Tcl Source Code

.TH binary n 8.0 Tcl "Tcl Built-In Commands"
.so man.macros
.BS
'\" Note:  do not modify the .SH NAME line immediately below!
.SH NAME
binary \- Insert and extract fields from binary strings
.SH SYNOPSIS
.VS 8.6
\fBbinary decode \fIformat\fR ?\fI\-option value ...\fR? \fIdata\fR
.br
\fBbinary encode \fIformat\fR ?\fI\-option value ...\fR? \fIdata\fR
.br
.VE 8.6
\fBbinary format \fIformatString \fR?\fIarg arg ...\fR?
.br
\fBbinary scan \fIstring formatString \fR?\fIvarName varName ...\fR?
.BE
.SH DESCRIPTION
.PP
This command provides facilities for manipulating binary data.  The
subcommand \fBbinary format\fR creates a binary string from normal
Tcl values.  For example, given the values 16 and 22, on a 32-bit
architecture, it might produce an 8-byte binary string consisting of
two 4-byte integers, one for each of the numbers.  The subcommand
\fBbinary scan\fR, does the opposite: it extracts data
from a binary string and returns it as ordinary Tcl string values.
.VS 8.6
The \fBbinary encode\fR and \fBbinary decode\fR subcommands convert
binary data to or from string encodings such as base64 (used in MIME
messages for example).
.VE 8.6
.PP
Note that other operations on binary data, such as taking a subsequence of it,
getting its length, or reinterpreting it as a string in some encoding, are
done by other Tcl commands (respectively \fBstring range\fR,
\fBstring length\fR and \fBencoding convertfrom\fR in the example cases).  A
binary string in Tcl is merely one where all the characters it contains are in
the range \eu0000\-\eu00FF.
.SH "BINARY ENCODE AND DECODE"
.VS 8.6
.PP
When encoding binary data as a readable string, the starting binary data is
passed to the \fBbinary encode\fR command, together with the name of the
encoding to use and any encoding-specific options desired. Data which has been
encoded can be converted back to binary form using \fBbinary decode\fR. The
following formats and options are supported.
.TP
................................................................................
.
Instructs the decoder to throw an error if it encounters unexpected whitespace
characters. Otherwise it ignores them.
.PP
Note that neither the encoder nor the decoder handle the header and footer of
the uuencode format.
.RE
.VE 8.6
.SH "BINARY FORMAT"
.PP
The \fBbinary format\fR command generates a binary string whose layout
is specified by the \fIformatString\fR and whose contents come from
the additional arguments.  The resulting binary value is returned.
.PP
The \fIformatString\fR consists of a sequence of zero or more field
................................................................................
specifiers separated by zero or more spaces.  Each field specifier is
a single type character followed by an optional flag character followed
by an optional numeric \fIcount\fR.
Most field specifiers consume one argument to obtain the value to be
formatted.  The type character specifies how the value is to be
formatted.  The \fIcount\fR typically indicates how many items of the
specified type are taken from the value.  If present, the \fIcount\fR
is a non-negative decimal integer or \fB*\fR, which normally indicates


that all of the items in the value are to be used.  If the number of
arguments does not match the number of fields in the format string
that consume arguments, then an error is generated. The flag character
is ignored for \fBbinary format\fR.
.PP
Here is a small example to clarify the relation between the field
specifiers and the arguments:

.CS
\fBbinary format\fR d3d {1.0 2.0 3.0 4.0} 0.1
.CE
.PP
The first argument is a list of four numbers, but because of the count
of 3 for the associated field specifier, only the first three will be
used. The second argument is associated with the second field
................................................................................
the \fBencoding convertto\fR command should be used first to change
the string into an external representation
if this truncation is not desired (i.e. if the characters are
not part of the ISO 8859\-1 character set.)
If \fIarg\fR has fewer than \fIcount\fR bytes, then additional zero
bytes are used to pad out the field.  If \fIarg\fR is longer than the
specified length, the extra characters will be ignored.  If


\fIcount\fR is \fB*\fR, then all of the bytes in \fIarg\fR will be
formatted.  If \fIcount\fR is omitted, then one character will be
formatted.  For example,
.RS

.CS
\fBbinary format\fR a7a*a alpha bravo charlie
.CE




will return a string equivalent to \fBalpha\e000\e000bravoc\fR,




.CS
\fBbinary format\fR a* [encoding convertto utf-8 \eu20ac]
.CE




will return a string equivalent to \fB\e342\e202\e254\fR (which is the



UTF-8 byte sequence for a Euro-currency character) and

.CS
\fBbinary format\fR a* [encoding convertto iso8859-15 \eu20ac]
.CE







will return a string equivalent to \fB\e244\fR (which is the ISO
8859\-15 byte sequence for a Euro-currency character). Contrast these
last two with:

.CS
\fBbinary format\fR a* \eu20ac
.CE







which returns a string equivalent to \fB\e254\fR (i.e. \fB\exac\fR) by
truncating the high-bits of the character, and which is probably not
what is desired.
.RE
.IP \fBA\fR 5
This form is the same as \fBa\fR except that spaces are used for
padding instead of nulls.  For example,
.RS

.CS
\fBbinary format\fR A6A*A alpha bravo charlie
.CE




will return \fBalpha bravoc\fR.

.RE
.IP \fBb\fR 5
Stores a string of \fIcount\fR binary digits in low-to-high order
within each byte in the output string.  \fIArg\fR must contain a
sequence of \fB1\fR and \fB0\fR characters.  The resulting bytes are
emitted in first to last order with the bits being formatted in
low-to-high order within each byte.  If \fIarg\fR has fewer than
\fIcount\fR digits, then zeros will be used for the remaining bits.
If \fIarg\fR has more than the specified number of digits, the extra
digits will be ignored.  If \fIcount\fR is \fB*\fR, then all of the


digits in \fIarg\fR will be formatted.  If \fIcount\fR is omitted,
then one digit will be formatted.  If the number of bits formatted
does not end at a byte boundary, the remaining bits of the last byte
will be zeros.  For example,
.RS

.CS
\fBbinary format\fR b5b* 11100 111000011010
.CE




will return a string equivalent to \fB\ex07\ex87\ex05\fR.

.RE
.IP \fBB\fR 5
This form is the same as \fBb\fR except that the bits are stored in
high-to-low order within each byte.  For example,
.RS

.CS
\fBbinary format\fR B5B* 11100 111000011010
.CE




will return a string equivalent to \fB\exe0\exe1\exa0\fR.

.RE
.IP \fBH\fR 5
Stores a string of \fIcount\fR hexadecimal digits in high-to-low
within each byte in the output string.  \fIArg\fR must contain a
sequence of characters in the set
.QW 0123456789abcdefABCDEF .
The resulting bytes are emitted in first to last order with the hex digits
being formatted in high-to-low order within each byte.  If \fIarg\fR
has fewer than \fIcount\fR digits, then zeros will be used for the
remaining digits.  If \fIarg\fR has more than the specified number of
digits, the extra digits will be ignored.  If \fIcount\fR is

\fB*\fR, then all of the digits in \fIarg\fR will be formatted.  If
\fIcount\fR is omitted, then one digit will be formatted.  If the
number of digits formatted does not end at a byte boundary, the
remaining bits of the last byte will be zeros.  For example,
.RS

.CS
\fBbinary format\fR H3H*H2 ab DEF 987
.CE




will return a string equivalent to \fB\exab\ex00\exde\exf0\ex98\fR.

.RE
.IP \fBh\fR 5
This form is the same as \fBH\fR except that the digits are stored in
low-to-high order within each byte. This is seldom required. For example,
.RS

.CS
\fBbinary format\fR h3h*h2 AB def 987
.CE




will return a string equivalent to \fB\exba\ex00\exed\ex0f\ex89\fR.

.RE
.IP \fBc\fR 5
Stores one or more 8-bit integer values in the output string.  If no
\fIcount\fR is specified, then \fIarg\fR must consist of an integer
value. If \fIcount\fR is specified, \fIarg\fR must consist of a list
containing at least that many integers. The low-order 8 bits of each integer
are stored as a one-byte value at the cursor position.  If \fIcount\fR

is \fB*\fR, then all of the integers in the list are formatted. If the
number of elements in the list is greater
than \fIcount\fR, then the extra elements are ignored.  For example,
.RS

.CS
\fBbinary format\fR c3cc* {3 -3 128 1} 260 {2 5}
.CE

will return a string equivalent to


\fB\ex03\exfd\ex80\ex04\ex02\ex05\fR, whereas




.CS
\fBbinary format\fR c {2 5}
.CE

will generate an error.
.RE
.IP \fBs\fR 5
This form is the same as \fBc\fR except that it stores one or more
16-bit integers in little-endian byte order in the output string.  The
low-order 16-bits of each integer are stored as a two-byte value at
the cursor position with the least significant byte stored first.  For
example,
.RS

.CS
\fBbinary format\fR s3 {3 -3 258 1}
.CE

will return a string equivalent to


\fB\ex03\ex00\exfd\exff\ex02\ex01\fR.

.RE
.IP \fBS\fR 5
This form is the same as \fBs\fR except that it stores one or more
16-bit integers in big-endian byte order in the output string.  For
example,
.RS

.CS
\fBbinary format\fR S3 {3 -3 258 1}
.CE

will return a string equivalent to


\fB\ex00\ex03\exff\exfd\ex01\ex02\fR.

.RE
.IP \fBt\fR 5
This form (mnemonically \fItiny\fR) is the same as \fBs\fR and \fBS\fR
except that it stores the 16-bit integers in the output string in the
native byte order of the machine where the Tcl script is running.
To determine what the native byte order of the machine is, refer to
the \fBbyteOrder\fR element of the \fBtcl_platform\fR array.
.IP \fBi\fR 5
This form is the same as \fBc\fR except that it stores one or more
32-bit integers in little-endian byte order in the output string.  The
low-order 32-bits of each integer are stored as a four-byte value at
the cursor position with the least significant byte stored first.  For
example,
.RS

.CS
\fBbinary format\fR i3 {3 -3 65536 1}
.CE

will return a string equivalent to


\fB\ex03\ex00\ex00\ex00\exfd\exff\exff\exff\ex00\ex00\ex01\ex00\fR

.RE
.IP \fBI\fR 5
This form is the same as \fBi\fR except that it stores one or more one
or more 32-bit integers in big-endian byte order in the output string.
For example,
.RS

.CS
\fBbinary format\fR I3 {3 -3 65536 1}
.CE

will return a string equivalent to


\fB\ex00\ex00\ex00\ex03\exff\exff\exff\exfd\ex00\ex01\ex00\ex00\fR

.RE
.IP \fBn\fR 5
This form (mnemonically \fInumber\fR or \fInormal\fR) is the same as
\fBi\fR and \fBI\fR except that it stores the 32-bit integers in the
output string in the native byte order of the machine where the Tcl
script is running.
To determine what the native byte order of the machine is, refer to
................................................................................
the \fBbyteOrder\fR element of the \fBtcl_platform\fR array.
.IP \fBw\fR 5
This form is the same as \fBc\fR except that it stores one or more
64-bit integers in little-endian byte order in the output string.  The
low-order 64-bits of each integer are stored as an eight-byte value at
the cursor position with the least significant byte stored first.  For
example,
.RS

.CS
\fBbinary format\fR w 7810179016327718216
.CE

will return the string \fBHelloTcl\fR
.RE
.IP \fBW\fR 5
This form is the same as \fBw\fR except that it stores one or more one
or more 64-bit integers in big-endian byte order in the output string.
For example,
.RS

.CS
\fBbinary format\fR Wc 4785469626960341345 110
.CE

will return the string \fBBigEndian\fR
.RE
.IP \fBm\fR 5
This form (mnemonically the mirror of \fBw\fR) is the same as \fBw\fR
and \fBW\fR except that it stores the 64-bit integers in the output
string in the native byte order of the machine where the Tcl script is
running.
To determine what the native byte order of the machine is, refer to
................................................................................
point number may vary across architectures, so the number of bytes
that are generated may vary.  If the value overflows the
machine's native representation, then the value of FLT_MAX
as defined by the system will be used instead.  Because Tcl uses
double-precision floating point numbers internally, there may be some
loss of precision in the conversion to single-precision.  For example,
on a Windows system running on an Intel Pentium processor,
.RS

.CS
\fBbinary format\fR f2 {1.6 3.4}
.CE

will return a string equivalent to


\fB\excd\excc\excc\ex3f\ex9a\ex99\ex59\ex40\fR.

.RE
.IP \fBr\fR 5
This form (mnemonically \fIreal\fR) is the same as \fBf\fR except that
it stores the single-precision floating point numbers in little-endian
order.  This conversion only produces meaningful output when used on
machines which use the IEEE floating point representation (very
common, but not universal.)
................................................................................
This form is the same as \fBr\fR except that it stores the
single-precision floating point numbers in big-endian order.
.IP \fBd\fR 5
This form is the same as \fBf\fR except that it stores one or more one
or more double-precision floating point numbers in the machine's native
representation in the output string.  For example, on a
Windows system running on an Intel Pentium processor,
.RS

.CS
\fBbinary format\fR d1 {1.6}
.CE

will return a string equivalent to


\fB\ex9a\ex99\ex99\ex99\ex99\ex99\exf9\ex3f\fR.

.RE
.IP \fBq\fR 5
This form (mnemonically the mirror of \fBd\fR) is the same as \fBd\fR
except that it stores the double-precision floating point numbers in
little-endian order.  This conversion only produces meaningful output
when used on machines which use the IEEE floating point representation
(very common, but not universal.)
.IP \fBQ\fR 5
This form is the same as \fBq\fR except that it stores the
double-precision floating point numbers in big-endian order.
.IP \fBx\fR 5
Stores \fIcount\fR null bytes in the output string.  If \fIcount\fR is
not specified, stores one null byte.  If \fIcount\fR is \fB*\fR,

generates an error.  This type does not consume an argument.  For
example,
.RS

.CS
\fBbinary format\fR a3xa3x2a3 abc def ghi
.CE




will return a string equivalent to \fBabc\e000def\e000\e000ghi\fR.

.RE
.IP \fBX\fR 5
Moves the cursor back \fIcount\fR bytes in the output string.  If


\fIcount\fR is \fB*\fR or is larger than the current cursor position,
then the cursor is positioned at location 0 so that the next byte
stored will be the first byte in the result string.  If \fIcount\fR is
omitted then the cursor is moved back one byte.  This type does not
consume an argument.  For example,
.RS

.CS
\fBbinary format\fR a3X*a3X2a3 abc def ghi
.CE

will return \fBdghi\fR.
.RE
.IP \[email protected]\fR 5
Moves the cursor to the absolute location in the output string
specified by \fIcount\fR.  Position 0 refers to the first byte in the
output string.  If \fIcount\fR refers to a position beyond the last
byte stored so far, then null bytes will be placed in the uninitialized
locations and the cursor will be placed at the specified location.  If


\fIcount\fR is \fB*\fR, then the cursor is moved to the current end of
the output string.  If \fIcount\fR is omitted, then an error will be
generated.  This type does not consume an argument. For example,
.RS

.CS
\fBbinary format\fR [email protected]@*[email protected] abcde f ghi j
.CE




will return \fBabfdeghi\e000\e000j\fR.

.RE
.SH "BINARY SCAN"
.PP
The \fBbinary scan\fR command parses fields from a binary string,
returning the number of conversions performed.  \fIString\fR gives the
input bytes to be parsed (one byte per character, and characters not
representable as a byte have their high bits chopped)
................................................................................
spaces.  Each field specifier is a single type character followed by
an optional flag character followed by an optional numeric \fIcount\fR.
Most field specifiers consume one
argument to obtain the variable into which the scanned values should
be placed.  The type character specifies how the binary data is to be
interpreted.  The \fIcount\fR typically indicates how many items of
the specified type are taken from the data.  If present, the
\fIcount\fR is a non-negative decimal integer or \fB*\fR, which

normally indicates that all of the remaining items in the data are to
be used.  If there are not enough bytes left after the current cursor
position to satisfy the current field specifier, then the
corresponding variable is left untouched and \fBbinary scan\fR returns
immediately with the number of variables that were set.  If there are
not enough arguments for all of the fields in the format string that
consume arguments, then an error is generated. The flag character
.QW u
may be given to cause some types to be read as unsigned values. The flag
is accepted for all field types but is ignored for non-integer fields.
.PP
A similar example as with \fBbinary format\fR should explain the
relation between field specifiers and arguments in case of the binary
scan subcommand:

.CS
\fBbinary scan\fR $bytes s3s first second
.CE
.PP
This command (provided the binary string in the variable \fIbytes\fR
is long enough) assigns a list of three integers to the variable
\fIfirst\fR and assigns a single value to the variable \fIsecond\fR.
If \fIbytes\fR contains fewer than 8 bytes (i.e. four 2-byte
integers), no assignment to \fIsecond\fR will be made, and if
\fIbytes\fR contains fewer than 6 bytes (i.e. three 2-byte integers),
no assignment to \fIfirst\fR will be made.  Hence:

.CS
puts [\fBbinary scan\fR abcdefg s3s first second]
puts $first
puts $second
.CE

will print (assuming neither variable is set previously):

.CS
1
25185 25699 26213
can't read "second": no such variable
.CE
.PP
It is \fIimportant\fR to note that the \fBc\fR, \fBs\fR, and \fBS\fR
(and \fBi\fR and \fBI\fR on 64bit systems) will be scanned into
long data size values.  In doing this, values that have their high
bit set (0x80 for chars, 0x8000 for shorts, 0x80000000 for ints),
will be sign extended.  Thus the following will occur:

.CS
set signShort [\fBbinary format\fR s1 0x8000]
\fBbinary scan\fR $signShort s1 val; \fI# val == 0xFFFF8000\fR
.CE

If you require unsigned values you can include the
.QW u
flag character following
the field type. For example, to read an unsigned short value:

.CS
set signShort [\fBbinary format\fR s1 0x8000]
\fBbinary scan\fR $signShort su1 val; \fI# val == 0x00008000\fR
.CE
.PP
Each type-count pair moves an imaginary cursor through the binary data,
reading bytes from the current position.  The cursor is initially
at position 0 at the beginning of the data.  The type may be any one of
the following characters:
.IP \fBa\fR 5
The data is a byte string of length \fIcount\fR.  If \fIcount\fR

is \fB*\fR, then all of the remaining bytes in \fIstring\fR will be
scanned into the variable.  If \fIcount\fR is omitted, then one
byte will be scanned.
All bytes scanned will be interpreted as being characters in the
range \eu0000-\eu00ff so the \fBencoding convertfrom\fR command will be
needed if the string is not a binary string or a string encoded in ISO
8859\-1.
For example,
.RS

.CS
\fBbinary scan\fR abcde\e000fghi a6a10 var1 var2
.CE

will return \fB1\fR with the string equivalent to \fBabcde\e000\fR
stored in \fIvar1\fR and \fIvar2\fR left unmodified, and

.CS
\fBbinary scan\fR \e342\e202\e254 a* var1
set var2 [encoding convertfrom utf-8 $var1]
.CE

will store a Euro-currency character in \fIvar2\fR.
.RE
.IP \fBA\fR 5
This form is the same as \fBa\fR, except trailing blanks and nulls are stripped from
the scanned value before it is stored in the variable.  For example,
.RS

.CS
\fBbinary scan\fR "abc efghi  \e000" A* var1
.CE

will return \fB1\fR with \fBabc efghi\fR stored in \fIvar1\fR.
.RE
.IP \fBb\fR 5
The data is turned into a string of \fIcount\fR binary digits in
low-to-high order represented as a sequence of
.QW 1
and
.QW 0
characters.  The data bytes are scanned in first to last order with
the bits being taken in low-to-high order within each byte.  Any extra
bits in the last byte are ignored.  If \fIcount\fR is \fB*\fR, then

all of the remaining bits in \fIstring\fR will be scanned.  If
\fIcount\fR is omitted, then one bit will be scanned.  For example,
.RS

.CS
\fBbinary scan\fR \ex07\ex87\ex05 b5b* var1 var2
.CE

will return \fB2\fR with \fB11100\fR stored in \fIvar1\fR and
\fB1110000110100000\fR stored in \fIvar2\fR.
.RE
.IP \fBB\fR 5
This form is the same as \fBb\fR, except the bits are taken in
high-to-low order within each byte.  For example,
.RS

.CS
\fBbinary scan\fR \ex70\ex87\ex05 B5B* var1 var2
.CE

will return \fB2\fR with \fB01110\fR stored in \fIvar1\fR and
\fB1000011100000101\fR stored in \fIvar2\fR.
.RE
.IP \fBH\fR 5
The data is turned into a string of \fIcount\fR hexadecimal digits in
high-to-low order represented as a sequence of characters in the set
.QW 0123456789abcdef .
The data bytes are scanned in first to last
order with the hex digits being taken in high-to-low order within each
byte. Any extra bits in the last byte are ignored. If \fIcount\fR is

\fB*\fR, then all of the remaining hex digits in \fIstring\fR will be
scanned. If \fIcount\fR is omitted, then one hex digit will be
scanned. For example,
.RS

.CS
\fBbinary scan\fR \ex07\exC6\ex05\ex1f\ex34 H3H* var1 var2
.CE

will return \fB2\fR with \fB07c\fR stored in \fIvar1\fR and
\fB051f34\fR stored in \fIvar2\fR.
.RE
.IP \fBh\fR 5
This form is the same as \fBH\fR, except the digits are taken in
reverse (low-to-high) order within each byte. For example,
.RS

.CS
\fBbinary scan\fR \ex07\ex86\ex05\ex12\ex34 h3h* var1 var2
.CE

will return \fB2\fR with \fB706\fR stored in \fIvar1\fR and
\fB502143\fR stored in \fIvar2\fR.
.PP
Note that most code that wishes to parse the hexadecimal digits from
multiple bytes in order should use the \fBH\fR format.
.RE
.IP \fBc\fR 5
The data is turned into \fIcount\fR 8-bit signed integers and stored
in the corresponding variable as a list. If \fIcount\fR is \fB*\fR,


then all of the remaining bytes in \fIstring\fR will be scanned.  If
\fIcount\fR is omitted, then one 8-bit integer will be scanned.  For
example,
.RS

.CS
\fBbinary scan\fR \ex07\ex86\ex05 c2c* var1 var2
.CE

will return \fB2\fR with \fB7 -122\fR stored in \fIvar1\fR and \fB5\fR
stored in \fIvar2\fR.  Note that the integers returned are signed, but
they can be converted to unsigned 8-bit quantities using an expression
like:
.CS
set num [expr { $num & 0xff }]
.CE

.RE
.IP \fBs\fR 5
The data is interpreted as \fIcount\fR 16-bit signed integers
represented in little-endian byte order.  The integers are stored in

the corresponding variable as a list.  If \fIcount\fR is \fB*\fR, then

all of the remaining bytes in \fIstring\fR will be scanned.  If
\fIcount\fR is omitted, then one 16-bit integer will be scanned.  For
example,
.RS

.CS
\fBbinary scan\fR \ex05\ex00\ex07\ex00\exf0\exff s2s* var1 var2
.CE

will return \fB2\fR with \fB5 7\fR stored in \fIvar1\fR and \fB\-16\fR
stored in \fIvar2\fR.  Note that the integers returned are signed, but
they can be converted to unsigned 16-bit quantities using an expression
like:
.CS
set num [expr { $num & 0xffff }]
.CE

.RE
.IP \fBS\fR 5
This form is the same as \fBs\fR except that the data is interpreted
as \fIcount\fR 16-bit signed integers represented in big-endian byte
order.  For example,
.RS

.CS
\fBbinary scan\fR \ex00\ex05\ex00\ex07\exff\exf0 S2S* var1 var2
.CE

will return \fB2\fR with \fB5 7\fR stored in \fIvar1\fR and \fB\-16\fR
stored in \fIvar2\fR.
.RE
.IP \fBt\fR 5
The data is interpreted as \fIcount\fR 16-bit signed integers
represented in the native byte order of the machine running the Tcl

script.  It is otherwise identical to \fBs\fR and \fBS\fR.
To determine what the native byte order of the machine is, refer to
the \fBbyteOrder\fR element of the \fBtcl_platform\fR array.
.IP \fBi\fR 5
The data is interpreted as \fIcount\fR 32-bit signed integers
represented in little-endian byte order.  The integers are stored in

the corresponding variable as a list.  If \fIcount\fR is \fB*\fR, then

all of the remaining bytes in \fIstring\fR will be scanned.  If
\fIcount\fR is omitted, then one 32-bit integer will be scanned.  For
example,
.RS

.CS
set str \ex05\ex00\ex00\ex00\ex07\ex00\ex00\ex00\exf0\exff\exff\exff
\fBbinary scan\fR $str i2i* var1 var2
.CE

will return \fB2\fR with \fB5 7\fR stored in \fIvar1\fR and \fB\-16\fR
stored in \fIvar2\fR.  Note that the integers returned are signed, but
they can be converted to unsigned 32-bit quantities using an expression
like:
.CS
set num [expr { $num & 0xffffffff }]
.CE
.RE
.IP \fBI\fR 5
This form is the same as \fBI\fR except that the data is interpreted
as \fIcount\fR 32-bit signed integers represented in big-endian byte

order.  For example,
.RS

.CS
set str \ex00\ex00\ex00\ex05\ex00\ex00\ex00\ex07\exff\exff\exff\exf0
\fBbinary scan\fR $str I2I* var1 var2
.CE

will return \fB2\fR with \fB5 7\fR stored in \fIvar1\fR and \fB\-16\fR
stored in \fIvar2\fR.
.RE
.IP \fBn\fR 5
The data is interpreted as \fIcount\fR 32-bit signed integers
represented in the native byte order of the machine running the Tcl

script.  It is otherwise identical to \fBi\fR and \fBI\fR.
To determine what the native byte order of the machine is, refer to
the \fBbyteOrder\fR element of the \fBtcl_platform\fR array.
.IP \fBw\fR 5
The data is interpreted as \fIcount\fR 64-bit signed integers
represented in little-endian byte order.  The integers are stored in

the corresponding variable as a list.  If \fIcount\fR is \fB*\fR, then

all of the remaining bytes in \fIstring\fR will be scanned.  If
\fIcount\fR is omitted, then one 64-bit integer will be scanned.  For
example,
.RS

.CS
set str \ex05\ex00\ex00\ex00\ex07\ex00\ex00\ex00\exf0\exff\exff\exff
\fBbinary scan\fR $str wi* var1 var2
.CE

will return \fB2\fR with \fB30064771077\fR stored in \fIvar1\fR and
\fB\-16\fR stored in \fIvar2\fR.  Note that the integers returned are
signed and cannot be represented by Tcl as unsigned values.
.RE
.IP \fBW\fR 5
This form is the same as \fBw\fR except that the data is interpreted
as \fIcount\fR 64-bit signed integers represented in big-endian byte

order.  For example,
.RS

.CS
set str \ex00\ex00\ex00\ex05\ex00\ex00\ex00\ex07\exff\exff\exff\exf0
\fBbinary scan\fR $str WI* var1 var2
.CE

will return \fB2\fR with \fB21474836487\fR stored in \fIvar1\fR and \fB\-16\fR
stored in \fIvar2\fR.
.RE
.IP \fBm\fR 5
The data is interpreted as \fIcount\fR 64-bit signed integers
represented in the native byte order of the machine running the Tcl

script.  It is otherwise identical to \fBw\fR and \fBW\fR.
To determine what the native byte order of the machine is, refer to
the \fBbyteOrder\fR element of the \fBtcl_platform\fR array.
.IP \fBf\fR 5
The data is interpreted as \fIcount\fR single-precision floating point
numbers in the machine's native representation.  The floating point
numbers are stored in the corresponding variable as a list.  If


\fIcount\fR is \fB*\fR, then all of the remaining bytes in
\fIstring\fR will be scanned.  If \fIcount\fR is omitted, then one
single-precision floating point number will be scanned.  The size of a
floating point number may vary across architectures, so the number of
bytes that are scanned may vary.  If the data does not represent a
valid floating point number, the resulting value is undefined and
compiler dependent.  For example, on a Windows system running on an
Intel Pentium processor,
.RS

.CS
\fBbinary scan\fR \ex3f\excc\excc\excd f var1
.CE

will return \fB1\fR with \fB1.6000000238418579\fR stored in
\fIvar1\fR.
.RE
.IP \fBr\fR 5
This form is the same as \fBf\fR except that the data is interpreted
as \fIcount\fR single-precision floating point number in little-endian
order.  This conversion is not portable to the minority of systems not
................................................................................
order.  This conversion is not portable to the minority of systems not
using IEEE floating point representations.
.IP \fBd\fR 5
This form is the same as \fBf\fR except that the data is interpreted
as \fIcount\fR double-precision floating point numbers in the
machine's native representation. For example, on a Windows system
running on an Intel Pentium processor,
.RS

.CS
\fBbinary scan\fR \ex9a\ex99\ex99\ex99\ex99\ex99\exf9\ex3f d var1
.CE

will return \fB1\fR with \fB1.6000000000000001\fR
stored in \fIvar1\fR.
.RE
.IP \fBq\fR 5
This form is the same as \fBd\fR except that the data is interpreted
as \fIcount\fR double-precision floating point number in little-endian
order.  This conversion is not portable to the minority of systems not
................................................................................
.IP \fBQ\fR 5
This form is the same as \fBd\fR except that the data is interpreted
as \fIcount\fR double-precision floating point number in big-endian
order.  This conversion is not portable to the minority of systems not
using IEEE floating point representations.
.IP \fBx\fR 5
Moves the cursor forward \fIcount\fR bytes in \fIstring\fR.  If


\fIcount\fR is \fB*\fR or is larger than the number of bytes after the
current cursor position, then the cursor is positioned after
the last byte in \fIstring\fR.  If \fIcount\fR is omitted, then the
cursor is moved forward one byte.  Note that this type does not
consume an argument.  For example,
.RS

.CS
\fBbinary scan\fR \ex01\ex02\ex03\ex04 x2H* var1
.CE

will return \fB1\fR with \fB0304\fR stored in \fIvar1\fR.
.RE
.IP \fBX\fR 5
Moves the cursor back \fIcount\fR bytes in \fIstring\fR.  If


\fIcount\fR is \fB*\fR or is larger than the current cursor position,
then the cursor is positioned at location 0 so that the next byte
scanned will be the first byte in \fIstring\fR.  If \fIcount\fR
is omitted then the cursor is moved back one byte.  Note that this
type does not consume an argument.  For example,
.RS

.CS
\fBbinary scan\fR \ex01\ex02\ex03\ex04 c2XH* var1 var2
.CE

will return \fB2\fR with \fB1 2\fR stored in \fIvar1\fR and \fB020304\fR
stored in \fIvar2\fR.
.RE
.IP \[email protected]\fR 5
Moves the cursor to the absolute location in the data string specified
by \fIcount\fR.  Note that position 0 refers to the first byte in
\fIstring\fR.  If \fIcount\fR refers to a position beyond the end of
\fIstring\fR, then the cursor is positioned after the last byte.  If
\fIcount\fR is omitted, then an error will be generated.  For example,
.RS

.CS
\fBbinary scan\fR \ex01\ex02\ex03\ex04 [email protected]* var1 var2
.CE

will return \fB2\fR with \fB1 2\fR stored in \fIvar1\fR and \fB020304\fR
stored in \fIvar2\fR.
.RE
.SH "PORTABILITY ISSUES"
.PP
The \fBr\fR, \fBR\fR, \fBq\fR and \fBQ\fR conversions will only work
reliably for transferring data between computers which are all using

.TH binary n 8.0 Tcl "Tcl Built-In Commands"
.so man.macros
.BS
'\" Note:  do not modify the .SH NAME line immediately below!
.SH NAME
binary \- Insert and extract fields from binary strings
.SH SYNOPSIS

\fBbinary decode \fIformat\fR ?\fI\-option value ...\fR? \fIdata\fR
.br
\fBbinary encode \fIformat\fR ?\fI\-option value ...\fR? \fIdata\fR
.br

\fBbinary format \fIformatString \fR?\fIarg arg ...\fR?
.br
\fBbinary scan \fIstring formatString \fR?\fIvarName varName ...\fR?
.BE
.SH DESCRIPTION
.PP
This command provides facilities for manipulating binary data.  The
subcommand \fBbinary format\fR creates a binary string from normal
Tcl values.  For example, given the values 16 and 22, on a 32-bit
architecture, it might produce an 8-byte binary string consisting of
two 4-byte integers, one for each of the numbers.  The subcommand
\fBbinary scan\fR, does the opposite: it extracts data
from a binary string and returns it as ordinary Tcl string values.

The \fBbinary encode\fR and \fBbinary decode\fR subcommands convert
binary data to or from string encodings such as base64 (used in MIME
messages for example).

.PP
Note that other operations on binary data, such as taking a subsequence of it,
getting its length, or reinterpreting it as a string in some encoding, are
done by other Tcl commands (respectively \fBstring range\fR,
\fBstring length\fR and \fBencoding convertfrom\fR in the example cases).  A
binary string in Tcl is merely one where all the characters it contains are in
the range \eu0000\-\eu00FF.
.SH "BINARY ENCODE AND DECODE"

.PP
When encoding binary data as a readable string, the starting binary data is
passed to the \fBbinary encode\fR command, together with the name of the
encoding to use and any encoding-specific options desired. Data which has been
encoded can be converted back to binary form using \fBbinary decode\fR. The
following formats and options are supported.
.TP
................................................................................
.
Instructs the decoder to throw an error if it encounters unexpected whitespace
characters. Otherwise it ignores them.
.PP
Note that neither the encoder nor the decoder handle the header and footer of
the uuencode format.
.RE

.SH "BINARY FORMAT"
.PP
The \fBbinary format\fR command generates a binary string whose layout
is specified by the \fIformatString\fR and whose contents come from
the additional arguments.  The resulting binary value is returned.
.PP
The \fIformatString\fR consists of a sequence of zero or more field
................................................................................
specifiers separated by zero or more spaces.  Each field specifier is
a single type character followed by an optional flag character followed
by an optional numeric \fIcount\fR.
Most field specifiers consume one argument to obtain the value to be
formatted.  The type character specifies how the value is to be
formatted.  The \fIcount\fR typically indicates how many items of the
specified type are taken from the value.  If present, the \fIcount\fR
is a non-negative decimal integer or
.QW \fB*\fR ,
which normally indicates
that all of the items in the value are to be used.  If the number of
arguments does not match the number of fields in the format string
that consume arguments, then an error is generated. The flag character
is ignored for \fBbinary format\fR.
.PP
Here is a small example to clarify the relation between the field
specifiers and the arguments:
.PP
.CS
\fBbinary format\fR d3d {1.0 2.0 3.0 4.0} 0.1
.CE
.PP
The first argument is a list of four numbers, but because of the count
of 3 for the associated field specifier, only the first three will be
used. The second argument is associated with the second field
................................................................................
the \fBencoding convertto\fR command should be used first to change
the string into an external representation
if this truncation is not desired (i.e. if the characters are
not part of the ISO 8859\-1 character set.)
If \fIarg\fR has fewer than \fIcount\fR bytes, then additional zero
bytes are used to pad out the field.  If \fIarg\fR is longer than the
specified length, the extra characters will be ignored.  If
\fIcount\fR is
.QW \fB*\fR ,
then all of the bytes in \fIarg\fR will be
formatted.  If \fIcount\fR is omitted, then one character will be
formatted.  For example, the command:
.RS 
.PP
.CS
\fBbinary format\fR a7a*a alpha bravo charlie
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fBalpha\e000\e000bravoc\fR
.CE
.PP
the command:
.PP
.CS
\fBbinary format\fR a* [encoding convertto utf-8 \eu20ac]
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\e342\e202\e254\fR
.CE
.PP
(which is the
UTF-8 byte sequence for a Euro-currency character), and the command:
.PP
.CS
\fBbinary format\fR a* [encoding convertto iso8859-15 \eu20ac]
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\e244\fR
.CE
.PP
(which is the ISO
8859\-15 byte sequence for a Euro-currency character). Contrast these
last two with: 
.PP
.CS
\fBbinary format\fR a* \eu20ac
.CE 
.PP
which returns a binary string equivalent to:
.PP
.CS
\fB\e254\fR
.CE
.PP
(i.e. \fB\exac\fR) by
truncating the high-bits of the character, and which is probably not
what is desired.
.RE
.IP \fBA\fR 5
This form is the same as \fBa\fR except that spaces are used for
padding instead of nulls.  For example,
.RS 
.PP
.CS
\fBbinary format\fR A6A*A alpha bravo charlie
.CE 
.PP
will return
.PP
.CS
\fBalpha bravoc\fR
.CE
.RE
.IP \fBb\fR 5
Stores a string of \fIcount\fR binary digits in low-to-high order
within each byte in the output binary string.  \fIArg\fR must contain a
sequence of \fB1\fR and \fB0\fR characters.  The resulting bytes are
emitted in first to last order with the bits being formatted in
low-to-high order within each byte.  If \fIarg\fR has fewer than
\fIcount\fR digits, then zeros will be used for the remaining bits.
If \fIarg\fR has more than the specified number of digits, the extra
digits will be ignored.  If \fIcount\fR is
.QW \fB*\fR ,
then all of the
digits in \fIarg\fR will be formatted.  If \fIcount\fR is omitted,
then one digit will be formatted.  If the number of bits formatted
does not end at a byte boundary, the remaining bits of the last byte
will be zeros.  For example,
.RS 
.PP
.CS
\fBbinary format\fR b5b* 11100 111000011010
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\ex07\ex87\ex05\fR
.CE
.RE
.IP \fBB\fR 5
This form is the same as \fBb\fR except that the bits are stored in
high-to-low order within each byte.  For example,
.RS 
.PP
.CS
\fBbinary format\fR B5B* 11100 111000011010
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\exe0\exe1\exa0\fR
.CE
.RE
.IP \fBH\fR 5
Stores a string of \fIcount\fR hexadecimal digits in high-to-low
within each byte in the output binary string.  \fIArg\fR must contain a
sequence of characters in the set
.QW 0123456789abcdefABCDEF .
The resulting bytes are emitted in first to last order with the hex digits
being formatted in high-to-low order within each byte.  If \fIarg\fR
has fewer than \fIcount\fR digits, then zeros will be used for the
remaining digits.  If \fIarg\fR has more than the specified number of
digits, the extra digits will be ignored.  If \fIcount\fR is
.QW \fB*\fR ,
then all of the digits in \fIarg\fR will be formatted.  If
\fIcount\fR is omitted, then one digit will be formatted.  If the
number of digits formatted does not end at a byte boundary, the
remaining bits of the last byte will be zeros.  For example,
.RS 
.PP
.CS
\fBbinary format\fR H3H*H2 ab DEF 987
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\exab\ex00\exde\exf0\ex98\fR
.CE
.RE
.IP \fBh\fR 5
This form is the same as \fBH\fR except that the digits are stored in
low-to-high order within each byte. This is seldom required. For example,
.RS 
.PP
.CS
\fBbinary format\fR h3h*h2 AB def 987
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\exba\ex00\exed\ex0f\ex89\fR
.CE
.RE
.IP \fBc\fR 5
Stores one or more 8-bit integer values in the output string.  If no
\fIcount\fR is specified, then \fIarg\fR must consist of an integer
value. If \fIcount\fR is specified, \fIarg\fR must consist of a list
containing at least that many integers. The low-order 8 bits of each integer
are stored as a one-byte value at the cursor position.  If \fIcount\fR is
.QW \fB*\fR ,
then all of the integers in the list are formatted. If the
number of elements in the list is greater
than \fIcount\fR, then the extra elements are ignored.  For example,
.RS 
.PP
.CS
\fBbinary format\fR c3cc* {3 -3 128 1} 260 {2 5}
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\ex03\exfd\ex80\ex04\ex02\ex05\fR
.CE
.PP
whereas: 
.PP
.CS
\fBbinary format\fR c {2 5}
.CE 
.PP
will generate an error.
.RE
.IP \fBs\fR 5
This form is the same as \fBc\fR except that it stores one or more
16-bit integers in little-endian byte order in the output string.  The
low-order 16-bits of each integer are stored as a two-byte value at
the cursor position with the least significant byte stored first.  For
example,
.RS 
.PP
.CS
\fBbinary format\fR s3 {3 -3 258 1}
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\ex03\ex00\exfd\exff\ex02\ex01\fR
.CE
.RE
.IP \fBS\fR 5
This form is the same as \fBs\fR except that it stores one or more
16-bit integers in big-endian byte order in the output string.  For
example,
.RS 
.PP
.CS
\fBbinary format\fR S3 {3 -3 258 1}
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\ex00\ex03\exff\exfd\ex01\ex02\fR
.CE
.RE
.IP \fBt\fR 5
This form (mnemonically \fItiny\fR) is the same as \fBs\fR and \fBS\fR
except that it stores the 16-bit integers in the output string in the
native byte order of the machine where the Tcl script is running.
To determine what the native byte order of the machine is, refer to
the \fBbyteOrder\fR element of the \fBtcl_platform\fR array.
.IP \fBi\fR 5
This form is the same as \fBc\fR except that it stores one or more
32-bit integers in little-endian byte order in the output string.  The
low-order 32-bits of each integer are stored as a four-byte value at
the cursor position with the least significant byte stored first.  For
example,
.RS 
.PP
.CS
\fBbinary format\fR i3 {3 -3 65536 1}
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\ex03\ex00\ex00\ex00\exfd\exff\exff\exff\ex00\ex00\ex01\ex00\fR
.CE
.RE
.IP \fBI\fR 5
This form is the same as \fBi\fR except that it stores one or more one
or more 32-bit integers in big-endian byte order in the output string.
For example,
.RS 
.PP
.CS
\fBbinary format\fR I3 {3 -3 65536 1}
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\ex00\ex00\ex00\ex03\exff\exff\exff\exfd\ex00\ex01\ex00\ex00\fR
.CE
.RE
.IP \fBn\fR 5
This form (mnemonically \fInumber\fR or \fInormal\fR) is the same as
\fBi\fR and \fBI\fR except that it stores the 32-bit integers in the
output string in the native byte order of the machine where the Tcl
script is running.
To determine what the native byte order of the machine is, refer to
................................................................................
the \fBbyteOrder\fR element of the \fBtcl_platform\fR array.
.IP \fBw\fR 5
This form is the same as \fBc\fR except that it stores one or more
64-bit integers in little-endian byte order in the output string.  The
low-order 64-bits of each integer are stored as an eight-byte value at
the cursor position with the least significant byte stored first.  For
example,
.RS 
.PP
.CS
\fBbinary format\fR w 7810179016327718216
.CE 
.PP
will return the binary string \fBHelloTcl\fR.
.RE
.IP \fBW\fR 5
This form is the same as \fBw\fR except that it stores one or more one
or more 64-bit integers in big-endian byte order in the output string.
For example,
.RS 
.PP
.CS
\fBbinary format\fR Wc 4785469626960341345 110
.CE 
.PP
will return the binary string \fBBigEndian\fR
.RE
.IP \fBm\fR 5
This form (mnemonically the mirror of \fBw\fR) is the same as \fBw\fR
and \fBW\fR except that it stores the 64-bit integers in the output
string in the native byte order of the machine where the Tcl script is
running.
To determine what the native byte order of the machine is, refer to
................................................................................
point number may vary across architectures, so the number of bytes
that are generated may vary.  If the value overflows the
machine's native representation, then the value of FLT_MAX
as defined by the system will be used instead.  Because Tcl uses
double-precision floating point numbers internally, there may be some
loss of precision in the conversion to single-precision.  For example,
on a Windows system running on an Intel Pentium processor,
.RS 
.PP
.CS
\fBbinary format\fR f2 {1.6 3.4}
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\excd\excc\excc\ex3f\ex9a\ex99\ex59\ex40\fR
.CE
.RE
.IP \fBr\fR 5
This form (mnemonically \fIreal\fR) is the same as \fBf\fR except that
it stores the single-precision floating point numbers in little-endian
order.  This conversion only produces meaningful output when used on
machines which use the IEEE floating point representation (very
common, but not universal.)
................................................................................
This form is the same as \fBr\fR except that it stores the
single-precision floating point numbers in big-endian order.
.IP \fBd\fR 5
This form is the same as \fBf\fR except that it stores one or more one
or more double-precision floating point numbers in the machine's native
representation in the output string.  For example, on a
Windows system running on an Intel Pentium processor,
.RS 
.PP
.CS
\fBbinary format\fR d1 {1.6}
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fB\ex9a\ex99\ex99\ex99\ex99\ex99\exf9\ex3f\fR
.CE
.RE
.IP \fBq\fR 5
This form (mnemonically the mirror of \fBd\fR) is the same as \fBd\fR
except that it stores the double-precision floating point numbers in
little-endian order.  This conversion only produces meaningful output
when used on machines which use the IEEE floating point representation
(very common, but not universal.)
.IP \fBQ\fR 5
This form is the same as \fBq\fR except that it stores the
double-precision floating point numbers in big-endian order.
.IP \fBx\fR 5
Stores \fIcount\fR null bytes in the output string.  If \fIcount\fR is
not specified, stores one null byte.  If \fIcount\fR is
.QW \fB*\fR ,
generates an error.  This type does not consume an argument.  For
example,
.RS 
.PP
.CS
\fBbinary format\fR a3xa3x2a3 abc def ghi
.CE 
.PP
will return a binary string equivalent to:
.PP
.CS
\fBabc\e000def\e000\e000ghi\fR
.CE
.RE
.IP \fBX\fR 5
Moves the cursor back \fIcount\fR bytes in the output string.  If
\fIcount\fR is
.QW \fB*\fR
or is larger than the current cursor position,
then the cursor is positioned at location 0 so that the next byte
stored will be the first byte in the result string.  If \fIcount\fR is
omitted then the cursor is moved back one byte.  This type does not
consume an argument.  For example,
.RS 
.PP
.CS
\fBbinary format\fR a3X*a3X2a3 abc def ghi
.CE 
.PP
will return \fBdghi\fR.
.RE
.IP \[email protected]\fR 5
Moves the cursor to the absolute location in the output string
specified by \fIcount\fR.  Position 0 refers to the first byte in the
output string.  If \fIcount\fR refers to a position beyond the last
byte stored so far, then null bytes will be placed in the uninitialized
locations and the cursor will be placed at the specified location.  If
\fIcount\fR is
.QW \fB*\fR ,
then the cursor is moved to the current end of
the output string.  If \fIcount\fR is omitted, then an error will be
generated.  This type does not consume an argument. For example,
.RS 
.PP
.CS
\fBbinary format\fR [email protected]@*[email protected] abcde f ghi j
.CE 
.PP
will return
.PP
.CS
\fBabfdeghi\e000\e000j\fR
.CE
.RE
.SH "BINARY SCAN"
.PP
The \fBbinary scan\fR command parses fields from a binary string,
returning the number of conversions performed.  \fIString\fR gives the
input bytes to be parsed (one byte per character, and characters not
representable as a byte have their high bits chopped)
................................................................................
spaces.  Each field specifier is a single type character followed by
an optional flag character followed by an optional numeric \fIcount\fR.
Most field specifiers consume one
argument to obtain the variable into which the scanned values should
be placed.  The type character specifies how the binary data is to be
interpreted.  The \fIcount\fR typically indicates how many items of
the specified type are taken from the data.  If present, the
\fIcount\fR is a non-negative decimal integer or
.QW \fB*\fR ,
which normally indicates that all of the remaining items in the data are to
be used.  If there are not enough bytes left after the current cursor
position to satisfy the current field specifier, then the
corresponding variable is left untouched and \fBbinary scan\fR returns
immediately with the number of variables that were set.  If there are
not enough arguments for all of the fields in the format string that
consume arguments, then an error is generated. The flag character
.QW u
may be given to cause some types to be read as unsigned values. The flag
is accepted for all field types but is ignored for non-integer fields.
.PP
A similar example as with \fBbinary format\fR should explain the
relation between field specifiers and arguments in case of the binary
scan subcommand: 
.PP
.CS
\fBbinary scan\fR $bytes s3s first second
.CE
.PP
This command (provided the binary string in the variable \fIbytes\fR
is long enough) assigns a list of three integers to the variable
\fIfirst\fR and assigns a single value to the variable \fIsecond\fR.
If \fIbytes\fR contains fewer than 8 bytes (i.e. four 2-byte
integers), no assignment to \fIsecond\fR will be made, and if
\fIbytes\fR contains fewer than 6 bytes (i.e. three 2-byte integers),
no assignment to \fIfirst\fR will be made.  Hence: 
.PP
.CS
puts [\fBbinary scan\fR abcdefg s3s first second]
puts $first
puts $second
.CE 
.PP
will print (assuming neither variable is set previously): 
.PP
.CS
1
25185 25699 26213
can't read "second": no such variable
.CE
.PP
It is \fIimportant\fR to note that the \fBc\fR, \fBs\fR, and \fBS\fR
(and \fBi\fR and \fBI\fR on 64bit systems) will be scanned into
long data size values.  In doing this, values that have their high
bit set (0x80 for chars, 0x8000 for shorts, 0x80000000 for ints),
will be sign extended.  Thus the following will occur: 
.PP
.CS
set signShort [\fBbinary format\fR s1 0x8000]
\fBbinary scan\fR $signShort s1 val; \fI# val == 0xFFFF8000\fR
.CE 
.PP
If you require unsigned values you can include the
.QW u
flag character following
the field type. For example, to read an unsigned short value: 
.PP
.CS
set signShort [\fBbinary format\fR s1 0x8000]
\fBbinary scan\fR $signShort su1 val; \fI# val == 0x00008000\fR
.CE
.PP
Each type-count pair moves an imaginary cursor through the binary data,
reading bytes from the current position.  The cursor is initially
at position 0 at the beginning of the data.  The type may be any one of
the following characters:
.IP \fBa\fR 5
The data is a byte string of length \fIcount\fR.  If \fIcount\fR is
.QW \fB*\fR ,
then all of the remaining bytes in \fIstring\fR will be
scanned into the variable.  If \fIcount\fR is omitted, then one
byte will be scanned.
All bytes scanned will be interpreted as being characters in the
range \eu0000-\eu00ff so the \fBencoding convertfrom\fR command will be
needed if the string is not a binary string or a string encoded in ISO
8859\-1.
For example,
.RS 
.PP
.CS
\fBbinary scan\fR abcde\e000fghi a6a10 var1 var2
.CE 
.PP
will return \fB1\fR with the string equivalent to \fBabcde\e000\fR
stored in \fIvar1\fR and \fIvar2\fR left unmodified, and 
.PP
.CS
\fBbinary scan\fR \e342\e202\e254 a* var1
set var2 [encoding convertfrom utf-8 $var1]
.CE 
.PP
will store a Euro-currency character in \fIvar2\fR.
.RE
.IP \fBA\fR 5
This form is the same as \fBa\fR, except trailing blanks and nulls are stripped from
the scanned value before it is stored in the variable.  For example,
.RS 
.PP
.CS
\fBbinary scan\fR "abc efghi  \e000" A* var1
.CE 
.PP
will return \fB1\fR with \fBabc efghi\fR stored in \fIvar1\fR.
.RE
.IP \fBb\fR 5
The data is turned into a string of \fIcount\fR binary digits in
low-to-high order represented as a sequence of
.QW 1
and
.QW 0
characters.  The data bytes are scanned in first to last order with
the bits being taken in low-to-high order within each byte.  Any extra
bits in the last byte are ignored.  If \fIcount\fR is
.QW \fB*\fR ,
then all of the remaining bits in \fIstring\fR will be scanned.  If
\fIcount\fR is omitted, then one bit will be scanned.  For example,
.RS 
.PP
.CS
\fBbinary scan\fR \ex07\ex87\ex05 b5b* var1 var2
.CE 
.PP
will return \fB2\fR with \fB11100\fR stored in \fIvar1\fR and
\fB1110000110100000\fR stored in \fIvar2\fR.
.RE
.IP \fBB\fR 5
This form is the same as \fBb\fR, except the bits are taken in
high-to-low order within each byte.  For example,
.RS 
.PP
.CS
\fBbinary scan\fR \ex70\ex87\ex05 B5B* var1 var2
.CE 
.PP
will return \fB2\fR with \fB01110\fR stored in \fIvar1\fR and
\fB1000011100000101\fR stored in \fIvar2\fR.
.RE
.IP \fBH\fR 5
The data is turned into a string of \fIcount\fR hexadecimal digits in
high-to-low order represented as a sequence of characters in the set
.QW 0123456789abcdef .
The data bytes are scanned in first to last
order with the hex digits being taken in high-to-low order within each
byte. Any extra bits in the last byte are ignored. If \fIcount\fR is
.QW \fB*\fR ,
then all of the remaining hex digits in \fIstring\fR will be
scanned. If \fIcount\fR is omitted, then one hex digit will be
scanned. For example,
.RS 
.PP
.CS
\fBbinary scan\fR \ex07\exC6\ex05\ex1f\ex34 H3H* var1 var2
.CE 
.PP
will return \fB2\fR with \fB07c\fR stored in \fIvar1\fR and
\fB051f34\fR stored in \fIvar2\fR.
.RE
.IP \fBh\fR 5
This form is the same as \fBH\fR, except the digits are taken in
reverse (low-to-high) order within each byte. For example,
.RS 
.PP
.CS
\fBbinary scan\fR \ex07\ex86\ex05\ex12\ex34 h3h* var1 var2
.CE 
.PP
will return \fB2\fR with \fB706\fR stored in \fIvar1\fR and
\fB502143\fR stored in \fIvar2\fR.
.PP
Note that most code that wishes to parse the hexadecimal digits from
multiple bytes in order should use the \fBH\fR format.
.RE
.IP \fBc\fR 5
The data is turned into \fIcount\fR 8-bit signed integers and stored
in the corresponding variable as a list, or as unsigned if \fBu\fR is placed
immediately after the \fBc\fR. If \fIcount\fR is
.QW \fB*\fR ,
then all of the remaining bytes in \fIstring\fR will be scanned.  If
\fIcount\fR is omitted, then one 8-bit integer will be scanned.  For
example,
.RS 
.PP
.CS
\fBbinary scan\fR \ex07\ex86\ex05 c2c* var1 var2
.CE 
.PP
will return \fB2\fR with \fB7 -122\fR stored in \fIvar1\fR and \fB5\fR
stored in \fIvar2\fR.  Note that the integers returned are signed unless





\fBcu\fR in place of \fBc\fR.
.RE
.IP \fBs\fR 5
The data is interpreted as \fIcount\fR 16-bit signed integers
represented in little-endian byte order, or as unsigned if \fBu\fR is placed
immediately after the \fBs\fR.  The integers are stored in
the corresponding variable as a list.  If \fIcount\fR is
.QW \fB*\fR ,
then all of the remaining bytes in \fIstring\fR will be scanned.  If
\fIcount\fR is omitted, then one 16-bit integer will be scanned.  For
example,
.RS 
.PP
.CS
\fBbinary scan\fR \ex05\ex00\ex07\ex00\exf0\exff s2s* var1 var2
.CE 
.PP
will return \fB2\fR with \fB5 7\fR stored in \fIvar1\fR and \fB\-16\fR
stored in \fIvar2\fR.  Note that the integers returned are signed unless





\fBsu\fR is used in place of \fBs\fR.
.RE
.IP \fBS\fR 5
This form is the same as \fBs\fR except that the data is interpreted
as \fIcount\fR 16-bit integers represented in big-endian byte
order.  For example,
.RS 
.PP
.CS
\fBbinary scan\fR \ex00\ex05\ex00\ex07\exff\exf0 S2S* var1 var2
.CE 
.PP
will return \fB2\fR with \fB5 7\fR stored in \fIvar1\fR and \fB\-16\fR
stored in \fIvar2\fR.
.RE
.IP \fBt\fR 5
The data is interpreted as \fIcount\fR 16-bit signed integers
represented in the native byte order of the machine running the Tcl
script, or as unsigned if \fBu\fR is placed
immediately after the \fBt\fR.  It is otherwise identical to \fBs\fR and \fBS\fR.
To determine what the native byte order of the machine is, refer to
the \fBbyteOrder\fR element of the \fBtcl_platform\fR array.
.IP \fBi\fR 5
The data is interpreted as \fIcount\fR 32-bit signed integers
represented in little-endian byte order, or as unsigned if \fBu\fR is placed
immediately after the \fBi\fR.  The integers are stored in
the corresponding variable as a list.  If \fIcount\fR is
.QW \fB*\fR ,
then all of the remaining bytes in \fIstring\fR will be scanned.  If
\fIcount\fR is omitted, then one 32-bit integer will be scanned.  For
example,
.RS 
.PP
.CS
set str \ex05\ex00\ex00\ex00\ex07\ex00\ex00\ex00\exf0\exff\exff\exff
\fBbinary scan\fR $str i2i* var1 var2
.CE 
.PP
will return \fB2\fR with \fB5 7\fR stored in \fIvar1\fR and \fB\-16\fR
stored in \fIvar2\fR.  Note that the integers returned are signed unless

\fBiu\fR is used in place of \fBi\fR.



.RE
.IP \fBI\fR 5
This form is the same as \fBI\fR except that the data is interpreted
as \fIcount\fR 32-bit signed integers represented in big-endian byte
order, or as unsigned if \fBu\fR is placed
immediately after the \fBI\fR.  For example,
.RS 
.PP
.CS
set str \ex00\ex00\ex00\ex05\ex00\ex00\ex00\ex07\exff\exff\exff\exf0
\fBbinary scan\fR $str I2I* var1 var2
.CE 
.PP
will return \fB2\fR with \fB5 7\fR stored in \fIvar1\fR and \fB\-16\fR
stored in \fIvar2\fR.
.RE
.IP \fBn\fR 5
The data is interpreted as \fIcount\fR 32-bit signed integers
represented in the native byte order of the machine running the Tcl
script, or as unsigned if \fBu\fR is placed
immediately after the \fBn\fR.  It is otherwise identical to \fBi\fR and \fBI\fR.
To determine what the native byte order of the machine is, refer to
the \fBbyteOrder\fR element of the \fBtcl_platform\fR array.
.IP \fBw\fR 5
The data is interpreted as \fIcount\fR 64-bit signed integers
represented in little-endian byte order, or as unsigned if \fBu\fR is placed
immediately after the \fBw\fR.  The integers are stored in
the corresponding variable as a list.  If \fIcount\fR is
.QW \fB*\fR ,
then all of the remaining bytes in \fIstring\fR will be scanned.  If
\fIcount\fR is omitted, then one 64-bit integer will be scanned.  For
example,
.RS 
.PP
.CS
set str \ex05\ex00\ex00\ex00\ex07\ex00\ex00\ex00\exf0\exff\exff\exff
\fBbinary scan\fR $str wi* var1 var2
.CE 
.PP
will return \fB2\fR with \fB30064771077\fR stored in \fIvar1\fR and
\fB\-16\fR stored in \fIvar2\fR.

.RE
.IP \fBW\fR 5
This form is the same as \fBw\fR except that the data is interpreted
as \fIcount\fR 64-bit signed integers represented in big-endian byte
order, or as unsigned if \fBu\fR is placed
immediately after the \fBW\fR.  For example,
.RS 
.PP
.CS
set str \ex00\ex00\ex00\ex05\ex00\ex00\ex00\ex07\exff\exff\exff\exf0
\fBbinary scan\fR $str WI* var1 var2
.CE 
.PP
will return \fB2\fR with \fB21474836487\fR stored in \fIvar1\fR and \fB\-16\fR
stored in \fIvar2\fR.
.RE
.IP \fBm\fR 5
The data is interpreted as \fIcount\fR 64-bit signed integers
represented in the native byte order of the machine running the Tcl
script, or as unsigned if \fBu\fR is placed
immediately after the \fBm\fR.  It is otherwise identical to \fBw\fR and \fBW\fR.
To determine what the native byte order of the machine is, refer to
the \fBbyteOrder\fR element of the \fBtcl_platform\fR array.
.IP \fBf\fR 5
The data is interpreted as \fIcount\fR single-precision floating point
numbers in the machine's native representation.  The floating point
numbers are stored in the corresponding variable as a list.  If
\fIcount\fR is
.QW \fB*\fR ,
then all of the remaining bytes in
\fIstring\fR will be scanned.  If \fIcount\fR is omitted, then one
single-precision floating point number will be scanned.  The size of a
floating point number may vary across architectures, so the number of
bytes that are scanned may vary.  If the data does not represent a
valid floating point number, the resulting value is undefined and
compiler dependent.  For example, on a Windows system running on an
Intel Pentium processor,
.RS 
.PP
.CS
\fBbinary scan\fR \ex3f\excc\excc\excd f var1
.CE 
.PP
will return \fB1\fR with \fB1.6000000238418579\fR stored in
\fIvar1\fR.
.RE
.IP \fBr\fR 5
This form is the same as \fBf\fR except that the data is interpreted
as \fIcount\fR single-precision floating point number in little-endian
order.  This conversion is not portable to the minority of systems not
................................................................................
order.  This conversion is not portable to the minority of systems not
using IEEE floating point representations.
.IP \fBd\fR 5
This form is the same as \fBf\fR except that the data is interpreted
as \fIcount\fR double-precision floating point numbers in the
machine's native representation. For example, on a Windows system
running on an Intel Pentium processor,
.RS 
.PP
.CS
\fBbinary scan\fR \ex9a\ex99\ex99\ex99\ex99\ex99\exf9\ex3f d var1
.CE 
.PP
will return \fB1\fR with \fB1.6000000000000001\fR
stored in \fIvar1\fR.
.RE
.IP \fBq\fR 5
This form is the same as \fBd\fR except that the data is interpreted
as \fIcount\fR double-precision floating point number in little-endian
order.  This conversion is not portable to the minority of systems not
................................................................................
.IP \fBQ\fR 5
This form is the same as \fBd\fR except that the data is interpreted
as \fIcount\fR double-precision floating point number in big-endian
order.  This conversion is not portable to the minority of systems not
using IEEE floating point representations.
.IP \fBx\fR 5
Moves the cursor forward \fIcount\fR bytes in \fIstring\fR.  If
\fIcount\fR is
.QW \fB*\fR
or is larger than the number of bytes after the
current cursor position, then the cursor is positioned after
the last byte in \fIstring\fR.  If \fIcount\fR is omitted, then the
cursor is moved forward one byte.  Note that this type does not
consume an argument.  For example,
.RS 
.PP
.CS
\fBbinary scan\fR \ex01\ex02\ex03\ex04 x2H* var1
.CE 
.PP
will return \fB1\fR with \fB0304\fR stored in \fIvar1\fR.
.RE
.IP \fBX\fR 5
Moves the cursor back \fIcount\fR bytes in \fIstring\fR.  If
\fIcount\fR is
.QW \fB*\fR
or is larger than the current cursor position,
then the cursor is positioned at location 0 so that the next byte
scanned will be the first byte in \fIstring\fR.  If \fIcount\fR
is omitted then the cursor is moved back one byte.  Note that this
type does not consume an argument.  For example,
.RS 
.PP
.CS
\fBbinary scan\fR \ex01\ex02\ex03\ex04 c2XH* var1 var2
.CE 
.PP
will return \fB2\fR with \fB1 2\fR stored in \fIvar1\fR and \fB020304\fR
stored in \fIvar2\fR.
.RE
.IP \[email protected]\fR 5
Moves the cursor to the absolute location in the data string specified
by \fIcount\fR.  Note that position 0 refers to the first byte in
\fIstring\fR.  If \fIcount\fR refers to a position beyond the end of
\fIstring\fR, then the cursor is positioned after the last byte.  If
\fIcount\fR is omitted, then an error will be generated.  For example,
.RS 
.PP
.CS
\fBbinary scan\fR \ex01\ex02\ex03\ex04 [email protected]* var1 var2
.CE 
.PP
will return \fB2\fR with \fB1 2\fR stored in \fIvar1\fR and \fB020304\fR
stored in \fIvar2\fR.
.RE
.SH "PORTABILITY ISSUES"
.PP
The \fBr\fR, \fBR\fR, \fBq\fR and \fBQ\fR conversions will only work
reliably for transferring data between computers which are all using

\fB<<\0\0>>\fR
.
Left and right shift.  Valid for integers.
A right shift always propagates the sign bit.
.TP 20
\fB<\0\0>\0\0<=\0\0>=\fR
.







Boolean less than, greater than, less than or equal, and greater than or equal.




.TP 20
\fB==\0\0!=\fR
.
Boolean equal and not equal.
.TP 20
\fBeq\0\0ne\fR
.
................................................................................
.CE
.PP
A string comparison whose result is 1:
.PP
.CS
\fBexpr\fR {"0y" > "0x12"}
.CE








.PP
Define a procedure that computes an
.QW interesting
mathematical function:
.PP
.CS
proc tcl::mathfunc::calc {x y} {
................................................................................
.SH "SEE ALSO"
array(n), for(n), if(n), mathfunc(n), mathop(n), namespace(n), proc(n),
string(n), Tcl(n), while(n)
.SH KEYWORDS
arithmetic, boolean, compare, expression, fuzzy comparison
.SH COPYRIGHT
.nf
Copyright (c) 1993 The Regents of the University of California.
Copyright (c) 1994-2000 Sun Microsystems Incorporated.
Copyright (c) 2005 by Kevin B. Kenny <[email protected]>. All rights reserved.
.fi
'\" Local Variables:
'\" mode: nroff
'\" End:

\fB<<\0\0>>\fR
.
Left and right shift.  Valid for integers.
A right shift always propagates the sign bit.
.TP 20
\fB<\0\0>\0\0<=\0\0>=\fR
.
Boolean numeric-preferring comparisons: less than, greater than, less than or
equal, and greater than or equal. If either argument is not numeric, the
comparison is done using UNICODE string comparison, as with the string
comparison operators below, which have the same precedence.
.TP 20
\fBlt\0\0gt\0\0le\0\0ge\fR
.VS "8.7, TIP461"
Boolean string comparisons: less than, greater than, less than or equal, and
greater than or equal. These always compare values using their UNICODE strings
(also see \fBstring compare\fR), unlike with the numeric-preferring
comparisons abov, which have the same precedence.
.VE "8.7, TIP461"
.TP 20
\fB==\0\0!=\fR
.
Boolean equal and not equal.
.TP 20
\fBeq\0\0ne\fR
.
................................................................................
.CE
.PP
A string comparison whose result is 1:
.PP
.CS
\fBexpr\fR {"0y" > "0x12"}
.CE
.PP
.VS "8.7, TIP461"
A forced string comparison whose result is 0:
.PP
.CS
\fBexpr\fR {"0x03" gt "2"}
.CE
.VE "8.7, TIP461"
.PP
Define a procedure that computes an
.QW interesting
mathematical function:
.PP
.CS
proc tcl::mathfunc::calc {x y} {
................................................................................
.SH "SEE ALSO"
array(n), for(n), if(n), mathfunc(n), mathop(n), namespace(n), proc(n),
string(n), Tcl(n), while(n)
.SH KEYWORDS
arithmetic, boolean, compare, expression, fuzzy comparison
.SH COPYRIGHT
.nf
Copyright \(co 1993 The Regents of the University of California.
Copyright \(co 1994-2000 Sun Microsystems Incorporated.
Copyright \(co 2005 by Kevin B. Kenny <[email protected]>. All rights reserved.
.fi
'\" Local Variables:
'\" mode: nroff
'\" End:

'\"
'\" Copyright (c) 2018 by Kevin B. Kenny <[email protected]>. All rights reserved
'\" Copyright (c) 2019 by Donal Fellows
'\"
'\" See the file "license.terms" for information on usage and redistribution
'\" of this file, and for a DISCLAIMER OF ALL WARRANTIES.
'\"
.TH fpclassify n 8.7 Tcl "Tcl Float Classifier"
.so man.macros
.BS
'\" Note:  do not modify the .SH NAME line immediately below!
.SH NAME
fpclassify \- Floating point number classification of Tcl values
.SH SYNOPSIS
package require \fBTcl 8.7\fR
.sp
\fBfpclassify \fIvalue\fR
.BE
.SH DESCRIPTION
The \fBfpclassify\fR command takes a floating point number, \fIvalue\fR, and
returns one of the following strings that describe it:
.TP
\fBzero\fR
.
\fIvalue\fR is a floating point zero.
.TP
\fBsubnormal\fR
.
\fIvalue\fR is the result of a gradual underflow.
.TP
\fBnormal\fR
.
\fIvalue\fR is an ordinary floating-point number (not zero, subnormal,
infinite, nor NaN).
.TP
\fBinfinite\fR
.
\fIvalue\fR is a floating-point infinity.
.TP
\fBnan\fR
.
\fIvalue\fR is Not-a-Number.
.PP
The \fBfpclassify\fR command throws an error if value is not a floating-point
value and cannot be converted to one.
.SH EXAMPLE
.PP
This shows how to check whether the result of a computation is numerically
safe or not. (Note however that it does not guard against numerical errors;
just against representational problems.)
.PP
.CS
set value [command-that-computes-a-value]
switch [\fBfpclassify\fR $value] {
    normal - zero {
        puts "Result is $value"
    }
    infinite {
        puts "Result is infinite"
    }
    subnormal {
        puts "Result is $value - WARNING! precision lost"
    }
    nan {
        puts "Computation completely failed"
    }
}
.CE
.SH "SEE ALSO"
expr(n), mathfunc(n)
.SH KEYWORDS
floating point
.SH STANDARDS
This command depends on the \fBfpclassify\fR() C macro conforming to
.QW "ISO C99"
(i.e., to ISO/IEC 9899:1999).
.SH COPYRIGHT
.nf
Copyright \(co 2018 by Kevin B. Kenny <[email protected]>. All rights reserved
.fi
'\" Local Variables:
'\" mode: nroff
'\" End:

\fB::tcl::mathfunc::floor\fR \fIarg\fR
.br
\fB::tcl::mathfunc::fmod\fR \fIx\fR \fIy\fR
.br
\fB::tcl::mathfunc::hypot\fR \fIx\fR \fIy\fR
.br
\fB::tcl::mathfunc::int\fR \fIarg\fR
.br










\fB::tcl::mathfunc::isqrt\fR \fIarg\fR






.br
\fB::tcl::mathfunc::log\fR \fIarg\fR
.br
\fB::tcl::mathfunc::log10\fR \fIarg\fR
.br
\fB::tcl::mathfunc::max\fR \fIarg\fR ?\fIarg\fR ...?
.br
................................................................................
namespace \fB::tcl::mathfunc\fR; these functions are also available
for code apart from \fBexpr\fR, by invoking the given commands
directly.
.PP
Tcl supports the following mathematical functions in expressions, all
of which work solely with floating-point numbers unless otherwise noted:
.DS
.ta 3c 6c 9c
\fBabs\fR	\fBacos\fR	\fBasin\fR	\fBatan\fR
\fBatan2\fR	\fBbool\fR	\fBceil\fR	\fBcos\fR
\fBcosh\fR	\fBdouble\fR	\fBentier\fR	\fBexp\fR
\fBfloor\fR	\fBfmod\fR	\fBhypot\fR	\fBint\fR


\fBisqrt\fR	\fBlog\fR	\fBlog10\fR	\fBmax\fR
\fBmin\fR	\fBpow\fR	\fBrand\fR	\fBround\fR
\fBsin\fR	\fBsinh\fR	\fBsqrt\fR	\fBsrand\fR
\fBtan\fR	\fBtanh\fR	\fBwide\fR
.DE
.PP
In addition to these predefined functions, applications may
define additional functions by using \fBproc\fR (or any other method,
such as \fBinterp alias\fR or \fBTcl_CreateObjCommand\fR) to define
new commands in the \fBtcl::mathfunc\fR namespace.
.SS "DETAILED DEFINITIONS"
................................................................................
.
The argument may be any numeric value.  The integer part of \fIarg\fR
is determined, and then the low order bits of that integer value up
to the machine word size are returned as an integer value.  For reference,
the number of bytes in the machine word are stored in the \fBwordSize\fR
element of the \fBtcl_platform\fR array.
.TP




























\fBisqrt \fIarg\fR
.
Computes the integer part of the square root of \fIarg\fR.  \fIArg\fR must be
a positive value, either an integer or a floating point number.
Unlike \fBsqrt\fR, which is limited to the precision of a floating point
number, \fIisqrt\fR will return a result of arbitrary precision.

















.TP
\fBlog \fIarg\fR
.
Returns the natural logarithm of \fIarg\fR.  \fIArg\fR must be a
positive value.
.TP
\fBlog10 \fIarg\fR
................................................................................
.TP
\fBwide \fIarg\fR
.
The argument may be any numeric value.  The integer part of \fIarg\fR
is determined, and then the low order 64 bits of that integer value
are returned as an integer value.
.SH "SEE ALSO"
expr(n), mathop(n), namespace(n)
.SH "COPYRIGHT"
.nf
Copyright (c) 1993 The Regents of the University of California.
Copyright (c) 1994-2000 Sun Microsystems Incorporated.
Copyright (c) 2005, 2006 by Kevin B. Kenny <[email protected]>.
.fi
'\" Local Variables:
'\" mode: nroff
'\" fill-column: 78
'\" End:

\fB::tcl::mathfunc::floor\fR \fIarg\fR
.br
\fB::tcl::mathfunc::fmod\fR \fIx\fR \fIy\fR
.br
\fB::tcl::mathfunc::hypot\fR \fIx\fR \fIy\fR
.br
\fB::tcl::mathfunc::int\fR \fIarg\fR
.br
.VS "8.7, TIP 521"
\fB::tcl::mathfunc::isfinite\fR \fIarg\fR
.br
\fB::tcl::mathfunc::isinf\fR \fIarg\fR
.br
\fB::tcl::mathfunc::isnan\fR \fIarg\fR
.br
\fB::tcl::mathfunc::isnormal\fR \fIarg\fR
.VE "8.7, TIP 521"
.br
\fB::tcl::mathfunc::isqrt\fR \fIarg\fR
.br
.VS "8.7, TIP 521"
\fB::tcl::mathfunc::issubnormal\fR \fIarg\fR
.br
\fB::tcl::mathfunc::isunordered\fR \fIx y\fR
.VE "8.7, TIP 521"
.br
\fB::tcl::mathfunc::log\fR \fIarg\fR
.br
\fB::tcl::mathfunc::log10\fR \fIarg\fR
.br
\fB::tcl::mathfunc::max\fR \fIarg\fR ?\fIarg\fR ...?
.br
................................................................................
namespace \fB::tcl::mathfunc\fR; these functions are also available
for code apart from \fBexpr\fR, by invoking the given commands
directly.
.PP
Tcl supports the following mathematical functions in expressions, all
of which work solely with floating-point numbers unless otherwise noted:
.DS
.ta 3.2c 6.4c 9.6c
\fBabs\fR	\fBacos\fR	\fBasin\fR	\fBatan\fR
\fBatan2\fR	\fBbool\fR	\fBceil\fR	\fBcos\fR
\fBcosh\fR	\fBdouble\fR	\fBentier\fR	\fBexp\fR
\fBfloor\fR	\fBfmod\fR	\fBhypot\fR	\fBint\fR
\fBisfinite\fR	\fBisinf\fR	\fBisnan\fR	\fBisnormal\fR
\fBisqrt\fR	\fBissubnormal\fR	\fBisunordered\fR	\fBlog\fR
\fBlog10\fR	\fBmax\fR	\fBmin\fR	\fBpow\fR
\fBrand\fR	\fBround\fR	\fBsin\fR	\fBsinh\fR
\fBsqrt\fR	\fBsrand\fR	\fBtan\fR	\fBtanh\fR
\fBwide\fR
.DE
.PP
In addition to these predefined functions, applications may
define additional functions by using \fBproc\fR (or any other method,
such as \fBinterp alias\fR or \fBTcl_CreateObjCommand\fR) to define
new commands in the \fBtcl::mathfunc\fR namespace.
.SS "DETAILED DEFINITIONS"
................................................................................
.
The argument may be any numeric value.  The integer part of \fIarg\fR
is determined, and then the low order bits of that integer value up
to the machine word size are returned as an integer value.  For reference,
the number of bytes in the machine word are stored in the \fBwordSize\fR
element of the \fBtcl_platform\fR array.
.TP
\fBisfinite \fIarg\fR
.VS "8.7, TIP 521"
Returns 1 if the floating-point number \fIarg\fR is finite. That is, if it is
zero, subnormal, or normal. Returns 0 if the number is infinite or NaN. Throws
an error if \fIarg\fR cannot be promoted to a floating-point value.
.VE "8.7, TIP 521"
.TP
\fBisinf \fIarg\fR
.VS "8.7, TIP 521"
Returns 1 if the floating-point number \fIarg\fR is infinite. Returns 0 if the
number is finite or NaN. Throws an error if \fIarg\fR cannot be promoted to a
floating-point value.
.VE "8.7, TIP 521"
.TP
\fBisnan \fIarg\fR
.VS "8.7, TIP 521"
Returns 1 if the floating-point number \fIarg\fR is Not-a-Number. Returns 0 if
the number is finite or infinite. Throws an error if \fIarg\fR cannot be
promoted to a floating-point value.
.VE "8.7, TIP 521"
.TP
\fBisnormal \fIarg\fR
.VS "8.7, TIP 521"
Returns 1 if the floating-point number \fIarg\fR is normal. Returns 0 if the
number is zero, subnormal, infinite or NaN. Throws an error if \fIarg\fR
cannot be promoted to a floating-point value.
.VE "8.7, TIP 521"
.TP
\fBisqrt \fIarg\fR
.
Computes the integer part of the square root of \fIarg\fR.  \fIArg\fR must be
a positive value, either an integer or a floating point number.
Unlike \fBsqrt\fR, which is limited to the precision of a floating point
number, \fIisqrt\fR will return a result of arbitrary precision.
.TP
\fBissubnormal \fIarg\fR
.VS "8.7, TIP 521"
Returns 1 if the floating-point number \fIarg\fR is subnormal, i.e., the
result of gradual underflow. Returns 0 if the number is zero, normal, infinite
or NaN. Throws an error if \fIarg\fR cannot be promoted to a floating-point
value.
.VE "8.7, TIP 521"
.TP
\fBisunordered \fIx y\fR
.VS "8.7, TIP 521"
Returns 1 if \fIx\fR and \fIy\fR cannot be compared for ordering, that is, if
either one is NaN. Returns 0 if both values can be ordered, that is, if they
are both chosen from among the set of zero, subnormal, normal and infinite
values. Throws an error if either \fIx\fR or \fIy\fR cannot be promoted to a
floating-point value.
.VE "8.7, TIP 521"
.TP
\fBlog \fIarg\fR
.
Returns the natural logarithm of \fIarg\fR.  \fIArg\fR must be a
positive value.
.TP
\fBlog10 \fIarg\fR
................................................................................
.TP
\fBwide \fIarg\fR
.
The argument may be any numeric value.  The integer part of \fIarg\fR
is determined, and then the low order 64 bits of that integer value
are returned as an integer value.
.SH "SEE ALSO"
expr(n), fpclassify(n), mathop(n), namespace(n)
.SH "COPYRIGHT"
.nf
Copyright \(co 1993 The Regents of the University of California.
Copyright \(co 1994-2000 Sun Microsystems Incorporated.
Copyright \(co 2005, 2006 by Kevin B. Kenny <[email protected]>.
.fi
'\" Local Variables:
'\" mode: nroff
'\" fill-column: 78
'\" End:

\fB::tcl::mathop::>=\fR ?\fIarg\fR ...?
.br
\fB::tcl::mathop::>\fR ?\fIarg\fR ...?
.br
\fB::tcl::mathop::eq\fR ?\fIarg\fR ...?
.br
\fB::tcl::mathop::ne\fR \fIarg arg\fR










.br
\fB::tcl::mathop::in\fR \fIarg list\fR
.br
\fB::tcl::mathop::ni\fR \fIarg list\fR
.sp
.BE
.SH DESCRIPTION
................................................................................
The following operator commands are supported:
.DS
.ta 2c 4c 6c 8c
\fB~\fR	\fB!\fR	\fB+\fR	\fB\-\fR	\fB*\fR
\fB/\fR	\fB%\fR	\fB**\fR	\fB&\fR	\fB|\fR
\fB^\fR	\fB>>\fR	\fB<<\fR	\fB==\fR	\fBeq\fR
\fB!=\fR	\fBne\fR	\fB<\fR	\fB<=\fR	\fB>\fR
\fB>=\fR	\fBin\fR	\fBni\fR

.DE
.SS "MATHEMATICAL OPERATORS"
.PP
The behaviors of the mathematical operator commands are as follows:
.TP
\fB!\fR \fIboolean\fR
.
................................................................................
\fB<\fR ?\fIarg\fR ...?
.
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be strictly more than the one preceding it.
Comparisons are performed preferentially on the numeric values, and are
otherwise performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value. When the
arguments are numeric but should be compared as strings, the \fBstring
compare\fR command should be used instead.
.TP
\fB<=\fR ?\fIarg\fR ...?
.
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be equal to or more than the one preceding it.
Comparisons are performed preferentially on the numeric values, and are
otherwise performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value. When the
arguments are numeric but should be compared as strings, the \fBstring
compare\fR command should be used instead.
.TP
\fB>\fR ?\fIarg\fR ...?
.
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be strictly less than the one preceding it.
Comparisons are performed preferentially on the numeric values, and are
otherwise performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value. When the
arguments are numeric but should be compared as strings, the \fBstring
compare\fR command should be used instead.
.TP
\fB>=\fR ?\fIarg\fR ...?
.
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be equal to or less than the one preceding it.
Comparisons are performed preferentially on the numeric values, and are
otherwise performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value. When the
arguments are numeric but should be compared as strings, the \fBstring
compare\fR command should be used instead.
































.SS "BIT-WISE OPERATORS"
.PP
The behaviors of the bit-wise operator commands (all of which only operate on
integral arguments) are as follows:
.TP
\fB~\fR \fInumber\fR
.
................................................................................

\fI# Test for list membership\fR
set gotIt [\fBin\fR 3 $list]

\fI# Test to see if a value is within some defined range\fR
set inRange [\fB<=\fR 1 $x 5]

\fI# Test to see if a list is sorted\fR
set sorted [\fB<=\fR {*}$list]




.CE
.SH "SEE ALSO"
expr(n), mathfunc(n), namespace(n)
.SH KEYWORDS
command, expression, operator
'\" Local Variables:
'\" mode: nroff
'\" End:

\fB::tcl::mathop::>=\fR ?\fIarg\fR ...?
.br
\fB::tcl::mathop::>\fR ?\fIarg\fR ...?
.br
\fB::tcl::mathop::eq\fR ?\fIarg\fR ...?
.br
\fB::tcl::mathop::ne\fR \fIarg arg\fR
.br
.VS "8.7, TIP461"
\fB::tcl::mathop::lt\fR ?\fIarg\fR ...?
.br
\fB::tcl::mathop::le\fR ?\fIarg\fR ...?
.br
\fB::tcl::mathop::gt\fR ?\fIarg\fR ...?
.br
\fB::tcl::mathop::ge\fR ?\fIarg\fR ...?
.VE "8.7, TIP461"
.br
\fB::tcl::mathop::in\fR \fIarg list\fR
.br
\fB::tcl::mathop::ni\fR \fIarg list\fR
.sp
.BE
.SH DESCRIPTION
................................................................................
The following operator commands are supported:
.DS
.ta 2c 4c 6c 8c
\fB~\fR	\fB!\fR	\fB+\fR	\fB\-\fR	\fB*\fR
\fB/\fR	\fB%\fR	\fB**\fR	\fB&\fR	\fB|\fR
\fB^\fR	\fB>>\fR	\fB<<\fR	\fB==\fR	\fBeq\fR
\fB!=\fR	\fBne\fR	\fB<\fR	\fB<=\fR	\fB>\fR
\fB>=\fR	\fBin\fR	\fBni\fR	\fBlt\fR	\fBle\fR
\fBgt\fR	\fBge\fR
.DE
.SS "MATHEMATICAL OPERATORS"
.PP
The behaviors of the mathematical operator commands are as follows:
.TP
\fB!\fR \fIboolean\fR
.
................................................................................
\fB<\fR ?\fIarg\fR ...?
.
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be strictly more than the one preceding it.
Comparisons are performed preferentially on the numeric values, and are
otherwise performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value. When the
arguments are numeric but should be compared as strings, the \fBlt\fR
operator or the \fBstring compare\fR command should be used instead.
.TP
\fB<=\fR ?\fIarg\fR ...?
.
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be equal to or more than the one preceding it.
Comparisons are performed preferentially on the numeric values, and are
otherwise performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value. When the
arguments are numeric but should be compared as strings,  the \fBle\fR
operator or the \fBstring compare\fR command should be used instead.
.TP
\fB>\fR ?\fIarg\fR ...?
.
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be strictly less than the one preceding it.
Comparisons are performed preferentially on the numeric values, and are
otherwise performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value. When the
arguments are numeric but should be compared as strings, the \fBgt\fR
operator or the \fBstring compare\fR command should be used instead.
.TP
\fB>=\fR ?\fIarg\fR ...?
.
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be equal to or less than the one preceding it.
Comparisons are performed preferentially on the numeric values, and are
otherwise performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value. When the
arguments are numeric but should be compared as strings, the \fBge\fR
operator or the \fBstring compare\fR command should be used instead.
.TP
\fBlt\fR ?\fIarg\fR ...?
.VS "8.7, TIP461"
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be strictly more than the one preceding it.
Comparisons are performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value.
.VE "8.7, TIP461"
.TP
\fBle\fR ?\fIarg\fR ...?
.VS "8.7, TIP461"
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be equal to or strictly more than the one preceding it.
Comparisons are performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value.
.VE "8.7, TIP461"
.TP
\fBgt\fR ?\fIarg\fR ...?
.VS "8.7, TIP461"
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be strictly less than the one preceding it.
Comparisons are performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value.
.VE "8.7, TIP461"
.TP
\fBge\fR ?\fIarg\fR ...?
.VS "8.7, TIP461"
Returns whether the arbitrarily-many arguments are ordered, with each argument
after the first having to be equal to or strictly less than the one preceding it.
Comparisons are performed using UNICODE string comparison. If fewer than two
arguments are present, this operation always returns a true value.
.VE "8.7, TIP461"
.SS "BIT-WISE OPERATORS"
.PP
The behaviors of the bit-wise operator commands (all of which only operate on
integral arguments) are as follows:
.TP
\fB~\fR \fInumber\fR
.
................................................................................

\fI# Test for list membership\fR
set gotIt [\fBin\fR 3 $list]

\fI# Test to see if a value is within some defined range\fR
set inRange [\fB<=\fR 1 $x 5]

\fI# Test to see if a list is numerically sorted\fR
set sorted [\fB<=\fR {*}$list]

\fI# Test to see if a list is lexically sorted\fR
set alphaList {a b c d e f}
set sorted [\fBle\fR {*}$alphaList]
.CE
.SH "SEE ALSO"
expr(n), mathfunc(n), namespace(n)
.SH KEYWORDS
command, expression, operator
'\" Local Variables:
'\" mode: nroff
'\" End:

    {"strcaseLower",	ASSEM_1BYTE,	INST_STR_LOWER,		1,	1},
    {"strcaseTitle",	ASSEM_1BYTE,	INST_STR_TITLE,		1,	1},
    {"strcaseUpper",	ASSEM_1BYTE,	INST_STR_UPPER,		1,	1},
    {"strcmp",		ASSEM_1BYTE,	INST_STR_CMP,		2,	1},
    {"strcat",		ASSEM_CONCAT1,	INST_STR_CONCAT1,	INT_MIN,1},
    {"streq",		ASSEM_1BYTE,	INST_STR_EQ,		2,	1},
    {"strfind",		ASSEM_1BYTE,	INST_STR_FIND,		2,	1},


    {"strindex",	ASSEM_1BYTE,	INST_STR_INDEX,		2,	1},

    {"strlen",		ASSEM_1BYTE,	INST_STR_LEN,		1,	1},

    {"strmap",		ASSEM_1BYTE,	INST_STR_MAP,		3,	1},
    {"strmatch",	ASSEM_BOOL,	INST_STR_MATCH,		2,	1},
    {"strneq",		ASSEM_1BYTE,	INST_STR_NEQ,		2,	1},
    {"strrange",	ASSEM_1BYTE,	INST_STR_RANGE,		3,	1},
    {"strreplace",	ASSEM_1BYTE,	INST_STR_REPLACE,	4,	1},
    {"strrfind",	ASSEM_1BYTE,	INST_STR_FIND_LAST,	2,	1},
    {"strtrim",		ASSEM_1BYTE,	INST_STR_TRIM,		2,	1},
................................................................................
    INST_COROUTINE_NAME,					/* 149 */
    INST_NS_CURRENT,						/* 151 */
    INST_INFO_LEVEL_NUM,					/* 152 */
    INST_RESOLVE_COMMAND,					/* 154 */
    INST_STR_TRIM, INST_STR_TRIM_LEFT, INST_STR_TRIM_RIGHT,	/* 166-168 */
    INST_CONCAT_STK,						/* 169 */
    INST_STR_UPPER, INST_STR_LOWER, INST_STR_TITLE,		/* 170-172 */
    INST_NUM_TYPE						/* 180 */

};

/*
 * Helper macros.
 */

#if defined(TCL_DEBUG_ASSEMBLY) && defined(__GNUC__) && __GNUC__ > 2

    {"strcaseLower",	ASSEM_1BYTE,	INST_STR_LOWER,		1,	1},
    {"strcaseTitle",	ASSEM_1BYTE,	INST_STR_TITLE,		1,	1},
    {"strcaseUpper",	ASSEM_1BYTE,	INST_STR_UPPER,		1,	1},
    {"strcmp",		ASSEM_1BYTE,	INST_STR_CMP,		2,	1},
    {"strcat",		ASSEM_CONCAT1,	INST_STR_CONCAT1,	INT_MIN,1},
    {"streq",		ASSEM_1BYTE,	INST_STR_EQ,		2,	1},
    {"strfind",		ASSEM_1BYTE,	INST_STR_FIND,		2,	1},
    {"strge",		ASSEM_1BYTE,	INST_STR_GE,		2,	1},
    {"strgt",		ASSEM_1BYTE,	INST_STR_GT,		2,	1},
    {"strindex",	ASSEM_1BYTE,	INST_STR_INDEX,		2,	1},
    {"strle",		ASSEM_1BYTE,	INST_STR_LE,		2,	1},
    {"strlen",		ASSEM_1BYTE,	INST_STR_LEN,		1,	1},
    {"strlt",		ASSEM_1BYTE,	INST_STR_LT,		2,	1},
    {"strmap",		ASSEM_1BYTE,	INST_STR_MAP,		3,	1},
    {"strmatch",	ASSEM_BOOL,	INST_STR_MATCH,		2,	1},
    {"strneq",		ASSEM_1BYTE,	INST_STR_NEQ,		2,	1},
    {"strrange",	ASSEM_1BYTE,	INST_STR_RANGE,		3,	1},
    {"strreplace",	ASSEM_1BYTE,	INST_STR_REPLACE,	4,	1},
    {"strrfind",	ASSEM_1BYTE,	INST_STR_FIND_LAST,	2,	1},
    {"strtrim",		ASSEM_1BYTE,	INST_STR_TRIM,		2,	1},
................................................................................
    INST_COROUTINE_NAME,					/* 149 */
    INST_NS_CURRENT,						/* 151 */
    INST_INFO_LEVEL_NUM,					/* 152 */
    INST_RESOLVE_COMMAND,					/* 154 */
    INST_STR_TRIM, INST_STR_TRIM_LEFT, INST_STR_TRIM_RIGHT,	/* 166-168 */
    INST_CONCAT_STK,						/* 169 */
    INST_STR_UPPER, INST_STR_LOWER, INST_STR_TITLE,		/* 170-172 */
    INST_NUM_TYPE,						/* 180 */
    INST_STR_LT, INST_STR_GT, INST_STR_LE, INST_STR_GE		/* 191-194 */
};

/*
 * Helper macros.
 */

#if defined(TCL_DEBUG_ASSEMBLY) && defined(__GNUC__) && __GNUC__ > 2

#include "tclInt.h"
#include "tclOOInt.h"
#include "tclCompile.h"
#include "tommath.h"
#include <math.h>
#include <assert.h>




#define INTERP_STACK_INITIAL_SIZE 2000
#define CORO_STACK_INITIAL_SIZE    200

/*
 * Determine whether we're using IEEE floating point
 */
................................................................................
static Tcl_ObjCmdProc	ExprBinaryFunc;
static Tcl_ObjCmdProc	ExprBoolFunc;
static Tcl_ObjCmdProc	ExprCeilFunc;
static Tcl_ObjCmdProc	ExprDoubleFunc;
static Tcl_ObjCmdProc	ExprFloorFunc;
static Tcl_ObjCmdProc	ExprIntFunc;
static Tcl_ObjCmdProc	ExprIsqrtFunc;






static Tcl_ObjCmdProc	ExprMaxFunc;
static Tcl_ObjCmdProc	ExprMinFunc;
static Tcl_ObjCmdProc	ExprRandFunc;
static Tcl_ObjCmdProc	ExprRoundFunc;
static Tcl_ObjCmdProc	ExprSqrtFunc;
static Tcl_ObjCmdProc	ExprSrandFunc;
static Tcl_ObjCmdProc	ExprUnaryFunc;
static Tcl_ObjCmdProc	ExprWideFunc;

static void		MathFuncWrongNumArgs(Tcl_Interp *interp, int expected,
			    int actual, Tcl_Obj *const *objv);
static Tcl_NRPostProc	NRCoroutineCallerCallback;
static Tcl_NRPostProc	NRCoroutineExitCallback;
static Tcl_NRPostProc	NRCommand;

static void		ProcessUnexpectedResult(Tcl_Interp *interp,
................................................................................
    {"coroutine",	NULL,			NULL,			TclNRCoroutineObjCmd,	CMD_IS_SAFE},
    {"error",		Tcl_ErrorObjCmd,	TclCompileErrorCmd,	NULL,	CMD_IS_SAFE},
    {"eval",		Tcl_EvalObjCmd,		NULL,			TclNREvalObjCmd,	CMD_IS_SAFE},
    {"expr",		Tcl_ExprObjCmd,		TclCompileExprCmd,	TclNRExprObjCmd,	CMD_IS_SAFE},
    {"for",		Tcl_ForObjCmd,		TclCompileForCmd,	TclNRForObjCmd,	CMD_IS_SAFE},
    {"foreach",		Tcl_ForeachObjCmd,	TclCompileForeachCmd,	TclNRForeachCmd,	CMD_IS_SAFE},
    {"format",		Tcl_FormatObjCmd,	TclCompileFormatCmd,	NULL,	CMD_IS_SAFE},

    {"global",		Tcl_GlobalObjCmd,	TclCompileGlobalCmd,	NULL,	CMD_IS_SAFE},
    {"if",		Tcl_IfObjCmd,		TclCompileIfCmd,	TclNRIfObjCmd,	CMD_IS_SAFE},
    {"incr",		Tcl_IncrObjCmd,		TclCompileIncrCmd,	NULL,	CMD_IS_SAFE},
    {"join",		Tcl_JoinObjCmd,		NULL,			NULL,	CMD_IS_SAFE},
    {"lappend",		Tcl_LappendObjCmd,	TclCompileLappendCmd,	NULL,	CMD_IS_SAFE},
    {"lassign",		Tcl_LassignObjCmd,	TclCompileLassignCmd,	NULL,	CMD_IS_SAFE},
    {"lindex",		Tcl_LindexObjCmd,	TclCompileLindexCmd,	NULL,	CMD_IS_SAFE},
................................................................................
    { "double",	ExprDoubleFunc,	NULL			},
    { "entier",	ExprIntFunc,	NULL			},
    { "exp",	ExprUnaryFunc,	(ClientData) exp	},
    { "floor",	ExprFloorFunc,	NULL			},
    { "fmod",	ExprBinaryFunc,	(ClientData) fmod	},
    { "hypot",	ExprBinaryFunc,	(ClientData) hypot	},
    { "int",	ExprIntFunc,	NULL			},




    { "isqrt",	ExprIsqrtFunc,	NULL			},


    { "log",	ExprUnaryFunc,	(ClientData) log	},
    { "log10",	ExprUnaryFunc,	(ClientData) log10	},
    { "max",	ExprMaxFunc,	NULL			},
    { "min",	ExprMinFunc,	NULL			},
    { "pow",	ExprBinaryFunc,	(ClientData) pow	},
    { "rand",	ExprRandFunc,	NULL			},
    { "round",	ExprRoundFunc,	NULL			},
................................................................................
    { ">",	TclSortingOpCmd,	TclCompileGreaterOpCmd,
		/* unused */ {0},	NULL},
    { ">=",	TclSortingOpCmd,	TclCompileGeqOpCmd,
		/* unused */ {0},	NULL},
    { "==",	TclSortingOpCmd,	TclCompileEqOpCmd,
		/* unused */ {0},	NULL},
    { "eq",	TclSortingOpCmd,	TclCompileStreqOpCmd,








		/* unused */ {0},	NULL},
    { NULL,	NULL,			NULL,
		{0},			NULL}
};
 
/*
 *----------------------------------------------------------------------
................................................................................
     * To avoid duplicating the random number generation code we simply clean
     * up our state and call the real random number function. That function
     * will always succeed.
     */

    return ExprRandFunc(clientData, interp, 1, objv);
}





















































































































































































































































































































 
/*
 *----------------------------------------------------------------------
 *
 * MathFuncWrongNumArgs --
 *
 *	Generate an error message when a math function presents the wrong

#include "tclInt.h"
#include "tclOOInt.h"
#include "tclCompile.h"
#include "tommath.h"
#include <math.h>
#include <assert.h>
#ifndef fpclassify /* Older MSVC */
#include <float.h>
#endif /* !fpclassify */

#define INTERP_STACK_INITIAL_SIZE 2000
#define CORO_STACK_INITIAL_SIZE    200

/*
 * Determine whether we're using IEEE floating point
 */
................................................................................
static Tcl_ObjCmdProc	ExprBinaryFunc;
static Tcl_ObjCmdProc	ExprBoolFunc;
static Tcl_ObjCmdProc	ExprCeilFunc;
static Tcl_ObjCmdProc	ExprDoubleFunc;
static Tcl_ObjCmdProc	ExprFloorFunc;
static Tcl_ObjCmdProc	ExprIntFunc;
static Tcl_ObjCmdProc	ExprIsqrtFunc;
static Tcl_ObjCmdProc   ExprIsFiniteFunc;
static Tcl_ObjCmdProc   ExprIsInfinityFunc;
static Tcl_ObjCmdProc   ExprIsNaNFunc;
static Tcl_ObjCmdProc   ExprIsNormalFunc;
static Tcl_ObjCmdProc   ExprIsSubnormalFunc;
static Tcl_ObjCmdProc   ExprIsUnorderedFunc;
static Tcl_ObjCmdProc	ExprMaxFunc;
static Tcl_ObjCmdProc	ExprMinFunc;
static Tcl_ObjCmdProc	ExprRandFunc;
static Tcl_ObjCmdProc	ExprRoundFunc;
static Tcl_ObjCmdProc	ExprSqrtFunc;
static Tcl_ObjCmdProc	ExprSrandFunc;
static Tcl_ObjCmdProc	ExprUnaryFunc;
static Tcl_ObjCmdProc	ExprWideFunc;
static Tcl_ObjCmdProc   FloatClassifyObjCmd;
static void		MathFuncWrongNumArgs(Tcl_Interp *interp, int expected,
			    int actual, Tcl_Obj *const *objv);
static Tcl_NRPostProc	NRCoroutineCallerCallback;
static Tcl_NRPostProc	NRCoroutineExitCallback;
static Tcl_NRPostProc	NRCommand;

static void		ProcessUnexpectedResult(Tcl_Interp *interp,
................................................................................
    {"coroutine",	NULL,			NULL,			TclNRCoroutineObjCmd,	CMD_IS_SAFE},
    {"error",		Tcl_ErrorObjCmd,	TclCompileErrorCmd,	NULL,	CMD_IS_SAFE},
    {"eval",		Tcl_EvalObjCmd,		NULL,			TclNREvalObjCmd,	CMD_IS_SAFE},
    {"expr",		Tcl_ExprObjCmd,		TclCompileExprCmd,	TclNRExprObjCmd,	CMD_IS_SAFE},
    {"for",		Tcl_ForObjCmd,		TclCompileForCmd,	TclNRForObjCmd,	CMD_IS_SAFE},
    {"foreach",		Tcl_ForeachObjCmd,	TclCompileForeachCmd,	TclNRForeachCmd,	CMD_IS_SAFE},
    {"format",		Tcl_FormatObjCmd,	TclCompileFormatCmd,	NULL,	CMD_IS_SAFE},
    {"fpclassify",      FloatClassifyObjCmd,    NULL,                   NULL,   CMD_IS_SAFE},
    {"global",		Tcl_GlobalObjCmd,	TclCompileGlobalCmd,	NULL,	CMD_IS_SAFE},
    {"if",		Tcl_IfObjCmd,		TclCompileIfCmd,	TclNRIfObjCmd,	CMD_IS_SAFE},
    {"incr",		Tcl_IncrObjCmd,		TclCompileIncrCmd,	NULL,	CMD_IS_SAFE},
    {"join",		Tcl_JoinObjCmd,		NULL,			NULL,	CMD_IS_SAFE},
    {"lappend",		Tcl_LappendObjCmd,	TclCompileLappendCmd,	NULL,	CMD_IS_SAFE},
    {"lassign",		Tcl_LassignObjCmd,	TclCompileLassignCmd,	NULL,	CMD_IS_SAFE},
    {"lindex",		Tcl_LindexObjCmd,	TclCompileLindexCmd,	NULL,	CMD_IS_SAFE},
................................................................................
    { "double",	ExprDoubleFunc,	NULL			},
    { "entier",	ExprIntFunc,	NULL			},
    { "exp",	ExprUnaryFunc,	(ClientData) exp	},
    { "floor",	ExprFloorFunc,	NULL			},
    { "fmod",	ExprBinaryFunc,	(ClientData) fmod	},
    { "hypot",	ExprBinaryFunc,	(ClientData) hypot	},
    { "int",	ExprIntFunc,	NULL			},
    { "isfinite", ExprIsFiniteFunc, NULL        	},
    { "isinf",	ExprIsInfinityFunc, NULL        	},
    { "isnan",	ExprIsNaNFunc,	NULL            	},
    { "isnormal", ExprIsNormalFunc, NULL        	},
    { "isqrt",	ExprIsqrtFunc,	NULL			},
    { "issubnormal", ExprIsSubnormalFunc, NULL,         },
    { "isunordered", ExprIsUnorderedFunc, NULL,         },
    { "log",	ExprUnaryFunc,	(ClientData) log	},
    { "log10",	ExprUnaryFunc,	(ClientData) log10	},
    { "max",	ExprMaxFunc,	NULL			},
    { "min",	ExprMinFunc,	NULL			},
    { "pow",	ExprBinaryFunc,	(ClientData) pow	},
    { "rand",	ExprRandFunc,	NULL			},
    { "round",	ExprRoundFunc,	NULL			},
................................................................................
    { ">",	TclSortingOpCmd,	TclCompileGreaterOpCmd,
		/* unused */ {0},	NULL},
    { ">=",	TclSortingOpCmd,	TclCompileGeqOpCmd,
		/* unused */ {0},	NULL},
    { "==",	TclSortingOpCmd,	TclCompileEqOpCmd,
		/* unused */ {0},	NULL},
    { "eq",	TclSortingOpCmd,	TclCompileStreqOpCmd,
		/* unused */ {0},	NULL},
    { "lt",	TclSortingOpCmd,	TclCompileStrLtOpCmd,
		/* unused */ {0},	NULL},
    { "le",	TclSortingOpCmd,	TclCompileStrLeOpCmd,
		/* unused */ {0},	NULL},
    { "gt",	TclSortingOpCmd,	TclCompileStrGtOpCmd,
		/* unused */ {0},	NULL},
    { "ge",	TclSortingOpCmd,	TclCompileStrGeOpCmd,
		/* unused */ {0},	NULL},
    { NULL,	NULL,			NULL,
		{0},			NULL}
};
 
/*
 *----------------------------------------------------------------------
................................................................................
     * To avoid duplicating the random number generation code we simply clean
     * up our state and call the real random number function. That function
     * will always succeed.
     */

    return ExprRandFunc(clientData, interp, 1, objv);
}
 
/*
 *----------------------------------------------------------------------
 *
 * Double Classification Functions --
 *
 *	This page contains the functions that implement all of the built-in
 *	math functions for classifying IEEE doubles.
 *
 *      These have to be a little bit careful while Tcl_GetDoubleFromObj()
 *      rejects NaN values, which these functions *explicitly* accept.
 *
 * Results:
 *	Each function returns TCL_OK if it succeeds and pushes an Tcl object
 *	holding the result. If it fails it returns TCL_ERROR and leaves an
 *	error message in the interpreter's result.
 *
 * Side effects:
 *	None.
 *
 *----------------------------------------------------------------------
 */

/*
 * Older MSVC is supported by Tcl, but doesn't have fpclassify(). Of course.
 * But it does have _fpclass() which does almost the same job.
 *
 * This makes it conform to the C99 standard API, and just delegates to the
 * standard macro on platforms that do it correctly.
 */

static inline int
ClassifyDouble(
    double d)
{
#ifdef fpclassify
    return fpclassify(d);
#else /* !fpclassify */
#define FP_ZERO 0
#define FP_NORMAL 1
#define FP_SUBNORMAL 2
#define FP_INFINITE 3
#define FP_NAN 4

    switch (_fpclass(d)) {
    case _FPCLASS_NZ:
    case _FPCLASS_PZ:
        return FP_ZERO;
    case _FPCLASS_NN:
    case _FPCLASS_PN:
        return FP_NORMAL;
    case _FPCLASS_ND:
    case _FPCLASS_PD:
        return FP_SUBNORMAL;
    case _FPCLASS_NINF:
    case _FPCLASS_PINF:
        return FP_INFINITE;
    default:
        Tcl_Panic("result of _fpclass() outside documented range!");
    case _FPCLASS_QNAN:
    case _FPCLASS_SNAN:
        return FP_NAN;
    }
#endif /* fpclassify */
}

static int
ExprIsFiniteFunc(
    ClientData ignored,
    Tcl_Interp *interp,		/* The interpreter in which to execute the
				 * function. */
    int objc,			/* Actual parameter count */
    Tcl_Obj *const *objv)	/* Actual parameter list */
{
    double d;
    ClientData ptr;
    int type, result = 0;

    if (objc != 2) {
	MathFuncWrongNumArgs(interp, 2, objc, objv);
	return TCL_ERROR;
    }

    if (TclGetNumberFromObj(interp, objv[1], &ptr, &type) != TCL_OK) {
        return TCL_ERROR;
    }
    if (type != TCL_NUMBER_NAN) {
        if (Tcl_GetDoubleFromObj(interp, objv[1], &d) != TCL_OK) {
            return TCL_ERROR;
        }
        type = ClassifyDouble(d);
        result = (type != FP_INFINITE && type != FP_NAN);
    }
    Tcl_SetObjResult(interp, Tcl_NewBooleanObj(result));
    return TCL_OK;
}

static int
ExprIsInfinityFunc(
    ClientData ignored,
    Tcl_Interp *interp,		/* The interpreter in which to execute the
				 * function. */
    int objc,			/* Actual parameter count */
    Tcl_Obj *const *objv)	/* Actual parameter list */
{
    double d;
    ClientData ptr;
    int type, result = 0;

    if (objc != 2) {
	MathFuncWrongNumArgs(interp, 2, objc, objv);
	return TCL_ERROR;
    }

    if (TclGetNumberFromObj(interp, objv[1], &ptr, &type) != TCL_OK) {
        return TCL_ERROR;
    }
    if (type != TCL_NUMBER_NAN) {
        if (Tcl_GetDoubleFromObj(interp, objv[1], &d) != TCL_OK) {
            return TCL_ERROR;
        }
        result = (ClassifyDouble(d) == FP_INFINITE);
    }
    Tcl_SetObjResult(interp, Tcl_NewBooleanObj(result));
    return TCL_OK;
}

static int
ExprIsNaNFunc(
    ClientData ignored,
    Tcl_Interp *interp,		/* The interpreter in which to execute the
				 * function. */
    int objc,			/* Actual parameter count */
    Tcl_Obj *const *objv)	/* Actual parameter list */
{
    double d;
    ClientData ptr;
    int type, result = 1;

    if (objc != 2) {
	MathFuncWrongNumArgs(interp, 2, objc, objv);
	return TCL_ERROR;
    }

    if (TclGetNumberFromObj(interp, objv[1], &ptr, &type) != TCL_OK) {
        return TCL_ERROR;
    }
    if (type != TCL_NUMBER_NAN) {
        if (Tcl_GetDoubleFromObj(interp, objv[1], &d) != TCL_OK) {
            return TCL_ERROR;
        }
        result = (ClassifyDouble(d) == FP_NAN);
    }
    Tcl_SetObjResult(interp, Tcl_NewBooleanObj(result));
    return TCL_OK;
}

static int
ExprIsNormalFunc(
    ClientData ignored,
    Tcl_Interp *interp,		/* The interpreter in which to execute the
				 * function. */
    int objc,			/* Actual parameter count */
    Tcl_Obj *const *objv)	/* Actual parameter list */
{
    double d;
    ClientData ptr;
    int type, result = 0;

    if (objc != 2) {
	MathFuncWrongNumArgs(interp, 2, objc, objv);
	return TCL_ERROR;
    }

    if (TclGetNumberFromObj(interp, objv[1], &ptr, &type) != TCL_OK) {
        return TCL_ERROR;
    }
    if (type != TCL_NUMBER_NAN) {
        if (Tcl_GetDoubleFromObj(interp, objv[1], &d) != TCL_OK) {
            return TCL_ERROR;
        }
        result = (ClassifyDouble(d) == FP_NORMAL);
    }
    Tcl_SetObjResult(interp, Tcl_NewBooleanObj(result));
    return TCL_OK;
}

static int
ExprIsSubnormalFunc(
    ClientData ignored,
    Tcl_Interp *interp,		/* The interpreter in which to execute the
				 * function. */
    int objc,			/* Actual parameter count */
    Tcl_Obj *const *objv)	/* Actual parameter list */
{
    double d;
    ClientData ptr;
    int type, result = 0;

    if (objc != 2) {
	MathFuncWrongNumArgs(interp, 2, objc, objv);
	return TCL_ERROR;
    }

    if (TclGetNumberFromObj(interp, objv[1], &ptr, &type) != TCL_OK) {
        return TCL_ERROR;
    }
    if (type != TCL_NUMBER_NAN) {
        if (Tcl_GetDoubleFromObj(interp, objv[1], &d) != TCL_OK) {
            return TCL_ERROR;
        }
        result = (ClassifyDouble(d) == FP_SUBNORMAL);
    }
    Tcl_SetObjResult(interp, Tcl_NewBooleanObj(result));
    return TCL_OK;
}

static int
ExprIsUnorderedFunc(
    ClientData ignored,
    Tcl_Interp *interp,		/* The interpreter in which to execute the
				 * function. */
    int objc,			/* Actual parameter count */
    Tcl_Obj *const *objv)	/* Actual parameter list */
{
    double d;
    ClientData ptr;
    int type, result = 0;

    if (objc != 3) {
	MathFuncWrongNumArgs(interp, 3, objc, objv);
	return TCL_ERROR;
    }

    if (TclGetNumberFromObj(interp, objv[1], &ptr, &type) != TCL_OK) {
        return TCL_ERROR;
    }
    if (type == TCL_NUMBER_NAN) {
        result = 1;
    } else {
        d = *((const double *) ptr);
        result = (ClassifyDouble(d) == FP_NAN);
    }

    if (TclGetNumberFromObj(interp, objv[2], &ptr, &type) != TCL_OK) {
        return TCL_ERROR;
    }
    if (type == TCL_NUMBER_NAN) {
        result |= 1;
    } else {
        d = *((const double *) ptr);
        result |= (ClassifyDouble(d) == FP_NAN);
    }

    Tcl_SetObjResult(interp, Tcl_NewBooleanObj(result));
    return TCL_OK;
}

static int
FloatClassifyObjCmd(
    ClientData ignored,
    Tcl_Interp *interp,		/* The interpreter in which to execute the
				 * function. */
    int objc,			/* Actual parameter count */
    Tcl_Obj *const *objv)	/* Actual parameter list */
{
    double d;
    Tcl_Obj *objPtr;
    ClientData ptr;
    int type;

    if (objc != 2) {
        Tcl_WrongNumArgs(interp, 1, objv, "floatValue");
	return TCL_ERROR;
    }

    if (TclGetNumberFromObj(interp, objv[1], &ptr, &type) != TCL_OK) {
        return TCL_ERROR;
    }
    if (type == TCL_NUMBER_NAN) {
        goto gotNaN;
    } else if (Tcl_GetDoubleFromObj(interp, objv[1], &d) != TCL_OK) {
        return TCL_ERROR;
    }
    switch (ClassifyDouble(d)) {
    case FP_INFINITE:
        TclNewLiteralStringObj(objPtr, "infinite");
        break;
    case FP_NAN:
    gotNaN:
        TclNewLiteralStringObj(objPtr, "nan");
        break;
    case FP_NORMAL:
        TclNewLiteralStringObj(objPtr, "normal");
        break;
    case FP_SUBNORMAL:
        TclNewLiteralStringObj(objPtr, "subnormal");
        break;
    case FP_ZERO:
        TclNewLiteralStringObj(objPtr, "zero");
        break;
    default:
        Tcl_SetObjResult(interp, Tcl_ObjPrintf(
                "unable to classify number: %f", d));
        return TCL_ERROR;
    }
    Tcl_SetObjResult(interp, objPtr);
    return TCL_OK;
}
 
/*
 *----------------------------------------------------------------------
 *
 * MathFuncWrongNumArgs --
 *
 *	Generate an error message when a math function presents the wrong

    Tcl_Parse *parsePtr,
    Command *cmdPtr,		/* Points to defintion of command being
				 * compiled. */
    CompileEnv *envPtr)
{
    return CompileComparisonOpCmd(interp, parsePtr, INST_STR_EQ, envPtr);
}












































 
int
TclCompileMinusOpCmd(
    Tcl_Interp *interp,
    Tcl_Parse *parsePtr,
    Command *cmdPtr,		/* Points to defintion of command being
				 * compiled. */

    Tcl_Parse *parsePtr,
    Command *cmdPtr,		/* Points to defintion of command being
				 * compiled. */
    CompileEnv *envPtr)
{
    return CompileComparisonOpCmd(interp, parsePtr, INST_STR_EQ, envPtr);
}

int
TclCompileStrLtOpCmd(
    Tcl_Interp *interp,
    Tcl_Parse *parsePtr,
    Command *cmdPtr,		/* Points to defintion of command being
				 * compiled. */
    CompileEnv *envPtr)
{
    return CompileComparisonOpCmd(interp, parsePtr, INST_STR_LT, envPtr);
}

int
TclCompileStrLeOpCmd(
    Tcl_Interp *interp,
    Tcl_Parse *parsePtr,
    Command *cmdPtr,		/* Points to defintion of command being
				 * compiled. */
    CompileEnv *envPtr)
{
    return CompileComparisonOpCmd(interp, parsePtr, INST_STR_LE, envPtr);
}

int
TclCompileStrGtOpCmd(
    Tcl_Interp *interp,
    Tcl_Parse *parsePtr,
    Command *cmdPtr,		/* Points to defintion of command being
				 * compiled. */
    CompileEnv *envPtr)
{
    return CompileComparisonOpCmd(interp, parsePtr, INST_STR_GT, envPtr);
}

int
TclCompileStrGeOpCmd(
    Tcl_Interp *interp,
    Tcl_Parse *parsePtr,
    Command *cmdPtr,		/* Points to defintion of command being
				 * compiled. */
    CompileEnv *envPtr)
{
    return CompileComparisonOpCmd(interp, parsePtr, INST_STR_GE, envPtr);
}
 
int
TclCompileMinusOpCmd(
    Tcl_Interp *interp,
    Tcl_Parse *parsePtr,
    Command *cmdPtr,		/* Points to defintion of command being
				 * compiled. */

				 * for us. In the end though, a close paren is
				 * not really a binary operator, and some
				 * special coding in ParseExpr() make sure we
				 * never put an actual CLOSE_PAREN node in the
				 * parse tree. The sub-expression between
				 * parens becomes the single argument of the
				 * matching OPEN_PAREN unary operator. */




#define END		(BINARY | 28)
				/* This lexeme represents the end of the
				 * string being parsed. Treating it as a
				 * binary operator follows the same logic as
				 * the CLOSE_PAREN lexeme and END pairs with
				 * START, in the same way that CLOSE_PAREN
				 * pairs with OPEN_PAREN. */

................................................................................
    PREC_OR,		/* OR */
    PREC_EQUAL,		/* STREQ */
    PREC_EQUAL,		/* STRNEQ */
    PREC_EXPON,		/* EXPON */
    PREC_EQUAL,		/* IN_LIST */
    PREC_EQUAL,		/* NOT_IN_LIST */
    PREC_CLOSE_PAREN,	/* CLOSE_PAREN */




    PREC_END,		/* END */
    /* Expansion room for more binary operators */
    0,  0,  0,
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,
    /* Unary operator lexemes */
    PREC_UNARY,		/* UNARY_PLUS */
    PREC_UNARY,		/* UNARY_MINUS */
    PREC_UNARY,		/* FUNCTION */
    PREC_START,		/* START */
    PREC_OPEN_PAREN,	/* OPEN_PAREN */
    PREC_UNARY,		/* NOT*/
................................................................................
    0,			/* OR */
    INST_STR_EQ,	/* STREQ */
    INST_STR_NEQ,	/* STRNEQ */
    INST_EXPON,		/* EXPON */
    INST_LIST_IN,	/* IN_LIST */
    INST_LIST_NOT_IN,	/* NOT_IN_LIST */
    0,			/* CLOSE_PAREN */




    0,			/* END */
    /* Expansion room for more binary operators */
    0,  0,  0,
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,
    /* Unary operator lexemes */
    INST_UPLUS,		/* UNARY_PLUS */
    INST_UMINUS,	/* UNARY_MINUS */
    0,			/* FUNCTION */
    0,			/* START */
    0,			/* OPEN_PAREN */
    INST_LNOT,		/* NOT*/
................................................................................
		*lexemePtr = STRNEQ;
		return 2;
	    case 'i':
		*lexemePtr = NOT_IN_LIST;
		return 2;
	    }
	}





























    }

    literal = Tcl_NewObj();
    if (TclParseNumber(NULL, literal, NULL, start, numBytes, &end,
	    TCL_PARSE_NO_WHITESPACE) == TCL_OK) {
	if (end < start + numBytes && !TclIsBareword(*end)) {

................................................................................
}
 
/*
 *----------------------------------------------------------------------
 *
 * TclSortingOpCmd --
 *	Implements the commands:
 *		<, <=, >, >=, ==, eq
 *	in the ::tcl::mathop namespace. These commands are defined for
 *	arbitrary number of arguments by computing the AND of the base
 *	operator applied to all neighbor argument pairs.
 *
 * Results:
 *	A standard Tcl return code and result left in interp.
 *

				 * for us. In the end though, a close paren is
				 * not really a binary operator, and some
				 * special coding in ParseExpr() make sure we
				 * never put an actual CLOSE_PAREN node in the
				 * parse tree. The sub-expression between
				 * parens becomes the single argument of the
				 * matching OPEN_PAREN unary operator. */
#define STR_LT		(BINARY | 28)
#define STR_GT		(BINARY | 29)
#define STR_LEQ		(BINARY | 30)
#define STR_GEQ		(BINARY | 31)
#define END		(BINARY | 32)
				/* This lexeme represents the end of the
				 * string being parsed. Treating it as a
				 * binary operator follows the same logic as
				 * the CLOSE_PAREN lexeme and END pairs with
				 * START, in the same way that CLOSE_PAREN
				 * pairs with OPEN_PAREN. */

................................................................................
    PREC_OR,		/* OR */
    PREC_EQUAL,		/* STREQ */
    PREC_EQUAL,		/* STRNEQ */
    PREC_EXPON,		/* EXPON */
    PREC_EQUAL,		/* IN_LIST */
    PREC_EQUAL,		/* NOT_IN_LIST */
    PREC_CLOSE_PAREN,	/* CLOSE_PAREN */
    PREC_COMPARE,	/* STR_LT */
    PREC_COMPARE,	/* STR_GT */
    PREC_COMPARE,	/* STR_LEQ */
    PREC_COMPARE,	/* STR_GEQ */
    PREC_END,		/* END */
    /* Expansion room for more binary operators */

    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,

    /* Unary operator lexemes */
    PREC_UNARY,		/* UNARY_PLUS */
    PREC_UNARY,		/* UNARY_MINUS */
    PREC_UNARY,		/* FUNCTION */
    PREC_START,		/* START */
    PREC_OPEN_PAREN,	/* OPEN_PAREN */
    PREC_UNARY,		/* NOT*/
................................................................................
    0,			/* OR */
    INST_STR_EQ,	/* STREQ */
    INST_STR_NEQ,	/* STRNEQ */
    INST_EXPON,		/* EXPON */
    INST_LIST_IN,	/* IN_LIST */
    INST_LIST_NOT_IN,	/* NOT_IN_LIST */
    0,			/* CLOSE_PAREN */
    INST_STR_LT,	/* STR_LT */
    INST_STR_GT,	/* STR_GT */
    INST_STR_LE,	/* STR_LEQ */
    INST_STR_GE,	/* STR_GEQ */
    0,			/* END */
    /* Expansion room for more binary operators */

    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,
    0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,  0,

    /* Unary operator lexemes */
    INST_UPLUS,		/* UNARY_PLUS */
    INST_UMINUS,	/* UNARY_MINUS */
    0,			/* FUNCTION */
    0,			/* START */
    0,			/* OPEN_PAREN */
    INST_LNOT,		/* NOT*/
................................................................................
		*lexemePtr = STRNEQ;
		return 2;
	    case 'i':
		*lexemePtr = NOT_IN_LIST;
		return 2;
	    }
	}
	break;

    case 'l':
	if ((numBytes > 1)
		&& ((numBytes == 2) || start[2] & 0x80 || !isalpha(UCHAR(start[2])))) {
	    switch (start[1]) {
	    case 't':
		*lexemePtr = STR_LT;
		return 2;
	    case 'e':
		*lexemePtr = STR_LEQ;
		return 2;
	    }
	}
	break;
	
    case 'g':
	if ((numBytes > 1)
		&& ((numBytes == 2) || start[2] & 0x80 || !isalpha(UCHAR(start[2])))) {
	    switch (start[1]) {
	    case 't':
		*lexemePtr = STR_GT;
		return 2;
	    case 'e':
		*lexemePtr = STR_GEQ;
		return 2;
	    }
	}
	break;
    }

    literal = Tcl_NewObj();
    if (TclParseNumber(NULL, literal, NULL, start, numBytes, &end,
	    TCL_PARSE_NO_WHITESPACE) == TCL_OK) {
	if (end < start + numBytes && !TclIsBareword(*end)) {

................................................................................
}
 
/*
 *----------------------------------------------------------------------
 *
 * TclSortingOpCmd --
 *	Implements the commands:
 *		<, <=, >, >=, ==, eq, lt, le, gt, ge
 *	in the ::tcl::mathop namespace. These commands are defined for
 *	arbitrary number of arguments by computing the AND of the base
 *	operator applied to all neighbor argument pairs.
 *
 * Results:
 *	A standard Tcl return code and result left in interp.
 *

	/* The top word is the default, the next op4 words (min 1) are a key
	 * path into the dictionary just below the keys on the stack, and all
	 * those values are replaced by the value read out of that key-path
	 * (like [dict get]) except if there is no such key, when instead the
	 * default is pushed instead.
	 * Stack:  ... dict key1 ... keyN default => ... value */










    {NULL, 0, 0, 0, {OPERAND_NONE}}
};
 
/*
 * Prototypes for procedures defined later in this file:
 */

	/* The top word is the default, the next op4 words (min 1) are a key
	 * path into the dictionary just below the keys on the stack, and all
	 * those values are replaced by the value read out of that key-path
	 * (like [dict get]) except if there is no such key, when instead the
	 * default is pushed instead.
	 * Stack:  ... dict key1 ... keyN default => ... value */

    {"strlt",		  1,   -1,         0,	{OPERAND_NONE}},
	/* String Less:			push (stknext < stktop) */
    {"strgt",		  1,   -1,         0,	{OPERAND_NONE}},
	/* String Greater:		push (stknext > stktop) */
    {"strle",		  1,   -1,         0,	{OPERAND_NONE}},
	/* String Less or equal:	push (stknext <= stktop) */
    {"strge",		  1,   -1,         0,	{OPERAND_NONE}},
	/* String Greater or equal:	push (stknext >= stktop) */

    {NULL, 0, 0, 0, {OPERAND_NONE}}
};
 
/*
 * Prototypes for procedures defined later in this file:
 */

    INST_LAPPEND_LIST_ARRAY,
    INST_LAPPEND_LIST_ARRAY_STK,
    INST_LAPPEND_LIST_STK,

    INST_CLOCK_READ,

    INST_DICT_GET_DEF,







    /* The last opcode */
    LAST_INST_OPCODE
};
 
/*
 * Table describing the Tcl bytecode instructions: their name (for displaying

    INST_LAPPEND_LIST_ARRAY,
    INST_LAPPEND_LIST_ARRAY_STK,
    INST_LAPPEND_LIST_STK,

    INST_CLOCK_READ,

    INST_DICT_GET_DEF,

	/* TIP 461 */
	INST_STR_LT,
	INST_STR_GT,
	INST_STR_LE,
	INST_STR_GE,

    /* The last opcode */
    LAST_INST_OPCODE
};
 
/*
 * Table describing the Tcl bytecode instructions: their name (for displaying

     * -----------------------------------------------------------------
     *	   Start of string-related instructions.
     */

    case INST_STR_EQ:
    case INST_STR_NEQ:		/* String (in)equality check */
    case INST_STR_CMP:		/* String compare. */




    stringCompare:
	value2Ptr = OBJ_AT_TOS;
	valuePtr = OBJ_UNDER_TOS;

	{
	    int checkEq = ((*pc == INST_EQ) || (*pc == INST_NEQ)
		    || (*pc == INST_STR_EQ) || (*pc == INST_STR_NEQ));
................................................................................
		match = (match == 0);
		break;
	    case INST_STR_NEQ:
	    case INST_NEQ:
		match = (match != 0);
		break;
	    case INST_LT:

		match = (match < 0);
		break;
	    case INST_GT:

		match = (match > 0);
		break;
	    case INST_LE:

		match = (match <= 0);
		break;
	    case INST_GE:

		match = (match >= 0);
		break;
	    }
	}

	TRACE(("\"%.20s\" \"%.20s\" => %d\n", O2S(valuePtr), O2S(value2Ptr),
		(match < 0 ? -1 : match > 0 ? 1 : 0)));

     * -----------------------------------------------------------------
     *	   Start of string-related instructions.
     */

    case INST_STR_EQ:
    case INST_STR_NEQ:		/* String (in)equality check */
    case INST_STR_CMP:		/* String compare. */
    case INST_STR_LT:
    case INST_STR_GT:
    case INST_STR_LE:
    case INST_STR_GE:
    stringCompare:
	value2Ptr = OBJ_AT_TOS;
	valuePtr = OBJ_UNDER_TOS;

	{
	    int checkEq = ((*pc == INST_EQ) || (*pc == INST_NEQ)
		    || (*pc == INST_STR_EQ) || (*pc == INST_STR_NEQ));
................................................................................
		match = (match == 0);
		break;
	    case INST_STR_NEQ:
	    case INST_NEQ:
		match = (match != 0);
		break;
	    case INST_LT:
	    case INST_STR_LT:
		match = (match < 0);
		break;
	    case INST_GT:
	    case INST_STR_GT:
		match = (match > 0);
		break;
	    case INST_LE:
	    case INST_STR_LE:
		match = (match <= 0);
		break;
	    case INST_GE:
	    case INST_STR_GE:
		match = (match >= 0);
		break;
	    }
	}

	TRACE(("\"%.20s\" \"%.20s\" => %d\n", O2S(valuePtr), O2S(value2Ptr),
		(match < 0 ? -1 : match > 0 ? 1 : 0)));

			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);
MODULE_SCOPE int	TclCompileEqOpCmd(Tcl_Interp *interp,
			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);
MODULE_SCOPE int	TclCompileStreqOpCmd(Tcl_Interp *interp,
			    Tcl_Parse *parsePtr, Command *cmdPtr,












			    struct CompileEnv *envPtr);

MODULE_SCOPE int	TclCompileAssembleCmd(Tcl_Interp *interp,
			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);

/*

			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);
MODULE_SCOPE int	TclCompileEqOpCmd(Tcl_Interp *interp,
			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);
MODULE_SCOPE int	TclCompileStreqOpCmd(Tcl_Interp *interp,
			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);
MODULE_SCOPE int	TclCompileStrLtOpCmd(Tcl_Interp *interp,
			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);
MODULE_SCOPE int	TclCompileStrLeOpCmd(Tcl_Interp *interp,
			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);
MODULE_SCOPE int	TclCompileStrGtOpCmd(Tcl_Interp *interp,
			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);
MODULE_SCOPE int	TclCompileStrGeOpCmd(Tcl_Interp *interp,
			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);

MODULE_SCOPE int	TclCompileAssembleCmd(Tcl_Interp *interp,
			    Tcl_Parse *parsePtr, Command *cmdPtr,
			    struct CompileEnv *envPtr);

/*

		 * substituted part of the command (as a list) as their
		 * content!
		 */

		if (isNew) {
		    Tcl_Obj *cmdObj, *cmdPrefixObj;

#if 0
		    TclNewObj(cmdObj);
		    Tcl_AppendStringsToObj(cmdObj, nsPtr->fullName,
			    (nsPtr->parentPtr ? "::" : ""), nsCmdName, NULL);
#else
		    cmdObj = Tcl_NewStringObj(nsCmdName, -1);
#endif
		    cmdPrefixObj = Tcl_NewListObj(1, &cmdObj);

		 * substituted part of the command (as a list) as their
		 * content!
		 */

		if (isNew) {
		    Tcl_Obj *cmdObj, *cmdPrefixObj;

#if 1
		    TclNewObj(cmdObj);
		    Tcl_AppendStringsToObj(cmdObj, nsPtr->fullName,
			    (nsPtr->parentPtr ? "::" : ""), nsCmdName, NULL);
#else
		    cmdObj = Tcl_NewStringObj(nsCmdName, -1);
#endif
		    cmdPrefixObj = Tcl_NewListObj(1, &cmdObj);

    /*
     * b = bw * 2**b2 * 5**b5
     * mminus = 5**m5
     */

    TclInitBignumFromWideUInt(&b, bw);
    mp_init_set_int(&mminus, 1);
    MulPow5(&b, b5, &b);
    mp_mul_2d(&b, b2, &b);

    /*
     * Adjust if the logarithm was guessed wrong.
     */

................................................................................
    /*
     * b = bw * 2**b2 * 5**b5
     * S = 2**s2 * 5*s5
     */

    TclInitBignumFromWideUInt(&b, bw);
    mp_mul_2d(&b, b2, &b);
    mp_init_set_int(&S, 1);
    MulPow5(&S, s5, &S); mp_mul_2d(&S, s2, &S);

    /*
     * Handle the case where we guess the position of the decimal point wrong.
     */

    if (mp_cmp_mag(&b, &S) == MP_LT) {
................................................................................
	--k;
    }

    /*
     * mminus = 2**m2minus * 5**m5
     */

    mp_init_set_int(&mminus, minit);
    mp_mul_2d(&mminus, m2minus, &mminus);
    if (m2plus > m2minus) {
	mp_init_copy(&mplus, &mminus);
	mp_mul_2d(&mplus, m2plus-m2minus, &mplus);
    }

    /*
................................................................................
     * b = bw * 2**b2 * 5**b5
     * S = 2**s2 * 5*s5
     */

    mp_init_multi(&dig, NULL);
    TclInitBignumFromWideUInt(&b, bw);
    mp_mul_2d(&b, b2, &b);
    mp_init_set_int(&S, 1);
    MulPow5(&S, s5, &S); mp_mul_2d(&S, s2, &S);

    /*
     * Handle the case where we guess the position of the decimal point wrong.
     */

    if (mp_cmp_mag(&b, &S) == MP_LT) {

    /*
     * b = bw * 2**b2 * 5**b5
     * mminus = 5**m5
     */

    TclInitBignumFromWideUInt(&b, bw);
    mp_init_set(&mminus, 1);
    MulPow5(&b, b5, &b);
    mp_mul_2d(&b, b2, &b);

    /*
     * Adjust if the logarithm was guessed wrong.
     */

................................................................................
    /*
     * b = bw * 2**b2 * 5**b5
     * S = 2**s2 * 5*s5
     */

    TclInitBignumFromWideUInt(&b, bw);
    mp_mul_2d(&b, b2, &b);
    mp_init_set(&S, 1);
    MulPow5(&S, s5, &S); mp_mul_2d(&S, s2, &S);

    /*
     * Handle the case where we guess the position of the decimal point wrong.
     */

    if (mp_cmp_mag(&b, &S) == MP_LT) {
................................................................................
	--k;
    }

    /*
     * mminus = 2**m2minus * 5**m5
     */

    mp_init_set(&mminus, minit);
    mp_mul_2d(&mminus, m2minus, &mminus);
    if (m2plus > m2minus) {
	mp_init_copy(&mplus, &mminus);
	mp_mul_2d(&mplus, m2plus-m2minus, &mplus);
    }

    /*
................................................................................
     * b = bw * 2**b2 * 5**b5
     * S = 2**s2 * 5*s5
     */

    mp_init_multi(&dig, NULL);
    TclInitBignumFromWideUInt(&b, bw);
    mp_mul_2d(&b, b2, &b);
    mp_init_set(&S, 1);
    MulPow5(&S, s5, &S); mp_mul_2d(&S, s2, &S);

    /*
     * Handle the case where we guess the position of the decimal point wrong.
     */

    if (mp_cmp_mag(&b, &S) == MP_LT) {

    global xxx
    set xxx ""
    12days 1 1 1
    set result [string length $xxx]
    unset xxx
    return $result
}

# start of tests

catch {unset a b i x}

test expr-1.1 {TclCompileExprCmd: no expression} {
    list [catch {expr  } msg] $msg
} {1 {wrong # args: should be "expr arg ?arg ...?"}}
................................................................................
} -returnCodes error -match glob -result *
test expr-8.34 {expr edge cases} -body {
    expr {1E+}
} -returnCodes error -match glob -result *
test expr-8.35 {expr edge cases} -body {
    expr {1ea}
} -returnCodes error -match glob -result *





















test expr-9.1 {CompileRelationalExpr: just shift expr} {expr 3<<2} 12
test expr-9.2 {CompileRelationalExpr: just shift expr} {expr 0xff>>2} 63
test expr-9.3 {CompileRelationalExpr: just shift expr} {expr -1>>2} -1
test expr-9.4 {CompileRelationalExpr: just shift expr} {expr {1<<3}} 8
test expr-9.5 {CompileRelationalExpr: shift expr producing LONG_MIN} {
    expr {int(1<<63)}
................................................................................
} {
	set a [list one two three]
	list [expr {$a eq {}}] [expr {$a < {}}] [expr {$a > {}}] [
		string match {*no string representation*} [
		::tcl::unsupported::representation $a]]
} {0 0 1 1}






















# cleanup
if {[info exists a]} {










































































































    unset a





}
catch {unset min}
catch {unset max}







::tcltest::cleanupTests
return

# Local Variables:
# mode: tcl
# End:

    global xxx
    set xxx ""
    12days 1 1 1
    set result [string length $xxx]
    unset xxx
    return $result
}
 
# start of tests

catch {unset a b i x}

test expr-1.1 {TclCompileExprCmd: no expression} {
    list [catch {expr  } msg] $msg
} {1 {wrong # args: should be "expr arg ?arg ...?"}}
................................................................................
} -returnCodes error -match glob -result *
test expr-8.34 {expr edge cases} -body {
    expr {1E+}
} -returnCodes error -match glob -result *
test expr-8.35 {expr edge cases} -body {
    expr {1ea}
} -returnCodes error -match glob -result *
test expr-8.36 {CompileEqualtyExpr: string comparison ops} {
    set x 012
    set y 0x0
    list [expr {$x < $y}] [expr {$x lt $y}] [expr {$x lt $x}]
} {0 1 0}
test expr-8.37 {CompileEqualtyExpr: string comparison ops} {
    set x 012
    set y 0x0
    list [expr {$x <= $y}] [expr {$x le $y}] [expr {$x le $x}]
} {0 1 1}
test expr-8.38 {CompileEqualtyExpr: string comparison ops} {
    set x 012
    set y 0x0
    list [expr {$x > $y}] [expr {$x gt $y}] [expr {$x gt $x}]
} {1 0 0}
test expr-8.39 {CompileEqualtyExpr: string comparison ops} {
    set x 012
    set y 0x0
    list [expr {$x >= $y}] [expr {$x ge $y}] [expr {$x ge $x}]
} {1 0 1}

test expr-9.1 {CompileRelationalExpr: just shift expr} {expr 3<<2} 12
test expr-9.2 {CompileRelationalExpr: just shift expr} {expr 0xff>>2} 63
test expr-9.3 {CompileRelationalExpr: just shift expr} {expr -1>>2} -1
test expr-9.4 {CompileRelationalExpr: just shift expr} {expr {1<<3}} 8
test expr-9.5 {CompileRelationalExpr: shift expr producing LONG_MIN} {
    expr {int(1<<63)}
................................................................................
} {
	set a [list one two three]
	list [expr {$a eq {}}] [expr {$a < {}}] [expr {$a > {}}] [
		string match {*no string representation*} [
		::tcl::unsupported::representation $a]]
} {0 0 1 1}

foreach func {isfinite isinf isnan isnormal issubnormal} {
    test expr-53.1.$func {float classification: basic arg handling} -body {
	expr ${func}()
    } -returnCodes error -result "too few arguments for math function \"$func\""
    test expr-53.2.$func {float classification: basic arg handling} -body {
	expr ${func}(1,2)
    } -returnCodes error -result "too many arguments for math function \"$func\""
    test expr-53.3.$func {float classification: basic arg handling} -body {
	expr ${func}(true)
    } -returnCodes error -result {expected number but got "true"}
    test expr-53.4.$func {float classification: basic arg handling} -body {
	expr ${func}("gorp")
    } -returnCodes error -result {expected number but got "gorp"}
    test expr-53.5.$func {float classification: basic arg handling} -body {
	expr ${func}(1.0)
    } -match glob -result *
    test expr-53.6.$func {float classification: basic arg handling} -body {
	expr ${func}(0x123)
    } -match glob -result *
}



test expr-54.0 {float classification: isfinite} {expr {isfinite(1.0)}} 1
test expr-54.1 {float classification: isfinite} {expr {isfinite(-1.0)}} 1
test expr-54.2 {float classification: isfinite} {expr {isfinite(0.0)}} 1
test expr-54.3 {float classification: isfinite} {expr {isfinite(-0.0)}} 1
test expr-54.4 {float classification: isfinite} {expr {isfinite(1/Inf)}} 1
test expr-54.5 {float classification: isfinite} {expr {isfinite(-1/Inf)}} 1
test expr-54.6 {float classification: isfinite} {expr {isfinite(1e-314)}} 1
test expr-54.7 {float classification: isfinite} {expr {isfinite(inf)}} 0
test expr-54.8 {float classification: isfinite} {expr {isfinite(-inf)}} 0
test expr-54.9 {float classification: isfinite} {expr {isfinite(NaN)}} 0

test expr-55.0 {float classification: isinf} {expr {isinf(1.0)}} 0
test expr-55.1 {float classification: isinf} {expr {isinf(-1.0)}} 0
test expr-55.2 {float classification: isinf} {expr {isinf(0.0)}} 0
test expr-55.3 {float classification: isinf} {expr {isinf(-0.0)}} 0
test expr-55.4 {float classification: isinf} {expr {isinf(1/Inf)}} 0
test expr-55.5 {float classification: isinf} {expr {isinf(-1/Inf)}} 0
test expr-55.6 {float classification: isinf} {expr {isinf(1e-314)}} 0
test expr-55.7 {float classification: isinf} {expr {isinf(inf)}} 1
test expr-55.8 {float classification: isinf} {expr {isinf(-inf)}} 1
test expr-55.9 {float classification: isinf} {expr {isinf(NaN)}} 0

test expr-56.0 {float classification: isnan} {expr {isnan(1.0)}} 0
test expr-56.1 {float classification: isnan} {expr {isnan(-1.0)}} 0
test expr-56.2 {float classification: isnan} {expr {isnan(0.0)}} 0
test expr-56.3 {float classification: isnan} {expr {isnan(-0.0)}} 0
test expr-56.4 {float classification: isnan} {expr {isnan(1/Inf)}} 0
test expr-56.5 {float classification: isnan} {expr {isnan(-1/Inf)}} 0
test expr-56.6 {float classification: isnan} {expr {isnan(1e-314)}} 0
test expr-56.7 {float classification: isnan} {expr {isnan(inf)}} 0
test expr-56.8 {float classification: isnan} {expr {isnan(-inf)}} 0
test expr-56.9 {float classification: isnan} {expr {isnan(NaN)}} 1

test expr-57.0 {float classification: isnormal} {expr {isnormal(1.0)}} 1
test expr-57.1 {float classification: isnormal} {expr {isnormal(-1.0)}} 1
test expr-57.2 {float classification: isnormal} {expr {isnormal(0.0)}} 0
test expr-57.3 {float classification: isnormal} {expr {isnormal(-0.0)}} 0
test expr-57.4 {float classification: isnormal} {expr {isnormal(1/Inf)}} 0
test expr-57.5 {float classification: isnormal} {expr {isnormal(-1/Inf)}} 0
test expr-57.6 {float classification: isnormal} {expr {isnormal(1e-314)}} 0
test expr-57.7 {float classification: isnormal} {expr {isnormal(inf)}} 0
test expr-57.8 {float classification: isnormal} {expr {isnormal(-inf)}} 0
test expr-57.9 {float classification: isnormal} {expr {isnormal(NaN)}} 0

test expr-58.0 {float classification: issubnormal} {expr {issubnormal(1.0)}} 0
test expr-58.1 {float classification: issubnormal} {expr {issubnormal(-1.0)}} 0
test expr-58.2 {float classification: issubnormal} {expr {issubnormal(0.0)}} 0
test expr-58.3 {float classification: issubnormal} {expr {issubnormal(-0.0)}} 0
test expr-58.4 {float classification: issubnormal} {expr {issubnormal(1/Inf)}} 0
test expr-58.5 {float classification: issubnormal} {expr {issubnormal(-1/Inf)}} 0
test expr-58.6 {float classification: issubnormal} {expr {issubnormal(1e-314)}} 1
test expr-58.7 {float classification: issubnormal} {expr {issubnormal(inf)}} 0
test expr-58.8 {float classification: issubnormal} {expr {issubnormal(-inf)}} 0
test expr-58.9 {float classification: issubnormal} {expr {issubnormal(NaN)}} 0

test expr-59.0 {float classification: fpclassify} {fpclassify 1.0} normal
test expr-59.1 {float classification: fpclassify} {fpclassify -1.0} normal
test expr-59.2 {float classification: fpclassify} {fpclassify 0.0} zero
test expr-59.3 {float classification: fpclassify} {fpclassify -0.0} zero
test expr-59.4 {float classification: fpclassify} {fpclassify [expr 1/Inf]} zero
test expr-59.5 {float classification: fpclassify} {fpclassify [expr -1/Inf]} zero
test expr-59.6 {float classification: fpclassify} {fpclassify 1e-314} subnormal
test expr-59.7 {float classification: fpclassify} {fpclassify inf} infinite
test expr-59.8 {float classification: fpclassify} {fpclassify -inf} infinite
test expr-59.9 {float classification: fpclassify} {fpclassify NaN} nan
test expr-59.10 {float classification: fpclassify} -returnCodes error -body {
    fpclassify
} -result {wrong # args: should be "fpclassify floatValue"}
test expr-59.11 {float classification: fpclassify} -returnCodes error -body {
    fpclassify a b
} -result {wrong # args: should be "fpclassify floatValue"}
test expr-59.12 {float classification: fpclassify} -returnCodes error -body {
    fpclassify gorp
} -result {expected number but got "gorp"}

test expr-60.1 {float classification: basic arg handling} -body {
    expr isunordered()
} -returnCodes error -result {too few arguments for math function "isunordered"}
test expr-60.2 {float classification: basic arg handling} -body {
    expr isunordered(1)
} -returnCodes error -result {too few arguments for math function "isunordered"}
test expr-60.3 {float classification: basic arg handling} -body {
    expr {isunordered(1, 2, 3)}
} -returnCodes error -result {too many arguments for math function "isunordered"}
test expr-60.4 {float classification: basic arg handling} -body {
    expr {isunordered(true, 1.0)}
} -returnCodes error -result {expected number but got "true"}
test expr-60.5 {float classification: basic arg handling} -body {
    expr {isunordered("gorp", 1.0)}
} -returnCodes error -result {expected number but got "gorp"}
test expr-60.6 {float classification: basic arg handling} -body {
    expr {isunordered(0x123, 1.0)}
} -match glob -result *
test expr-60.7 {float classification: basic arg handling} -body {
    expr {isunordered(1.0, true)}
} -returnCodes error -result {expected number but got "true"}
test expr-60.8 {float classification: basic arg handling} -body {
    expr {isunordered(1.0, "gorp")}
} -returnCodes error -result {expected number but got "gorp"}
test expr-60.9 {float classification: basic arg handling} -body {
    expr {isunordered(1.0, 0x123)}
} -match glob -result *

# Big matrix of comparisons, but it's just a binary isinf()
set values {1.0 -1.0 0.0 -0.0 1e-314 Inf -Inf NaN}
set results {0 0 0 0 0 0 0 1}
set ctr 0
foreach v1 $values r1 $results {
    foreach v2 $values r2 $results {
	test expr-61.[incr ctr] "float classification: isunordered($v1,$v2)" {
	    expr {isunordered($v1, $v2)}
	} [expr {$r1 || $r2}]
    }


}
unset -nocomplain values results ctr
 
# cleanup
unset -nocomplain a
unset -nocomplain min
unset -nocomplain max
::tcltest::cleanupTests
return

# Local Variables:
# mode: tcl
# End:

    catch {namespace delete x}
} -body {
    namespace eval x info vars foo
} -cleanup {
    namespace delete x
} -result {}

set functions {abs acos asin atan atan2 bool ceil cos cosh double entier exp floor fmod hypot int isqrt log log10 max min pow rand round sin sinh sqrt srand tan tanh wide}
# Check whether the extra testing functions are defined...
if {!([catch {expr T1()} msg] && ($msg eq {invalid command name "tcl::mathfunc::T1"}))} {
    set functions "T1 T2 T3 $functions"  ;# A lazy way of prepending!
}
test info-20.1 {info functions option} {info functions sin} sin
test info-20.2 {info functions option} {lsort [info functions]} $functions
test info-20.3 {info functions option} {

    catch {namespace delete x}
} -body {
    namespace eval x info vars foo
} -cleanup {
    namespace delete x
} -result {}

set functions {abs acos asin atan atan2 bool ceil cos cosh double entier exp floor fmod hypot int isfinite isinf isnan isnormal isqrt issubnormal isunordered log log10 max min pow rand round sin sinh sqrt srand tan tanh wide}
# Check whether the extra testing functions are defined...
if {!([catch {expr T1()} msg] && ($msg eq {invalid command name "tcl::mathfunc::T1"}))} {
    set functions "T1 T2 T3 $functions"  ;# A lazy way of prepending!
}
test info-20.1 {info functions option} {info functions sin} sin
test info-20.2 {info functions option} {lsort [info functions]} $functions
test info-20.3 {info functions option} {

        set res2 [lindex $results $i+1]
        if {$res1 ne $res2} {
            return "$i:($res1 != $res2)"
        }
    }
    return [lindex $results 0]
}

# start of tests

namespace eval ::testmathop {
    namespace path ::tcl::mathop
    variable op ;# stop surprises!

    test mathop-1.1 {compiled +} { + } 0
................................................................................
    lappend res [TestOp - 0 -9223372036854775808]         ;# -2**63
    lappend res [TestOp / -9223372036854775808 -1]
    lappend res [TestOp * 2147483648 2]
    lappend res [TestOp * 9223372036854775808 2]
    set res
} [list 2147483648 9223372036854775808 9223372036854775808 4294967296 18446744073709551616]









































if 0 {
    # Compare ops to expr bytecodes
    namespace import ::tcl::mathop::*
    proc _X {a b c} {
        set x [+ $a [- $b $c]]
        set y [expr {$a + ($b - $c)}]
        set z [< $a $b $c]
    }
    set ::tcl_traceCompile 2
    _X 3 4 5
    set ::tcl_traceCompile 0
}

# cleanup
namespace delete ::testmathop
namespace delete ::testmathop2
::tcltest::cleanupTests
return

# Local Variables:
# mode: tcl
# End:

        set res2 [lindex $results $i+1]
        if {$res1 ne $res2} {
            return "$i:($res1 != $res2)"
        }
    }
    return [lindex $results 0]
}
 
# start of tests

namespace eval ::testmathop {
    namespace path ::tcl::mathop
    variable op ;# stop surprises!

    test mathop-1.1 {compiled +} { + } 0
................................................................................
    lappend res [TestOp - 0 -9223372036854775808]         ;# -2**63
    lappend res [TestOp / -9223372036854775808 -1]
    lappend res [TestOp * 2147483648 2]
    lappend res [TestOp * 9223372036854775808 2]
    set res
} [list 2147483648 9223372036854775808 9223372036854775808 4294967296 18446744073709551616]

test mathop-27.1 {lt operator} {::tcl::mathop::lt} 1
test mathop-27.2 {lt operator} {::tcl::mathop::lt a} 1
test mathop-27.3 {lt operator} {::tcl::mathop::lt a b} 1
test mathop-27.4 {lt operator} {::tcl::mathop::lt b a} 0
test mathop-27.5 {lt operator} {::tcl::mathop::lt a a} 0
test mathop-27.6 {lt operator} {::tcl::mathop::lt a b c} 1
test mathop-27.7 {lt operator} {::tcl::mathop::lt b a c} 0
test mathop-27.8 {lt operator} {::tcl::mathop::lt a c b} 0
test mathop-27.9 {lt operator} {::tcl::mathop::lt 012 0x0} 1

test mathop-28.1 {le operator} {::tcl::mathop::le} 1
test mathop-28.2 {le operator} {::tcl::mathop::le a} 1
test mathop-28.3 {le operator} {::tcl::mathop::le a b} 1
test mathop-28.4 {le operator} {::tcl::mathop::le b a} 0
test mathop-28.5 {le operator} {::tcl::mathop::le a a} 1
test mathop-28.6 {le operator} {::tcl::mathop::le a b c} 1
test mathop-28.7 {le operator} {::tcl::mathop::le b a c} 0
test mathop-28.8 {le operator} {::tcl::mathop::le a c b} 0
test mathop-28.9 {le operator} {::tcl::mathop::le 012 0x0} 1

test mathop-29.1 {gt operator} {::tcl::mathop::gt} 1
test mathop-29.2 {gt operator} {::tcl::mathop::gt a} 1
test mathop-29.3 {gt operator} {::tcl::mathop::gt a b} 0
test mathop-29.4 {gt operator} {::tcl::mathop::gt b a} 1
test mathop-29.5 {gt operator} {::tcl::mathop::gt a a} 0
test mathop-29.6 {gt operator} {::tcl::mathop::gt c b a} 1
test mathop-29.7 {gt operator} {::tcl::mathop::gt b a c} 0
test mathop-29.8 {gt operator} {::tcl::mathop::gt a c b} 0
test mathop-29.9 {gt operator} {::tcl::mathop::gt 0x0 012} 1

test mathop-30.1 {ge operator} {::tcl::mathop::ge} 1
test mathop-30.2 {ge operator} {::tcl::mathop::ge a} 1
test mathop-30.3 {ge operator} {::tcl::mathop::ge a b} 0
test mathop-30.4 {ge operator} {::tcl::mathop::ge b a} 1
test mathop-30.5 {ge operator} {::tcl::mathop::ge a a} 1
test mathop-30.6 {ge operator} {::tcl::mathop::ge c b a} 1
test mathop-30.7 {ge operator} {::tcl::mathop::ge b a c} 0
test mathop-30.8 {ge operator} {::tcl::mathop::ge a c b} 0
test mathop-30.9 {ge operator} {::tcl::mathop::ge 0x0 012} 1

if 0 {
    # Compare ops to expr bytecodes
    namespace import ::tcl::mathop::*
    proc _X {a b c} {
        set x [+ $a [- $b $c]]
        set y [expr {$a + ($b - $c)}]
        set z [< $a $b $c]
    }
    set ::tcl_traceCompile 2
    _X 3 4 5
    set ::tcl_traceCompile 0
}
 
# cleanup
namespace delete ::testmathop
namespace delete ::testmathop2
::tcltest::cleanupTests
return

# Local Variables:
# mode: tcl
# End:

	proc x2 {} {format 2}
	proc x3 {} {format 3}
	namespace ensemble create
    }
    list [ns x0 z] [ns x1] [ns x2] [ns x3]
} -cleanup {
    namespace delete ns
} -result {{1 z} 1 2 3}
test namespace-42.8 {
    ensembles: [Bug 1670091], panic due to pointer to a deallocated List
    struct.
} -setup {
    proc demo args {}
    variable target [list [namespace which demo] x]
    proc trial args {variable target; string length $target}
................................................................................
    set result {}
    lappend result [catch {ns a b c} msg] $msg
    lappend result [catch {ns a b c} msg] $msg
    lappend result [catch {ns b c d} msg] $msg
    lappend result [catch {ns c d e} msg] $msg
    lappend result [catch {ns Magic foo bar spong wibble} msg] $msg
    list $result [lsort [info commands ::ns::*]] $log [namespace delete ns]
} {{0 2 0 2 0 2 0 2 1 {unknown or protected subcommand "Magic"}} {::ns::Magic ::ns::a ::ns::b ::ns::c} {{making a} {running a b c} {running a b c} {making b} {running b c d} {making c} {running c d e} {unknown Magic - args = foo bar spong wibble}} {}}
test namespace-47.2 {ensemble: unknown handler} {
    namespace eval ns {
	namespace export {[a-z]*}
	proc Magic {ensemble subcmd args} {
	    error foobar
	}
	namespace ensemble create -unknown ::ns::Magic
................................................................................
} -cleanup {
    namespace delete ns
} -result\
   {0 0\
    1 {wrong # args: should be "ns z1 x a1"}\
    1 {wrong # args: should be "ns z2 x a1 a2"}\
    1 {wrong # args: should be "ns z2 x a1 a2"}\
    1 {wrong # args: should be "z0"}\
    0 {1 v}\
    1 {wrong # args: should be "ns v x z2 a2"}\
    0 {2 v v2}}
test namespace-53.11 {ensembles: nested rewrite} -setup {
    namespace eval ns {
	namespace export x
	namespace eval x {
................................................................................
	    return [lsort $gone]
	} finally {
	    namespace delete ::testing
	}
    }
} {::testing::abc::def ::testing::abc::ghi}

test namespace-56.4 {bug 16fe1b5807: names starting with ":"} {
namespace eval : {
    namespace ensemble create
    namespace export *
    proc p1 {} {
	    return 16fe1b5807
    }
}

: p1
} 16fe1b5807













 
# cleanup
catch {rename cmd1 {}}
catch {unset l}
catch {unset msg}
catch {unset trigger}
namespace delete {*}[namespace children :: test_ns_*]
::tcltest::cleanupTests
return

# Local Variables:
# mode: tcl
# End:

	proc x2 {} {format 2}
	proc x3 {} {format 3}
	namespace ensemble create
    }
    list [ns x0 z] [ns x1] [ns x2] [ns x3]
} -cleanup {
    namespace delete ns
} -result {{1 ::ns::x0::z} 1 2 3}
test namespace-42.8 {
    ensembles: [Bug 1670091], panic due to pointer to a deallocated List
    struct.
} -setup {
    proc demo args {}
    variable target [list [namespace which demo] x]
    proc trial args {variable target; string length $target}
................................................................................
    set result {}
    lappend result [catch {ns a b c} msg] $msg
    lappend result [catch {ns a b c} msg] $msg
    lappend result [catch {ns b c d} msg] $msg
    lappend result [catch {ns c d e} msg] $msg
    lappend result [catch {ns Magic foo bar spong wibble} msg] $msg
    list $result [lsort [info commands ::ns::*]] $log [namespace delete ns]
} {{0 2 0 2 0 2 0 2 1 {unknown or protected subcommand "Magic"}} {::ns::Magic ::ns::a ::ns::b ::ns::c} {{making a} {running ::ns::a b c} {running ::ns::a b c} {making b} {running ::ns::b c d} {making c} {running ::ns::c d e} {unknown Magic - args = foo bar spong wibble}} {}}
test namespace-47.2 {ensemble: unknown handler} {
    namespace eval ns {
	namespace export {[a-z]*}
	proc Magic {ensemble subcmd args} {
	    error foobar
	}
	namespace ensemble create -unknown ::ns::Magic
................................................................................
} -cleanup {
    namespace delete ns
} -result\
   {0 0\
    1 {wrong # args: should be "ns z1 x a1"}\
    1 {wrong # args: should be "ns z2 x a1 a2"}\
    1 {wrong # args: should be "ns z2 x a1 a2"}\
    1 {wrong # args: should be "::ns::x::z0"}\
    0 {1 v}\
    1 {wrong # args: should be "ns v x z2 a2"}\
    0 {2 v v2}}
test namespace-53.11 {ensembles: nested rewrite} -setup {
    namespace eval ns {
	namespace export x
	namespace eval x {
................................................................................
	    return [lsort $gone]
	} finally {
	    namespace delete ::testing
	}
    }
} {::testing::abc::def ::testing::abc::ghi}

test namespace-56.4 {bug 16fe1b5807: names starting with ":"} knownBug {
namespace eval : {
    namespace ensemble create
    namespace export *
    proc p1 {} {
	    return 16fe1b5807
    }
}

: p1
} 16fe1b5807

test namespace-56.5 {Bug 8b9854c3d8} -setup {
    namespace eval namespace-56.5 {
	proc cmd {} {string match ::* [lindex [[string cat info] level 0] 0]}
	namespace export *
	namespace ensemble create
    }
} -body {
    namespace-56.5 cmd
} -cleanup {
    namespace delete namespace-56.5
} -result 1

 
# cleanup
catch {rename cmd1 {}}
catch {unset l}
catch {unset msg}
catch {unset trigger}
namespace delete {*}[namespace children :: test_ns_*]
::tcltest::cleanupTests
return

# Local Variables:
# mode: tcl
# End: