Updated 2014-04-09 03:25:08 by AMG

[ Scripted Compiler :: Lexing C :: --> Parsing C ]

An important part of any Scripted Compiler is the ability actually process the system language underlying the scripting language. In the case of Tcl this is the C Language.

The first step is always to separate the input stream into tokens, each representing one semantic atom. In compiler speak, lexing.

The following script lexes a string containing C source into a list of tokens. It assumes that the sources are free of preprocessor statements like "#include", "#define", etc.

Also note that the script is built upon the base package provided in Scripted Lexing. While this means the code shown here is quite tailored to parsing for a compiler the general principle used is broad enough to allow for many variations. Examples:

  • Keep the whitespace as tokens. Might be required for a pretty-printer.
  • Treat comments as whitespace and remove them. True compiler. Keeping the comments, but not other whitespace as in the script below is more something for a code analyzer looking for additional data (meta-data) in comments. See Source Navigator for a tool in this area.
  • Modify the definitions, convert the keywords and punctuation into single byte codes, and refrain from splitting/listifying the result. Sort of a special method for compressing C sources.

The next step will be parsing, i.e. adding structure to the token stream under control of a grammar. An existing tool for that is Yeti. See the C Language for grammar references.

I believe that the method I have used below can be used to lex any system language currently in use today, Pascal, Modula, FORTRAN, C++, ... Again this is something of interest to Source Navigator.

Notes

The lexer base from Scripted Lexing is possibly not optimal, but fairly ok in my book so far. Example result:
 [andreask@pliers trans]$ ./driver -noraw -notoken tclIO.c
 __________________________________________________
 tclIO.c:
        242918 characters
        Lexing in 13446065 microseconds
               =  13.446065 seconds
               =  55.35227937 usec/char
 __________________________________________________

Not bad for a lexer written in a scripting language IMHO.

TODO

  • Read up on C syntax. I believe that I currently do not recognize all possible types of numbers.

clex.tcl (The code, finally :)
# -*- tcl -*-
# Lexing C

package require lexbase
package provide clex 2.0

namespace eval clex {
    # Define the lexer symbols for the language 'C', as an example.

    namespace import ::lexbase::*

    DefStart

    DefP (   LPAREN      ; DefP )  RPAREN    ; DefP ->  DEREF
    DefP <   LT          ; DefP <= LE        ; DefP ==  EQ
    DefP >   GT          ; DefP >= GE        ; DefP !=  NE
    DefP \[  LBRACKET    ; DefP \] RBRACKET  ; DefP =   ASSIGN
    DefP \{  LBRACE      ; DefP \} RBRACE    ; DefP *=  MUL_ASSIGN
    DefP .   DOT         ; DefP ,  COMMA     ; DefP /=  DIV_ASSIGN
    DefP ++  INCR_OP     ; DefP -- DECR_OP   ; DefP %=  REM_ASSIGN
    DefP &   ADDR_BITAND ; DefP *  MULT_STAR ; DefP +=  PLUS_ASSIGN
    DefP +   PLUS        ; DefP -  MINUS     ; DefP -=  MINUS_ASSIGN
    DefP ~   BITNOT      ; DefP !  LOGNOT    ; DefP <<= LSHIFT_ASSIGN
    DefP /   DIV         ; DefP %  REM       ; DefP >>= RSHIFT_ASSIGN
    DefP <<  LSHIFT      ; DefP >> RSHIFT    ; DefP &=  BITAND_ASSIGN
    DefP ^   BITEOR      ; DefP && LOGAND    ; DefP ^=  BITEOR_ASSIGN
    DefP |   BITOR       ; DefP || LOGOR     ; DefP |=  BITOR_ASSIGN
    DefP ?   QUERY       ; DefP :  COLON     ; DefP \;  SEMICOLON
    DefP ... ELLIPSIS    ; DefP ~= BITNOT_ASSIGN

    DefK typedef ; DefK extern   ; DefK static ; DefK auto ; DefK register
    DefK void    ; DefK char     ; DefK short  ; DefK int  ; DefK long
    DefK float   ; DefK double   ; DefK signed ; DefK unsigned
    DefK goto    ; DefK continue ; DefK break  ; DefK return
    DefK case    ; DefK default  ; DefK switch
    DefK struct  ; DefK union    ; DefK enum
    DefK while   ; DefK do       ; DefK for
    DefK const   ; DefK volatile
    DefK if      ; DefK else
    DefK sizeof

    DefM COMMENT        ::clex::C_comment_begin   ::clex::C_comment_end
    DefM COMMENT        ::clex::C99_comment_begin ::clex::C99_comment_end
    DefM STRING_LITERAL ::clex::C_string_begin    ::clex::C_string_end
    DefM STRING_LITERAL ::clex::C_char_begin      ::clex::C_char_end  

    # Floats containing '.'s have to be matched early because the '.'
    # is later seen as punctuation.

    DefM CONSTANT ::clex::C_floatA_begin ::clex::C_floatA_end
    DefM CONSTANT ::clex::C_floatB_begin ::clex::C_floatB_end

    DefI IDENT
    DefWS {[ \t\v\f\r\n]+}

    DefRxM {^0x[[:xdigit:]]+} CONSTANT
    DefRxM {^\d+}             CONSTANT

    DefEnd
}

proc ::clex::C_comment_begin {string start} {
    return [string first "/*" $string $start]
}

proc ::clex::C_comment_end {string start} {
    incr start 2 ; # Skip behind /*
    set  stop [string first "*/" $string $start]
    incr stop 1 ; # Skip to /
    return $stop
}

proc ::clex::C99_comment_begin {string start} {
    string first // $string $start
}

proc ::clex::C99_comment_end {string start} {
    regexp -indices -start $start {//(?:\\.|[^\n\\])*(?:\n|$)} $string range
    lindex $range 1
}

proc ::clex::C_string_begin {string start} {
    return [string first "\"" $string $start]
}

proc ::clex::C_string_end {string start} {
    # The next vari-sized thing is a "-quoted string.
    # Finding its end is bit more difficult, because we have
    # to accept \" as one character inside of the string. "

    set from $start
    while 1 {
        incr from
        set stop  [string first "\"" $string $from]

        # Note that we do not use [string first] to look for a \",
        # but simply check the preceding character. That is less
        # expensive than possibly running through the whole string.

        incr stop -1
        if {[string equal [string index $string $stop] "\\"]} {
            incr stop 2
            set from $stop
            continue
        }
        incr stop
        break
    }
    return $stop
}

proc ::clex::C_char_begin {string start} {
    return [string first "'" $string $start]
}

proc ::clex::C_char_end {string start} {
    # The next vari-sized thing is a '-quoted string.
    # Finding its end is bit more difficult, because we have
    # to accept \' as one character inside of the string. "

    set from $start
    while 1 {
        incr from
        set stop  [string first "'"   $string $from]

        # Note that we do not use [string first] to look for a \",
        # but simply check the preceding character. That is less
        # expensive than possibly running through the whole string.

        incr stop -1
        if {[string equal [string index $string $stop] "\\"]} {
            incr stop 2
            set from $stop
            continue
        }
        incr stop
        break
    }
    return $stop
}

proc ::clex::C_floatA_begin {string start} {
    upvar stash stash
    if {[regexp -indices -start $start {\W([0-9]*\.[0-9]+([eEdD][+-]?[0-9]+)?)\W} $string -> match]} {
        #puts a==[string range $string [lindex $match 0] [lindex $match 1]]
        set stash(float-a) [lindex $match 1]
        return [lindex $match 0]
    }
    return -1
}

proc ::clex::C_floatA_end {string start} {
    upvar   stash stash
    return $stash(float-a)
}

proc ::clex::C_floatB_begin {string start} {
    upvar stash stash
    if {[regexp -indices -start $start {\W([0-9]+\.[0-9]*([eEdD][+-]?[0-9]+)?)\W} $string -> match]} {
        #puts b==[string range $string [lindex $match 0] [lindex $match 1]]
        set stash(float-b) [lindex $match 1]
        return [lindex $match 0]
    }
    return -1
}

proc ::clex::C_floatB_end {string start} {
    upvar   stash stash
    return $stash(float-b)
    return -1
}

driver
#!/usr/bin/env tclsh
# -*- tcl -*-

set time  1
set token 1
set raw   1
while {1} {
    switch -exact -- [lindex $argv 0] {
        -notime  {set time  0}
        -notoken {set token 0}
        -noraw   {set raw   0}
        default {break}
    }
    set argv [lrange $argv 1 end]
}


source lexbase.tcl
source clex.tcl

# Read file, lex it, time the execution to measure performance

set data [read [set fh [open [set fname [lindex $argv 0]]]]][close $fh]
set len  [string length $data]
set usec [lindex [time {set res [lexbase::lex $data]}] 0]

foreach {sym attr}  $res  break
foreach {aidx aval} $attr break

if {$time} {
    # Write performance statistics.
    puts __________________________________________________
    puts "$fname:"
    puts "\t$len characters"
    puts "\tLexing in $usec microseconds"
    puts "\t       =  [expr {double($usec)/1000000}] seconds"
    puts "\t       =  [expr {double($usec)/$len}] usec/char"
}

if {$token} {
    # Generate tokenized listing of the input, using the lexing results as input.
    puts __________________________________________________
    set av 0
    foreach s $sym {
        switch -glob -- $s {
            *- {puts "$s <<[lindex $aval [lindex $aidx $av]]>>" ; incr av}
            *  {puts "$s"}
        }
    }
}

if {$raw} {
    # Dump the raw lexer result.
    puts __________________________________________________
    puts Symbols___________________________________________
    puts $sym
    puts ""
    puts Attribute-Indices_________________________________
    puts $aidx
    puts ""
    puts Attribute-Data____________________________________
    puts \{[join $aval "\} \{"]\}
    puts ""
    puts __________________________________________________
}

puts __________________________________________________

AMG: Here's another lexer (I say "scanner") for C that uses ylex:
# cscanner.tcl
package require ylex

# Create the object used to assemble the scanner.
yeti::ylex CScannerFactory -name CScanner

# On error, print the filename, line number, and column number.
CScannerFactory code error {
    if {$file ne {}} {
        puts -nonewline $verbout $file:
    }
    puts $verbout "$line:$column: $yyerrmsg"
}

# Define public variables and methods.
CScannerFactory code public {
    variable file {}        ;# Current file name, or empty string if none.
    variable line 1         ;# Current line number.
    variable column 1       ;# Current column number.
    variable typeNames {}   ;# List of TYPE_NAME tokens.

    # addTypeName --
    # Adds a typedef name to the list of names treated as TYPE_NAME.
    method addTypeName {name} {
        lappend typeNames $name
    }
}

# Define internal methods.
CScannerFactory code private {
    # result --
    # Common result handler for matches.  Updates the line and column counts,
    # and returns the arguments if provided.
    method result {args} {
        set text [string map {\r ""} $yytext]
        set start 0
        while {$start < [string length $text]} {
            regexp -start $start {([^\n\t]*)([\n\t]?)} $text chunk body space
            incr column [string length $body]
            if {$space eq "\n"} {
                set column 1
                incr line
            } elseif {$space eq "\t"} {
                set column [expr {(($column + 7) & ~3) + 1}]
            }
            incr start [string length $chunk]
        }
        if {[llength $args]} {
            return -level 2 $args
        }
    }

    # lineDirective --
    # Processes #line directives.
    method lineDirective {} {
        if {[regexp {^\s*#line (\d+)(?: "(.+)")?\n$} $yytext _ line newFile]
         && $newFile ne ""} {
            set file [subst -nocommands -novariables $newFile]
        }
    }

    # tokenType --
    # Decides if a token is TYPE_NAME or IDENTIFIER according to $typeNames.
    method tokenType {} {
        if {$yytext in $typeNames} {
            return TYPE_NAME
        } else {
            return IDENTIFIER
        }
    }

    # scanChar --
    # Converts character literals to integers.
    method scanChar {char} {
        set char [subst -nocommands -novariables $char]
        if {[string length $char] != 1} {
            error "multi-character constants not supported"
        }
        scan $char %c
    }

    # scanStr --
    # Converts string literals to Tcl strings.
    method scanStr {string} {
        subst -nocommands -novariables $string
    }
}

# Define useful abbreviations for upcoming regular expressions.
CScannerFactory macro {
    C   {(?://(?:\\.|[^\n\\])*(?:\n|$))}
    E   {(?:[eE][+-]?\d+)}
    FS  {[fFlL]}
    IS  {(?:[uU]?[lL]{0,2}|[lL]{0,2}[uU]?)}
}

# Generate a regular expression matching any simple token.  The value of such
# tokens is the uppercase version of the token string itself.
foreach token {
    auto bool break case char const continue default do double else enum extern
    float for goto if int long register return short signed sizeof static
    struct switch typedef union unsigned void volatile while ...  >>= <<= +=
    -= *= /= %= &= ^= |= >> << ++ -- -> && || <= >= == != ; \{ \} , : = ( ) [
    ] .  & ! ~ - + * / % < > ^ | ?
} {
    lappend pattern [regsub -all {[][*+?{}()|.^$]} $token {\\&}]
}
set pattern (?:[join $pattern |])

# Match simple tokens.
CScannerFactory add $pattern {result [string toupper $yytext]}

# Match and decode more complex tokens.
CScannerFactory add {
    {[ \t\v\n\f]}                {result}
    {/\*.*?\*/}                  {result}
    {<C>}                        {result}
    {(?n)^\s*#line[^\n]*\n}      {lineDirective}
    {[a-zA-Z_]\w*\M}             {result [tokenType]    $yytext}
    {0[xX]([[:xdigit:]]+)<IS>\M} {result CONSTANT       [scan $1 %x]}
    {0([0-7]+)<IS>\M}            {result CONSTANT       [scan $1 %o]}
    {(\d+)<IS>\M}                {result CONSTANT       [scan $1 %d]}
    {L?'((?:[^\\']|\\.)+)'}      {result CONSTANT       [scanChar $1]}
    {(\d+<E>)<FS>?\M}            {result CONSTANT       [scan $1 %f]}
    {(\d*\.\d+<E>?)<FS>?\M}      {result CONSTANT       [scan $1 %f]}
    {(\d+\.\d*<E>?)<FS>?\M}      {result CONSTANT       [scan $1 %f]}
    {L?"((?:[^\\"]|\\.)+)"}      {result STRING_LITERAL [scanStr $1]}
    {.}                          {error "invalid character \"$yytext\""}
}

# Create the CScanner class.  You might want to cache the generated script to
# avoid dependency on ylex and to improve startup time.
eval [CScannerFactory dump]
itcl::delete object CScannerFactory

It's quite different than the code given at the top of this page. The primary difference is that it directly uses the various symbols like "+" as the terminal names. Since we're using Tcl, I don't see a problem with this. I find that it makes the grammar much more readable.