Hi -
> > [...]
> > The fixes are obvious, proper, self-contained, tested, working.
>
> I referred to just the one commit, above. It is functionally incomplete.
Absolutely not. That commit is functionally 100% complete. Its
function was to ensure that pmwebd and pmmgr survive a crash, such as
if another mystery memory leak bloats them into OOM oblivion.
> The resolution to ...
> http://www.pcp.io/pipermail/pcp/2016-May/010466.html
> http://www.pcp.io/pipermail/pcp/2016-May/010467.html
> http://www.pcp.io/pipermail/pcp/2016-May/010519.html
> ... remains at large, but wanted (dead or alive).
Those are steps toward a different goal - your goal.
> > Incremental change is a good thing. Embrace it.
>
> As PCP maintainers we expect a certain level of completeness [...]
You are missing the point. You are redefining "completeness" by
threatening to reject (or revert!) a contributor's work, even though
it helps the project, unless they do more and more work toward *your*
goals.
I've explained before that a maintainer's role to a large extent is
cajoling contributions from the community. Eagerly merging things is
one form of cajoling. Threats are the opposite of cajoling.
- FChE
|