Add ARM assembly lexer (#1057)

This commit adds a lexer for ARM assembly code. This is for the syntax 
used by ARM DS-5, Keil, RealView, ADS, SDT, objasm, asasm and aasm 
toolchains and is **not** the same as that shared by GNU AS and Clang's 
integrated assembler.

lib/rouge/demos/armasm

@@ -0,0 +1,12 @@
+            GET     common.s
+
+RetVal      *       0x123 :SHL: 4
+
+            AREA    |Area$$Name|, CODE, READONLY
+
+MyFunction  ROUT             ; This is a comment
+            ASSERT  RetVal <> 0
+1           MOVW    r0, #RetVal
+            BX      lr
+
+            END

lib/rouge/lexers/armasm.rb

@@ -0,0 +1,143 @@
+# -*- coding: utf-8 -*- #
+# frozen_string_literal: true
+
+module Rouge
+  module Lexers
+    class ArmAsm < RegexLexer
+      title "ArmAsm"
+      desc "Arm assembly syntax"
+      tag 'armasm'
+      filenames '*.s'
+
+      def self.preproc_keyword
+        @preproc_keyword ||= %w(
+          define elif else endif error if ifdef ifndef include line pragma undef warning
+        )
+      end
+
+      def self.file_directive
+        @file_directive ||= %w(
+          BIN GET INCBIN INCLUDE LNK
+        )
+      end
+
+      def self.general_directive
+        @general_directive ||= %w(
+          ALIAS ALIGN AOF AOUT AREA ARM ASSERT ATTR CN CODE16 CODE32 COMMON CP
+          DATA DCB DCD DCDO DCDU DCFD DCFDU DCFH DCFHU DCFS DCFSU DCI DCI.N DCI.W
+          DCQ DCQU DCW DCWU DN ELIF ELSE END ENDFUNC ENDIF ENDP ENTRY EQU EXPORT
+          EXPORTAS EXTERN FIELD FILL FN FRAME FUNCTION GBLA GBLL GBLS GLOBAL IF
+          IMPORT INFO KEEP LCLA LCLL LCLS LEADR LEAF LTORG MACRO MAP MEND MEXIT
+          NOFP OPT ORG PRESERVE8 PROC QN RELOC REQUIRE REQUIRE8 RLIST RN ROUT
+          SETA SETL SETS SN SPACE STRONG SUBT THUMB THUMBX TTL WEND WHILE
+          \[ \] [|!#*=%&^]
+        )
+      end
+
+      def self.shift_or_condition
+        @shift_or_condition ||= %w(
+          ASR LSL LSR ROR RRX AL CC CS EQ GE GT HI HS LE LO LS LT MI NE PL VC VS
+          asr lsl lsr ror rrx al cc cs eq ge gt hi hs le lo ls lt mi ne pl vc vs
+        )
+      end
+
+      def self.builtin
+        @builtin ||= %w(
+          ARCHITECTURE AREANAME ARMASM_VERSION CODESIZE COMMANDLINE CONFIG CPU
+          ENDIAN FALSE FPIC FPU INPUTFILE INTER LINENUM LINENUMUP LINENUMUPPER
+          OBJASM_VERSION OPT PC PCSTOREOFFSET REENTRANT ROPI RWPI TRUE VAR
+        )
+      end
+
+      def self.operator
+        @operator ||= %w(
+          AND BASE CC CC_ENCODING CHR DEF EOR FATTR FEXEC FLOAD FSIZE INDEX LAND
+          LEFT LEN LEOR LNOT LOR LOWERCASE MOD NOT OR RCONST REVERSE_CC RIGHT ROL
+          ROR SHL SHR STR TARGET_ARCH_[0-9A-Z_]+ TARGET_FEATURE_[0-9A-Z_]+
+          TARGET_FPU_[A-Z_] TARGET_PROFILE_[ARM] UAL UPPERCASE
+        )
+      end
+
+      state :root do
+        rule %r/\n/, Text
+        rule %r/^[ \t]*#[ \t]*(?:(?:#{ArmAsm.preproc_keyword.join('|')})(?:[ \t].*)?)?\n/, Comment::Preproc
+        rule %r/[ \t]+/, Text, :command
+        rule %r/;.*/, Comment
+        rule %r/\$[a-z_]\w*\.?/i, Name::Namespace # variable substitution or macro argument
+        rule %r/\w+|\|[^|\n]+\|/, Name::Label
+      end
+
+      state :command do
+        rule %r/\n/, Text, :pop!
+        rule %r/[ \t]+/ do |m|
+          token Text
+          goto :args
+        end
+        rule %r/;.*/, Comment, :pop!
+        rule %r/(?:#{ArmAsm.file_directive.join('|')})\b/ do |m|
+          token Keyword
+          goto :filespec
+        end
+        rule %r/(?:#{ArmAsm.general_directive.join('|')})(?=[; \t\n])/, Keyword
+        rule %r/(?:[A-Z][\dA-Z]*|[a-z][\da-z]*)(?:\.[NWnw])?(?:\.[DFIPSUdfipsu]?(?:8|16|32|64)?){,3}\b/, Name::Builtin # rather than attempt to list all opcodes, rely on all-uppercase or all-lowercase rule
+        rule %r/[a-z_]\w*|\|[^|\n]+\|/i, Name::Function # probably a macro name
+        rule %r/\$[a-z]\w*\.?/i, Name::Namespace
+      end
+
+      state :args do
+        rule %r/\n/, Text, :pop!
+        rule %r/[ \t]+/, Text
+        rule %r/;.*/, Comment, :pop!
+        rule %r/(?:#{ArmAsm.shift_or_condition.join('|')})\b/, Name::Builtin
+        rule %r/[a-z_]\w*|\|[^|\n]+\|/i, Name::Variable # various types of symbol
+        rule %r/%[bf]?[at]?\d+(?:[a-z_]\w*)?/i, Name::Label
+        rule %r/(?:&|0x)\h+(?!p)/i, Literal::Number::Hex
+        rule %r/(?:&|0x)[.\h]+(?:p[-+]?\d+)?/i, Literal::Number::Float
+        rule %r/0f_\h{8}|0d_\h{16}/i, Literal::Number::Float
+        rule %r/(?:2_[01]+|3_[0-2]+|4_[0-3]+|5_[0-4]+|6_[0-5]+|7_[0-6]+|8_[0-7]+|9_[0-8]+|\d+)(?!e)/i, Literal::Number::Integer
+        rule %r/(?:2_[.01]+|3_[.0-2]+|4_[.0-3]+|5_[.0-4]+|6_[.0-5]+|7_[.0-6]+|8_[.0-7]+|9_[.0-8]+|[.\d]+)(?:e[-+]?\d+)?/i, Literal::Number::Float
+        rule %r/[@:](?=[ \t]*(?:8|16|32|64|128|256)[^\d])/, Operator
+        rule %r/[.@]|\{(?:#{ArmAsm.builtin.join('|')})\}/, Name::Constant
+        rule %r/[-!#%&()*+,\/<=>?^{|}]|\[|\]|!=|&&|\/=|<<|<=|<>|==|><|>=|>>|\|\||:(?:#{ArmAsm.operator.join('|')}):/, Operator
+        rule %r/\$[a-z]\w*\.?/i, Name::Namespace
+        rule %r/'/ do |m|
+          token Literal::String::Char
+          goto :singlequoted
+        end
+        rule %r/"/ do |m|
+          token Literal::String::Double
+          goto :doublequoted
+        end
+      end
+
+      state :singlequoted do
+        rule %r/\n/, Text, :pop!
+        rule %r/\$\$/, Literal::String::Char
+        rule %r/\$[a-z]\w*\.?/i, Name::Namespace
+        rule %r/'/ do |m|
+          token Literal::String::Char
+          goto :args
+        end
+        rule %r/[^$'\n]+/, Literal::String::Char
+      end
+
+      state :doublequoted do
+        rule %r/\n/, Text, :pop!
+        rule %r/\$\$/, Literal::String::Double
+        rule %r/\$[a-z]\w*\.?/i, Name::Namespace
+        rule %r/"/ do |m|
+          token Literal::String::Double
+          goto :args
+        end
+        rule %r/[^$"\n]+/, Literal::String::Double
+      end
+
+      state :filespec do
+        rule %r/\n/, Text, :pop!
+        rule %r/\$\$/, Literal::String::Other
+        rule %r/\$[a-z]\w*\.?/i, Name::Namespace
+        rule %r/[^$\n]+/, Literal::String::Other
+      end
+    end
+  end
+end

spec/lexers/armasm_spec.rb

@@ -0,0 +1,14 @@
+# -*- coding: utf-8 -*- #
+# frozen_string_literal: true
+
+describe Rouge::Lexers::ArmAsm do
+  let(:subject) { Rouge::Lexers::ArmAsm.new }
+
+  describe 'guessing' do
+    include Support::Guessing
+
+    it 'guesses by filename' do
+      assert_guess :filename => 'foo.s'
+    end
+  end
+end

spec/visual/samples/armasm

@@ -0,0 +1,90 @@
+#ifndef Version
+    #include "Version.h"
+#endif
+
+; it shouldn't have any problem with apostrophes in comments
+; or "quotation marks" or more ;s
+
+        GET     otherfile.s
+
+        ^       0,r12 ; typical data structure layout
+offset1 #       4
+        #       4 ; sometimes no label assigned
+offset2 #       0
+size    *       :INDEX:@
+
+ IF ?offset1 = 4
+        ! 1, "This is an assembly-time assert"
+ ELIF -1<>:NOT:0
+        ! 1, "This is another assert"
+ ELSE
+        ! 0, "But this is a warning", 1
+ ENDIF
+
+my_r0   RN      0 ; non-standard register name symbols
+
+sixteen * 4 * 4
+
+        GBLS    VBar
+VBar    SETS    "|"
+
+        MACRO
+$label  MyMacro$suffix $arg1 = default, $arg2
+        LCLS    thing
+ [ "$arg2"=""
+thing   SETS    "wibble ; this isn't a comment"
+ |
+thing   SETS    $arg2
+ ]
+        LCLL    boolean
+boolean SETL    {TRUE}
+        EXPORT  $arg1
+$arg1
+$thing  MOV     pc, #0
+        MEND
+
+
+        ORG     0
+
+        ARM
+
+        AREA    |Area$$Name|, CODE, READONLY
+
+MyFunc  ROUT
+
+just_a_label
+label_and ; comment
+        LDR     my_r0, =just_a_statement
+label   AND     r0, r1, r2 ; and a comment
+        MOV     r0, r1, r2, LSL r3
+        MOV.W   r0, r1, r2, ROR #1
+        ADDS.N  r0, r1
+        IT      MI
+        MOVMI   r0, r1, RRX
+        UND     #0x10-&10
+        PUSH    {r0,r2-r4,ip,lr}
+        LDF     F0, =25E-1
+        VMOVEQ.F64 d0, #-.7e1
+
+Symbols_may_contain_d161ts_and_underscores
+but_must_start_with_a_letter
+and_are_case_SenSitIve
+01 ; this, by contrast, is a local label
+
+        LDR     r0, here
+        B       .+8
+here    DATA
+        DCD     1
+        BNE     %BT01
+
+        DCB 1,2,3,'A',';','"' ; bytes
+        = "This is a string with embedded $$ dollar and "" double quote characters", 0
+        = "$VBar.not_part_of_variable_name", 0
+        DCD     -1 ; words
+        &       -2,:INDEX:offset1 ; more words
+        %       16
+same_as SPACE   16
+or      FILL    16, 0
+        INCBIN  include.bin
+
+        END