SciML style does not expect new lines to be placed with every argument. This puts it out of reach for almost any compliable Julia function so it essentially only applies to arrays.

remove ci for versions < 1.6

* add max_short_function_length

* bump

* add show method for Options

* add force_long_function_def

* document force_long_function_def

* set version

---------

Co-authored-by: Dominique Luna <[email protected]>

…ing compat) (#865)

Co-authored-by: CompatHelper Julia <[email protected]>

Co-authored-by: Edward Jiang <[email protected]>

Update docstrings of BlueStyle, SciMLStyle, and YASStyle to contain the
computed list of options that are different from DefaultStyle.

* use JuliaSyntax for document + state

* CSTParser is bad at this part

* wip

* same api

* wip

* first format working

* rm DataStructures

* options

* good progress

* works

* ok

* Disable for now

* fixes

* progress - unicode is being an issue

* ok

* ok

* more tests pass

* ok i think it's time to use julia syntax for more now all tests pass except some alignment ones and that's due to unicode length/ncodeunits stuff

* k

* precompile

* hmmm

* progress

* wip

* wip

* wip

* fmt

* wio

* lineage

* lfg

* wip

* coming together

* dwadaw

* wow

* parameters working better

* more tests passing

* almost got initial tests passing

* wip

* let passing

* format

* ok

* more tests passing

* fmt

* fix

* foo

* format

* more stuff working

* wip

* wip

* almost there

* fmt

* hmmmm

* looking good

* almost

* tests pass

* save

* add warning to skip file if it's a parse error

* ok

* ok

* ok

* progress

* ok

* another binary shortcircuit test

* ok

* blue

* test

* ok

* passes

* pass

* fix

* kwargs fix

* wip

* src

* ok

* ok

* lazy op

* more fixes

* standalone binary

* comment out

* jfmt src

* for

* format

* another test

* kw

* simplify

* parameters

* Cleanup

* src done

* another fix

* fmt

* a

* fixes

* fmt

* fmt test files

* works

* rm line

* fix

* ok

* ok

* precompile

* merged from master

* ok

* ok

* ok

* wip

* updates docs

* fmt

* int/uint

* methods

* version check

* version checks

* fix

* fmt

* test

* ok

* format

* ok

* ok

* fixes

* using jet

* jet

* fmt

* wip

* nospecialize

* ok

* more stuff

* ok

* reduce precompilation time by a decent amount

* remove kwargs

* rename

* ok

* cleanup

* ok

* precompile

* progress

* allow hash eq comments to be placed whereever

* fix

* ok

* try this

* stuff

* this too

* more

* ok

* shouldn't go crazy with @nospecialize

* get rid of precompietools for now

* seems to be working

* ok

* seems good

* hmm

* jet

* hmm

* fmt

* allocs

* update

* type stable

* kwdef

* ok

* ok

* -1 -> 0

* ok

* style update

* another format change

* fix short-circuit to if pass ?

* format works

* ok

* update

* ok

* docs update

* simplify

* fix

* format

* this seems to be it

* keep semicolons in quote/begin blocks

* some updates

* format

* improve runtime of strings and docstring formatting

* remove @nospecialize

* reduce compilation time

* greater workload

* rm old format.jl file

* workload

* minor edit

* more precompilation stuff

* format

* fix test

* 1.10 is the new LTS

* tuple blocks

* format

Following links resulted on my machine (running NixOS and using devenv)
in precompilation attempting to format read-only .jl files (because
`.devenv/profile` links to readonly `/nix/store/…` directories). Since I
don't see a case where you may want JuliaFormatter to jump to far-away
directory by following links (looks rather dangerous to me, actually?),
I simply disabled following links.

Note this could be improved by checking whether the link points to a
file within the currently-to-be-formatted directory.

* fixes kwargs issues and returns being added to do expressions - we're just not doing that completely anymore

* fix more issues

* fix bug related to the whitespace_ops_in_indices option

* revert multiline comments handling for now

* add some support for `public` keyword (#886)

* add some support for public keyword

* fix the kset memberships for `public`

* add a simple test

* run public keyword test on 1.11+

* add always return for do blocks with yas style

---------

Co-authored-by: Mirek Kratochvil <[email protected]>

Since this style has been in use all over the SciML project for a long time I think it is no longer considered work-in-progress.