improved init_phase and use ordered table in code eval so the order o…

…f the files included matters

changelog.md

@@ -1,5 +1,22 @@
 Changelog for Nimjl. Date in format YYYY_MM_DD
 
+Release v0.8.1 - 2024_03_23
+===========================
+* Repo moved to SciNim
+* Added activate option in init to start-up Julia in a virtual environment
+* Added Hook function at init to execute code : BEFORE jl_init() is called; AFTER jl_init() and Pkg.activate() is called but BEFORE Pkg.add calls and BEFORE the embedded code is passed through ``eval`` function.
+* Order of update is now relevant
+
+Release v0.8.0 - 2023_09_30
+===========================
+* Clean-up some code
+* Added more type in evaluation of Dict and Tuples
+* Improve loading a Package at init when the Package is already present (speed up init phase)
+
+Release v0.7.6 - 2023_02_22
+===========================
+* Small CI change
+
 Release v0.7.5 - 2022_07_07
 ===========================
 * Fixed https://github.com/Clonkk/nimjl/issues/18
@@ -34,15 +51,15 @@ Release v0.7.1 - 2022_01_04
 Release v0.7.0 - 2022_01_04
 ===========================
 * Add Julia.useModule alias for jlUseModule
-* Add Julia.includeFile (include is reserved keyword) alias for jlInclude 
+* Add Julia.includeFile (include is reserved keyword) alias for jlInclude
 * Add mechanism to embed julia files at compile-time and run the code at init for an easy way to distribute binary with Julia code contained
 * Add Pkg template to easily install new Julia package during init ; it is also compatible with the embedding stuff :
 * See ex09
   ```nim
-    Julia.init: 
-      Pkg: 
+    Julia.init:
+      Pkg:
         add("LinearAlgebra")
-      Embed: 
+      Embed:
         file("myfile.jl")
   ```
 
@@ -60,7 +77,7 @@ Release v0.6.1 - 2021_10_14
 
 Release v0.6.0 - 2021_10_08
 ===========================
-* Add --gc:orc to CI 
+* Add --gc:orc to CI
 
 Release v0.5.9 - 2021_10_05
 ===========================
@@ -157,4 +174,4 @@ Release v0.4.0 - 2021_03_08
 * Add ``toJlVal`` / ``to`` proc to make conversion betwen Julia types and Nim types "smooth"
 * Added Julia exception handler from Nim
 * Examples in the examples folder
-* Tess suite and memory leak suite
+* Test suite and memory leak suite

examples/ex01_helloworld.nim

@@ -1,15 +1,18 @@
 import nimjl
 
-Julia.init() # Initialize Julia VM. This should be done once in the lifetime of your program.
+proc main() =
+  Julia.init() # Initialize Julia VM. This should be done once in the lifetime of your program.
 
-# Calling Julia function from Nim will always return a JlValue
-# This JlValue can be "nothing"
-# Therefore, Julia function who do not return a value can be discarded
-var res = Julia.println("Hello world")
-echo res # nothing
-# Check that res is actually nothing
-if res == JlNothing:
-  echo "Julia.println returned nothing"
+  # Calling Julia function from Nim will always return a JlValue
+  # This JlValue can be "nothing"
+  # Therefore, Julia function who do not return a value can be discarded
+  var res = Julia.println("Hello world")
+  echo res # nothing
+  # Check that res is actually nothing
+  if res == JlNothing:
+    echo "Julia.println returned nothing"
 
-discard Julia.println("This also works")
+  discard Julia.println("This also works")
 
+when isMainModule:
+  main()

examples/ex02_sqrt.nim

@@ -1,13 +1,16 @@
 import nimjl
 import std/math
 
-jlVmInit() # Initialize Julia VM. This should be done once in the lifetime of your program.
+proc main() =
+  jlVmInit() # Initialize Julia VM. This should be done once in the lifetime of your program.
 
-var myval = 4.0'f64
-# Call Julia function "sqrt" and convert the result to a float
-# This syntax also works to call a function directly from a Julia modfule
-var res = JlBase.sqrt(myval).to(float64)
-# Echo on JlValue calls println from Julia
-echo res # 2.0
-doAssert res == sqrt(myval)
+  var myval = 4.0'f64
+  # Call Julia function "sqrt" and convert the result to a float
+  # This syntax also works to call a function directly from a Julia modfule
+  var res = JlBase.sqrt(myval).to(float64)
+  # Echo on JlValue calls println from Julia
+  echo res # 2.0
+  doAssert res == sqrt(myval)
 
+when isMainModule:
+  main()

examples/ex03_sugar.nim

@@ -1,17 +1,20 @@
 import nimjl
 import std/sequtils
 
-Julia.init()
-# Use Module handle
-discard JlMain.println(@[1, 2, 3])
-discard Julia.println(toSeq(0..5))
+proc main =
+ Julia.init()
+ # Use Module handle
+ discard JlMain.println(@[1, 2, 3])
+ discard Julia.println(toSeq(0..5))
 
-let arr = [1.0, 2.0, 3.0].toJlArray()
-# You can now use echo to call println for you on Julia type !
-echo jltypeof(arr)
-echo arr
+ let arr = [1.0, 2.0, 3.0].toJlArray()
+ # You can now use echo to call println for you on Julia type !
+ echo jltypeof(arr)
+ echo arr
 
-# You can also call proc from the value directly
-echo arr.stride(1)
-echo arr.strides()
+ # You can also call proc from the value directly
+ echo arr.stride(1)
+ echo arr.strides()
 
+when isMainModule:
+  main()

examples/ex04_dict.nim

@@ -12,29 +12,31 @@ proc popMe[U, V](tab: var JlValue, key: U): V =
   # Therefore "pop!" function will not modify tab
   result = jlCall("pop!", tab, key).to(V)
 
-
-jlVmInit() # Initialize Julia VM. This should be done once in the lifetime of your program.
-
-var mytab: Table[int64, float64] = {1'i64: 0.90'f64, 2'i64: 0.80'f64, 3'i64: 0.70'f64}.toTable
-block:
-  let key = 1
-  var poppedValue = popMe(mytab, key)
-  # mytab has not been changed due to copy when passing value to Julia
-  doAssert key in mytab
-  doAssert poppedValue == mytab[key]
-
-block:
-  var myJlTab = toJlVal(mytab) # Convert myJlTab to JlValue
-  let key = 1
-  var poppedValue = popMe[int64, float64](myJlTab, key)
-  doAssert not myJlTab.to(Table[int64, float64]).contains(key) # Value was removed from myJlTab
-  doAssert poppedValue == mytab[key]
-
-block:
-  # You can use [] on Julia dict as well
-  var jldict = toJlVal({"alpha": 1.1, "beta": 2.2}.toTable)
-  doAssert jldict["alpha"].to(float) == 1.1
-  doAssert jldict["beta"].to(float) == 2.2
-  jldict["alpha"] = 3.3
-  doAssert jldict["alpha"].to(float) == 3.3
-
+proc main() =
+  jlVmInit() # Initialize Julia VM. This should be done once in the lifetime of your program.
+
+  var mytab: Table[int64, float64] = {1'i64: 0.90'f64, 2'i64: 0.80'f64, 3'i64: 0.70'f64}.toTable
+  block:
+    let key = 1
+    var poppedValue = popMe(mytab, key)
+    # mytab has not been changed due to copy when passing value to Julia
+    doAssert key in mytab
+    doAssert poppedValue == mytab[key]
+
+  block:
+    var myJlTab = toJlVal(mytab) # Convert myJlTab to JlValue
+    let key = 1
+    var poppedValue = popMe[int64, float64](myJlTab, key)
+    doAssert not myJlTab.to(Table[int64, float64]).contains(key) # Value was removed from myJlTab
+    doAssert poppedValue == mytab[key]
+
+  block:
+    # You can use [] on Julia dict as well
+    var jldict = toJlVal({"alpha": 1.1, "beta": 2.2}.toTable)
+    doAssert jldict["alpha"].to(float) == 1.1
+    doAssert jldict["beta"].to(float) == 2.2
+    jldict["alpha"] = 3.3
+    doAssert jldict["alpha"].to(float) == 3.3
+
+when isMainModule:
+  main()

examples/ex05_module.nim

@@ -7,63 +7,66 @@ type
     y: float
     z: string
 
-Julia.init() # Initialize Julia VM. This should be done once in the lifetime of your program.
-# Include Julia file
-jlInclude("localmodule.jl")
-# Use the module. If you're confused by the syntax, go and read through Julia's Manual where module usage is explained
-jlUseModule(".nimjlExample")
+proc main() =
+  Julia.init() # Initialize Julia VM. This should be done once in the lifetime of your program.
+  # Include Julia file
+  jlInclude("localmodule.jl")
+  # Use the module. If you're confused by the syntax, go and read through Julia's Manual where module usage is explained
+  jlUseModule(".nimjlExample")
 
-block: # Tuple handling
-  # Look, you can pass Nim tuple to Julia
-  var mytup: MyTuple = (nimTupKey1: 1, nimTupKey2: 2)
-  # Convert Nim tuple to Julia tuple automatically
-  var res = Julia.customFunction(mytup)
-  # Convert Julia tuple to Nim tuple
-  var nimres = res.to(MyTuple)
+  block: # Tuple handling
+    # Look, you can pass Nim tuple to Julia
+    var mytup: MyTuple = (nimTupKey1: 1, nimTupKey2: 2)
+    # Convert Nim tuple to Julia tuple automatically
+    var res = Julia.customFunction(mytup)
+    # Convert Julia tuple to Nim tuple
+    var nimres = res.to(MyTuple)
 
-  echo myTup
-  echo nimres
+    echo myTup
+    echo nimres
 
-  doAssert myTup.nimTupKey1+1 == nimres.nimTupKey1
-  doAssert myTup.nimTupKey2+1 == nimres.nimTupKey2
+    doAssert myTup.nimTupKey1+1 == nimres.nimTupKey1
+    doAssert myTup.nimTupKey2+1 == nimres.nimTupKey2
 
-block: # Object manipulation
-  # Call constructor
-  var foo = Julia.makeFoo()
-  # Access fields with dot syntax
-  # Calls getproperty in Julia side
-  echo foo.x
-  echo foo.y
-  echo foo.z
-  # Modify fields with .= syntax
-  # Calls setproperty! on Julia side
-  foo.x = 2
-  foo.y = 3.14
-  foo.z = "General Kenobi !"
-  # Yay this has been modified
-  echo foo
+  block: # Object manipulation
+    # Call constructor
+    var foo = Julia.makeFoo()
+    # Access fields with dot syntax
+    # Calls getproperty in Julia side
+    echo foo.x
+    echo foo.y
+    echo foo.z
+    # Modify fields with .= syntax
+    # Calls setproperty! on Julia side
+    foo.x = 2
+    foo.y = 3.14
+    foo.z = "General Kenobi !"
+    # Yay this has been modified
+    echo foo
 
-  # You can use dot syntax on Julia value to call proc as well
-  discard foo.applyToFoo()
-  echo foo
+    # You can use dot syntax on Julia value to call proc as well
+    discard foo.applyToFoo()
+    echo foo
 
-block:
-  var foo = Foo(x: 12, y: 15, z: "This string comes from Nim")
-  # You can convert Nim to Julia object if :
-  # * fields have the same name and type (fieldName becomdes Julia symbol)
-  # * The Julia type have an empty constructor -- Nim needs to initialize the Julia variable before calling setproperty! providing a default empty constructor is the easiest way of doing it
-  var jlfoo = toJlVal(foo)
-  echo jlfoo
-  echo jltypeof(jlfoo) # This echo "Foo" -> Julia sees this as a Foo mutable struct type
+  block:
+    var foo = Foo(x: 12, y: 15, z: "This string comes from Nim")
+    # You can convert Nim to Julia object if :
+    # * fields have the same name and type (fieldName becomdes Julia symbol)
+    # * The Julia type have an empty constructor -- Nim needs to initialize the Julia variable before calling setproperty! providing a default empty constructor is the easiest way of doing it
+    var jlfoo = toJlVal(foo)
+    echo jlfoo
+    echo jltypeof(jlfoo) # This echo "Foo" -> Julia sees this as a Foo mutable struct type
 
-  discard jlfoo.applyToFoo()
-  # Object are copid during conversions so modifying jlfoo does not modify foo
-  # There is an exception to this for Array fields -- see ex_arrays for explanation
-  echo jlfoo
-  echo foo
+    discard jlfoo.applyToFoo()
+    # Object are copid during conversions so modifying jlfoo does not modify foo
+    # There is an exception to this for Array fields -- see ex_arrays for explanation
+    echo jlfoo
+    echo foo
 
-block:
-  let res = Julia.returnDoubleValue()
-  echo res.myint
-  echo res.mystr
+  block:
+    let res = Julia.returnDoubleValue()
+    echo res.myint
+    echo res.mystr
 
+when isMainModule:
+  main()

examples/ex11_external_deps.nim

@@ -1,9 +1,18 @@
 import nimjl
 
-## See https://pkgdocs.julialang.org/dev/api/#Pkg.add for more info
-Julia.init(1):
-  Pkg:
-    add(name="Polynomials", version="3.0.0")
-    add(name="LinearAlgebra")
-    add("DSP")
-    add(name="Wavelets", version="0.9.4")
+proc main() =
+  ## See https://pkgdocs.julialang.org/dev/api/#Pkg.add for more info
+  Julia.init(1):
+    Pkg:
+      add(name="Polynomials", version="3.0.0")
+      add(name="LinearAlgebra")
+      add("DSP")
+
+  Julia.useModule("Pkg")
+  let jlpkg = Julia.getModule("Pkg")
+  discard jlpkg.status()
+
+  Julia.exit()
+
+when isMainModule:
+  main()

nimjl/cores.nim

@@ -82,13 +82,14 @@ proc jlVmExit*(exit_code: cint = 0.cint) =
 #   discard jlEval(fmt"exit_save_sysimage({fpath})")
 
 #########################################
-var staticContents: Table[string, string]
+var staticContents: OrderedTable[string, string]
 
 import std/logging
 
 proc loadJlRessources*() =
   for key, content in staticContents.pairs():
-    info("> Nimjl loading Julia ressource: ", key, ".jl")
+    info("> Nimjl loading Julia ressource: ", key)
+    # debugEcho("> Nimjl loading Julia ressource: ", key)
     JlCode(content)
 
 # Init & Exit function