hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stack <st...@duboce.net>
Subject Re: NOTICE: removing doBulkLoad(Path hfofDir, final HTable table) though it has not been through a complete deprecation cycle?
Date Mon, 23 May 2016 15:04:40 GMT
Thanks lads for the feedback.
St.Ack

On Sun, May 22, 2016 at 10:55 PM, Enis Söztutar <enis.soz@gmail.com> wrote:

> It sounds good. HTable is already deprecated, so removing a method with
> that arg should be fine. We can deprecate it still in 1.3 and later just in
> case.
>
> Enis
>
> On Sun, May 22, 2016 at 8:01 PM, Heng Chen <heng.chen.1986@gmail.com>
> wrote:
>
> > +1.
> >
> > 2016-05-23 1:29 GMT+08:00 Stack <stack@duboce.net>:
> >
> > > HBASE-15875 wants to remove HTable and HTableInterface, classes that
> have
> > > been deprecated since before hbase-1.0.0 release. Unfortunately, we
> > missed
> > > deprecating the method doBulkLoad(Path hfofDir, final HTable table) in
> > > LoadIncrementalHFiles, a public, stable class.
> > >
> > > My thinking is that its removal is 'ok' -- with proper notice -- since
> > this
> > > a method used by offline tooling. My thinking is that as long as proper
> > > notice, folks can prepare their tooling ahead of time changing them to
> > use
> > > the alternative that takes a Table instance in time for the upgrade.
> This
> > > method is also 'damaged' given it has a param that was properly
> > deprecated
> > > -- i.e. the HTable instance -- but nonetheless, its removal will be a
> > > breaking change in 2.0.
> > >
> > > You lot all good w/ this?
> > > Thanks,
> > > St.
> > >
> >
>

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