Applications that worked with Accumulo 1.4 may or may not work with 1.6 already (we've made a lot of changes to the InputFormat, for example) so trying to promise compatibility with 2.0 sounds like a very losing battle.
On Wed, Oct 8, 2014 at 3:44 PM, Keith Turner <ke...@deenlo.com> wrote: > On Tue, Oct 7, 2014 at 3:03 PM, Bill Havanki <bhava...@clouderagovt.com> > wrote: > > > I took a look at Christopher's commits for ACCUMULO-3197 and they all > look > > fine to me. Any other reviewers may like to add "?w=1" to the URL for > each > > commit to ignore whitespace-only changes in the view, e.g.: > > > > > > > https://github.com/ctubbsii/accumulo/commit/dc1332b5fb5f358f3fff432a1a0fef4f56c1628e > > *?w=1* > > > > Going forward, it'd be nice to have a rule of thumb for how long a > > deprecated item will linger: some possibilities: > > > > - 2 minor releases or the next major release, whichever comes first > > - always until the next major release (this may make sense starting with > > 2.0.0) > > > > I like the idea of a tool to find use of deprecated calls; it appears > that > > Eclipse and Sonar can do that: > > > > > > > http://stackoverflow.com/questions/14490021/scanning-code-base-for-use-of-deprecated-methods > > > > Overall, +1 to removing deprecations from 1.4 and earlier. > > > > So this in effect making the statement that Accumulo apps that worked w/ > 1.4 may not work w/ 2.0.0. Is that what we want? If this would cause > someone to not Adopt 2.0.0, is that what we would want? Do we want to be > able to say that if your app worked w/ 1.4, it will work with 2.0.0? If > so, 2.0.0 does not have to exist forever. Eventually we can release 3.0.0 > and break 1.4 apps. > > > > > > > Bill > > > > > > On Mon, Oct 6, 2014 at 11:18 PM, Christopher <ctubb...@apache.org> > wrote: > > > > > On Mon, Oct 6, 2014 at 9:45 PM, Adam Fuchs <afu...@apache.org> wrote: > > > > > > > So, I think we can make a general argument to set policy, and when > > > removing > > > > a specific method we should make a specific argument. Personally, I > > would > > > > set the bar at identifying the specific harm cause by the retention > of > > > the > > > > method, as well as polling the community and considering objections. > > > > > > > > Christopher, you made an argument about people misunderstanding the > > > > semantics of the method and using it incorrectly. Is that not solved > by > > > > just deprecating the method? > > > > > > > > > > > Clearly no, since mistakes are still occurring in 1.7.0-SNAPSHOT and it > > was > > > deprecated in 1.6.0. Further, it was hard to notice because: > > > > > > 1) it's the only way to currently get that information from the API to > > the > > > RPC layer (see ACCUMULO-3199) > > > (In my proposed commit[1], I offer a temporary workaround which > involves > > > better naming, and limits the API to the ZooKeeperInstance only) > > > 2) the use of the method occurred in a somewhat badly named utility > > method > > > which suppressed deprecation warnings > > > > > > Until ACCUMULO-3199 is fixed to address the shortcoming of being able > to > > > get the user-provided client RPC config to the RPC layer, this method > is > > > going to be prone to abuse. > > > > > > [1] https://github.com/ctubbsii/accumulo/commit/52806b6?diff=split > > > > > > -- > > > Christopher L Tubbs II > > > http://gravatar.com/ctubbsii > > > > > > > > > > > -- > > // Bill Havanki > > // Solutions Architect, Cloudera Govt Solutions > > // 443.686.9283 > > >
