > >> `console`.
> > >>
> > >> I have tried to keep other backwards compatibility by accepting `-i`
> > >> (aka `install`) and a yaml file. If the yaml file is the only
> > >> parameter, the server will continue to start in the foreground.
> > >>
> > >> If there are no objections, then @spmallette will assume lazy
> > >> consensus after 72 hours and do his thing. I'm guessing. ;-)
> > >>
> > >> --
> > >> Robert Dale
> > >>
> > >
> > >
> >
>
Re: [DISCUSS] breaking gremlin-server.sh just a little
Posted by Dylan Millikin <dy...@gmail.com>.
I would wait for 3.3.0. I'm a little confused about the versions at the
moment as I thought our minor version 3.2.x should not be breaking (but
that doesn't seem to be the case since 3.2.2 was breaking compared to
3.2.0, though maybe that was a mishap).
Other than that the PR is a really nice one.
On Tue, Oct 4, 2016 at 8:28 AM, Stephen Mallette <sp...@gmail.com>
wrote:
> The more I think about this, the more I think I would just prefer to merge
> this on 3.3.0. We open that branch this weekend, so it's not as though
> we're pushing PR out too far. Not sure if that changes anyone's silence on
> this one, but that's what I'm thinking at this point.
>
> On Fri, Sep 30, 2016 at 12:30 PM, Stephen Mallette <sp...@gmail.com>
> wrote:
>
> > This is a very neat change I think - our gremlin-server.sh looks pretty
> > legit now. I'd just add one clarification that the proposal here is to
> push
> > this breaking change into the next release of 3.2.3. the question is
> > whether we wait for 3.3.0 which we would presumably start on pretty soon
> > (couple of weeks) to avoid pushing a breaking change into the 3.2.x line.
> > any thoughts on the matter?
> >
> > On Fri, Sep 30, 2016 at 12:22 PM, Robert Dale <ro...@gmail.com> wrote:
> >
> >> In relation to PR https://github.com/apache/tinkerpop/pull/439 , I'm
> >> proposing that gremlin-server.sh will be an init script. As an init
> >> script, the expectation of not providing any parameters is to display
> >> usage/help:
> >>
> >> Usage: gremlin-server.sh {start|stop|restart|status|console|install
> >> <group> <artifact> <version>|<conf file>}
> >>
> >> This breaks the current usage of starting the server in the foreground
> >> with the default yaml file. Instead, a user would provide the command
> >> `console`.
> >>
> >> I have tried to keep other backwards compatibility by accepting `-i`
> >> (aka `install`) and a yaml file. If the yaml file is the only
> >> parameter, the server will continue to start in the foreground.
> >>
> >> If there are no objections, then @spmallette will assume lazy
> >> consensus after 72 hours and do his thing. I'm guessing. ;-)
> >>
> >> --
> >> Robert Dale
> >>
> >
> >
>