Adding MIPS Lexer

pygments/lexers/_mapping.py

@@ -273,6 +273,7 @@
     'LuaLexer': ('pygments.lexers.scripting', 'Lua', ('lua',), ('*.lua', '*.wlua'), ('text/x-lua', 'application/x-lua')),
     'MCFunctionLexer': ('pygments.lexers.mcfunction', 'MCFunction', ('mcfunction', 'mcf'), ('*.mcfunction',), ('text/mcfunction',)),
     'MIMELexer': ('pygments.lexers.mime', 'MIME', ('mime',), (), ('multipart/mixed', 'multipart/related', 'multipart/alternative')),
+    'MIPSLexer': ('pygments.lexers.mips', 'MIPS', ('mips',), ('*.s', '*.asm', '*.mips', '*.S', '*.ASM', '*.MIPS'), ()),
     'MOOCodeLexer': ('pygments.lexers.scripting', 'MOOCode', ('moocode', 'moo'), ('*.moo',), ('text/x-moocode',)),
     'MSDOSSessionLexer': ('pygments.lexers.shell', 'MSDOS Session', ('doscon',), (), ()),
     'Macaulay2Lexer': ('pygments.lexers.macaulay2', 'Macaulay2', ('macaulay2',), ('*.m2',), ()),

pygments/lexers/mips.py

@@ -0,0 +1,126 @@
+"""
+    pygments.lexers.mips
+    ~~~~~~~~~~~~~~~~~~~~
+    Lexers for MIPS assembly.
+    :copyright: Copyright 2006-2022 by the Pygments team, see AUTHORS.
+    :license: BSD, see LICENSE for details.
+"""
+
+import re
+from pygments.lexer import RegexLexer, words
+from pygments.token import Whitespace, Comment, String, Keyword, Name, Text
+
+__all__ = ["MIPSLexer"]
+
+
+class MIPSLexer(RegexLexer):
+    """
+    A MIPS Assembly Lexer.
+
+    Based on the Emacs major mode by hlissner:
+    https://github.com/hlissner/emacs-mips-mode
+    """
+
+    name = 'MIPS'
+    aliases = ['mips']
+    # TODO: add '*.s' and '*.asm', which will require designing an analyse_text
+    # method for this lexer and refactoring those from Gas and Nasm in order to
+    # have relatively reliable detection
+    filenames = ['*.mips', '*.MIPS']
+
+    keywords = [
+        # Arithmetic insturctions
+        "add", "sub", "subu", "addi", "subi", "addu", "addiu",
+        # Multiplication/division
+        "mul", "mult", "multu", "mulu", "madd", "maddu", "msub", "msubu", "div", "divu",
+        # Bitwise operations
+        "and", "or", "nor", "xor", "andi", "ori", "xori", "clo", "clz",
+        # Shifts
+        "sll", "srl", "sllv", "srlv", "sra", "srav",
+        # Comparisons
+        "slt", "sltu", "slti", "sltiu",
+        # Move data
+        "mfhi", "mthi", "mflo", "mtlo", "movn", "movz", "movf", "movt",
+        # Jump
+        "j", "jal", "jalr", "jr",
+        # branch
+        "bc1f", "bc1t", "beq", "bgez", "bgezal", "bgtz", "blez", "bltzal", "bltz", "bne",
+        # Load
+        "lui", "lb", "lbu", "lh", "lhu", "lw", "lwcl", "lwl", "lwr",
+        # Store
+        "sb", "sh", "sw", "swl", "swr", # coproc: swc1 sdc1
+        # Concurrent load/store
+        "ll", "sc",
+        # Trap handling
+        "teq", "teqi", "tne", "tneqi", "tge", "tgeu", "tgei", "tgeiu", "tlt", "tltu", "tlti",
+        "tltiu",
+        # Exception / Interrupt
+        "eret", "break", "bop", "syscall",
+        #--- Floats -----------------------------------------------------
+        # Arithmetic
+        "add.s", "add.d", "sub.s", "sub.d", "mul.s", "mul.d", "div.s", "div.d", "neg.d",
+        "neg.s",
+        # Comparison
+        "c.e.d", "c.e.s", "c.le.d", "c.le.s", "c.lt.s", "c.lt.d", # "c.gt.s", "c.gt.d",
+        "madd.s", "madd.d", "msub.s", "msub.d",
+        # Move Floats
+        "mov.d", "move.s", "movf.d", "movf.s", "movt.d", "movt.s", "movn.d", "movn.s",
+        "movnzd", "movz.s", "movz.d",
+        # Conversion
+        "cvt.d.s", "cvt.d.w", "cvt.s.d", "cvt.s.w", "cvt.w.d", "cvt.w.s", "trunc.w.d",
+        "trunc.w.s",
+        # Math
+        "abs.s", "abs.d", "sqrt.s", "sqrt.d", "ceil.w.d", "ceil.w.s", "floor.w.d",
+        "floor.w.s", "round.w.d", "round.w.s",
+    ]
+
+    pseudoinstructions = [
+        # Arithmetic & logical
+        "rem", "remu", "mulo", "mulou", "abs", "neg", "negu", "not", "rol", "ror",
+        # branches
+        "b", "beqz", "bge", "bgeu", "bgt", "bgtu", "ble", "bleu", "blt", "bltu", "bnez",
+        # loads
+        "la", "li", "ld", "ulh", "ulhu", "ulw",
+        # Store
+        "sd", "ush", "usw",
+        # move
+        "move", # coproc: "mfc1.d",
+        # comparisons
+        "sgt", "sgtu", "sge", "sgeu", "sle", "sleu", "sne", "seq",
+        #--- Floats -----------------------------------------------------
+        # load-store
+        "l.d", "l.s", "s.d", "s.s",
+    ]
+
+    directives = [
+        ".align", ".ascii", ".asciiz", ".byte", ".data", ".double", ".extern", ".float",
+        ".globl", ".half", ".kdata", ".ktext", ".space", ".text", ".word",
+    ]
+
+    deprecated = [
+        "beql", "bnel", "bgtzl", "bgezl", "bltzl", "blezl", "bltzall", "bgezall",
+    ]
+
+    tokens = {
+        'root': [
+            (r'\s+', Whitespace),
+            (r'#.*', Comment),
+            (r'"', String, 'string'),
+            (r'-?[0-9]+?', Keyword.Constant),
+            (r'\w*:', Name.Function),
+            (words(deprecated, suffix=r'\b'), Keyword.Pseudo), # need warning face
+            (words(pseudoinstructions, suffix=r'\b'), Name.Variable),
+            (words(keywords, suffix=r'\b'), Keyword),
+            (r'[slm][ftwd]c[0-9]([.]d)?', Keyword),
+            (r'\$(f?[0-2][0-9]|f?3[01]|[ft]?[0-9]|[vk][01]|a[0-3]|s[0-7]|[gsf]p|ra|at|zero)', Keyword.Type),
+            (words(directives, suffix=r'\b'), Name.Entity), # Preprocessor?
+            (r':|,|;|\{|\}|=>|@|\$|=', Name.Builtin),
+            (r'\w+', Text),
+            (r'.', Text),
+        ],
+        'string': [
+            (r'\\.', String.Escape),
+            (r'"', String, '#pop'),
+            (r'[^\\"]+', String),
+        ],
+    }

tests/examplefiles/mips/function_with_stack.mips

@@ -0,0 +1,84 @@
+# Simple routine to demo functions
+# USING a stack in this example to preserve
+# values of calling function
+
+# ------------------------------------------------------------------
+
+    .text
+
+    .globl  main
+main:
+    # Register assignments
+    # $s0 = x
+    # $s1 = y
+
+    # Initialize registers
+    lw  $s0, x      # Reg $s0 = x
+    lw  $s1, y      # Reg $s1 = y
+
+    # Call function
+    move    $a0, $s0    # Argument 1: x ($s0)
+    jal fun     # Save current PC in $ra, and jump to fun
+    move    $s1,$v0     # Return value saved in $v0. This is y ($s1)
+
+    # Print msg1
+    li  $v0, 4      # print_string syscall code = 4
+    la  $a0, msg1
+    syscall
+
+    # Print result (y)
+    li  $v0,1       # print_int syscall code = 1
+    move    $a0, $s1    # Load integer to print in $a0
+    syscall
+
+    # Print newline
+    li  $v0,4       # print_string syscall code = 4
+    la  $a0, lf
+    syscall
+
+    # Exit
+    li  $v0,10      # exit
+    syscall
+
+# ------------------------------------------------------------------
+
+    # FUNCTION: int fun(int a)
+    # Arguments are stored in $a0
+    # Return value is stored in $v0
+    # Return address is stored in $ra (put there by jal instruction)
+    # Typical function operation is:
+
+fun:    # This function overwrites $s0 and $s1
+    # We should save those on the stack
+    # This is PUSH'ing onto the stack
+    addi $sp,$sp,-4     # Adjust stack pointer
+    sw $s0,0($sp)       # Save $s0
+    addi $sp,$sp,-4     # Adjust stack pointer
+    sw $s1,0($sp)       # Save $s1
+
+    # Do the function math
+    li $s0, 3
+    mul $s1,$s0,$a0     # s1 = 3*$a0  (i.e. 3*a)
+    addi $s1,$s1,5      # 3*a+5
+
+    # Save the return value in $v0
+    move $v0,$s1
+
+    # Restore saved register values from stack in opposite order
+    # This is POP'ing from the stack
+    lw $s1,0($sp)       # Restore $s1
+    addi $sp,$sp,4      # Adjust stack pointer
+    lw $s0,0($sp)       # Restore $s0
+    addi $sp,$sp,4      # Adjust stack pointer
+
+    # Return from function
+    jr $ra          # Jump to addr stored in $ra
+
+# ------------------------------------------------------------------
+
+    # Start .data segment (data!)
+    .data
+x:  .word 5
+y:  .word 0
+msg1:   .asciiz "y="
+lf:     .asciiz "\n"