TST: Add unit tests for older timezone issues

[mroeschke]

• closes BUG: DatetimeIndex.shift(freq=...) raises near DST boundary #8616
• closes TST/BUG: Improve error handling when parsing iso 8601 strings #8910
• closes BUG: Timestamp.replace() behaves naively at DST boundaries (plus bonus segfault!) #7825
• closes ERR: raise in tz incompat with DatetimeIndex ops #8306
• closes bug in tz_localize when summer/winter time change #8531
• tests added / passed
• passes git diff upstream/master -u -- "*.py" | flake8 --diff

In the spirit of your #21407 (comment) @jreback, cleaned up some old timezone issues (but not necessarily covered by #21407) that appear to have been solved

[jreback]

@mroeschke looks awesome!

so can you either add these issue numbers to the whatsnew for #21407, or if its a 'different' issue (even though it was solved via the same), add an appropriate whatnew e.g. a resampling fixed deserves its own whatsnew (in bug fixes for all of these i think)

[codecov]

Codecov Report

Merging #21491 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master   #21491   +/-   ##
=======================================
  Coverage   91.92%   91.92%           
=======================================
  Files         153      153           
  Lines       49587    49587           
=======================================
  Hits        45583    45583           
  Misses       4004     4004

Flag	Coverage Δ
#multiple	90.32% <ø> (ø)	⬆️
#single	41.89% <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 9e982e1...6d20d6a. Read the comment docs.

[mroeschke]

Thanks @jreback. Yeah I don't think #21407 was directly related to any of these older issues, so made separate entries for each.

[jreback]

thanks @mroeschke !

[jorisvandenbossche]

I don't think we should put those in 0.23.2, as we didn't fix them in there.

[jreback]

see #21407

[mroeschke]

These all didn't get fixed by #21407 per se. These issues may have been fixed by other prior PR's in other releases. What should our whatsnew policy be on cleaning up these older issues that have already been fixed?

[jorisvandenbossche]

I don't know if it is needed to list them at all. But if we list them, I would rather add them eg to the 0.24.0 (or 0.23.0) file than to the 0.23.2 file. For those bug fix releases, the list is much shorter and it is often more valuable to actually look which bugs/regressions have been fixed compared to 0.23.0 (in the 0.24.0 or 0.23.0 files the list is huge anyways ...)

[jreback]

it’s fine to move some to 0.24 whatsnew
though if they r indeed fixed by #21407 are ok here

[mroeschke]

Sounds good. I have another timezones issues PR in the works so I'll keep that in mind