\chapter{Lexical analysis\label{lexical}}

A Python program is read by a \emph{parser}. Input to the parser is a

stream of \emph{tokens}, generated by the \emph{lexical analyzer}. This

chapter describes how the lexical analyzer breaks a file into tokens.

\index{lexical analysis}

\index{parser}

\index{token}

Python uses the 7-bit \ASCII{} character set for program text.

\versionadded[An encoding declaration can be used to indicate that

string literals and comments use an encoding different from ASCII]{2.3}

For compatibility with older versions, Python only warns if it finds

8-bit characters; those warnings should be corrected by either declaring

an explicit encoding, or using escape sequences if those bytes are binary

data, instead of characters.

The run-time character set depends on the I/O devices connected to the

program but is generally a superset of \ASCII.

\strong{Future compatibility note:} It may be tempting to assume that the

character set for 8-bit characters is ISO Latin-1 (an \ASCII{}

superset that covers most western languages that use the Latin

alphabet), but it is possible that in the future Unicode text editors

will become common. These generally use the UTF-8 encoding, which is

also an \ASCII{} superset, but with very different use for the

characters with ordinals 128-255. While there is no consensus on this

subject yet, it is unwise to assume either Latin-1 or UTF-8, even

though the current implementation appears to favor Latin-1. This

applies both to the source character set and the run-time character

set.

\section{Line structure\label{line-structure}}

A Python program is divided into a number of \emph{logical lines}.

\index{line structure}

\subsection{Logical lines\label{logical}}

The end of

a logical line is represented by the token NEWLINE. Statements cannot

cross logical line boundaries except where NEWLINE is allowed by the

syntax (e.g., between statements in compound statements).

A logical line is constructed from one or more \emph{physical lines}

by following the explicit or implicit \emph{line joining} rules.

\index{logical line}

\index{physical line}

\index{line joining}

\index{NEWLINE token}

\subsection{Physical lines\label{physical}}

A physical line is a sequence of characters terminated by an end-of-line

sequence. In source files, any of the standard platform line

termination sequences can be used - the \UNIX form using \ASCII{} LF

(linefeed), the Windows form using the \ASCII{} sequence CR LF (return

followed by linefeed), or the Macintosh form using the \ASCII{} CR

(return) character. All of these forms can be used equally, regardless

of platform.

When embedding Python, source code strings should be passed to Python

APIs using the standard C conventions for newline characters (the

\code{\e n} character, representing \ASCII{} LF, is the line

terminator).

\subsection{Comments\label{comments}}

A comment starts with a hash character (\code{\#}) that is not part of

a string literal, and ends at the end of the physical line. A comment

signifies the end of the logical line unless the implicit line joining

rules are invoked.

Comments are ignored by the syntax; they are not tokens.

\index{comment}

\index{hash character}

\subsection{Encoding declarations\label{encodings}}

\index{source character set}

\index{encodings}

If a comment in the first or second line of the Python script matches

the regular expression \regexp{coding[=:]\e s*([-\e w.]+)}, this comment is

processed as an encoding declaration; the first group of this

expression names the encoding of the source code file. The recommended

forms of this expression are

\begin{verbatim}

# -*- coding: <encoding-name> -*-

\end{verbatim}

which is recognized also by GNU Emacs, and

\begin{verbatim}

# vim:fileencoding=<encoding-name>

\end{verbatim}

which is recognized by Bram Moolenaar's VIM. In addition, if the first

bytes of the file are the UTF-8 byte-order mark

(\code{'\e xef\e xbb\e xbf'}), the declared file encoding is UTF-8

(this is supported, among others, by Microsoft's \program{notepad}).

If an encoding is declared, the encoding name must be recognized by

Python. % XXX there should be a list of supported encodings.

The encoding is used for all lexical analysis, in particular to find

the end of a string, and to interpret the contents of Unicode literals.

String literals are converted to Unicode for syntactical analysis,

then converted back to their original encoding before interpretation

starts. The encoding declaration must appear on a line of its own.

\subsection{Explicit line joining\label{explicit-joining}}

Two or more physical lines may be joined into logical lines using

backslash characters (\code{\e}), as follows: when a physical line ends

in a backslash that is not part of a string literal or comment, it is

joined with the following forming a single logical line, deleting the

backslash and the following end-of-line character. For example:

\index{physical line}

\index{line joining}

\index{line continuation}

\index{backslash character}

%

\begin{verbatim}

if 1900 < year < 2100 and 1 <= month <= 12 \

and 1 <= day <= 31 and 0 <= hour < 24 \

and 0 <= minute < 60 and 0 <= second < 60: # Looks like a valid date

return 1

\end{verbatim}

A line ending in a backslash cannot carry a comment. A backslash does

not continue a comment. A backslash does not continue a token except

for string literals (i.e., tokens other than string literals cannot be

split across physical lines using a backslash). A backslash is

illegal elsewhere on a line outside a string literal.

\subsection{Implicit line joining\label{implicit-joining}}

Expressions in parentheses, square brackets or curly braces can be

split over more than one physical line without using backslashes.

For example:

\begin{verbatim}

month_names = ['Januari', 'Februari', 'Maart', # These are the

'April', 'Mei', 'Juni', # Dutch names

'Juli', 'Augustus', 'September', # for the months

'Oktober', 'November', 'December'] # of the year

\end{verbatim}

Implicitly continued lines can carry comments. The indentation of the

continuation lines is not important. Blank continuation lines are

allowed. There is no NEWLINE token between implicit continuation

lines. Implicitly continued lines can also occur within triple-quoted

strings (see below); in that case they cannot carry comments.

\subsection{Blank lines \label{blank-lines}}

\index{blank line}

A logical line that contains only spaces, tabs, formfeeds and possibly

a comment, is ignored (i.e., no NEWLINE token is generated). During

interactive input of statements, handling of a blank line may differ

depending on the implementation of the read-eval-print loop. In the

standard implementation, an entirely blank logical line (i.e.\ one

containing not even whitespace or a comment) terminates a multi-line

statement.

\subsection{Indentation\label{indentation}}

Leading whitespace (spaces and tabs) at the beginning of a logical

line is used to compute the indentation level of the line, which in

turn is used to determine the grouping of statements.

\index{indentation}

\index{whitespace}

\index{leading whitespace}

\index{space}

\index{tab}

\index{grouping}

\index{statement grouping}

First, tabs are replaced (from left to right) by one to eight spaces

such that the total number of characters up to and including the

replacement is a multiple of

eight (this is intended to be the same rule as used by \UNIX). The

total number of spaces preceding the first non-blank character then

determines the line's indentation. Indentation cannot be split over

multiple physical lines using backslashes; the whitespace up to the

first backslash determines the indentation.

\strong{Cross-platform compatibility note:} because of the nature of

text editors on non-UNIX platforms, it is unwise to use a mixture of

spaces and tabs for the indentation in a single source file. It

should also be noted that different platforms may explicitly limit the

maximum indentation level.

A formfeed character may be present at the start of the line; it will

be ignored for the indentation calculations above. Formfeed

characters occurring elsewhere in the leading whitespace have an

undefined effect (for instance, they may reset the space count to

zero).

The indentation levels of consecutive lines are used to generate

INDENT and DEDENT tokens, using a stack, as follows.

\index{INDENT token}

\index{DEDENT token}

Before the first line of the file is read, a single zero is pushed on

the stack; this will never be popped off again. The numbers pushed on

the stack will always be strictly increasing from bottom to top. At

the beginning of each logical line, the line's indentation level is

compared to the top of the stack. If it is equal, nothing happens.

If it is larger, it is pushed on the stack, and one INDENT token is

generated. If it is smaller, it \emph{must} be one of the numbers

occurring on the stack; all numbers on the stack that are larger are

popped off, and for each number popped off a DEDENT token is

generated. At the end of the file, a DEDENT token is generated for

each number remaining on the stack that is larger than zero.

Here is an example of a correctly (though confusingly) indented piece

of Python code:

\begin{verbatim}

def perm(l):

# Compute the list of all permutations of l

if len(l) <= 1:

return [l]

r = []

for i in range(len(l)):

s = l[:i] + l[i+1:]

p = perm(s)

for x in p:

r.append(l[i:i+1] + x)

return r

\end{verbatim}

The following example shows various indentation errors:

\begin{verbatim}

def perm(l): # error: first line indented

for i in range(len(l)): # error: not indented

s = l[:i] + l[i+1:]

p = perm(l[:i] + l[i+1:]) # error: unexpected indent

for x in p:

r.append(l[i:i+1] + x)

return r # error: inconsistent dedent

\end{verbatim}

(Actually, the first three errors are detected by the parser; only the

last error is found by the lexical analyzer --- the indentation of

\code{return r} does not match a level popped off the stack.)

\subsection{Whitespace between tokens\label{whitespace}}

Except at the beginning of a logical line or in string literals, the

whitespace characters space, tab and formfeed can be used

interchangeably to separate tokens. Whitespace is needed between two

tokens only if their concatenation could otherwise be interpreted as a

different token (e.g., ab is one token, but a b is two tokens).

\section{Other tokens\label{other-tokens}}

Besides NEWLINE, INDENT and DEDENT, the following categories of tokens

exist: \emph{identifiers}, \emph{keywords}, \emph{literals},

\emph{operators}, and \emph{delimiters}.

Whitespace characters (other than line terminators, discussed earlier)

are not tokens, but serve to delimit tokens.

Where

ambiguity exists, a token comprises the longest possible string that

forms a legal token, when read from left to right.

\section{Identifiers and keywords\label{identifiers}}

Identifiers (also referred to as \emph{names}) are described by the following

lexical definitions:

\index{identifier}

\index{name}

\begin{productionlist}

\production{identifier}

{(\token{letter}|"_") (\token{letter} | \token{digit} | "_")*}

\production{letter}

{\token{lowercase} | \token{uppercase}}

\production{lowercase}

{"a"..."z"}

\production{uppercase}

{"A"..."Z"}

\production{digit}

{"0"..."9"}

\end{productionlist}

Identifiers are unlimited in length. Case is significant.

\subsection{Keywords\label{keywords}}

The following identifiers are used as reserved words, or

\emph{keywords} of the language, and cannot be used as ordinary

identifiers. They must be spelled exactly as written here:%

\index{keyword}%

\index{reserved word}

\begin{verbatim}

and del from not while

as elif global or with

assert else if pass yield

break except import print

class exec in raise

continue finally is return

def for lambda try

\end{verbatim}

% When adding keywords, use reswords.py for reformatting

\versionchanged[\constant{None} became a constant and is now

recognized by the compiler as a name for the built-in object

\constant{None}. Although it is not a keyword, you cannot assign

a different object to it]{2.4}

\versionchanged[Both \keyword{as} and \keyword{with} are only recognized

when the \code{with_statement} future feature has been enabled.

It will always be enabled in Python 2.6. See section~\ref{with} for

details. Note that using \keyword{as} and \keyword{with} as identifiers

will always issue a warning, even when the \code{with_statement} future

directive is not in effect]{2.5}

\subsection{Reserved classes of identifiers\label{id-classes}}

Certain classes of identifiers (besides keywords) have special

meanings. These classes are identified by the patterns of leading and

trailing underscore characters:

\begin{description}

\item[\code{_*}]

Not imported by \samp{from \var{module} import *}. The special

identifier \samp{_} is used in the interactive interpreter to store

the result of the last evaluation; it is stored in the

\module{__builtin__} module. When not in interactive mode, \samp{_}

has no special meaning and is not defined.

See section~\ref{import}, ``The \keyword{import} statement.''

\note{The name \samp{_} is often used in conjunction with

internationalization; refer to the documentation for the

\ulink{\module{gettext} module}{../lib/module-gettext.html} for more

information on this convention.}

\item[\code{__*__}]

System-defined names. These names are defined by the interpreter

and its implementation (including the standard library);

applications should not expect to define additional names using this

convention. The set of names of this class defined by Python may be

extended in future versions.

See section~\ref{specialnames}, ``Special method names.''

\item[\code{__*}]

Class-private names. Names in this category, when used within the

context of a class definition, are re-written to use a mangled form

to help avoid name clashes between ``private'' attributes of base

and derived classes.

See section~\ref{atom-identifiers}, ``Identifiers (Names).''

\end{description}

\section{Literals\label{literals}}

Literals are notations for constant values of some built-in types.

\index{literal}

\index{constant}

\subsection{String literals\label{strings}}

String literals are described by the following lexical definitions:

\index{string literal}

\index{ASCII@\ASCII}

\begin{productionlist}

\production{stringliteral}

{[\token{stringprefix}](\token{shortstring} | \token{longstring})}

\production{stringprefix}

{"r" | "u" | "ur" | "R" | "U" | "UR" | "Ur" | "uR"}

\production{shortstring}

{"'" \token{shortstringitem}* "'"

| '"' \token{shortstringitem}* '"'}

\production{longstring}

{"'''" \token{longstringitem}* "'''"}

\productioncont{| '"""' \token{longstringitem}* '"""'}

\production{shortstringitem}

{\token{shortstringchar} | \token{escapeseq}}

\production{longstringitem}

{\token{longstringchar} | \token{escapeseq}}

\production{shortstringchar}

{<any source character except "\e" or newline or the quote>}

\production{longstringchar}

{<any source character except "\e">}

\production{escapeseq}

{"\e" <any ASCII character>}

\end{productionlist}

One syntactic restriction not indicated by these productions is that

whitespace is not allowed between the \grammartoken{stringprefix} and

the rest of the string literal. The source character set is defined

by the encoding declaration; it is \ASCII{} if no encoding declaration

is given in the source file; see section~\ref{encodings}.

\index{triple-quoted string}

\index{Unicode Consortium}

\index{string!Unicode}

In plain English: String literals can be enclosed in matching single

quotes (\code{'}) or double quotes (\code{"}). They can also be

enclosed in matching groups of three single or double quotes (these

are generally referred to as \emph{triple-quoted strings}). The

backslash (\code{\e}) character is used to escape characters that

otherwise have a special meaning, such as newline, backslash itself,

or the quote character. String literals may optionally be prefixed

with a letter \character{r} or \character{R}; such strings are called

\dfn{raw strings}\index{raw string} and use different rules for interpreting

backslash escape sequences. A prefix of \character{u} or \character{U}

makes the string a Unicode string. Unicode strings use the Unicode character

set as defined by the Unicode Consortium and ISO~10646. Some additional

escape sequences, described below, are available in Unicode strings.

The two prefix characters may be combined; in this case, \character{u} must

appear before \character{r}.

In triple-quoted strings,

unescaped newlines and quotes are allowed (and are retained), except

that three unescaped quotes in a row terminate the string. (A

``quote'' is the character used to open the string, i.e. either

\code{'} or \code{"}.)

Unless an \character{r} or \character{R} prefix is present, escape

sequences in strings are interpreted according to rules similar

to those used by Standard C. The recognized escape sequences are:

\index{physical line}

\index{escape sequence}

\index{Standard C}

\index{C}

\begin{tableiii}{l|l|c}{code}{Escape Sequence}{Meaning}{Notes}

\lineiii{\e\var{newline}} {Ignored}{}

\lineiii{\e\e} {Backslash (\code{\e})}{}

\lineiii{\e'} {Single quote (\code{'})}{}

\lineiii{\e"} {Double quote (\code{"})}{}

\lineiii{\e a} {\ASCII{} Bell (BEL)}{}

\lineiii{\e b} {\ASCII{} Backspace (BS)}{}

\lineiii{\e f} {\ASCII{} Formfeed (FF)}{}

\lineiii{\e n} {\ASCII{} Linefeed (LF)}{}

\lineiii{\e N\{\var{name}\}}

{Character named \var{name} in the Unicode database (Unicode only)}{}

\lineiii{\e r} {\ASCII{} Carriage Return (CR)}{}

\lineiii{\e t} {\ASCII{} Horizontal Tab (TAB)}{}

\lineiii{\e u\var{xxxx}}

{Character with 16-bit hex value \var{xxxx} (Unicode only)}{(1)}

\lineiii{\e U\var{xxxxxxxx}}

{Character with 32-bit hex value \var{xxxxxxxx} (Unicode only)}{(2)}

\lineiii{\e v} {\ASCII{} Vertical Tab (VT)}{}

\lineiii{\e\var{ooo}} {Character with octal value \var{ooo}}{(3,5)}

\lineiii{\e x\var{hh}} {Character with hex value \var{hh}}{(4,5)}

\end{tableiii}

\index{ASCII@\ASCII}

\noindent

Notes:

\begin{itemize}

\item[(1)]

Individual code units which form parts of a surrogate pair can be

encoded using this escape sequence.

\item[(2)]

Any Unicode character can be encoded this way, but characters

outside the Basic Multilingual Plane (BMP) will be encoded using a

surrogate pair if Python is compiled to use 16-bit code units (the

default). Individual code units which form parts of a surrogate

pair can be encoded using this escape sequence.

\item[(3)]

As in Standard C, up to three octal digits are accepted.

\item[(4)]

Unlike in Standard C, at most two hex digits are accepted.

\item[(5)]

In a string literal, hexadecimal and octal escapes denote the

byte with the given value; it is not necessary that the byte

encodes a character in the source character set. In a Unicode

literal, these escapes denote a Unicode character with the given

value.

\end{itemize}

Unlike Standard \index{unrecognized escape sequence}C,

all unrecognized escape sequences are left in the string unchanged,

i.e., \emph{the backslash is left in the string}. (This behavior is

useful when debugging: if an escape sequence is mistyped, the

resulting output is more easily recognized as broken.) It is also

important to note that the escape sequences marked as ``(Unicode

only)'' in the table above fall into the category of unrecognized

escapes for non-Unicode string literals.

When an \character{r} or \character{R} prefix is present, a character

following a backslash is included in the string without change, and \emph{all

backslashes are left in the string}. For example, the string literal

\code{r"\e n"} consists of two characters: a backslash and a lowercase

\character{n}. String quotes can be escaped with a backslash, but the

backslash remains in the string; for example, \code{r"\e""} is a valid string

literal consisting of two characters: a backslash and a double quote;

\code{r"\e"} is not a valid string literal (even a raw string cannot

end in an odd number of backslashes). Specifically, \emph{a raw

string cannot end in a single backslash} (since the backslash would

escape the following quote character). Note also that a single

backslash followed by a newline is interpreted as those two characters

as part of the string, \emph{not} as a line continuation.

When an \character{r} or \character{R} prefix is used in conjunction

with a \character{u} or \character{U} prefix, then the \code{\e uXXXX}

and \code{\e UXXXXXXXX} escape sequences are processed while

\emph{all other backslashes are left in the string}.

For example, the string literal

\code{ur"\e{}u0062\e n"} consists of three Unicode characters: `LATIN

SMALL LETTER B', `REVERSE SOLIDUS', and `LATIN SMALL LETTER N'.

Backslashes can be escaped with a preceding backslash; however, both

remain in the string. As a result, \code{\e uXXXX} escape sequences

are only recognized when there are an odd number of backslashes.

\subsection{String literal concatenation\label{string-catenation}}

Multiple adjacent string literals (delimited by whitespace), possibly

using different quoting conventions, are allowed, and their meaning is

the same as their concatenation. Thus, \code{"hello" 'world'} is

equivalent to \code{"helloworld"}. This feature can be used to reduce

the number of backslashes needed, to split long strings conveniently

across long lines, or even to add comments to parts of strings, for

example:

\begin{verbatim}

re.compile("[A-Za-z_]" # letter or underscore

"[A-Za-z0-9_]*" # letter, digit or underscore

)

\end{verbatim}

Note that this feature is defined at the syntactical level, but

implemented at compile time. The `+' operator must be used to

concatenate string expressions at run time. Also note that literal

concatenation can use different quoting styles for each component

(even mixing raw strings and triple quoted strings).

\subsection{Numeric literals\label{numbers}}

There are four types of numeric literals: plain integers, long

integers, floating point numbers, and imaginary numbers. There are no

complex literals (complex numbers can be formed by adding a real

number and an imaginary number).

\index{number}

\index{numeric literal}

\index{integer literal}

\index{plain integer literal}

\index{long integer literal}

\index{floating point literal}

\index{hexadecimal literal}

\index{octal literal}

\index{decimal literal}

\index{imaginary literal}

\index{complex!literal}

Note that numeric literals do not include a sign; a phrase like

\code{-1} is actually an expression composed of the unary operator

`\code{-}' and the literal \code{1}.

\subsection{Integer and long integer literals\label{integers}}

Integer and long integer literals are described by the following

lexical definitions:

\begin{productionlist}

\production{longinteger}

{\token{integer} ("l" | "L")}

\production{integer}

{\token{decimalinteger} | \token{octinteger} | \token{hexinteger}}

\production{decimalinteger}

{\token{nonzerodigit} \token{digit}* | "0"}

\production{octinteger}

{"0" \token{octdigit}+}

\production{hexinteger}

{"0" ("x" | "X") \token{hexdigit}+}

\production{nonzerodigit}

{"1"..."9"}

\production{octdigit}

{"0"..."7"}

\production{hexdigit}

{\token{digit} | "a"..."f" | "A"..."F"}

\end{productionlist}

Although both lower case \character{l} and upper case \character{L} are

allowed as suffix for long integers, it is strongly recommended to always

use \character{L}, since the letter \character{l} looks too much like the

digit \character{1}.

Plain integer literals that are above the largest representable plain

integer (e.g., 2147483647 when using 32-bit arithmetic) are accepted

as if they were long integers instead.\footnote{In versions of Python

prior to 2.4, octal and hexadecimal literals in the range just above

the largest representable plain integer but below the largest unsigned

32-bit number (on a machine using 32-bit arithmetic), 4294967296, were

taken as the negative plain integer obtained by subtracting 4294967296

from their unsigned value.} There is no limit for long integer

literals apart from what can be stored in available memory.

Some examples of plain integer literals (first row) and long integer

literals (second and third rows):

\begin{verbatim}

7 2147483647 0177

3L 79228162514264337593543950336L 0377L 0x100000000L

79228162514264337593543950336 0xdeadbeef

\end{verbatim}

\subsection{Floating point literals\label{floating}}

Floating point literals are described by the following lexical

definitions:

\begin{productionlist}

\production{floatnumber}

{\token{pointfloat} | \token{exponentfloat}}

\production{pointfloat}

{[\token{intpart}] \token{fraction} | \token{intpart} "."}

\production{exponentfloat}

{(\token{intpart} | \token{pointfloat})

\token{exponent}}

\production{intpart}

{\token{digit}+}

\production{fraction}

{"." \token{digit}+}

\production{exponent}

{("e" | "E") ["+" | "-"] \token{digit}+}

\end{productionlist}

Note that the integer and exponent parts of floating point numbers

can look like octal integers, but are interpreted using radix 10. For

example, \samp{077e010} is legal, and denotes the same number

as \samp{77e10}.

The allowed range of floating point literals is

implementation-dependent.

Some examples of floating point literals:

\begin{verbatim}

3.14 10. .001 1e100 3.14e-10 0e0

\end{verbatim}

Note that numeric literals do not include a sign; a phrase like

\code{-1} is actually an expression composed of the unary operator

\code{-} and the literal \code{1}.

\subsection{Imaginary literals\label{imaginary}}

Imaginary literals are described by the following lexical definitions:

\begin{productionlist}

\production{imagnumber}{(\token{floatnumber} | \token{intpart}) ("j" | "J")}

\end{productionlist}

An imaginary literal yields a complex number with a real part of

0.0. Complex numbers are represented as a pair of floating point

numbers and have the same restrictions on their range. To create a

complex number with a nonzero real part, add a floating point number

to it, e.g., \code{(3+4j)}. Some examples of imaginary literals:

\begin{verbatim}

3.14j 10.j 10j .001j 1e100j 3.14e-10j

\end{verbatim}

\section{Operators\label{operators}}

The following tokens are operators:

\index{operators}

\begin{verbatim}

+ - * ** / // %

<< >> & | ^ ~

< > <= >= == != <>

\end{verbatim}

The comparison operators \code{<>} and \code{!=} are alternate

spellings of the same operator. \code{!=} is the preferred spelling;

\code{<>} is obsolescent.

\section{Delimiters\label{delimiters}}

The following tokens serve as delimiters in the grammar:

\index{delimiters}

\begin{verbatim}

( ) [ ] { } @

, : . ` = ;

+= -= *= /= //= %=

&= |= ^= >>= <<= **=

\end{verbatim}

The period can also occur in floating-point and imaginary literals. A

sequence of three periods has a special meaning as an ellipsis in slices.

The second half of the list, the augmented assignment operators, serve

lexically as delimiters, but also perform an operation.

The following printing \ASCII{} characters have special meaning as part

of other tokens or are otherwise significant to the lexical analyzer:

\begin{verbatim}

' " # \

\end{verbatim}

The following printing \ASCII{} characters are not used in Python. Their

occurrence outside string literals and comments is an unconditional

error:

\index{ASCII@\ASCII}

\begin{verbatim}

$ ?

\end{verbatim}