CompatHelper: bump compat for JuliaSyntax to 1, (keep existing compat)

[github-actions]

This pull request changes the compat entry for the JuliaSyntax package from 0.4 to 0.4, 1.
This keeps the compat entries for earlier versions.

Note: I have not tested your package with this new compat entry.
It is your responsibility to make sure that your package tests pass before you merge this pull request.

[sunxd3]

This is not a trivial fix, for now, I'll write down some potential changes required, and come back when I have time.

• true, false are no longer tokenized as two separate kinds of tokens, but the same token K"Bool" (Replace K"true" and K"false" with K"Bool" JuliaLang/JuliaSyntax.jl#488)
• <- is parsed differently
  • v0.4.10

    julia> tokenize("p <- dnorm(0, 1)")
    12-element Vector{Token}:
     Token(JuliaSyntax.SyntaxHead(K"Identifier", 0x0000), 0x00000001:0x00000001)
     Token(JuliaSyntax.SyntaxHead(K"Whitespace", 0x0001), 0x00000002:0x00000002)
     Token(JuliaSyntax.SyntaxHead(K"<", 0x0000), 0x00000003:0x00000003)
     Token(JuliaSyntax.SyntaxHead(K"-", 0x0000), 0x00000004:0x00000004)
     Token(JuliaSyntax.SyntaxHead(K"Whitespace", 0x0001), 0x00000005:0x00000005)
     Token(JuliaSyntax.SyntaxHead(K"Identifier", 0x0000), 0x00000006:0x0000000a)
     Token(JuliaSyntax.SyntaxHead(K"(", 0x0001), 0x0000000b:0x0000000b)
     Token(JuliaSyntax.SyntaxHead(K"Integer", 0x0000), 0x0000000c:0x0000000c)
     Token(JuliaSyntax.SyntaxHead(K",", 0x0001), 0x0000000d:0x0000000d)
     Token(JuliaSyntax.SyntaxHead(K"Whitespace", 0x0001), 0x0000000e:0x0000000e)
     Token(JuliaSyntax.SyntaxHead(K"Integer", 0x0000), 0x0000000f:0x0000000f)
     Token(JuliaSyntax.SyntaxHead(K")", 0x0001), 0x00000010:0x00000010)

  • 1.0.2

    julia> tokenize("p <- dnorm(0, 1)")
    12-element Vector{Token}:
     Token(JuliaSyntax.SyntaxHead(K"Identifier", 0x0000), 0x00000001:0x00000001)
     Token(JuliaSyntax.SyntaxHead(K"Whitespace", 0x0001), 0x00000002:0x00000002)
     Token(JuliaSyntax.SyntaxHead(K"Identifier", 0x0000), 0x00000003:0x00000003)
     Token(JuliaSyntax.SyntaxHead(K"Identifier", 0x0000), 0x00000004:0x00000004)
     Token(JuliaSyntax.SyntaxHead(K"Whitespace", 0x0001), 0x00000005:0x00000005)
     Token(JuliaSyntax.SyntaxHead(K"Identifier", 0x0000), 0x00000006:0x0000000a)
     Token(JuliaSyntax.SyntaxHead(K"(", 0x0001), 0x0000000b:0x0000000b)
     Token(JuliaSyntax.SyntaxHead(K"Integer", 0x0000), 0x0000000c:0x0000000c)
     Token(JuliaSyntax.SyntaxHead(K",", 0x0001), 0x0000000d:0x0000000d)
     Token(JuliaSyntax.SyntaxHead(K"Whitespace", 0x0001), 0x0000000e:0x0000000e)
     Token(JuliaSyntax.SyntaxHead(K"Integer", 0x0000), 0x0000000f:0x0000000f)
     Token(JuliaSyntax.SyntaxHead(K")", 0x0001), 0x00000010:0x00000010)

  i.e., <- is now parsed as two K"Identifier" instead of K"<" and K"-", but this is something I relied on in the
  parser

Figure out:

• What version of JuliaSyntax is Julia 1.10 using? How about Julia 1.2? (This is potentially related to when I must fix this)
• Even if in a future version of Julia when JuliaSyntax 1.0 is used, must I update from JuliaSyntax 0.4, or both version can be used?

[github-actions]

JuliaBUGS.jl documentation for PR #287 is available at:
https://TuringLang.github.io/JuliaBUGS.jl/previews/PR287/

[coveralls]

Pull Request Test Coverage Report for Build 15043798846

Details

• 8 of 8 (100.0%) changed or added relevant lines in 1 file are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage increased (+0.006%) to 83.518%

---

Files with Coverage Reduction	New Missed Lines	%
src/parser/bugs_parser.jl	1	89.09%

Totals
Change from base Build 14887157894:	0.006%
Covered Lines:	2108
Relevant Lines:	2524

---

💛 - Coveralls

[github-actions]

Benchmark results on macOS (aarch64)

BridgeStan not found at location specified by $BRIDGESTAN environment variable, downloading version 2.6.2 to /Users/runner/.bridgestan/bridgestan-2.6.2
Done!

Stan results:

Model	Parameters	Density Time (µs)	Density+Gradient Time (µs)
rats	65	5.0916	7.22233
pumps	12	0.873136	0.966346
bones	13	79.541	90.458
oxford	244	11.521	13.6875
epil	303	36.0	39.542
lsat	1006	107.416	151.645
schools	133	222.25	351.188
beetles	2	0.722925	0.803941
air	5	0.783784	0.8986

JuliaBUGS Mooncake results:

Model	Parameters	Density Time (µs)	Density+Gradient Time (µs)
rats	65	1.16668	7.19775
pumps	12	0.458338	1.80988
bones	33	47.083	244.75
oxford	244	21.125	51.75
epil	303	7.146	34.75
lsat	1006	55.083	237.667
schools	133	90.75	1027.08
beetles	2	1.02441	3.97029
air	5	0.288527	1.16286

---

Benchmark results on Ubuntu (x64)

BridgeStan not found at location specified by $BRIDGESTAN environment variable, downloading version 2.6.2 to /home/runner/.bridgestan/bridgestan-2.6.2
Done!

Stan results:

Model	Parameters	Density Time (µs)	Density+Gradient Time (µs)
rats	65	5.1636	7.52733
pumps	12	0.960467	1.20379
bones	13	73.057	90.208
oxford	244	13.836	17.7935
epil	303	27.432	33.894
lsat	1006	153.547	195.155
schools	133	507.84	742.579
beetles	2	0.832429	0.9698
air	5	0.661909	0.788757

JuliaBUGS Mooncake results:

Model	Parameters	Density Time (µs)	Density+Gradient Time (µs)
rats	65	1.993	14.3015
pumps	12	0.595204	2.42367
bones	33	47.2535	344.486
oxford	244	23.424	72.876
epil	303	10.8005	55.935
lsat	1006	90.88	365.869
schools	133	103.309	1078.99
beetles	2	1.09823	4.68533
air	5	0.338333	1.3881

[sunxd3]

The two fixes down in this PR are:

1. by default, operators are parsed as K"Identifier" in v1.0: Parse non-syntactic operator tokens as K"Identifier" kind JuliaLang/JuliaSyntax.jl#523
2. K"true" and K"false" are replaced with K"Bool"

1 is solved by setting a keyword argument of tokenize function, 2 is specially handled.