mirror of
https://github.com/alecthomas/chroma.git
synced 2025-02-09 13:23:51 +02:00
Lexer support for the stas language (#699)
This commit is contained in:
l-m
GitHub
parent
be4aebd830
commit
71d4be6869
@ -54,7 +54,7 @@ O | Objective-C, OCaml, Octave, OnesEnterprise, OpenEdge ABL, OpenSCAD, Org Mode
|
||||
P | PacmanConf, Perl, PHP, PHTML, Pig, PkgConfig, PL/pgSQL, plaintext, Pony, PostgreSQL SQL dialect, PostScript, POVRay, PowerShell, Prolog, PromQL, Properties, Protocol Buffer, PSL, Puppet, Python 2, Python
|
||||
Q | QBasic
|
||||
R | R, Racket, Ragel, Raku, react, ReasonML, reg, reStructuredText, Rexx, Ruby, Rust
|
||||
S | SAS, Sass, Scala, Scheme, Scilab, SCSS, Sed, Smalltalk, Smarty, Snobol, Solidity, SPARQL, SQL, SquidConf, Standard ML, Stylus, Svelte, Swift, SYSTEMD, systemverilog
|
||||
S | SAS, Sass, Scala, Scheme, Scilab, SCSS, Sed, Smalltalk, Smarty, Snobol, Solidity, SPARQL, SQL, SquidConf, Standard ML, stas, Stylus, Svelte, Swift, SYSTEMD, systemverilog
|
||||
T | TableGen, TASM, Tcl, Tcsh, Termcap, Terminfo, Terraform, TeX, Thrift, TOML, TradingView, Transact-SQL, Turing, Turtle, Twig, TypeScript, TypoScript, TypoScriptCssData, TypoScriptHtmlData
|
||||
V | VB.net, verilog, VHDL, VHS, VimL, vue
|
||||
W | WDTE
|
||||
|
||||
85
lexers/embedded/stas.xml
Normal file
85
lexers/embedded/stas.xml
Normal file
@ -0,0 +1,85 @@
|
||||
<lexer>
|
||||
<config>
|
||||
<name>stas</name>
|
||||
<filename>*.stas</filename>
|
||||
</config>
|
||||
<rules>
|
||||
<state name="string-double-quoted">
|
||||
<rule pattern="\\.">
|
||||
<token type="LiteralStringEscape"/>
|
||||
</rule>
|
||||
<rule pattern="[^\\"]+">
|
||||
<token type="LiteralString"/>
|
||||
</rule>
|
||||
<rule pattern=""">
|
||||
<token type="LiteralString"/>
|
||||
<pop depth="1"/>
|
||||
</rule>
|
||||
</state>
|
||||
<state name="string-single-quoted">
|
||||
<rule pattern="\\.">
|
||||
<token type="LiteralStringEscape"/>
|
||||
</rule>
|
||||
<rule pattern="[^\\']+">
|
||||
<token type="LiteralString"/>
|
||||
</rule>
|
||||
<rule pattern="'">
|
||||
<token type="LiteralString"/>
|
||||
<pop depth="1"/>
|
||||
</rule>
|
||||
</state>
|
||||
<state name="string-char-literal">
|
||||
<rule pattern="\\.">
|
||||
<token type="LiteralStringEscape"/>
|
||||
</rule>
|
||||
<rule pattern="[^\\`]+">
|
||||
<token type="LiteralString"/>
|
||||
</rule>
|
||||
<rule pattern="`">
|
||||
<token type="LiteralStringChar"/>
|
||||
<pop depth="1"/>
|
||||
</rule>
|
||||
</state>
|
||||
<state name="root">
|
||||
<rule pattern="(\n|\s)+">
|
||||
<token type="Text"/>
|
||||
</rule>
|
||||
<rule pattern="(?<!\S)(fn|argc|argv|swap|dup|over|over2|rot|rot4|drop|w8|w16|w32|w64|r8|r16|r32|r64|syscall0|syscall1|syscall2|syscall3|syscall4|syscall5|syscall6|_breakpoint|assert|const|auto|reserve|pop|include|addr|if|else|elif|while|break|continue|ret)(?!\S)">
|
||||
<token type="Keyword"/>
|
||||
</rule>
|
||||
<rule pattern="(?<!\S)(\+|\-|\*|\/|\%|\%\%|\+\+|\-\-|>>|<<)(?!\S)">
|
||||
<token type="Operator"/>
|
||||
</rule>
|
||||
<rule pattern="(?<!\S)(\=|\!\=|>|<|>\=|<\=|>s|<s|>\=s|<\=s)(?!\S)">
|
||||
<token type="Operator"/>
|
||||
</rule>
|
||||
<rule pattern="(?<!\S)(\&|\||\^|\~|\!|-\>)(?!\S)">
|
||||
<token type="Operator"/>
|
||||
</rule>
|
||||
<rule pattern="(?<!\S)\-?(\d+)(?!\S)">
|
||||
<token type="LiteralNumber"/>
|
||||
</rule>
|
||||
<rule pattern="(?<!\S);.*(\S|\n)">
|
||||
<token type="Comment"/>
|
||||
</rule>
|
||||
<rule pattern="'">
|
||||
<token type="LiteralString"/>
|
||||
<push state="string-single-quoted"/>
|
||||
</rule>
|
||||
<rule pattern=""">
|
||||
<token type="LiteralString"/>
|
||||
<push state="string-double-quoted"/>
|
||||
</rule>
|
||||
<rule pattern="`">
|
||||
<token type="LiteralStringChar"/>
|
||||
<push state="string-char-literal"/>
|
||||
</rule>
|
||||
<rule pattern="(?<!\S)[{}](?!\S)">
|
||||
<token type="Punctuation"/>
|
||||
</rule>
|
||||
<rule pattern="(?<!\S)[^\s]+(?!\S)">
|
||||
<token type="Name"/>
|
||||
</rule>
|
||||
</state>
|
||||
</rules>
|
||||
</lexer>
|
||||
402
lexers/testdata/stas.actual
vendored
Normal file
402
lexers/testdata/stas.actual
vendored
Normal file
@ -0,0 +1,402 @@
|
||||
include 'std.stas'
|
||||
|
||||
reserve debug_symbols 1
|
||||
reserve verbose_mode 1
|
||||
auto backend_type 1
|
||||
|
||||
const StasBackend.fasm { 1 }
|
||||
const StasBackend.nasm { 2 }
|
||||
|
||||
; (StasBackend -- str len)
|
||||
fn StasBackend.to_str 1 2 {
|
||||
dup StasBackend.fasm = if {
|
||||
"fasm"
|
||||
} elif dup StasBackend.nasm = {
|
||||
"nasm"
|
||||
} else {
|
||||
0 0 0 assert -> 'unreachable'
|
||||
}
|
||||
rot rot drop
|
||||
}
|
||||
|
||||
include 'src/stringbuffer.stas' ; handling strings
|
||||
include 'src/tokens.stas' ; stas token definitions
|
||||
include 'src/util.stas' ; utility functions, error handling
|
||||
include 'src/scanner.stas' ; lexer/scanner, creates tokens
|
||||
include 'src/parserdefs.stas' ; stas parser definitions, very large file
|
||||
include 'src/eval.stas' ; constant evaluation
|
||||
include 'src/parser.stas' ; stas parser, creates IR instructions
|
||||
include 'src/write.stas' ; buffers + writing to files
|
||||
include 'src/dce.stas' ; dead code elimination compiler pass
|
||||
include 'src/x86.stas' ; stas codegen definitions and reg allocator
|
||||
include 'src/gen.stas' ; stas code generator, creates x86_64 asm
|
||||
|
||||
fn usage 0 0 {
|
||||
"stas 0.1.1 Copyright (C) 2022 l-m.dev\n\n" eputs
|
||||
"USAGE: ./stas [OPTIONS] [FILE]\n\n" eputs
|
||||
|
||||
" -o <output> Specify '-o -' to dump assembly to stdout\n" eputs
|
||||
" -g Debug info. Most effective with the `nasm` backend\n" eputs
|
||||
" -b <backend> Assemblers `nasm` or `fasm` as compiler backend\n" eputs
|
||||
" -r Execute file after compiling. Arguments after this\n" eputs
|
||||
" switch will ignored and passed to the program\n" eputs
|
||||
" -v, --verbose Activate verbose mode\n" eputs
|
||||
" --dump-tok Dump token information after scanning stage\n" eputs
|
||||
" --dump-ir Dump intermediate representation after parsing stage\n" eputs
|
||||
" -h, --help Show this message\n\n" eputs
|
||||
}
|
||||
|
||||
fn help_and_exit 0 0 {
|
||||
usage
|
||||
0 exit
|
||||
}
|
||||
|
||||
fn usage_and_exit 0 0 {
|
||||
usage
|
||||
1 exit
|
||||
}
|
||||
|
||||
fn usage_msg_and_exit 2 0 {
|
||||
error.generic_fatal_noexit
|
||||
usage_and_exit
|
||||
}
|
||||
|
||||
fn parse_backend_type 2 0 {
|
||||
over over "fasm" streq if {
|
||||
StasBackend.fasm pop backend_type
|
||||
} elif over over "nasm" streq {
|
||||
StasBackend.nasm pop backend_type
|
||||
} else {
|
||||
"unknown backend" usage_msg_and_exit
|
||||
}
|
||||
drop drop
|
||||
}
|
||||
|
||||
const sizeof(fasm_arg_buf) { sizeof(u64) 32 * }
|
||||
|
||||
; (infile.str infile.len outfile.str outfile.len is_blocking)
|
||||
fn execute_backend 5 0 {
|
||||
auto is_blocking 1 pop is_blocking
|
||||
auto outfile 2 pop outfile
|
||||
auto infile 2 pop infile
|
||||
|
||||
reserve arg_buf sizeof(fasm_arg_buf)
|
||||
|
||||
backend_type StasBackend.fasm = if {
|
||||
arg_buf dup "fasm" drop w64
|
||||
sizeof(u64) + dup infile drop w64
|
||||
sizeof(u64) + dup outfile drop w64
|
||||
sizeof(u64) + dup "-m" drop w64
|
||||
sizeof(u64) + dup "1048576" drop w64
|
||||
sizeof(u64) + NULL w64
|
||||
|
||||
"/usr/bin/fasm"
|
||||
} elif backend_type StasBackend.nasm = {
|
||||
arg_buf dup "nasm" drop w64
|
||||
sizeof(u64) + dup infile drop w64
|
||||
sizeof(u64) + dup "-o" drop w64
|
||||
sizeof(u64) + dup outfile drop w64
|
||||
sizeof(u64) + dup "-O0" drop w64
|
||||
sizeof(u64) + dup "-felf64" drop w64
|
||||
|
||||
debug_symbols r8 if {
|
||||
sizeof(u64) + dup "-Fdwarf" drop w64
|
||||
sizeof(u64) + dup "-g" drop w64
|
||||
}
|
||||
|
||||
sizeof(u64) + NULL w64
|
||||
"/usr/bin/nasm"
|
||||
}
|
||||
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
"`" eputs
|
||||
arg_buf argp_print
|
||||
"`\n" eputs
|
||||
}
|
||||
|
||||
arg_buf is_blocking child_execve_and_shut_up
|
||||
}
|
||||
|
||||
const ArgParseMode.none { 0 }
|
||||
const ArgParseMode.output { 1 }
|
||||
const ArgParseMode.backend { 2 }
|
||||
|
||||
fn main 0 0 {
|
||||
argc 1 = if {
|
||||
usage_and_exit
|
||||
}
|
||||
|
||||
reserve dump_ir 1
|
||||
reserve dump_tok 1
|
||||
reserve to_stdout 1
|
||||
|
||||
to_stdout 0 w8
|
||||
dump_ir 0 w8
|
||||
dump_tok 0 w8
|
||||
|
||||
auto run_exec_arg 1
|
||||
0 pop run_exec_arg
|
||||
|
||||
UINT64_MAX pop fwrite_buffer.fd_loc
|
||||
StasBackend.fasm pop backend_type
|
||||
|
||||
auto argparse_mode 1
|
||||
auto argstr 2
|
||||
|
||||
auto out_file 2
|
||||
auto in_file 2
|
||||
NULL 0 pop out_file
|
||||
NULL 0 pop in_file
|
||||
|
||||
ArgParseMode.none pop argparse_mode
|
||||
debug_symbols false w8
|
||||
|
||||
1
|
||||
while dup argc < {
|
||||
dup args[] pop argstr
|
||||
|
||||
argstr "-o" streq if {
|
||||
argparse_mode ArgParseMode.none != if {
|
||||
usage_and_exit
|
||||
}
|
||||
ArgParseMode.output pop argparse_mode
|
||||
} elif argstr "-b" streq {
|
||||
argparse_mode ArgParseMode.none != if {
|
||||
usage_and_exit
|
||||
}
|
||||
ArgParseMode.backend pop argparse_mode
|
||||
} elif argstr "-g" streq {
|
||||
argparse_mode ArgParseMode.none != if {
|
||||
usage_and_exit
|
||||
}
|
||||
debug_symbols r8 if {
|
||||
usage_and_exit
|
||||
}
|
||||
debug_symbols true w8
|
||||
} elif argstr "--verbose" streq argstr "-v" streq | {
|
||||
argparse_mode ArgParseMode.none != if {
|
||||
usage_and_exit
|
||||
}
|
||||
verbose_mode r8 if {
|
||||
usage_and_exit
|
||||
}
|
||||
verbose_mode true w8
|
||||
} elif argstr "-r" streq {
|
||||
argparse_mode ArgParseMode.none != if {
|
||||
usage_and_exit
|
||||
}
|
||||
pop run_exec_arg
|
||||
argc
|
||||
} elif argstr "--help" streq argstr "-h" streq | {
|
||||
help_and_exit
|
||||
} elif argstr "--dump-ir" streq argstr "-h" streq | {
|
||||
argparse_mode ArgParseMode.none != dump_ir r8 | dump_tok r8 | if {
|
||||
usage_and_exit
|
||||
}
|
||||
dump_ir true w8
|
||||
} elif argstr "--dump-tok" streq argstr "-h" streq | {
|
||||
argparse_mode ArgParseMode.none != dump_ir r8 | dump_tok r8 | if {
|
||||
usage_and_exit
|
||||
}
|
||||
dump_tok true w8
|
||||
} else {
|
||||
argparse_mode ArgParseMode.none = if {
|
||||
in_file drop NULL != if {
|
||||
usage_and_exit
|
||||
}
|
||||
argstr pop in_file
|
||||
} elif argparse_mode ArgParseMode.output = {
|
||||
out_file drop NULL != if {
|
||||
usage_and_exit
|
||||
}
|
||||
argstr pop out_file
|
||||
} elif argparse_mode ArgParseMode.backend = {
|
||||
argstr parse_backend_type
|
||||
} else {
|
||||
0 assert
|
||||
}
|
||||
ArgParseMode.none pop argparse_mode
|
||||
}
|
||||
++
|
||||
}
|
||||
drop
|
||||
|
||||
argparse_mode ArgParseMode.none != if {
|
||||
argparse_mode ArgParseMode.output = if {
|
||||
"supply output file" usage_msg_and_exit
|
||||
} elif argparse_mode ArgParseMode.backend = {
|
||||
"supply backend type" usage_msg_and_exit
|
||||
}
|
||||
}
|
||||
|
||||
in_file drop NULL = if {
|
||||
"supply stas file" usage_msg_and_exit
|
||||
}
|
||||
|
||||
out_file drop NULL = if {
|
||||
debug_symbols r8 backend_type StasBackend.nasm = | if {
|
||||
"a.o"
|
||||
} else {
|
||||
"a.out"
|
||||
}
|
||||
pop out_file
|
||||
} else {
|
||||
out_file "-" streq if {
|
||||
to_stdout true w8
|
||||
}
|
||||
}
|
||||
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
"scanning file `" eputs in_file eputs "`\n" eputs
|
||||
}
|
||||
|
||||
log.time.start
|
||||
in_file stas.scan_file
|
||||
"scanning took " log.time.end
|
||||
|
||||
dump_tok r8 if {
|
||||
token_stream.dump
|
||||
ret
|
||||
}
|
||||
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
"parsing " eputs token_stream.len eputu " tokens\n" eputs
|
||||
}
|
||||
log.time.start
|
||||
stas.parse
|
||||
"parsing took " log.time.end
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
functions.len eputu " functions, " eputs label_c ++ eputu " labels\n" eputs
|
||||
log.msg.start
|
||||
global_var_context.len eputu " global variables, " eputs toplevel_constants.len eputu " constants\n" eputs
|
||||
}
|
||||
|
||||
dump_ir r8 if {
|
||||
ir_stream.dump
|
||||
ret
|
||||
}
|
||||
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
"dce pass started\n" eputs
|
||||
}
|
||||
log.time.start
|
||||
stas.dce
|
||||
"dce took " log.time.end
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
used_functions eputu " used functions, of which " eputs inlined_functions eputu " are eligible for inline\n" eputs
|
||||
log.msg.start
|
||||
slits.len eputu " string literals\n" eputs
|
||||
}
|
||||
|
||||
auto out_file_asm_sv 1
|
||||
|
||||
to_stdout r8 ! if {
|
||||
out_file new_string_view
|
||||
dup ".tmp" push_string_view
|
||||
dup pop out_file_asm_sv
|
||||
|
||||
string_view_to_str fd_new_file_for_writing
|
||||
pop fwrite_buffer.fd_loc
|
||||
} else {
|
||||
stdout pop fwrite_buffer.fd_loc
|
||||
}
|
||||
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
"generating code from " eputs ir_stream.len eputu " IR instructions\n" eputs
|
||||
}
|
||||
log.time.start
|
||||
in_file stas.gen
|
||||
"gen took " log.time.end
|
||||
|
||||
to_stdout r8 if {
|
||||
ret
|
||||
}
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
"generated " eputs
|
||||
|
||||
fwrite_buffer.fd_loc fd_stat_size
|
||||
|
||||
dup 1024 / 0 > if {
|
||||
eputu " KiBs of code\n" eputs
|
||||
} else {
|
||||
eputu " bytes of code\n" eputs
|
||||
}
|
||||
}
|
||||
|
||||
fwrite_buffer.fd_loc close 0 <s if {
|
||||
"FATAL: Failed to close file descriptor\n" eputs
|
||||
1 exit
|
||||
}
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
"wrote code to `" eputs out_file_asm_sv string_view_to_str eputs "`\n" eputs
|
||||
}
|
||||
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
"executing assembler backend `" eputs backend_type StasBackend.to_str eputs "`\n" eputs
|
||||
}
|
||||
|
||||
log.time.start
|
||||
out_file_asm_sv string_view_to_str out_file run_exec_arg 0 != execute_backend
|
||||
"backend took " log.time.end
|
||||
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
"created binary `" eputs out_file eputs "`\n" eputs
|
||||
}
|
||||
|
||||
debug_symbols r8 backend_type StasBackend.nasm = | if {
|
||||
ret
|
||||
}
|
||||
|
||||
run_exec_arg 0 != if {
|
||||
reserve null_p sizeof(u64)
|
||||
null_p NULL w64
|
||||
|
||||
auto argp 1
|
||||
|
||||
{
|
||||
string_buffer string_buffer.len + pop argp
|
||||
|
||||
out_file drop string_buffer.generic_append_u64
|
||||
|
||||
run_exec_arg 1 +
|
||||
while dup argc < {
|
||||
dup sizeof(u64) * argv + r64 string_buffer.generic_append_u64
|
||||
++
|
||||
}
|
||||
drop
|
||||
|
||||
NULL string_buffer.generic_append_u64
|
||||
}
|
||||
|
||||
out_file drop
|
||||
argp
|
||||
null_p
|
||||
|
||||
verbose_mode r8 if {
|
||||
log.msg.start
|
||||
|
||||
"exceve binary `" eputs
|
||||
|
||||
argp
|
||||
argp_print
|
||||
|
||||
"`\n" eputs
|
||||
}
|
||||
|
||||
execve 0 <s if {
|
||||
"FATAL: Could not execve file\n" eputs
|
||||
1 exit
|
||||
}
|
||||
}
|
||||
}
|
||||
2081
lexers/testdata/stas.expected
vendored
Normal file
2081
lexers/testdata/stas.expected
vendored
Normal file
File diff suppressed because it is too large
Load Diff
Loading…
x
Reference in New Issue
Block a user