1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383
|
# ### ### ### ######### ######### #########
##
# (c) 2007 Andreas Kupries.
# WIP = Word Interpreter (Also a Work In Progress :). Especially while
# it is running :P
# Micro interpreter for lists of words. Domain specific languages
# based on this will have a bit of a Forth feel, with the input stream
# segmented into words and any other structuring left to whatever
# language. Note that we have here in essence only the core dispatch
# loop, and no actual commands whatsoever, making this definitely only
# a Forth feel and not an actual Forth.
# The idea is derived from Colin McCormack's treeql processor,
# modified to require less boiler plate within the command
# implementations, at the expense of, likely, execution speed. In
# addition the interface between processor core and commands is more
# complex too.
# ### ### ### ######### ######### #########
## Requisites
package require Tcl 8.5
# Use new Tcl 8.5a6+ features for specification of allowed packages.
# We can use either snit 2.2+, or 1.3+ if snit2 is not available.
package require snit 2.2 1.3
# The run_next_* methods use set operations (x in set)
package require struct::set
# For 8.5 we are using features like word-expansion to simplify the
# various evaluations. Otherwise this is identical to v1.
# ### ### ### ######### ######### #########
## API & Implementation
snit::type wip {
# ### ### ### ######### ######### #########
## API
constructor {e} {} ; # create processor
# Defining commands and where they dispatch to.
method def {name {cp {}}} {} ; # Define a DSL command.
method defl {names} {} ; # Def many, simple names (cp = name)
method defd {dict} {} ; # s.a. name/cp dict
method deflva {args} {} ; # s.a. defl, var arg form
method defdva {args} {} ; # s.a. defd, var arg form
# Execution of word lists.
method runl {alist} {} ; # execute list of words
method run {args} {} ; # ditto, words as varargs
method run_next {} {} ; # run the next command in the input.
method run_next_while {accept} {} ; # s.a., while acceptable command
method run_next_until {reject} {} ; # s.a., until rejectable command
# Manipulation of the input word list.
method peek {} {} ; # peek at next word in input
method next {} {} ; # pull next word from input
method insert {at args} {} ; # insert words back into the input
method push {args} {} ; # ditto, at == 0
# ### ### ### ######### ######### #########
## Processor construction.
constructor {e args} {
if {$e eq ""} {
return -code error "No engine specified"
}
set engine $e
$self Definitions $args
return
}
method Definitions {alist} {
# args = series of 'def name' and 'def name cp' statements.
# The code to handle them is in essence a WIP too, just
# hardcoded, as state machine.
set state expect-def
set n {}
set cp {}
foreach a $alist {
if {$state eq "expect-def"} {
if {$a ne "def"} {
return -code error "Expected \"def\", got \"$a\""
}
set state get-name
} elseif {$state eq "get-name"} {
set name $a
set state get-cp-or-def
} elseif {$state eq "get-cp-or-def"} {
# This means that 'def' cannot be a command prefix for
# DSL command.
if {$a eq "def"} {
# Short definition, name only, completed.
$self def $name
# We already have the first word of the next
# definition here, name is coming up next.
set state get-name
} else {
# Long definition, name + cp, completed.
$self def $name $a
# Must be followed by the next definition.
set state expect-def
}
}
}
if {$state eq "get-cp-or-def"} {
# Had a short definition last, now complete.
$self def $name
} elseif {$state eq "get-name"} {
# Incomplete definition at the end, bogus
return -code error "Incomplete definition at end, name missing."
}
return
}
# ### ### ### ######### ######### #########
## Processor state
## Handle of the object incoming commands are dispatched to.
## The currently active DSL code, i.e. word list.
variable engine {} ; # command
variable program {} ; # list (string)
variable arity -array {} ; # array (command name -> command arity)
variable cmd -array {} ; # array (command name -> method cmd prefix)
# ### ### ### ######### ######### #########
## API: DSL definition
## DSL words map to method-prefixes, i.e. method names + fixed
## arguments. We store them with the engine already added in front
## to make them regular command prefixes. No 'mymethod' however,
## that works only in engine code itself, not form the outside.
method def {name {mp {}}} {
if {$mp eq {}} {
# Derive method-prefix from DSL word.
set mp [list $name]
set m $name
set n 0
} else {
# No need to check for an empty method-prefix. That cannot
# happen, as it is diverted, see above.
set m [lindex $mp 0]
set n [expr {[llength $mp]-1}]
}
# Get method arguments, check for problems.
set a [$engine info args $m]
if {[lindex $a end] eq "args"} {
return -code error "Unable to handle Tcl varargs"
}
# The arity of the command is number of required arguments,
# with compensation for those already covered by the
# method-prefix.
set cmd($name) [linsert $mp 0 $engine]
set arity($name) [expr {[llength $a] - $n}]
return
}
method deflva {args} { $self defl $args ; return }
method defdva {args} { $self defd $args ; return }
method defl {names} { foreach n $names { $self def $n } ; return }
method defd {dict} {
if {[llength $dict]%2==1} {
return -code error "Expected a dictionary, got \"$dict\""
}
foreach {name mp} $dict {
$self def $name $mp
}
return
}
# ### ### ### ######### ######### #########
## API: DSL execution
#
## Consider moving the core implementation into procs, to reduce
## call overhead
method run {args} {
return [$self runl $args]
}
method runl {alist} {
# Note: We are saving the current program and restore it
# afterwards, this handles the possibility that this is a
# recursive call into the dispatcher.
set saved $program
set program $alist
set r {}
while {[llength $program]} {
set r [$self run_next]
}
set program $saved
return $r
}
method run_next_while {accept} {
set r {}
while {[struct::set contains $accept [$self peek]]} {
set r [$self run_next]
}
return $r
}
method run_next_until {reject} {
set r {}
while {![struct::set contains $reject [$self peek]]} {
set r [$self run_next]
}
return $r
}
method run_next {} {
# The first word in the list is the current command. Determine
# the number of its fixed arguments. This also checks command
# validity in general.
set c [lindex $program 0]
if {![info exists arity($c)]} {
return -code error \
"Unknown command \"$c\""
}
set n $arity($c)
set m $cmd($c)
# Take the fixed arguments from the input as well.
set cargs [lrange $program 1 $n]
incr n
# Remove the command to dispatch, and its fixed arguments from
# the program. This is done before the dispatch so that the
# command has access to the true current state of the input.
set program [lrange $program $n end]
# Now run the command with its arguments. Commands needing
# more than the declared fixed number of arguments are
# responsible for reading them from input via the method
# 'next' provided by the processor core.
# Note: m already has the engine at the front, it was stored
# that way, see 'def'.
return [{*}$m {*}$cargs]
}
# ### ### ### ######### ######### #########
## Input manipulation
# Get next word from the input (shift)
method next {} {
set w [lindex $program 0]
set program [lrange $program 1 end]
return $w
}
# Peek at the next word in the input
method peek {} {
return [lindex $program 0]
}
# Retrieve the whole current program
method peekall {} {
return $program
}
# Replace the current programm
method replace {args} {
set program $args
return
}
method replacel {alist} {
set program $alist
return
}
# Insert words into the input stream.
method insert {at args} {
set program [linsert $program $at {*}$args]
return
}
method insertl {at alist} {
set program [linsert $program $at {*}$alist]
return
}
# <=> insert 0
method push {args} {
set program [linsert $program 0 {*}$args]
return
}
method pushl {alist} {
set program [linsert $program 0 {*}$alist]
return
}
# <=> insert end
method add {args} {
set program [linsert $program end {*}$args]
return
}
method addl {alist} {
set program [linsert $program end {*}$alist]
return
}
##
# ### ### ### ######### ######### #########
}
# ### ### ### ######### ######### #########
##
# Macro to declare the method of a component as proc. We use this
# later to make access to a WIP processor simpler (no need to write
# the component reference on our own). And no, this is not the same as
# the standard delegation. Doing that simply replaces the component
# name in the call with '$self'. We remove the need to have this
# written in the call.
snit::macro wip::methodasproc {var method suffix} {
proc $method$suffix {args} [string map [list @v@ $var @m@ $method] {
upvar 1 {@v@} dst
return [$dst {@m@} {*}$args]
}]
}
# ### ### ### ######### ######### #########
## Ready
# ### ### ### ######### ######### #########
##
# Macro to install most of the boilerplate needed to setup and use a
# WIP. The only thing left is to call the method 'wip_setup' in the
# constructor of the class using WIP. This macro allows the creation
# of multiple wip's, through custom suffices.
snit::macro wip::dsl {{suffix {}}} {
if {$suffix ne ""} {set suffix _$suffix}
# Instance state, wip processor used to run the language
component wip$suffix
# Standard method to create the processor component. The user has
# to manually add a call of this method to the constructor.
method wip${suffix}_setup {} [string map [list @@ $suffix] {
install {wip@@} using wip "${selfns}::wip@@" $self
}]
# Procedures for easy access to the processor methods, without
# having to use self and wip. I.e. special delegation.
foreach {p} {
add addl def
defd defdva defl deflva
insert insertl replace replacel
push pushl run runl
next peek peekall run_next
run_next_until run_next_while
} {
wip::methodasproc wip$suffix $p $suffix
}
return
}
# ### ### ### ######### ######### #########
## Ready
package provide wip 2.0
|