Make sure the sh scripts all have only LFs (not CRLFs)

[xparq]

(Originally: xparq/Args#33)

The .gitattributes settings supposedly already ensure that it's that way in the repo.
However, I still had core.autocrlf enabled locally, and either because of that or it was just never changed, (some of?) the scripts still had CRLF (in the work tree at least).

I recently dos2unixed them all, and also did a git add --renormalize (which didn't change anything then, but suppressed Git's really inexplicable "...will change to CRLF" warning); let's see what happens next time I check something out and feed it e.g. to WSL's bash...

[xparq]

OK, since run_cases has no extension, .gitattributes doesn't work this way, it must be inverted to do LF by default (for *)!
Then only set .cmd to CRLF, and leave the rest as auto!

[xparq]

No, 09fca6a did nothing about run_cases, it still kept? converted? it to have CRLFs!... :-/ (As checked in a newly cloned repo.)

[xparq]

Still nothing... :-/ This is nuts. Git is so gross...

Reading https://stackoverflow.com/questions/2517190/how-do-i-force-git-to-use-lf-instead-of-crlf-under-windows/2517442
and https://stackoverflow.com/questions/55633058/why-does-git-want-to-correct-my-line-endings-to-crlf-even-though-i-want-them-to now...

And https://stackoverflow.com/questions/35801024/how-to-see-what-type-of-line-endings-are-already-in-git-repository, to check what's in the repo in the first place!

[xparq]

Wow, those idiotic "backwards" warnings about the impending LF->CRLF conversions DESPITE and AFTER setting autocrlf to false and specifically declaring the .sh files as eol=LF in .gitattributes, those are the ones that are "CRLF the most"! :D

It seems that .gitattributes did the exact opposite of what I assumed it would! :-o

git ls-files --eol

i/crlf  w/crlf  attr/                 	.gitattributes
i/lf    w/lf    attr/                 	.gitignore
i/lf    w/lf    attr/                 	README.txt
i/lf    w/crlf  attr/text             	misc/loc.sh
i/lf    w/crlf  attr/text             	test/ .cfg
i/lf    w/lf    attr/                 	test/Args.hpp
i/lf    w/lf    attr/                 	test/Makefile
i/lf    w/lf    attr/                 	test/README - SH compound commands.case/CASE
i/none  w/none  attr/                 	test/README - SH compound commands.case/Hi from the test case dir!
i/lf    w/lf    attr/                 	test/README - SH.case/case
i/lf    w/lf    attr/                 	test/README - SH.case/expect
i/lf    w/lf    attr/                 	test/README - interleaved EXPECTs control.case/case
i/lf    w/lf    attr/                 	test/README - interleaved EXPECTs control.case/expect
i/lf    w/lf    attr/                 	test/README - interleaved EXPECTs.case/case
i/lf    w/lf    attr/                 	test/README.case/case
i/lf    w/lf    attr/                 	test/README.txt
i/lf    w/lf    attr/                 	test/TC dir with no suffix/CASE
i/lf    w/lf    attr/                 	test/_engine/asset/TC-Makefile.gnumake
i/lf    w/lf    attr/                 	test/_engine/asset/TC-Makefile.nmake
i/lf    w/crlf  attr/text             	test/_engine/build.sh
i/lf    w/lf    attr/                 	test/_engine/download_file.ps1
i/lf    w/crlf  attr/text             	test/_engine/functions.sh
i/lf    w/crlf  attr/text             	test/_engine/init_once.sh
i/lf    w/crlf  attr/text             	test/_engine/init_per_case.sh
i/lf    w/crlf  attr/text             	test/_engine/run_case
i/lf    w/lf    attr/                 	test/_these are expected to fail, test them separately!/except.case
i/lf    w/lf    attr/                 	test/_todo/exename.case - fails today, was ok yesterday/CASE
i/lf    w/lf    attr/                 	test/_todo/exename.case - fails today, was ok yesterday/exename.cpp
i/lf    w/crlf  attr/text             	test/_todo/exename.case - fails today, was ok yesterday/t.sh
i/lf    w/lf    attr/                 	test/_todo/exename.case - gcc why different when run from script/CASE
i/lf    w/lf    attr/                 	test/_todo/exename.case - gcc why different when run from script/exename.cpp
i/lf    w/lf    attr/                 	test/expect file.case/CASE
i/lf    w/lf    attr/                 	test/expect file.case/EXPECT
i/lf    w/lf    attr/                 	test/expect unquoted.case
i/lf    w/lf    attr/                 	test/expect-test 1 (one-line).case
i/lf    w/lf    attr/                 	test/general.case
i/-text w/-text attr/                 	test/issue 10.case/case
i/lf    w/lf    attr/                 	test/issue 10.case/issue 10.cpp
i/lf    w/lf    attr/                 	test/long_basic.case
i/lf    w/lf    attr/                 	test/long_colon.case
i/lf    w/lf    attr/                 	test/long_multi.case
i/-text w/-text attr/                 	test/oneliner-example.case
i/lf    w/crlf  attr/text             	test/run_cases
i/lf    w/crlf  attr/text             	test/run_cases.cmd
i/lf    w/lf    attr/                 	test/short_single_no_param.case
i/lf    w/lf    attr/                 	test/subset subdir/subtree case 1.case
i/lf    w/lf    attr/                 	test/subset subdir/subtree case with build.case/CASE
i/lf    w/lf    attr/                 	test/subset subdir/subtree case with build.case/subtree case with build.cpp
i/lf    w/lf    attr/                 	test/test-demo.cpp

[xparq]

And https://git-scm.com/docs/gitattributes... It clearly says:

eol
...
Set to string value "lf"
This setting forces Git to normalize line endings to LF on checkin and prevents conversion to CRLF when the file is checked out.

[xparq]

OK, fukit, I gave in, and just set lf the default (-> also e.g. https://stackoverflow.com/a/42136008/1479945):

# So, perhaps "all LF by default" would finally convince it to forget about CRLFs:
* text=auto eol=lf

With the exceptions:

*.cmd	eol=CRLF
*.ps1	eol=CRLF

This seems to work now, finally.