Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Accumulo >> mail # dev >> Resource leak warnings


Copy link to this message
-
Re: Resource leak warnings
This conversation is way past dead at this point, but I wanted to respond
now that things have had a time to cool off. I didn't really pursue a
response last week because I felt like you [Christopher] had unilaterally
imposed your will on the rest of us and were unwilling to discuss things
further. I was pretty upset and decided to just step back from the
conversation completely.

For the record, you don't have to keep things rebased, you can do a "git
update-index --assume-unchanged" on the closeable files in your workspace
and that would magically solve all your warnings. (I love that git is
magical).

The larger discussion that we need to have is what we do about the API
problems, and the long-lived resources. There has been some discussion in
IRC, on various JIRAs, and sprinkled across email about the proper
solution, but I'm having a hard time mentally merging all of those
conversations, so I'll propose that we refocus on it here.

What are our invariants? What are the goals? What tools are available to
solve the problem?

Mike
On Fri, Dec 13, 2013 at 4:03 PM, Christopher <[EMAIL PROTECTED]> wrote:

> On Fri, Dec 13, 2013 at 4:34 PM, Sean Busbey <busbey+[EMAIL PROTECTED]>
> wrote:
> > Could you just do #4 as a patch on your own local repo? It should be
> > relatively easy to keep rebased given that it's just an interface.
> [snip]
>
> No, I cannot easily keep 3 branches rebase'd, nor do I wish to
> constantly rebase and reorder my other work that I wish to push,
> waiting on this issue. I do not see an issue with this (essentially)
> partial revert, of what was (in my view) incomplete work. The option
> that seemed to be favored (by myself, included) was #3. Proceeding
> with #4 in the meantime does not preclude #3 from proceeding, but it
> does address the immediate issues at hand. We make improvements on
> changes all the time. I don't see this as any different.
>
> --
> Christopher L Tubbs II
> http://gravatar.com/ctubbsii
>