Update ccall syntax to @ccall

[lgoettgens]

this PR was created by applying the following script to all relevant files.
I'll leave out some files to avoid conflicts with currently open PRs. To remember them for the future, the list of skipped files is:

• src/calcium/ca.jl, src/calcium/qqbar.jl due to Add conversion CalciumField, QQBar -> Float64 #1893
• src/flint/fq_abs_series.jl, src/flint/fq_rel_series.jl due to Split series over fq_nmod and fq into different files #1929

const pattern = r"""\(?ccall\s*\(\s*\(\s*:([a-zA-Z0-9_]+)\s*,\s*([a-zA-Z0-9_]+)\s*\),\s*([a-zA-Z0-9{}$()_]+)\s*,\s*\((([^()]|\{\(|\)\})*)\)\s*,\s*(.*)\s*\)\)?"""

function replacement(str)
    if first(str) == '(' && str[end-1:end] == "))"
        str_ = str[2:end-1]
    else
        str_ = str
    end
    match = Base.match(pattern, str_)
    fun_name = match.captures[1]
    lib_name = match.captures[2]
    ret_type = match.captures[3]
    arg_types = filter(!isempty, strip.(split(match.captures[4], ",")))
    args = filter(!isempty, strip.(split(match.captures[end], ",")))
    if length(arg_types) != length(args)
        @info "Error: number of argument types and arguments do not match" fun_name lib_name ret_type arg_types args
        return str
    end
    for arg in args
        if count('(', arg) != count(')', arg)
            @info "Error: unbalanced parentheses in argument" fun_name lib_name ret_type arg_types args
            return str
        end
    end

    # Constructing new syntax for arguments
    arg_str = join(["$(maybe_parentheses(arg))::$(arg_type)" for (arg, arg_type) in zip(args, arg_types)], ", ")
    
    ret = "@ccall $(lib_name).$(fun_name)($(arg_str))::$(ret_type)"
    if str != str_
        ret = "($ret)"
    end
    return ret
end

function maybe_parentheses(arg)
    if first(arg) == '(' && last(arg) == ')'
        return arg
    end
    if occursin("+", arg) || occursin("-", arg)
        return "($arg)"
    else
        return arg
    end
end

function transform_ccalls(file::String)
    # Read the content of the input file
    content = read(file, String)

    transformed_content = replace(content, pattern => replacement)

    open(file, "w") do f
        write(f, transformed_content)
    end
end

ping @fingolfin

[codecov]

Codecov Report

Attention: Patch coverage is 91.37127% with 185 lines in your changes missing coverage. Please review.

Project coverage is 87.46%. Comparing base (9df1993) to head (425e2ad).
Report is 3 commits behind head on master.

Files with missing lines	Patch %	Lines
src/HeckeMoreStuff.jl	6.06%	31 Missing ⚠️
src/arb/ArbTypes.jl	61.53%	25 Missing ⚠️
src/flint/fmpz_mat.jl	75.90%	20 Missing ⚠️
src/flint/fmpq_mpoly.jl	86.90%	11 Missing ⚠️
src/flint/fmpz.jl	91.73%	10 Missing ⚠️
src/arb/RealPoly.jl	82.92%	7 Missing ⚠️
src/arb/Complex.jl	95.45%	6 Missing ⚠️
src/arb/acb.jl	95.38%	6 Missing ⚠️
src/arb/arb.jl	96.27%	6 Missing ⚠️
src/arb/arb_poly.jl	84.61%	6 Missing ⚠️
... and 19 more

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1932      +/-   ##
==========================================
+ Coverage   87.38%   87.46%   +0.07%     
==========================================
  Files          97       97              
  Lines       35532    35533       +1     
==========================================
+ Hits        31050    31079      +29     
+ Misses       4482     4454      -28     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[thofma]

OK, but why?

[lgoettgens]

@fingolfin already wanted to do this for some time, but hadn't found the time to do it.
My main motivation is that it makes the ccalls a lot more readable. I spent too much time in the past with trying to line up the argument types and argument values with my eyes.
We have already changed some places by hand in PRs like #1906 #1905 #1900, but doing this by hand is a lot more tedious and error-prone than just applying some regex-based rewriting once

[fingolfin]

So I guess for that to work we'd need to "merge" (not rebase, not squash)

[lgoettgens]

I dropped this for now. It seems easier to do another small PR afterwards than to keep this hash right during all of the rebasing I am during currently.

[fingolfin]

Fine by me

[lgoettgens]

Since #1893 is basically finished, I'll update this PR later today to include the two left out files as well