MIT-licensed sparse() parent method and expert driver, take two

[Sacha0]

See #14798 (the original pull request) for discussion. Apologies for the close/reopen-glitch!

[tkelman]

mention SparseArrays.dropzeros! as the way to get rid of them when the user would like to do so?

[Sacha0]

Sounds wise. In fact, given dropzeros! should become increasingly important / regularly used with the shift to retaining numerical zeros by default, perhaps dropzeros! should now be exported?

[StefanKarpinski]

+1 for exporting dropzeros! and making other operations retain zeros (including the sparse function).

[Sacha0]

mention SparseArrays.dropzeros! as the way to get rid of them when the user would like to do so?

I added lines to this effect to sparse's documentation and NEWS.md (without the SparseArrays qualification, anticipating export of dropzeros!).

Assuming exporting dropzeros! seems advantageous to all thread participants, should I add a separate commit to this PR exporting dropzeros!? Or would opening a separate PR (to broaden the change's visibility and hence opportunity for comment) be better?

[tkelman]

Maybe do the export separately, I'd rather get this merged. Hopefully @jrevels and others will figure out the right way to standardize running package-based benchmarks over time.

When it does get exported, it should also be added to the docs and probably tested a bit more extensively.

[Sacha0]

Maybe do the export separately, I'd rather get this merged. Hopefully @jrevels and others will figure out the right way to standardize running package-based benchmarks over time.

When it does get exported, it should also be added to the docs and probably tested a bit more extensively.

Sounds great. Let me know if any other touch-ups are in order here. Thanks again, and best!

[tkelman]

apt-get travis issue addressed by #15249, restarted

dpkg: dependency problems prevent configuration of gfortran-5:i386:
 gfortran-5:i386 depends on gcc-5 (= 5.2.1-23ubuntu1~12.04); however:
  Package gcc-5:i386 is not configured yet.
dpkg: error processing gfortran-5:i386 (--configure):
 dependency problems - leaving unconfigured
Setting up libdpkg-perl (1.16.1.2ubuntu7.7) ...
Setting up zlib1g-dev:i386 (1:1.2.3.4.dfsg-3ubuntu4) ...
Setting up libssl-dev:i386 (1.0.1-4ubuntu5.34) ...
Setting up make:i386 (3.81-8.1ubuntu1.1) ...
Processing triggers for libc-bin ...
ldconfig deferred processing now taking place
Errors were encountered while processing:
 gcc-5:i386
 g++-5:i386
 gfortran-5:i386
E: Sub-process /usr/bin/dpkg returned an error code (1)

edit: since that was fixed by a .travis.yml change, it may need a rebase to actually apply on the PR build. This passes on win32 so I don't see why it wouldn't pass on linux 32 bit.

[Sacha0]

@tkelman Much thanks for the thorough review!