I've just issued a pull request to remove all of the headers that
include the opening comments and the licensing information from the
*individual source files* in the yt-4.0 branch. This means that
information will only be in the top-level repository, rather than in
It is, evidently, bad practice to include the copyright -- which in
any case has been out of date for several years anyway -- in each
individual file, as it's included in the top-level directory. The
leading comments in almost 100% of cases provided no useful
information; it may be the case that they should be re-added, but as
of present most are just copy/pasted.
I'm posting this here to make sure that it can be considered before
getting merged, although having inspected the diff I believe there are
no behavior changes. The question is basically if we want this
boilerplate, frequently-outdated info in every file or not.
(PS the pull request is super long.)
I've got a PR that removes py2 compat code:
I'm wondering if we had come to a decision on what was required for it
to go in. Are we going to block on getting answer testing re-enabled?
Or is passing the unit tests sufficient?
Nathan's Herculean unyt pull request into yt-4.0 is passing all the tests.
Since it's been under review for a fairly long time, I am going to
merge it tomorrow unless I hear otherwise. All of the changes are
really straightforward, and overall a net gain.
With the advent of the ytdata frontend, it seems that all usages of
the GeometryHandler / Index object's functions save_data,
_reload_data_file, load_object and get_data are all dead now.
Any reason they can't be dispensed with? I suspect-without-evidence
that even before ytdata, usage was close to nil.