paul@199 | 1 | Concepts
|
paul@199 | 2 | ========
|
paul@199 | 3 |
|
paul@201 | 4 | This document describes the underlying concepts employed in micropython.
|
paul@201 | 5 |
|
paul@201 | 6 | * Namespaces and attribute definition
|
paul@199 | 7 | * Contexts and values
|
paul@200 | 8 | * Tables, attributes and lookups
|
paul@199 | 9 | * Objects and structures
|
paul@200 | 10 | * Parameters and lookups
|
paul@200 | 11 | * Instantiation
|
paul@222 | 12 | * Register usage
|
paul@199 | 13 |
|
paul@201 | 14 | Namespaces and Attribute Definition
|
paul@201 | 15 | ===================================
|
paul@201 | 16 |
|
paul@201 | 17 | Namespaces are any objects which can retain attributes.
|
paul@201 | 18 |
|
paul@201 | 19 | * Module attributes are defined either at the module level or by global
|
paul@201 | 20 | statements.
|
paul@201 | 21 | * Class attributes are defined only within class statements.
|
paul@201 | 22 | * Instance attributes are defined only by assignments to attributes of self
|
paul@201 | 23 | within __init__ methods.
|
paul@201 | 24 |
|
paul@201 | 25 | These restrictions apply because such attributes are thus explicitly declared,
|
paul@201 | 26 | permitting the use of tables (described below). Module and class attributes
|
paul@201 | 27 | can also be finalised in this way in order to permit certain optimisations.
|
paul@201 | 28 |
|
paul@243 | 29 | An additional restriction required for the current implementation of tables
|
paul@243 | 30 | (as described below) applies to class definitions: each class must be defined
|
paul@243 | 31 | using a unique name; repeated definition of classes having the same name is
|
paul@243 | 32 | thus not permitted. This restriction arises from the use of the "full name" of
|
paul@243 | 33 | a class as a key to the object table, where the full name is a qualified path
|
paul@243 | 34 | via the module hierarchy ending with the name of the class.
|
paul@243 | 35 |
|
paul@201 | 36 | See rejected.txt for complicating mechanisms which could be applied to
|
paul@201 | 37 | mitigate the effects of these restrictions on optimisations.
|
paul@201 | 38 |
|
paul@199 | 39 | Contexts and Values
|
paul@199 | 40 | ===================
|
paul@199 | 41 |
|
paul@199 | 42 | Values are used as the common reference representation in micropython: as
|
paul@199 | 43 | stored representations of attributes (of classes, instances, modules, and
|
paul@199 | 44 | other objects supporting attribute-like entities) as well as the stored values
|
paul@199 | 45 | associated with names in functions and methods.
|
paul@199 | 46 |
|
paul@199 | 47 | Unlike other implementations, micropython does not create things like bound
|
paul@199 | 48 | method objects for individual instances. Instead, all objects are referenced
|
paul@199 | 49 | using a context, reference pair:
|
paul@199 | 50 |
|
paul@199 | 51 | Value Layout
|
paul@199 | 52 | ------------
|
paul@199 | 53 |
|
paul@199 | 54 | 0 1
|
paul@199 | 55 | context object
|
paul@199 | 56 | reference reference
|
paul@199 | 57 |
|
paul@199 | 58 | Specific implementations might reverse this ordering for optimisation
|
paul@199 | 59 | purposes.
|
paul@199 | 60 |
|
paul@199 | 61 | Rationale
|
paul@199 | 62 | ---------
|
paul@199 | 63 |
|
paul@199 | 64 | To reduce the number of created objects whilst retaining the ability to
|
paul@199 | 65 | support bound method invocations. The context indicates the context in which
|
paul@199 | 66 | an invocation is performed, typically the owner of the method.
|
paul@199 | 67 |
|
paul@199 | 68 | Usage
|
paul@199 | 69 | -----
|
paul@199 | 70 |
|
paul@199 | 71 | The context may be inserted as the first argument when a value is involved in
|
paul@199 | 72 | an invocation. This argument may then be omitted from the invocation if its
|
paul@199 | 73 | usage is not appropriate.
|
paul@199 | 74 |
|
paul@199 | 75 | See invocation.txt for details.
|
paul@199 | 76 |
|
paul@237 | 77 | Context Value Types
|
paul@237 | 78 | -------------------
|
paul@237 | 79 |
|
paul@237 | 80 | The following types of context value exist:
|
paul@237 | 81 |
|
paul@237 | 82 | Type Usage Transformations
|
paul@237 | 83 | ---- ----- ---------------
|
paul@237 | 84 |
|
paul@237 | 85 | Replaceable With functions (not methods) May be replaced with an
|
paul@237 | 86 | instance or a class when a
|
paul@237 | 87 | value is stored on an
|
paul@237 | 88 | instance or class
|
paul@237 | 89 |
|
paul@237 | 90 | Placeholder With classes May not be replaced
|
paul@237 | 91 |
|
paul@237 | 92 | Instance With instances (and constants) May not be replaced
|
paul@237 | 93 | or functions as methods
|
paul@237 | 94 |
|
paul@237 | 95 | Class With functions as methods May be replaced when a
|
paul@237 | 96 | value is loaded from a
|
paul@237 | 97 | class attribute via an
|
paul@237 | 98 | instance
|
paul@237 | 99 |
|
paul@199 | 100 | Contexts in Acquired Values
|
paul@199 | 101 | ---------------------------
|
paul@199 | 102 |
|
paul@237 | 103 | There are four classes of instructions which provide values:
|
paul@199 | 104 |
|
paul@199 | 105 | Instruction Purpose Context Operations
|
paul@199 | 106 | ----------- ------- ------------------
|
paul@199 | 107 |
|
paul@237 | 108 | 1) LoadConst Load module, constant Use loaded object with itself
|
paul@237 | 109 | as context
|
paul@199 | 110 |
|
paul@237 | 111 | 2) LoadFunction Load function Combine replaceable context
|
paul@237 | 112 | with loaded object
|
paul@223 | 113 |
|
paul@237 | 114 | 3) LoadClass Load class Combine placeholder context
|
paul@237 | 115 | with loaded object
|
paul@237 | 116 |
|
paul@237 | 117 | 4) LoadAddress* Load attribute from Preserve or override stored
|
paul@201 | 118 | LoadAttr* class, module, context (as described in
|
paul@201 | 119 | instance assignment.txt)
|
paul@199 | 120 |
|
paul@199 | 121 | In order to comply with traditional Python behaviour, contexts may or may not
|
paul@199 | 122 | represent the object from which an attribute has been acquired.
|
paul@199 | 123 |
|
paul@199 | 124 | See assignment.txt for details.
|
paul@199 | 125 |
|
paul@199 | 126 | Contexts in Stored Values
|
paul@199 | 127 | -------------------------
|
paul@199 | 128 |
|
paul@223 | 129 | There are two classes of instruction for storing values:
|
paul@199 | 130 |
|
paul@223 | 131 | Instruction Purpose Context Operations
|
paul@223 | 132 | ----------- ------- ------------------
|
paul@199 | 133 |
|
paul@223 | 134 | 1) StoreAddress Store attribute in a Preserve context; note that no
|
paul@223 | 135 | known object test for class attribute
|
paul@223 | 136 | assignment should be necessary
|
paul@223 | 137 | since this instruction should only
|
paul@223 | 138 | be generated for module globals
|
paul@199 | 139 |
|
paul@223 | 140 | StoreAttr Store attribute in an Preserve context; note that no
|
paul@223 | 141 | instance test for class attribute
|
paul@223 | 142 | assignment should be necessary
|
paul@223 | 143 | since this instruction should only
|
paul@223 | 144 | be generated for self accesses
|
paul@199 | 145 |
|
paul@223 | 146 | StoreAttrIndex Store attribute in an Preserve context; since the index
|
paul@223 | 147 | unknown object lookup could yield a class
|
paul@223 | 148 | attribute, a test of the nature of
|
paul@223 | 149 | the nature of the structure is
|
paul@223 | 150 | necessary in order to prevent
|
paul@223 | 151 | assignments to classes
|
paul@199 | 152 |
|
paul@223 | 153 | 2) StoreAddressContext Store attribute in a Override context if appropriate;
|
paul@237 | 154 | known object if the value has a replaceable
|
paul@237 | 155 | context, permit the target to
|
paul@237 | 156 | take ownership of the value
|
paul@199 | 157 |
|
paul@199 | 158 | See assignment.txt for details.
|
paul@199 | 159 |
|
paul@200 | 160 | Tables, Attributes and Lookups
|
paul@200 | 161 | ==============================
|
paul@199 | 162 |
|
paul@199 | 163 | Attribute lookups, where the exact location of an object attribute is deduced,
|
paul@199 | 164 | are performed differently in micropython than in other implementations.
|
paul@199 | 165 | Instead of providing attribute dictionaries, in which attributes are found,
|
paul@199 | 166 | attributes are located at fixed places in object structures (described below)
|
paul@199 | 167 | and their locations are stored using a special representation known as a
|
paul@199 | 168 | table.
|
paul@199 | 169 |
|
paul@199 | 170 | For a given program, a table can be considered as being like a matrix mapping
|
paul@199 | 171 | classes to attribute names. For example:
|
paul@199 | 172 |
|
paul@199 | 173 | class A:
|
paul@200 | 174 | # instances have attributes x, y
|
paul@199 | 175 |
|
paul@199 | 176 | class B(A):
|
paul@200 | 177 | # introduces attribute z for instances
|
paul@199 | 178 |
|
paul@199 | 179 | class C:
|
paul@200 | 180 | # instances have attributes a, b, z
|
paul@199 | 181 |
|
paul@200 | 182 | This would provide the following table, referred to as an object table in the
|
paul@200 | 183 | context of classes and instances:
|
paul@199 | 184 |
|
paul@199 | 185 | Class/attr a b x y z
|
paul@199 | 186 |
|
paul@199 | 187 | A 1 2
|
paul@199 | 188 | B 1 2 3
|
paul@199 | 189 | C 1 2 3
|
paul@199 | 190 |
|
paul@199 | 191 | A limitation of this representation is that instance attributes may not shadow
|
paul@199 | 192 | class attributes: if an attribute with a given name is not defined on an
|
paul@199 | 193 | instance, an attribute with the same name cannot be provided by the class of
|
paul@199 | 194 | the instance or any superclass of the instance's class.
|
paul@199 | 195 |
|
paul@199 | 196 | The table can be compacted using a representation known as a displacement
|
paul@200 | 197 | list (referred to as an object list in this context):
|
paul@199 | 198 |
|
paul@199 | 199 | Classes with attribute offsets
|
paul@199 | 200 |
|
paul@199 | 201 | classcode A
|
paul@199 | 202 | attrcode a b x y z
|
paul@199 | 203 |
|
paul@199 | 204 | B
|
paul@199 | 205 | a b x y z
|
paul@199 | 206 |
|
paul@199 | 207 | C
|
paul@199 | 208 | a b x y z
|
paul@199 | 209 |
|
paul@199 | 210 | List . . 1 2 1 2 3 1 2 . . 3
|
paul@199 | 211 |
|
paul@199 | 212 | Here, the classcode refers to the offset in the list at which a class's
|
paul@199 | 213 | attributes are defined, whereas the attrcode defines the offset within a
|
paul@199 | 214 | region of attributes corresponding to a single attribute of a given name.
|
paul@199 | 215 |
|
paul@200 | 216 | Attribute Locations
|
paul@200 | 217 | -------------------
|
paul@200 | 218 |
|
paul@200 | 219 | The locations stored in table/list elements are for instance attributes
|
paul@200 | 220 | relative to the location of the instance, whereas those for class attributes
|
paul@200 | 221 | and modules are absolute addresses (although these could also be changed to
|
paul@242 | 222 | object-relative locations). Thus, each occupied table cell has the following
|
paul@242 | 223 | structure:
|
paul@242 | 224 |
|
paul@242 | 225 | attrcode, uses-absolute-address, address (or location)
|
paul@200 | 226 |
|
paul@199 | 227 | Objects and Structures
|
paul@199 | 228 | ======================
|
paul@199 | 229 |
|
paul@199 | 230 | As well as references, micropython needs to have actual objects to refer to.
|
paul@199 | 231 | Since classes, functions and instances are all objects, it is desirable that
|
paul@199 | 232 | certain common features and operations are supported in the same way for all
|
paul@199 | 233 | of these things. To permit this, a common data structure format is used.
|
paul@199 | 234 |
|
paul@215 | 235 | Header.................................................... Attributes.................
|
paul@200 | 236 |
|
paul@215 | 237 | Identifier Identifier Address Identifier Size Object Object ...
|
paul@199 | 238 |
|
paul@215 | 239 | 0 1 2 3 4 5 6 7
|
paul@215 | 240 | classcode attrcode/ invocation funccode size __class__ attribute ...
|
paul@215 | 241 | instance reference reference reference
|
paul@215 | 242 | status
|
paul@199 | 243 |
|
paul@206 | 244 | Classcode
|
paul@206 | 245 | ---------
|
paul@206 | 246 |
|
paul@206 | 247 | Used in attribute lookup.
|
paul@206 | 248 |
|
paul@199 | 249 | Here, the classcode refers to the attribute lookup table for the object (as
|
paul@200 | 250 | described above). Classes and instances share the same classcode, and their
|
paul@200 | 251 | structures reflect this. Functions all belong to the same type and thus employ
|
paul@200 | 252 | the classcode for the function built-in type, whereas modules have distinct
|
paul@200 | 253 | types since they must support different sets of attributes.
|
paul@199 | 254 |
|
paul@206 | 255 | Attrcode
|
paul@206 | 256 | --------
|
paul@206 | 257 |
|
paul@206 | 258 | Used to test instances for membership of classes (or descendants of classes).
|
paul@206 | 259 |
|
paul@242 | 260 | Since, in traditional Python, classes are only ever instances of some generic
|
paul@242 | 261 | built-in type, support for testing such a relationship directly has been
|
paul@207 | 262 | removed and the attrcode is not specified for classes: the presence of an
|
paul@242 | 263 | attrcode indicates that a given object is an instance. In addition, support
|
paul@242 | 264 | has also been removed for testing modules in the same way, meaning that the
|
paul@242 | 265 | attrcode is also not specified for modules.
|
paul@206 | 266 |
|
paul@215 | 267 | See the "Testing Instance Compatibility with Classes (Attrcode)" section below
|
paul@215 | 268 | for details of attrcodes.
|
paul@214 | 269 |
|
paul@213 | 270 | Invocation Reference
|
paul@213 | 271 | --------------------
|
paul@213 | 272 |
|
paul@213 | 273 | Used when an object is called.
|
paul@213 | 274 |
|
paul@213 | 275 | This is the address of the code to be executed when an invocation is performed
|
paul@213 | 276 | on the object.
|
paul@213 | 277 |
|
paul@215 | 278 | Funccode
|
paul@215 | 279 | --------
|
paul@213 | 280 |
|
paul@215 | 281 | Used to look up argument positions by name.
|
paul@213 | 282 |
|
paul@215 | 283 | The strategy with keyword arguments in micropython is to attempt to position
|
paul@215 | 284 | such arguments in the invocation frame as it is being constructed.
|
paul@215 | 285 |
|
paul@215 | 286 | See the "Parameters and Lookups" section for more information.
|
paul@215 | 287 |
|
paul@215 | 288 | Size
|
paul@215 | 289 | ----
|
paul@215 | 290 |
|
paul@219 | 291 | Used to indicate the size of an object including attributes.
|
paul@213 | 292 |
|
paul@209 | 293 | Attributes
|
paul@209 | 294 | ----------
|
paul@209 | 295 |
|
paul@209 | 296 | For classes, modules and instances, the attributes in the structure correspond
|
paul@209 | 297 | to the attributes of each kind of object. For functions, however, the
|
paul@209 | 298 | attributes in the structure correspond to the default arguments for each
|
paul@209 | 299 | function, if any.
|
paul@209 | 300 |
|
paul@206 | 301 | Structure Types
|
paul@206 | 302 | ---------------
|
paul@206 | 303 |
|
paul@199 | 304 | Class C:
|
paul@199 | 305 |
|
paul@215 | 306 | 0 1 2 3 4 5 6 7
|
paul@215 | 307 | classcode (unused) __new__ funccode size class type attribute ...
|
paul@215 | 308 | for C reference for reference reference
|
paul@215 | 309 | instantiator
|
paul@199 | 310 |
|
paul@199 | 311 | Instance of C:
|
paul@199 | 312 |
|
paul@215 | 313 | 0 1 2 3 4 5 6 7
|
paul@215 | 314 | classcode attrcode C.__call__ funccode size class C attribute ...
|
paul@215 | 315 | for C for C reference for reference reference
|
paul@215 | 316 | (if exists) C.__call__
|
paul@199 | 317 |
|
paul@200 | 318 | Function f:
|
paul@199 | 319 |
|
paul@215 | 320 | 0 1 2 3 4 5 6 7
|
paul@215 | 321 | classcode attrcode code funccode size class attribute ...
|
paul@215 | 322 | for for reference function (default)
|
paul@215 | 323 | function function reference reference
|
paul@200 | 324 |
|
paul@200 | 325 | Module m:
|
paul@200 | 326 |
|
paul@215 | 327 | 0 1 2 3 4 5 6 7
|
paul@219 | 328 | classcode attrcode (unused) (unused) (unused) module type attribute ...
|
paul@215 | 329 | for m for m reference (global)
|
paul@215 | 330 | reference
|
paul@200 | 331 |
|
paul@200 | 332 | The __class__ Attribute
|
paul@200 | 333 | -----------------------
|
paul@200 | 334 |
|
paul@200 | 335 | All objects support the __class__ attribute and this is illustrated above with
|
paul@200 | 336 | the first attribute.
|
paul@200 | 337 |
|
paul@200 | 338 | Class: refers to the type class (type.__class__ also refers to the type class)
|
paul@200 | 339 | Function: refers to the function class
|
paul@200 | 340 | Instance: refers to the class instantiated to make the object
|
paul@200 | 341 |
|
paul@203 | 342 | Lists and Tuples
|
paul@203 | 343 | ----------------
|
paul@203 | 344 |
|
paul@203 | 345 | The built-in list and tuple sequences employ variable length structures using
|
paul@203 | 346 | the attribute locations to store their elements, where each element is a
|
paul@203 | 347 | reference to a separately stored object.
|
paul@203 | 348 |
|
paul@214 | 349 | Testing Instance Compatibility with Classes (Attrcode)
|
paul@200 | 350 | ------------------------------------------------------
|
paul@200 | 351 |
|
paul@200 | 352 | Although it would be possible to have a data structure mapping classes to
|
paul@200 | 353 | compatible classes, such as a matrix indicating the subclasses (or
|
paul@200 | 354 | superclasses) of each class, the need to retain the key to such a data
|
paul@200 | 355 | structure for each class might introduce a noticeable overhead.
|
paul@200 | 356 |
|
paul@200 | 357 | Instead of having a separate structure, descendant classes of each class are
|
paul@200 | 358 | inserted as special attributes into the object table. This requires an extra
|
paul@200 | 359 | key to be retained, since each class must provide its own attribute code such
|
paul@200 | 360 | that upon an instance/class compatibility test, the code may be obtained and
|
paul@200 | 361 | used in the object table.
|
paul@200 | 362 |
|
paul@200 | 363 | Invocation and Code References
|
paul@200 | 364 | ------------------------------
|
paul@200 | 365 |
|
paul@200 | 366 | Modules: there is no meaningful invocation reference since modules cannot be
|
paul@200 | 367 | explicitly called.
|
paul@200 | 368 |
|
paul@200 | 369 | Functions: a simple code reference is employed pointing to code implementing
|
paul@200 | 370 | the function. Note that the function locals are completely distinct from this
|
paul@200 | 371 | structure and are not comparable to attributes. Instead, attributes are
|
paul@200 | 372 | reserved for default parameter values, although they do not appear in the
|
paul@200 | 373 | object table described above, appearing instead in a separate parameter table
|
paul@200 | 374 | described below.
|
paul@200 | 375 |
|
paul@200 | 376 | Classes: given that classes must be invoked in order to create instances, a
|
paul@200 | 377 | reference must be provided in class structures. However, this reference does
|
paul@200 | 378 | not point directly at the __init__ method of the class. Instead, the
|
paul@200 | 379 | referenced code belongs to a special initialiser function, __new__, consisting
|
paul@200 | 380 | of the following instructions:
|
paul@200 | 381 |
|
paul@200 | 382 | create instance for C
|
paul@200 | 383 | call C.__init__(instance, ...)
|
paul@200 | 384 | return instance
|
paul@200 | 385 |
|
paul@200 | 386 | Instances: each instance employs a reference to any __call__ method defined in
|
paul@200 | 387 | the class hierarchy for the instance, thus maintaining its callable nature.
|
paul@200 | 388 |
|
paul@200 | 389 | Both classes and modules may contain code in their definitions - the former in
|
paul@200 | 390 | the "body" of the class, potentially defining attributes, and the latter as
|
paul@200 | 391 | the "top-level" code in the module, potentially defining attributes/globals -
|
paul@200 | 392 | but this code is not associated with any invocation target. It is thus
|
paul@200 | 393 | generated in order of appearance and is not referenced externally.
|
paul@200 | 394 |
|
paul@200 | 395 | Invocation Operation
|
paul@200 | 396 | --------------------
|
paul@200 | 397 |
|
paul@200 | 398 | Consequently, regardless of the object an invocation is always done as
|
paul@200 | 399 | follows:
|
paul@200 | 400 |
|
paul@200 | 401 | get invocation reference from the header
|
paul@200 | 402 | jump to reference
|
paul@200 | 403 |
|
paul@200 | 404 | Additional preparation is necessary before the above code: positional
|
paul@200 | 405 | arguments must be saved in the invocation frame, and keyword arguments must be
|
paul@200 | 406 | resolved and saved to the appropriate position in the invocation frame.
|
paul@200 | 407 |
|
paul@200 | 408 | See invocation.txt for details.
|
paul@200 | 409 |
|
paul@200 | 410 | Parameters and Lookups
|
paul@200 | 411 | ======================
|
paul@200 | 412 |
|
paul@200 | 413 | Since Python supports keyword arguments when making invocations, it becomes
|
paul@200 | 414 | necessary to record the parameter names associated with each function or
|
paul@200 | 415 | method. Just as object tables record attributes positions on classes and
|
paul@200 | 416 | instances, parameter tables record parameter positions in function or method
|
paul@200 | 417 | parameter lists.
|
paul@200 | 418 |
|
paul@200 | 419 | For a given program, a parameter table can be considered as being like a
|
paul@200 | 420 | matrix mapping functions/methods to parameter names. For example:
|
paul@200 | 421 |
|
paul@200 | 422 | def f(x, y, z):
|
paul@200 | 423 | pass
|
paul@200 | 424 |
|
paul@200 | 425 | def g(a, b, c):
|
paul@200 | 426 | pass
|
paul@200 | 427 |
|
paul@200 | 428 | def h(a, x):
|
paul@200 | 429 | pass
|
paul@200 | 430 |
|
paul@200 | 431 | This would provide the following table, referred to as a parameter table in
|
paul@200 | 432 | the context of functions and methods:
|
paul@200 | 433 |
|
paul@200 | 434 | Function/param a b c x y z
|
paul@200 | 435 |
|
paul@200 | 436 | f 1 2 3
|
paul@200 | 437 | g 1 2 3
|
paul@200 | 438 | h 1 2
|
paul@200 | 439 |
|
paul@233 | 440 | Confusion can occur when functions are adopted as methods, since the context
|
paul@233 | 441 | then occupies the first slot in the invocation frame:
|
paul@233 | 442 |
|
paul@233 | 443 | def f(x, y, z):
|
paul@233 | 444 | pass
|
paul@233 | 445 |
|
paul@233 | 446 | f(x=1, y=2, z=3) -> f(<context>, 1, 2, 3)
|
paul@233 | 447 | -> f(1, 2, 3)
|
paul@233 | 448 |
|
paul@233 | 449 | class C:
|
paul@233 | 450 | f = f
|
paul@233 | 451 |
|
paul@233 | 452 | def g(x, y, z):
|
paul@233 | 453 | pass
|
paul@233 | 454 |
|
paul@233 | 455 | c = C()
|
paul@233 | 456 |
|
paul@233 | 457 | c.f(y=2, z=3) -> f(<context>, 2, 3)
|
paul@233 | 458 | c.g(y=2, z=3) -> C.g(<context>, 2, 3)
|
paul@233 | 459 |
|
paul@200 | 460 | Just as with parameter tables, a displacement list can be prepared from a
|
paul@200 | 461 | parameter table:
|
paul@200 | 462 |
|
paul@200 | 463 | Functions with parameter (attribute) offsets
|
paul@200 | 464 |
|
paul@200 | 465 | funccode f
|
paul@200 | 466 | attrcode a b c x y z
|
paul@200 | 467 |
|
paul@200 | 468 | g
|
paul@200 | 469 | a b c x y z
|
paul@200 | 470 |
|
paul@200 | 471 | h
|
paul@200 | 472 | a b c x y z
|
paul@200 | 473 |
|
paul@200 | 474 | List . . . 1 2 3 1 2 3 1 . . 2 . .
|
paul@200 | 475 |
|
paul@200 | 476 | Here, the funccode refers to the offset in the list at which a function's
|
paul@200 | 477 | parameters are defined, whereas the attrcode defines the offset within a
|
paul@200 | 478 | region of attributes corresponding to a single parameter of a given name.
|
paul@200 | 479 |
|
paul@200 | 480 | Instantiation
|
paul@200 | 481 | =============
|
paul@200 | 482 |
|
paul@200 | 483 | When instantiating classes, memory must be reserved for the header of the
|
paul@200 | 484 | resulting instance, along with locations for the attributes of the instance.
|
paul@200 | 485 | Since the instance header contains data common to all instances of a class, a
|
paul@200 | 486 | template header is copied to the start of the newly reserved memory region.
|
paul@222 | 487 |
|
paul@222 | 488 | Register Usage
|
paul@222 | 489 | ==============
|
paul@222 | 490 |
|
paul@222 | 491 | During code generation, much of the evaluation produces results which are
|
paul@222 | 492 | implicitly recorded in the "active value" register, and various instructions
|
paul@222 | 493 | will consume the active value. In addition, some instructions will consume a
|
paul@222 | 494 | separate "active source value" from a register, typically those which are
|
paul@222 | 495 | assigning the result of an expression to an assignment target.
|
paul@222 | 496 |
|
paul@222 | 497 | Since values often need to be retained for later use, a set of temporary
|
paul@222 | 498 | storage locations are typically employed. However, optimisations may reduce
|
paul@222 | 499 | the need to use such temporary storage where instructions which provide the
|
paul@222 | 500 | "active value" can be re-executed and will produce the same result.
|