Lkx is a new strongly typed scripting language, simpler and faster than lua. It can easily interact with c/c++, the name Lkx comes from my daughter’s name (Li Kaixin), so this language is also a gift for my daughter.
supported type | description |
---|---|
bool | boolean value, true or false |
double | double precision floating point |
string | sequence of characters |
int8, int16, int32, int64 | signed int of 8~64 bits |
uint8, uint16, uint32, uint64 | unsigned int of 8~64 bits |
[int8], [int16], [int32], [int64] | array of int8~64 |
[uint8], [uint16], [uint32], [uint64] | array of uint8~64 |
[bool], [double], [string] | array of bool, double, string |
{uint8}, {uint16}, {uint32}, {uint64} | unordered set of uint8~64 |
{m uint8}, {m uint16}, {m uint32}, {m uint64} | unordered multi set of uint8~64 |
{o uint8}, {o uint16}, {o uint32}, {o uint64} | orderded set of uint8~64 |
{mo uint8}, {mo uint16}, {mo uint32}, {mo uint64} | ordered multi set of uint8~64 |
{string}, {m string} | unordered set of string, unordered multi set of string |
{o string}, {mo string} | ordered set of string, ordered multi set of string |
{uint32:string} | unordered map, key type: uint32, value type: string |
{string:string} | unordered map, key type: string, value type: string |
{m string:uint16} | unordered multi map |
{o uint8:uint16} | ordered map |
{mo string:[uint32]} | ordered multi map, value type: array of uint32 |
In Lkx, each source file is a module, and each file has a consistent code layout, the source file is logically divided into several parts, each part must be separated by a blank line.
The 1st part is the import statement, which imports shared variables and functions from other files:
import "family/lib/utils.lkx"
import "family/sister/info.lkx"
The 2nd part is the declaration and initialization of variables, the shared variable must be declared before the local variable:
shared string brother_name = "bob"
uint32 brother_height = 170
uint32 brother_weight = 55
uint32 brother_age = 31
The 3rd part is the hot reloading hook function, which is used to control whether the variables declared in the current module need to be reinitialized during hot reloading:
internal void __lkx__hot_reloading_hook__()
{
}
The 4th part is the definition of the function, shared functions must be defined before local functions, and each function must be separated by a blank line:
shared bool brother_is_giant()
{
return is_a_giant(brother_height, brother_weight)
}
shared void brother_grow_up()
{
brother_height += 10
brother_weight += 20
}
bool brother_older_than_sister()
{
return brother_age >= sister_age
}
The source code for this sample file (path: "family/brother/info.lkx"
):
import "family/lib/utils.lkx"
import "family/sister/info.lkx"
shared string brother_name = "bob"
uint32 brother_height = 170
uint32 brother_weight = 55
uint32 brother_age = 31
internal void __lkx__hot_reloading_hook__()
{
}
shared bool brother_is_tall()
{
if(brother_height < TALL_CM)
{
return false
}
return true
}
shared bool is_older_brother()
{
return brother_older_than_sister()
}
shared bool brother_is_giant()
{
return is_a_giant(brother_height, brother_weight)
}
shared void brother_grow_up()
{
brother_height += 10
brother_weight += 20
}
bool brother_older_than_sister()
{
return brother_age >= sister_age
}
Lkx code is usually called by a c/c++ host program. After we have developed a project and run it for some time, we may need to modify the variables and functions in the script to meet some new requirements. In this case, Lkx can reload the modified script code without restarting the c/c++ host program.
Before the hot reloading of Lkx script, each variable in the script has been given the corresponding value by many executed functions. During hot reloading, we need to decide which variables should remain unchanged and which variables must be reinitialized based on business logic. Fortunately, Lkx provides an internal hot reloading hook function: __lkx_hot_reloading_hook__()
, In this function, you can specify which variables should remain unchanged during the hot reloading process, as follows (path: "family/lib/utils.lkx"
):
shared uint32 TALL_CM = 190
shared uint32 FAT_KG = 90
uint32 GIANT_CM = 300
uint32 GIANT_KG = 500
uint32 call_giant_count = 0
internal void __lkx__hot_reloading_hook__()
{
call_giant_count = __LKX_REMAIN_UNCHANGED__
}
shared bool is_a_giant(uint32 height, uint32 weight)
{
if(height < GIANT_CM)
{
return false
}
if(weight < GIANT_KG)
{
return false
}
++call_giant_count
return true
}
The call_giant_count
variable is assigned the special value __LKX_REMAIN_UNCHANGED__
to tell the interpreter that this variable needs to remain unchanged during hot reloading.
In order to keep the simplicity of Lkx language, the syntax for defining a structure class with data and functions is not supported in script code, but you can define it in c/c++ code and then export it to the script.
In Lkx’s c/c++ api, there are two types of struct: inner struct and outer struct. There are two main differences between them:
The inner struct is usually created by the Lkx script, and its attribute members can be directly accessed; the outer struct is usually created by the c/c++ host, and its members can only be accessed through member functions in the script.
The inner struct is created by the script, so its GC should also be processed by the Lkx interpreter, while the outer struct is created by the c/c++ host, so the interpreter usually does not need to process its GC.
In addition, you can also assign an outer struct to a data member of an inner struct, the following example shows how to export the inner and outer struct to the Lkx script:
typedef int (*lkx_Func)(lkx_Hub*);
typedef struct lkx_Reg_Func {
const char *name;
lkx_Func func;
} lkx_Reg_Func;
typedef struct lkx_Reg_Attr {
const char *name;
const char *type;
} lkx_Reg_Attr;
int main()
{
lkx_Hub *lkx_hub = lkx_new_hub();
lkx_init(lkx_hub, "./deploy/script/family/hub.lkx");
lkx_load(lkx_hub);
//outer struct CUser
lkx_Reg_Func cuserfuncs[] =
{
{"get_id", lkx__CUser__get_id},
{"get_name", lkx__CUser__get_name},
{"send_msg", lkx__CUser__send_msg},
{"save_data", lkx__CUser__save_data},
{NULL, NULL}
};
//inner struct User
lkx_Reg_Attr userattrs[] = {
{"level", "uin32"},
{"cuser", "CUser"},
{"country", "string"},
{"money", "uint32"},
{"children", "[User]"},
{NULL, NULL}
};
//outer struct CUser
lkx_reg_outer_struct(lkx_hub, "CUser", cuserfuncs);
lkx_reg_struct_metafunc(lkx_hub, "CUser", "__hash__", lkx__CUser____hash__, 1);
lkx_reg_struct_metafunc(lkx_hub, "CUser", "__m_hash__", lkx__CUser____m_hash__, 2);
lkx_reg_struct_metafunc(lkx_hub, "CUser", "__o_compare__", lkx__CUser____o_compare__, 0);
lkx_reg_struct_metafunc(lkx_hub, "CUser", "__mo_compare__", lkx__CUser____mo_compare__, 0);
//inner struct User
lkx_reg_inner_struct(lkx_hub, "User", userattrs);
lkx_reg_struct_metafunc(lkx_hub, "User", "__hash__", lkx__User____hash__, 1);
lkx_reg_struct_metafunc(lkx_hub, "User", "__m_hash__", lkx__User____m_hash__, 2);
lkx_reg_struct_metafunc(lkx_hub, "User", "__o_compare__", lkx__User____o_compare__, 0);
lkx_reg_struct_metafunc(lkx_hub, "User", "__mo_compare__", lkx__User____mo_compare__, 0);
//other code in main......
......
}
Precedence of convention over complex syntax is a design principle of the Lkx language.
In Lkx, the file paths in the import
statement are all prefixed with a unified project related name, which can reduce name conflicts throughout the entire project, as follows:
import "family/father/info.lkx"
import "family/mother/info.lkx"
import "family/brother/info.lkx"
Because shared variables and functions can be accessed in all script files, it is best to add meaningful prefix characters when naming them. For example, in the following lkx script file (path: "family/brother/info.lkx"
), variables are prefixed with “brother_”:
import "family/lib/utils.lkx"
import "family/sister/info.lkx"
shared string brother_name = "bob"
uint32 brother_height = 170
uint32 brother_weight = 55
uint32 brother_age = 31
Lkx does not support the const keyword, if you want to define a constant, you can change the name of the variable to uppercase. In Lkx script, the uppercase variable name represents a constant. This is also the design philosophy of Lkx based on convention:
shared uint32 TALL_CM = 190
shared uint32 FAT_KG = 90
uint32 GIANT_CM = 300
uint32 GIANT_KG = 500
uint32 call_giant_count = 0
All variables declared in the Lkx file will be initialized to their corresponding values by the interpreter:
0
false
[]
{}
{:}
null
When the Lkx interpreter parses the import statement, it will find the corresponding source file according to the __lkx_import_path_from_c_workdir__
set in the hub file. In addition, the hub file must import all other source files of the project, because the first file read by the interpreter is the hub file, and then parse other imported files in turn.
[string] __lkx_import_path_from_c_workdir__ =
[
"./deploy/script/",
"./deploy/script/3rdlib/",
"./deploy/goolge/lkxlib/",
"./deploy/tesla/lkxlib/",
"./deploy/apple/lkxlib/"
]
import "family/father/info.lkx"
import "family/mother/info.lkx"
import "family/brother/info.lkx"
As a scripting language, Lkx itself does not provide any timer functions, but you can easily implement it. As shown below, you can periodically call a script function in the main loop of c/c++to implement a timer:
while(true)
{
sleep(1);
lkx_call(lkx_hub, "timer_func_per_second");
}
lkx_Hub is the execution environment in Lkx, and each script function call runs inside a specific lkx_Hub. Each lkx_hub is an independent sandbox, so you can run different lkx_Hubs on different threads to get multithreading support. However, multithreaded programming is not easy and requires you to be more careful.
if end
blockcimpl
c impl of Lkx, main part of my work.
csimpl
c# impl of Lkx, need your contribution.
javaimpl
java impl of Lkx, need your contribution.
pyimpl
python impl of Lkx, need your contribution.
goimpl
go impl of Lkx, need your contribution.
If you’re interested in Lkx language, join the telegram group: https://t.me/lkx_language