6.10: Internal Object Implementation Note

Last updated
Save as PDF

Page ID: 40139

\( \newcommand{\vecs}[1]{\overset { \scriptstyle \rightharpoonup} {\mathbf{#1}} } \) \( \newcommand{\vecd}[1]{\overset{-\!-\!\rightharpoonup}{\vphantom{a}\smash {#1}}} \)\(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\) \(\newcommand{\id}{\mathrm{id}}\) \( \newcommand{\Span}{\mathrm{span}}\) \( \newcommand{\kernel}{\mathrm{null}\,}\) \( \newcommand{\range}{\mathrm{range}\,}\) \( \newcommand{\RealPart}{\mathrm{Re}}\) \( \newcommand{\ImaginaryPart}{\mathrm{Im}}\) \( \newcommand{\Argument}{\mathrm{Arg}}\) \( \newcommand{\norm}[1]{\| #1 \|}\) \( \newcommand{\inner}[2]{\langle #1, #2 \rangle}\) \( \newcommand{\Span}{\mathrm{span}}\)\(\newcommand{\AA}{\unicode[.8,0]{x212B}}\)

Here is an implementation note for people that really want to go deep inside the way Pharo represents internally objects. The implementation distinguished between two different kinds of objects:

Objects with zero or more fields that are passed by reference and exist on the Pharo heap.
Immediate objects that are passed by value. Depending on version, these are a range of the integers called SmallInteger, all Character objects and possibly a sub-range of 64-bit floating-point numbers called SmallFloat64. In the implementation, such immediate objects occupy an object pointer, most of whose bits encode the immediate’s value and some of the bits encode the object’s class.

The first kind of object, an ordinary object, comes in a number of varieties:

Normal objects that have zero or more named instance variables, such as Point which has an x and a y instance variable. Each instance variable holds an object pointer, which can be a reference to another ordinary object or an immediate.
Indexable objects like arrays that have zero or more indexed instance variables numbered from 1 to N. Each indexed instance variable holds an object pointer, which can be a reference to another ordinary object or an immediate. Indexable objects are accessed using the messages at: and at:put:. For example ((Array new: 1) at: 1 put: 2; at: 1) answers 2.
Objects like Closure or Context that have both named instance variables and indexed instance variables. In the object, the indexed instance variables follow the named instance variables.
Objects like ByteString or Bitmap that have indexed instance variables numbered from 1 to N that contain raw data. Each datum may occupy 8, 16 or 32-bits, depending on its class definition. The data can be accessed as either integers, characters or floating-point numbers, depending on how methods at: and at:put: are implemented. The at: and at:put: methods convert between Pharo objects and raw data, hiding the internal representation, but allowing the system to represent efficiently data such as strings, and bitmaps.

A beauty of Pharo is that you normally don’t need to care about the differences between these three kinds of object.