bpo-40328: Add tool for generating cjk mapping headers

[corona10]

https://bugs.python.org/issue40328

[corona10]

@vstinner
If this repository would not be a proper place to store files under tool/data
Is it okay to create a new repository for this? (it's too huge)
But files need to be stored somewhere in our manageable place since
these files can be broken and loose when the times are passed.
(e.g http://www.info.gov.hk/digital21/eng/hkscs/download/big5-iso.txt is broken now.)

[vstinner]

I would prefer to put this script in Tools/unicode/ and don't store Unicode files in Modules/cjkcodecs/tools/data/CP932.TXT, but always download them: as done by Tools/unicode/makeunicodedata.py or some unit tests.

[corona10]

@vstinner Please take a look :)

[corona10]

@methane @serhiy-storchaka
Also, Can you please take look at this PR?

[methane]

I don't have time to understand the script. I only confirmed that this pull request doesn't change the mapping header files.

[corona10]

@methane
Thank you Naoki San!
The main objective of this PR is generating the same mapping header files :)
For example, we can make some change with this script when
something needs to change about supporting Japanese.
(I don't know the practical case)

[corona10]

@vstinner Can we merge this PR?

[vstinner]

Should we run this script in "make regen-all"? Or do you expect that these generated files will never ever change?

Currently, "make regen-all" has no 3rd party dependencies, whereas this script requires to download files from the Internet. In my experience, any additional external dependency makes Python CIs less stable :-(

Other "unicode scripts" are no run by "make regen-all", so I think that it's fine.

[vstinner]

@corona10: I let you decide if you are confident enough to merge this PR with no review.

Since this script is not run by any CI and only run manually, I don't see any risk of regression.

It seems like "Unicode experts" (including me) don't have the bandwidth to review your change.