Change Future to DeprecationWarning for make_block_same_class

[jorisvandenbossche]

xref #19434

@jreback I didn't want to further discuss on the PR, so let's do that here :-)

By having it as a FutureWarning, we only annoy users, and the fastparquet developers are already aware of it.
BTW, we do exactly the same for pyarrow's 'misuse' of internal API, we added a deprecationwarning for them.

[jreback]

ok I guess. The problem is fp is deep into the internals. This is going to break things and I don't want this to hold us back in any way from changing things.

[codecov]

Codecov Report

Merging #19442 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master   #19442   +/-   ##
=======================================
  Coverage    91.6%    91.6%           
=======================================
  Files         150      150           
  Lines       48724    48724           
=======================================
  Hits        44632    44632           
  Misses       4092     4092

Flag	Coverage Δ
#multiple	89.97% <ø> (ø)	⬆️
#single	41.74% <ø> (ø)	⬆️

Impacted Files	Coverage Δ
pandas/core/internals.py	95.47% <ø> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 272cabb...ee61d95. Read the comment docs.

[jorisvandenbossche]

The problem is fp is deep into the internals. This is going to break things and I don't want this to hold us back in any way from changing things.

I agree this is a problem, that we don't really have an api to interact with the internals, but it is a problem in general for packages that need to do this (pyarrow does the same).
We could think about declaring a very small subset of it as 'stable'