phobos/std/format.d
Andrei Alexandrescu 8ac32901db unlisted bug fix
2009-04-19 17:20:11 +00:00

2961 lines
86 KiB
D

// Written in the D programming language.
/**
* This module implements the workhorse functionality for string and
* I/O formatting. It's comparable to C99's vsprintf().
*
* Macros:
* WIKI = Phobos/StdFormat
*/
module std.format;
//debug=format; // uncomment to turn on debugging printf's
import std.algorithm;
import std.stdarg; // caller will need va_list
import std.array;
import std.utf;
import std.c.stdio;
import std.c.stdlib;
import std.c.string;
import std.string;
import std.ctype;
import std.conv;
import std.functional;
import std.traits;
import std.typetuple;
import std.contracts;
import std.system;
import std.range;
version(unittest) {
import std.stdio, std.typecons;
}
version (Windows) version (DigitalMars)
{
version = DigitalMarsC;
}
version (DigitalMarsC)
{
// This is DMC's internal floating point formatting function
extern (C)
{
extern char* function(int c, int flags, int precision,
in real* pdval,
char* buf, size_t* psl, int width) __pfloatfmt;
}
alias std.c.stdio._snprintf snprintf;
}
else
{
// Use C99 snprintf
extern (C) int snprintf(char* s, size_t n, in char* format, ...);
}
/**********************************************************************
* Signals a mismatch between a format and its corresponding argument.
*/
class FormatError : Error
{
this()
{
super("std.format");
}
this(string msg)
{
super("std.format " ~ msg);
}
}
enum Mangle : char
{
Tvoid = 'v',
Tbool = 'b',
Tbyte = 'g',
Tubyte = 'h',
Tshort = 's',
Tushort = 't',
Tint = 'i',
Tuint = 'k',
Tlong = 'l',
Tulong = 'm',
Tfloat = 'f',
Tdouble = 'd',
Treal = 'e',
Tifloat = 'o',
Tidouble = 'p',
Tireal = 'j',
Tcfloat = 'q',
Tcdouble = 'r',
Tcreal = 'c',
Tchar = 'a',
Twchar = 'u',
Tdchar = 'w',
Tarray = 'A',
Tsarray = 'G',
Taarray = 'H',
Tpointer = 'P',
Tfunction = 'F',
Tident = 'I',
Tclass = 'C',
Tstruct = 'S',
Tenum = 'E',
Ttypedef = 'T',
Tdelegate = 'D',
Tconst = 'x',
Tinvariant = 'y',
}
// return the TypeInfo for a primitive type and null otherwise. This
// is required since for arrays of ints we only have the mangled char
// to work from. If arrays always subclassed TypeInfo_Array this
// routine could go away.
private TypeInfo primitiveTypeInfo(Mangle m)
{
static TypeInfo[Mangle] dic;
if (!dic.length) {
dic = [
Mangle.Tvoid : typeid(void),
Mangle.Tbool : typeid(bool),
Mangle.Tbyte : typeid(byte),
Mangle.Tubyte : typeid(ubyte),
Mangle.Tshort : typeid(short),
Mangle.Tushort : typeid(ushort),
Mangle.Tint : typeid(int),
Mangle.Tuint : typeid(uint),
Mangle.Tlong : typeid(long),
Mangle.Tulong : typeid(ulong),
Mangle.Tfloat : typeid(float),
Mangle.Tdouble : typeid(double),
Mangle.Treal : typeid(real),
Mangle.Tifloat : typeid(ifloat),
Mangle.Tidouble : typeid(idouble),
Mangle.Tireal : typeid(ireal),
Mangle.Tcfloat : typeid(cfloat),
Mangle.Tcdouble : typeid(cdouble),
Mangle.Tcreal : typeid(creal),
Mangle.Tchar : typeid(char),
Mangle.Twchar : typeid(wchar),
Mangle.Tdchar : typeid(dchar)
];
}
auto p = m in dic;
return p ? *p : null;
}
/************************************
* Interprets variadic argument list pointed to by argptr whose types
* are given by arguments[], formats them according to embedded format
* strings in the variadic argument list, and sends the resulting
* characters to putc.
*
* The variadic arguments are consumed in order. Each is formatted
* into a sequence of chars, using the default format specification
* for its type, and the characters are sequentially passed to putc.
* If a $(D char[]), $(D wchar[]), or $(D dchar[]) argument is
* encountered, it is interpreted as a format string. As many
* arguments as specified in the format string are consumed and
* formatted according to the format specifications in that string and
* passed to putc. If there are too few remaining arguments, a
* FormatError is thrown. If there are more remaining arguments than
* needed by the format specification, the default processing of
* arguments resumes until they are all consumed.
*
* Params:
* putc = Output is sent do this delegate, character by character.
* arguments = Array of $(D TypeInfo)s, one for each argument to be formatted.
* argptr = Points to variadic argument list.
*
* Throws:
* Mismatched arguments and formats result in a $(D FormatError) being thrown.
*
* Format_String:
* <a name="format-string">$(I Format strings)</a>
* consist of characters interspersed with
* $(I format specifications). Characters are simply copied
* to the output (such as putc) after any necessary conversion
* to the corresponding UTF-8 sequence.
*
* A $(I format specification) starts with a '%' character,
* and has the following grammar:
<pre>
$(I FormatSpecification):
$(B '%%')
$(B '%') $(I Flags) $(I Width) $(I Precision) $(I FormatChar)
$(I Flags):
$(I empty)
$(B '-') $(I Flags)
$(B '+') $(I Flags)
$(B '#') $(I Flags)
$(B '0') $(I Flags)
$(B ' ') $(I Flags)
$(I Width):
$(I empty)
$(I Integer)
$(B '*')
$(I Precision):
$(I empty)
$(B '.')
$(B '.') $(I Integer)
$(B '.*')
$(I Integer):
$(I Digit)
$(I Digit) $(I Integer)
$(I Digit):
$(B '0')
$(B '1')
$(B '2')
$(B '3')
$(B '4')
$(B '5')
$(B '6')
$(B '7')
$(B '8')
$(B '9')
$(I FormatChar):
$(B 's')
$(B 'b')
$(B 'd')
$(B 'o')
$(B 'x')
$(B 'X')
$(B 'e')
$(B 'E')
$(B 'f')
$(B 'F')
$(B 'g')
$(B 'G')
$(B 'a')
$(B 'A')
</pre>
<dl>
<dt>$(I Flags)
<dl>
<dt>$(B '-')
<dd>
Left justify the result in the field.
It overrides any $(B 0) flag.
<dt>$(B '+')
<dd>Prefix positive numbers in a signed conversion with a $(B +).
It overrides any $(I space) flag.
<dt>$(B '#')
<dd>Use alternative formatting:
<dl>
<dt>For $(B 'o'):
<dd> Add to precision as necessary so that the first digit
of the octal formatting is a '0', even if both the argument
and the $(I Precision) are zero.
<dt> For $(B 'x') ($(B 'X')):
<dd> If non-zero, prefix result with $(B 0x) ($(B 0X)).
<dt> For floating point formatting:
<dd> Always insert the decimal point.
<dt> For $(B 'g') ($(B 'G')):
<dd> Do not elide trailing zeros.
</dl>
<dt>$(B '0')
<dd> For integer and floating point formatting when not nan or
infinity, use leading zeros
to pad rather than spaces.
Ignore if there's a $(I Precision).
<dt>$(B ' ')
<dd>Prefix positive numbers in a signed conversion with a space.
</dl>
<dt>$(I Width)
<dd>
Specifies the minimum field width.
If the width is a $(B *), the next argument, which must be
of type $(B int), is taken as the width.
If the width is negative, it is as if the $(B -) was given
as a $(I Flags) character.
<dt>$(I Precision)
<dd> Gives the precision for numeric conversions.
If the precision is a $(B *), the next argument, which must be
of type $(B int), is taken as the precision. If it is negative,
it is as if there was no $(I Precision).
<dt>$(I FormatChar)
<dd>
<dl>
<dt>$(B 's')
<dd>The corresponding argument is formatted in a manner consistent
with its type:
<dl>
<dt>$(B bool)
<dd>The result is <tt>'true'</tt> or <tt>'false'</tt>.
<dt>integral types
<dd>The $(B %d) format is used.
<dt>floating point types
<dd>The $(B %g) format is used.
<dt>string types
<dd>The result is the string converted to UTF-8.
A $(I Precision) specifies the maximum number of characters
to use in the result.
<dt>classes derived from $(B Object)
<dd>The result is the string returned from the class instance's
$(B .toString()) method.
A $(I Precision) specifies the maximum number of characters
to use in the result.
<dt>non-string static and dynamic arrays
<dd>The result is [s<sub>0</sub>, s<sub>1</sub>, ...]
where s<sub>k</sub> is the kth element
formatted with the default format.
</dl>
<dt>$(B 'b','d','o','x','X')
<dd> The corresponding argument must be an integral type
and is formatted as an integer. If the argument is a signed type
and the $(I FormatChar) is $(B d) it is converted to
a signed string of characters, otherwise it is treated as
unsigned. An argument of type $(B bool) is formatted as '1'
or '0'. The base used is binary for $(B b), octal for $(B o),
decimal
for $(B d), and hexadecimal for $(B x) or $(B X).
$(B x) formats using lower case letters, $(B X) uppercase.
If there are fewer resulting digits than the $(I Precision),
leading zeros are used as necessary.
If the $(I Precision) is 0 and the number is 0, no digits
result.
<dt>$(B 'e','E')
<dd> A floating point number is formatted as one digit before
the decimal point, $(I Precision) digits after, the $(I FormatChar),
&plusmn;, followed by at least a two digit exponent: $(I d.dddddd)e$(I &plusmn;dd).
If there is no $(I Precision), six
digits are generated after the decimal point.
If the $(I Precision) is 0, no decimal point is generated.
<dt>$(B 'f','F')
<dd> A floating point number is formatted in decimal notation.
The $(I Precision) specifies the number of digits generated
after the decimal point. It defaults to six. At least one digit
is generated before the decimal point. If the $(I Precision)
is zero, no decimal point is generated.
<dt>$(B 'g','G')
<dd> A floating point number is formatted in either $(B e) or
$(B f) format for $(B g); $(B E) or $(B F) format for
$(B G).
The $(B f) format is used if the exponent for an $(B e) format
is greater than -5 and less than the $(I Precision).
The $(I Precision) specifies the number of significant
digits, and defaults to six.
Trailing zeros are elided after the decimal point, if the fractional
part is zero then no decimal point is generated.
<dt>$(B 'a','A')
<dd> A floating point number is formatted in hexadecimal
exponential notation 0x$(I h.hhhhhh)p$(I &plusmn;d).
There is one hexadecimal digit before the decimal point, and as
many after as specified by the $(I Precision).
If the $(I Precision) is zero, no decimal point is generated.
If there is no $(I Precision), as many hexadecimal digits as
necessary to exactly represent the mantissa are generated.
The exponent is written in as few digits as possible,
but at least one, is in decimal, and represents a power of 2 as in
$(I h.hhhhhh)*2<sup>$(I &plusmn;d)</sup>.
The exponent for zero is zero.
The hexadecimal digits, x and p are in upper case if the
$(I FormatChar) is upper case.
</dl>
Floating point NaN's are formatted as $(B nan) if the
$(I FormatChar) is lower case, or $(B NAN) if upper.
Floating point infinities are formatted as $(B inf) or
$(B infinity) if the
$(I FormatChar) is lower case, or $(B INF) or $(B INFINITY) if upper.
</dl>
Example:
-------------------------
import std.c.stdio;
import std.format;
void myPrint(...)
{
void putc(char c)
{
fputc(c, stdout);
}
std.format.doFormat(&putc, _arguments, _argptr);
}
...
int x = 27;
// prints 'The answer is 27:6'
myPrint("The answer is %s:", x, 6);
------------------------
*/
void doFormat(void delegate(dchar) putc, TypeInfo[] arguments, va_list argptr)
{ int j;
TypeInfo ti;
Mangle m;
uint flags;
int field_width;
int precision;
enum : uint
{
FLdash = 1,
FLplus = 2,
FLspace = 4,
FLhash = 8,
FLlngdbl = 0x20,
FL0pad = 0x40,
FLprecision = 0x80,
}
static TypeInfo skipCI(TypeInfo valti)
{
for (;;)
{
if (valti.classinfo.name.length == 18 &&
valti.classinfo.name[9..18] == "Invariant")
valti = (cast(TypeInfo_Invariant)valti).next;
else if (valti.classinfo.name.length == 14 &&
valti.classinfo.name[9..14] == "Const")
valti = (cast(TypeInfo_Const)valti).next;
else
break;
}
return valti;
}
void formatArg(char fc)
{
bool vbit;
ulong vnumber;
char vchar;
dchar vdchar;
Object vobject;
real vreal;
creal vcreal;
Mangle m2;
int signed = 0;
uint base = 10;
int uc;
char[ulong.sizeof * 8] tmpbuf; // long enough to print long in binary
const(char)* prefix = "";
string s;
void putstr(const char[] s)
{
//printf("flags = x%x\n", flags);
int prepad = 0;
int postpad = 0;
int padding = field_width - (strlen(prefix) + toUCSindex(s, s.length));
if (padding > 0)
{
if (flags & FLdash)
postpad = padding;
else
prepad = padding;
}
if (flags & FL0pad)
{
while (*prefix)
putc(*prefix++);
while (prepad--)
putc('0');
}
else
{
while (prepad--)
putc(' ');
while (*prefix)
putc(*prefix++);
}
foreach (dchar c; s)
putc(c);
while (postpad--)
putc(' ');
}
void putreal(real v)
{
//printf("putreal %Lg\n", vreal);
switch (fc)
{
case 's':
fc = 'g';
break;
case 'f', 'F', 'e', 'E', 'g', 'G', 'a', 'A':
break;
default:
//printf("fc = '%c'\n", fc);
Lerror:
throw new FormatError("floating");
}
version (DigitalMarsC)
{
uint sl;
char[] fbuf = tmpbuf;
if (!(flags & FLprecision))
precision = 6;
while (1)
{
sl = fbuf.length;
prefix = (*__pfloatfmt)(fc, flags | FLlngdbl,
precision, &v, cast(char*)fbuf, &sl, field_width);
if (sl != -1)
break;
sl = fbuf.length * 2;
fbuf = (cast(char*)alloca(sl * char.sizeof))[0 .. sl];
}
putstr(fbuf[0 .. sl]);
}
else
{
int sl;
char[] fbuf = tmpbuf;
char[12] format;
format[0] = '%';
int i = 1;
if (flags & FLdash)
format[i++] = '-';
if (flags & FLplus)
format[i++] = '+';
if (flags & FLspace)
format[i++] = ' ';
if (flags & FLhash)
format[i++] = '#';
if (flags & FL0pad)
format[i++] = '0';
format[i + 0] = '*';
format[i + 1] = '.';
format[i + 2] = '*';
format[i + 3] = 'L';
format[i + 4] = fc;
format[i + 5] = 0;
if (!(flags & FLprecision))
precision = -1;
while (1)
{ int n;
sl = fbuf.length;
n = snprintf(fbuf.ptr, sl, format.ptr, field_width, precision, v);
//printf("format = '%s', n = %d\n", cast(char*)format, n);
if (n >= 0 && n < sl)
{ sl = n;
break;
}
if (n < 0)
sl = sl * 2;
else
sl = n + 1;
fbuf = (cast(char*)alloca(sl * char.sizeof))[0 .. sl];
}
putstr(fbuf[0 .. sl]);
}
return;
}
static Mangle getMan(TypeInfo ti)
{
auto m = cast(Mangle)ti.classinfo.name[9];
if (ti.classinfo.name.length == 20 &&
ti.classinfo.name[9..20] == "StaticArray")
m = cast(Mangle)'G';
return m;
}
void putArray(void* p, size_t len, TypeInfo valti)
{
//printf("\nputArray(len = %u), tsize = %u\n", len, valti.tsize());
putc('[');
valti = skipCI(valti);
size_t tsize = valti.tsize();
auto argptrSave = argptr;
auto tiSave = ti;
auto mSave = m;
ti = valti;
//printf("\n%.*s\n", valti.classinfo.name);
m = getMan(valti);
while (len--)
{
//doFormat(putc, (&valti)[0 .. 1], p);
argptr = p;
formatArg('s');
p += tsize;
if (len > 0) putc(',');
}
m = mSave;
ti = tiSave;
argptr = argptrSave;
putc(']');
}
void putAArray(ubyte[long] vaa, TypeInfo valti, TypeInfo keyti)
{
putc('[');
bool comma=false;
auto argptrSave = argptr;
auto tiSave = ti;
auto mSave = m;
valti = skipCI(valti);
keyti = skipCI(keyti);
foreach(inout fakevalue; vaa)
{
if (comma) putc(',');
comma = true;
// the key comes before the value
ubyte* key = &fakevalue - long.sizeof;
//doFormat(putc, (&keyti)[0..1], key);
argptr = key;
ti = keyti;
m = getMan(keyti);
formatArg('s');
putc(':');
auto keysize = keyti.tsize;
keysize = (keysize + 3) & ~3;
ubyte* value = key + keysize;
//doFormat(putc, (&valti)[0..1], value);
argptr = value;
ti = valti;
m = getMan(valti);
formatArg('s');
}
m = mSave;
ti = tiSave;
argptr = argptrSave;
putc(']');
}
//printf("formatArg(fc = '%c', m = '%c')\n", fc, m);
switch (m)
{
case Mangle.Tbool:
vbit = va_arg!(bool)(argptr);
if (fc != 's')
{ vnumber = vbit;
goto Lnumber;
}
putstr(vbit ? "true" : "false");
return;
case Mangle.Tchar:
vchar = va_arg!(char)(argptr);
if (fc != 's')
{ vnumber = vchar;
goto Lnumber;
}
L2:
putstr((&vchar)[0 .. 1]);
return;
case Mangle.Twchar:
vdchar = va_arg!(wchar)(argptr);
goto L1;
case Mangle.Tdchar:
vdchar = va_arg!(dchar)(argptr);
L1:
if (fc != 's')
{ vnumber = vdchar;
goto Lnumber;
}
if (vdchar <= 0x7F)
{ vchar = cast(char)vdchar;
goto L2;
}
else
{ if (!isValidDchar(vdchar))
throw new UtfException("invalid dchar in format", 0);
char[4] vbuf;
putstr(toUTF8(vbuf, vdchar));
}
return;
case Mangle.Tbyte:
signed = 1;
vnumber = va_arg!(byte)(argptr);
goto Lnumber;
case Mangle.Tubyte:
vnumber = va_arg!(ubyte)(argptr);
goto Lnumber;
case Mangle.Tshort:
signed = 1;
vnumber = va_arg!(short)(argptr);
goto Lnumber;
case Mangle.Tushort:
vnumber = va_arg!(ushort)(argptr);
goto Lnumber;
case Mangle.Tint:
signed = 1;
vnumber = va_arg!(int)(argptr);
goto Lnumber;
case Mangle.Tuint:
Luint:
vnumber = va_arg!(uint)(argptr);
goto Lnumber;
case Mangle.Tlong:
signed = 1;
vnumber = cast(ulong)va_arg!(long)(argptr);
goto Lnumber;
case Mangle.Tulong:
Lulong:
vnumber = va_arg!(ulong)(argptr);
goto Lnumber;
case Mangle.Tclass:
vobject = va_arg!(Object)(argptr);
if (vobject is null)
s = "null";
else
s = vobject.toString();
goto Lputstr;
case Mangle.Tpointer:
vnumber = cast(ulong)va_arg!(void*)(argptr);
uc = 1;
flags |= FL0pad;
if (!(flags & FLprecision))
{ flags |= FLprecision;
precision = (void*).sizeof;
}
base = 16;
goto Lnumber;
case Mangle.Tfloat:
case Mangle.Tifloat:
if (fc == 'x' || fc == 'X')
goto Luint;
vreal = va_arg!(float)(argptr);
goto Lreal;
case Mangle.Tdouble:
case Mangle.Tidouble:
if (fc == 'x' || fc == 'X')
goto Lulong;
vreal = va_arg!(double)(argptr);
goto Lreal;
case Mangle.Treal:
case Mangle.Tireal:
vreal = va_arg!(real)(argptr);
goto Lreal;
case Mangle.Tcfloat:
vcreal = va_arg!(cfloat)(argptr);
goto Lcomplex;
case Mangle.Tcdouble:
vcreal = va_arg!(cdouble)(argptr);
goto Lcomplex;
case Mangle.Tcreal:
vcreal = va_arg!(creal)(argptr);
goto Lcomplex;
case Mangle.Tsarray:
putArray(argptr, (cast(TypeInfo_StaticArray)ti).len, (cast(TypeInfo_StaticArray)ti).next);
return;
case Mangle.Tarray:
int mi = 10;
if (ti.classinfo.name.length == 14 &&
ti.classinfo.name[9..14] == "Array")
{ // array of non-primitive types
TypeInfo tn = (cast(TypeInfo_Array)ti).next;
tn = skipCI(tn);
switch (cast(Mangle)tn.classinfo.name[9])
{
case Mangle.Tchar: goto LarrayChar;
case Mangle.Twchar: goto LarrayWchar;
case Mangle.Tdchar: goto LarrayDchar;
default:
break;
}
void[] va = va_arg!(void[])(argptr);
putArray(va.ptr, va.length, tn);
return;
}
if (ti.classinfo.name.length == 25 &&
ti.classinfo.name[9..25] == "AssociativeArray")
{ // associative array
ubyte[long] vaa = va_arg!(ubyte[long])(argptr);
putAArray(vaa,
(cast(TypeInfo_AssociativeArray)ti).next,
(cast(TypeInfo_AssociativeArray)ti).key);
return;
}
while (1)
{
m2 = cast(Mangle)ti.classinfo.name[mi];
switch (m2)
{
case Mangle.Tchar:
LarrayChar:
s = va_arg!(string)(argptr);
goto Lputstr;
case Mangle.Twchar:
LarrayWchar:
wchar[] sw = va_arg!(wchar[])(argptr);
s = toUTF8(sw);
goto Lputstr;
case Mangle.Tdchar:
LarrayDchar:
auto sd = va_arg!(dstring)(argptr);
s = toUTF8(sd);
Lputstr:
if (fc != 's')
throw new FormatError("string");
if (flags & FLprecision && precision < s.length)
s = s[0 .. precision];
putstr(s);
break;
case Mangle.Tconst:
case Mangle.Tinvariant:
mi++;
continue;
default:
TypeInfo ti2 = primitiveTypeInfo(m2);
if (!ti2)
goto Lerror;
void[] va = va_arg!(void[])(argptr);
putArray(va.ptr, va.length, ti2);
}
return;
}
case Mangle.Ttypedef:
ti = (cast(TypeInfo_Typedef)ti).base;
m = cast(Mangle)ti.classinfo.name[9];
formatArg(fc);
return;
case Mangle.Tenum:
ti = (cast(TypeInfo_Enum)ti).base;
m = cast(Mangle)ti.classinfo.name[9];
formatArg(fc);
return;
case Mangle.Tstruct:
{ TypeInfo_Struct tis = cast(TypeInfo_Struct)ti;
if (tis.xtoString is null)
throw new FormatError("Can't convert " ~ tis.toString() ~ " to string: \"string toString()\" not defined");
s = tis.xtoString(argptr);
argptr += (tis.tsize() + 3) & ~3;
goto Lputstr;
}
default:
goto Lerror;
}
Lnumber:
switch (fc)
{
case 's':
case 'd':
if (signed)
{ if (cast(long)vnumber < 0)
{ prefix = "-";
vnumber = -vnumber;
}
else if (flags & FLplus)
prefix = "+";
else if (flags & FLspace)
prefix = " ";
}
break;
case 'b':
signed = 0;
base = 2;
break;
case 'o':
signed = 0;
base = 8;
break;
case 'X':
uc = 1;
if (flags & FLhash && vnumber)
prefix = "0X";
signed = 0;
base = 16;
break;
case 'x':
if (flags & FLhash && vnumber)
prefix = "0x";
signed = 0;
base = 16;
break;
default:
goto Lerror;
}
if (!signed)
{
switch (m)
{
case Mangle.Tbyte:
vnumber &= 0xFF;
break;
case Mangle.Tshort:
vnumber &= 0xFFFF;
break;
case Mangle.Tint:
vnumber &= 0xFFFFFFFF;
break;
default:
break;
}
}
if (flags & FLprecision && fc != 'p')
flags &= ~FL0pad;
if (vnumber < base)
{
if (vnumber == 0 && precision == 0 && flags & FLprecision &&
!(fc == 'o' && flags & FLhash))
{
putstr(null);
return;
}
if (precision == 0 || !(flags & FLprecision))
{ vchar = cast(char)('0' + vnumber);
if (vnumber < 10)
vchar = cast(char)('0' + vnumber);
else
vchar = cast(char)((uc ? 'A' - 10 : 'a' - 10) + vnumber);
goto L2;
}
}
int n = tmpbuf.length;
char c;
int hexoffset = uc ? ('A' - ('9' + 1)) : ('a' - ('9' + 1));
while (vnumber)
{
c = cast(char)((vnumber % base) + '0');
if (c > '9')
c += hexoffset;
vnumber /= base;
tmpbuf[--n] = c;
}
if (tmpbuf.length - n < precision && precision < tmpbuf.length)
{
int m = tmpbuf.length - precision;
tmpbuf[m .. n] = '0';
n = m;
}
else if (flags & FLhash && fc == 'o')
prefix = "0";
putstr(tmpbuf[n .. tmpbuf.length]);
return;
Lreal:
putreal(vreal);
return;
Lcomplex:
putreal(vcreal.re);
putc('+');
putreal(vcreal.im);
putc('i');
return;
Lerror:
throw new FormatError("formatArg");
}
for (j = 0; j < arguments.length; )
{
ti = arguments[j++];
//printf("test1: '%.*s' %d\n", ti.classinfo.name,
//ti.classinfo.name.length); ti.print();
flags = 0;
precision = 0;
field_width = 0;
ti = skipCI(ti);
int mi = 9;
do
{
if (ti.classinfo.name.length <= mi)
goto Lerror;
m = cast(Mangle)ti.classinfo.name[mi++];
} while (m == Mangle.Tconst || m == Mangle.Tinvariant);
if (m == Mangle.Tarray)
{
if (ti.classinfo.name.length == 14 &&
ti.classinfo.name[9..14] == "Array")
{
TypeInfo tn = (cast(TypeInfo_Array)ti).next;
tn = skipCI(tn);
switch (cast(Mangle)tn.classinfo.name[9])
{
case Mangle.Tchar:
case Mangle.Twchar:
case Mangle.Tdchar:
ti = tn;
mi = 9;
break;
default:
break;
}
}
L1:
Mangle m2 = cast(Mangle)ti.classinfo.name[mi];
string fmt; // format string
wstring wfmt;
dstring dfmt;
/* For performance reasons, this code takes advantage of the
* fact that most format strings will be ASCII, and that the
* format specifiers are always ASCII. This means we only need
* to deal with UTF in a couple of isolated spots.
*/
switch (m2)
{
case Mangle.Tchar:
fmt = va_arg!(string)(argptr);
break;
case Mangle.Twchar:
wfmt = va_arg!(wstring)(argptr);
fmt = toUTF8(wfmt);
break;
case Mangle.Tdchar:
dfmt = va_arg!(dstring)(argptr);
fmt = toUTF8(dfmt);
break;
case Mangle.Tconst:
case Mangle.Tinvariant:
mi++;
goto L1;
default:
formatArg('s');
continue;
}
for (size_t i = 0; i < fmt.length; )
{ dchar c = fmt[i++];
dchar getFmtChar()
{ // Valid format specifier characters will never be UTF
if (i == fmt.length)
throw new FormatError("invalid specifier");
return fmt[i++];
}
int getFmtInt()
{ int n;
while (1)
{
n = n * 10 + (c - '0');
if (n < 0) // overflow
throw new FormatError("int overflow");
c = getFmtChar();
if (c < '0' || c > '9')
break;
}
return n;
}
int getFmtStar()
{ Mangle m;
TypeInfo ti;
if (j == arguments.length)
throw new FormatError("too few arguments");
ti = arguments[j++];
m = cast(Mangle)ti.classinfo.name[9];
if (m != Mangle.Tint)
throw new FormatError("int argument expected");
return va_arg!(int)(argptr);
}
if (c != '%')
{
if (c > 0x7F) // if UTF sequence
{
i--; // back up and decode UTF sequence
c = std.utf.decode(fmt, i);
}
Lputc:
putc(c);
continue;
}
// Get flags {-+ #}
flags = 0;
while (1)
{
c = getFmtChar();
switch (c)
{
case '-': flags |= FLdash; continue;
case '+': flags |= FLplus; continue;
case ' ': flags |= FLspace; continue;
case '#': flags |= FLhash; continue;
case '0': flags |= FL0pad; continue;
case '%': if (flags == 0)
goto Lputc;
default: break;
}
break;
}
// Get field width
field_width = 0;
if (c == '*')
{
field_width = getFmtStar();
if (field_width < 0)
{ flags |= FLdash;
field_width = -field_width;
}
c = getFmtChar();
}
else if (c >= '0' && c <= '9')
field_width = getFmtInt();
if (flags & FLplus)
flags &= ~FLspace;
if (flags & FLdash)
flags &= ~FL0pad;
// Get precision
precision = 0;
if (c == '.')
{ flags |= FLprecision;
//flags &= ~FL0pad;
c = getFmtChar();
if (c == '*')
{
precision = getFmtStar();
if (precision < 0)
{ precision = 0;
flags &= ~FLprecision;
}
c = getFmtChar();
}
else if (c >= '0' && c <= '9')
precision = getFmtInt();
}
if (j == arguments.length)
goto Lerror;
ti = arguments[j++];
ti = skipCI(ti);
mi = 9;
do
{
m = cast(Mangle)ti.classinfo.name[mi++];
} while (m == Mangle.Tconst || m == Mangle.Tinvariant);
if (c > 0x7F) // if UTF sequence
goto Lerror; // format specifiers can't be UTF
formatArg(cast(char)c);
}
}
else
{
formatArg('s');
}
}
return;
Lerror:
throw new FormatError();
}
/* ======================== Unit Tests ====================================== */
unittest
{
int i;
string s;
debug(format) printf("std.format.format.unittest\n");
s = std.string.format("hello world! %s %s ", true, 57, 1_000_000_000, 'x', " foo");
assert(s == "hello world! true 57 1000000000x foo");
s = std.string.format(1.67, " %A ", -1.28, float.nan);
/* The host C library is used to format floats.
* C99 doesn't specify what the hex digit before the decimal point
* is for %A.
*/
version (linux)
assert(s == "1.67 -0XA.3D70A3D70A3D8P-3 nan");
else
assert(s == "1.67 -0X1.47AE147AE147BP+0 nan");
s = std.string.format("%x %X", 0x1234AF, 0xAFAFAFAF);
assert(s == "1234af AFAFAFAF");
s = std.string.format("%b %o", 0x1234AF, 0xAFAFAFAF);
assert(s == "100100011010010101111 25753727657");
s = std.string.format("%d %s", 0x1234AF, 0xAFAFAFAF);
assert(s == "1193135 2947526575");
s = std.string.format("%s", 1.2 + 3.4i);
assert(s == "1.2+3.4i");
s = std.string.format("%x %X", 1.32, 6.78f);
assert(s == "3ff51eb851eb851f 40D8F5C3");
s = std.string.format("%#06.*f",2,12.345);
assert(s == "012.35");
s = std.string.format("%#0*.*f",6,2,12.345);
assert(s == "012.35");
s = std.string.format("%7.4g:", 12.678);
assert(s == " 12.68:");
s = std.string.format("%7.4g:", 12.678L);
assert(s == " 12.68:");
s = std.string.format("%04f|%05d|%#05x|%#5x",-4.,-10,1,1);
assert(s == "-4.000000|-0010|0x001| 0x1");
i = -10;
s = std.string.format("%d|%3d|%03d|%1d|%01.4f",i,i,i,i,cast(double) i);
assert(s == "-10|-10|-10|-10|-10.0000");
i = -5;
s = std.string.format("%d|%3d|%03d|%1d|%01.4f",i,i,i,i,cast(double) i);
assert(s == "-5| -5|-05|-5|-5.0000");
i = 0;
s = std.string.format("%d|%3d|%03d|%1d|%01.4f",i,i,i,i,cast(double) i);
assert(s == "0| 0|000|0|0.0000");
i = 5;
s = std.string.format("%d|%3d|%03d|%1d|%01.4f",i,i,i,i,cast(double) i);
assert(s == "5| 5|005|5|5.0000");
i = 10;
s = std.string.format("%d|%3d|%03d|%1d|%01.4f",i,i,i,i,cast(double) i);
assert(s == "10| 10|010|10|10.0000");
s = std.string.format("%.0d", 0);
assert(s == "");
s = std.string.format("%.g", .34);
assert(s == "0.3");
s = std.string.format("%.0g", .34);
assert(s == "0.3");
s = std.string.format("%.2g", .34);
assert(s == "0.34");
s = std.string.format("%0.0008f", 1e-08);
assert(s == "0.00000001");
s = std.string.format("%0.0008f", 1e-05);
assert(s == "0.00001000");
s = "helloworld";
string r;
r = std.string.format("%.2s", s[0..5]);
assert(r == "he");
r = std.string.format("%.20s", s[0..5]);
assert(r == "hello");
r = std.string.format("%8s", s[0..5]);
assert(r == " hello");
byte[] arrbyte = new byte[4];
arrbyte[0] = 100;
arrbyte[1] = -99;
arrbyte[3] = 0;
r = std.string.format(arrbyte);
assert(r == "[100,-99,0,0]");
ubyte[] arrubyte = new ubyte[4];
arrubyte[0] = 100;
arrubyte[1] = 200;
arrubyte[3] = 0;
r = std.string.format(arrubyte);
assert(r == "[100,200,0,0]");
short[] arrshort = new short[4];
arrshort[0] = 100;
arrshort[1] = -999;
arrshort[3] = 0;
r = std.string.format(arrshort);
assert(r == "[100,-999,0,0]");
r = std.string.format("%s",arrshort);
assert(r == "[100,-999,0,0]");
ushort[] arrushort = new ushort[4];
arrushort[0] = 100;
arrushort[1] = 20_000;
arrushort[3] = 0;
r = std.string.format(arrushort);
assert(r == "[100,20000,0,0]");
int[] arrint = new int[4];
arrint[0] = 100;
arrint[1] = -999;
arrint[3] = 0;
r = std.string.format(arrint);
assert(r == "[100,-999,0,0]");
r = std.string.format("%s",arrint);
assert(r == "[100,-999,0,0]");
long[] arrlong = new long[4];
arrlong[0] = 100;
arrlong[1] = -999;
arrlong[3] = 0;
r = std.string.format(arrlong);
assert(r == "[100,-999,0,0]");
r = std.string.format("%s",arrlong);
assert(r == "[100,-999,0,0]");
ulong[] arrulong = new ulong[4];
arrulong[0] = 100;
arrulong[1] = 999;
arrulong[3] = 0;
r = std.string.format(arrulong);
assert(r == "[100,999,0,0]");
string[] arr2 = new string[4];
arr2[0] = "hello";
arr2[1] = "world";
arr2[3] = "foo";
r = std.string.format(arr2);
assert(r == "[hello,world,,foo]");
r = std.string.format("%.8d", 7);
assert(r == "00000007");
r = std.string.format("%.8x", 10);
assert(r == "0000000a");
r = std.string.format("%-3d", 7);
assert(r == "7 ");
r = std.string.format("%*d", -3, 7);
assert(r == "7 ");
r = std.string.format("%.*d", -3, 7);
assert(r == "7");
typedef int myint;
myint m = -7;
r = std.string.format(m);
assert(r == "-7");
r = std.string.format("abc"c);
assert(r == "abc");
r = std.string.format("def"w);
assert(r == "def");
r = std.string.format("ghi"d);
assert(r == "ghi");
void* p = cast(void*)0xDEADBEEF;
r = std.string.format(p);
assert(r == "DEADBEEF");
r = std.string.format("%#x", 0xabcd);
assert(r == "0xabcd");
r = std.string.format("%#X", 0xABCD);
assert(r == "0XABCD");
r = std.string.format("%#o", 012345);
assert(r == "012345");
r = std.string.format("%o", 9);
assert(r == "11");
r = std.string.format("%+d", 123);
assert(r == "+123");
r = std.string.format("%+d", -123);
assert(r == "-123");
r = std.string.format("% d", 123);
assert(r == " 123");
r = std.string.format("% d", -123);
assert(r == "-123");
r = std.string.format("%%");
assert(r == "%");
r = std.string.format("%d", true);
assert(r == "1");
r = std.string.format("%d", false);
assert(r == "0");
r = std.string.format("%d", 'a');
assert(r == "97");
wchar wc = 'a';
r = std.string.format("%d", wc);
assert(r == "97");
dchar dc = 'a';
r = std.string.format("%d", dc);
assert(r == "97");
byte b = byte.max;
r = std.string.format("%x", b);
assert(r == "7f");
r = std.string.format("%x", ++b);
assert(r == "80");
r = std.string.format("%x", ++b);
assert(r == "81");
short sh = short.max;
r = std.string.format("%x", sh);
assert(r == "7fff");
r = std.string.format("%x", ++sh);
assert(r == "8000");
r = std.string.format("%x", ++sh);
assert(r == "8001");
i = int.max;
r = std.string.format("%x", i);
assert(r == "7fffffff");
r = std.string.format("%x", ++i);
assert(r == "80000000");
r = std.string.format("%x", ++i);
assert(r == "80000001");
r = std.string.format("%x", 10);
assert(r == "a");
r = std.string.format("%X", 10);
assert(r == "A");
r = std.string.format("%x", 15);
assert(r == "f");
r = std.string.format("%X", 15);
assert(r == "F");
Object c = null;
r = std.string.format(c);
assert(r == "null");
enum TestEnum
{
Value1, Value2
}
r = std.string.format("%s", TestEnum.Value2);
assert(r == "1");
invariant(char[5])[int] aa = ([3:"hello", 4:"betty"]);
r = std.string.format("%s", aa.values);
assert(r == "[[h,e,l,l,o],[b,e,t,t,y]]");
r = std.string.format("%s", aa);
assert(r == "[3:[h,e,l,l,o],4:[b,e,t,t,y]]");
static const dchar[] ds = ['a','b'];
for (int j = 0; j < ds.length; ++j)
{
r = std.string.format(" %d", ds[j]);
if (j == 0)
assert(r == " 97");
else
assert(r == " 98");
}
r = std.string.format(">%14d<, ", 15, [1,2,3]);
assert(r == "> 15<, [1,2,3]");
assert(std.string.format("%8s", "bar") == " bar");
assert(std.string.format("%8s", "b\u00e9ll\u00f4") == " b\u00e9ll\u00f4");
}
// Andrei
//------------------------------------------------------------------------------
import std.bitmanip;
/*
* A compiled version of an individual writef format
* specifier. FormatInfo only focuses on representation, without
* assigning any semantics to the fields. */
struct FormatInfo
{
/** Special values for width and precision, DYNAMIC width or
* precision means that they were specified with '*' in the format
* string. */
enum short DYNAMIC = short.max;
/** Special value for precision */
enum short UNSPECIFIED = DYNAMIC - 1;
/** minimum width, default 0. */
short width = 0;
/** precision. */
short precision = UNSPECIFIED;
/** The actual format specifier, 's' by default. */
char spec = 's';
/** Index of the argument for positional parameters, from 1 to
* ubyte.max. (0 means not used) */
ubyte index;
/* Flags: flDash for '-', flZero for '0', flSpace for ' ', flPlus
* for '+', flHash for '#'. */
mixin(bitfields!(
bool, "flDash", 1,
bool, "flZero", 1,
bool, "flSpace", 1,
bool, "flPlus", 1,
bool, "flHash", 1,
ubyte, "", 3));
/* For arrays only: the trailing spec after the */
const(char)[] innerTrailing;
/*
* Given a string format specification fmt, parses a format
* specifier. The string is assumed to start with the character
* immediately following the '%'. The string is advanced to right
* after the end of the format specifier. */
static FormatInfo parse(S)(ref S fmt)
{
FormatInfo result;
if (!fmt.length) return result;
size_t i = 0;
for (;;)
switch (fmt[i])
{
case '(':
{
// embedded format specifier
auto j = i + 1;
void check(bool condition)
{
enforce(
condition,
text("Incorrect format specifier: %", fmt[i .. $]));
}
for (int innerParens;; ++j)
{
check(j < fmt.length);
if (fmt[j] == '(') ++innerParens;
else if (fmt[j] == ')')
{
if (innerParens-- == 0) break;
}
else if (fmt[j] == '%')
{
// skip the '%' and the character following
// it, whatever it is
++j;
}
}
auto innerFmtSpec = fmt[i + 1 .. j];
result = parse(innerFmtSpec);
result.innerTrailing = to!(typeof(innerTrailing))(innerFmtSpec);
// We practically found the format specifier
i = j + 1;
//result.raw = to!(const(char)[])(fmt[0 .. i]);
fmt = fmt[i .. $];
return result;
}
case '-': result.flDash = true; ++i; break;
case '+': result.flPlus = true; ++i; break;
case '#': result.flHash = true; ++i; break;
case '0': result.flZero = true; ++i; break;
case ' ': result.flSpace = true; ++i; break;
case '*':
if (isdigit(fmt[++i]))
{
// a '*' followed by digits and '$' is a positional format
fmt = fmt[1 .. $];
result.width = -.parse!(typeof(result.width))(fmt);
i = 0;
if (fmt[i++] != '$') throw new FormatError("$ expected");
}
else
{
// read result
result.width = DYNAMIC;
}
break;
case '1': case '2': case '3': case '4':
case '5': case '6': case '7': case '8': case '9':
auto tmp = fmt[i .. $];
const widthOrArgIndex = .parse!(ushort)(tmp);
assert(tmp.length,
text("Incorrect format specifier %", fmt[i .. $]));
i = tmp.ptr - fmt.ptr;
if (tmp.length && tmp[0] == '$')
{
// index!
result.index = to!(ubyte)(widthOrArgIndex);
++i;
}
else
{
// width
result.width = cast(short) widthOrArgIndex;
}
break;
case '.':
if (fmt[++i] == '*')
{
if (isdigit(fmt[++i]))
{
// a '.*' followed by digits and '$' is a
// positional precision
fmt = fmt[i .. $];
i = 0;
result.precision = cast(short) -.parse!(int)(fmt);
if (fmt[i++] != '$')
{
throw new FormatError("$ expected");
}
}
else
{
// read result
result.precision = DYNAMIC;
}
}
else if (fmt[i] == '-')
{
// negative precision, as good as 0
result.precision = 0;
auto tmp = fmt[i .. $];
.parse!(int)(tmp); // skip digits
i = tmp.ptr - fmt.ptr;
}
else if (isdigit(fmt[i]))
{
auto tmp = fmt[i .. $];
result.precision = .parse!(short)(tmp);
i = tmp.ptr - fmt.ptr;
}
else
{
// "." was specified, but nothing after it
result.precision = 0;
}
break;
default:
// this is the format char
result.spec = fmt[i++];
//result.raw = to!(const(char)[])(fmt[0 .. i]);
fmt = fmt[i .. $];
return result;
} // end switch and for
}
}
//------------------------------------------------------------------------------
// Writes characters in the format strings up to the first format
// specifier and updates the format specifier to remove the written
// portion The updated format fmt does not include the '%'
private void writeUpToFormatSpec(OutRange, S)(ref OutRange w, ref S fmt)
{
for (size_t i = 0; i < fmt.length; ++i)
{
if (fmt[i] != '%') continue;
if (fmt[++i] != '%')
{
// spec found, print and bailout
w.put(fmt[0 .. i - 1]);
fmt = fmt[i .. $];
return;
}
// doubled! Now print whatever we had, then update the string
// and move on
w.put(fmt[0 .. i]);
fmt = fmt[i + 1 .. $];
i = 0;
}
// no format spec found
w.put(fmt);
fmt = null;
}
unittest
{
Appender!(char[]) w;
string fmt = "abc%sdef%sghi";
writeUpToFormatSpec(w, fmt);
assert(w.data == "abc" && fmt == "sdef%sghi");
writeUpToFormatSpec(w, fmt);
assert(w.data == "abcsdef" && fmt == "sghi");
// test with embedded %%s
fmt = "ab%%cd%%ef%sg%%h%sij";
w.clear;
writeUpToFormatSpec(w, fmt);
assert(w.data == "ab%cd%ef" && fmt == "sg%%h%sij");
writeUpToFormatSpec(w, fmt);
assert(w.data == "ab%cd%efsg%h" && fmt == "sij");
}
/*
* Formats an integral number 'arg' according to 'f' and writes it to
* 'w'.
*/
private void formatImplInt(Writer, D)(ref Writer w, D argx, FormatInfo f)
//if (isIntegral!(D))
{
Unqual!(D) arg = argx;
if (f.spec == 'r')
{
// raw write, skip all else and write the thing
auto begin = cast(const char*) &arg;
if (std.system.endian == Endian.LittleEndian && f.flPlus
|| std.system.endian == Endian.BigEndian && f.flDash)
{
// must swap bytes
foreach_reverse (i; 0 .. arg.sizeof)
w.put(begin[i]);
}
else
{
foreach (i; 0 .. arg.sizeof)
w.put(begin[i]);
}
return;
}
if (f.precision == FormatInfo.UNSPECIFIED)
{
// default precision for integrals is 1
f.precision = 1;
}
else
{
// if a precision is specified, the '0' flag is ignored.
f.flZero = false;
}
char leftPad = void;
if (!f.flDash && !f.flZero)
leftPad = ' ';
else if (!f.flDash && f.flZero)
leftPad = '0';
else
leftPad = 0;
// format and write an integral argument
uint base =
f.spec == 'x' || f.spec == 'X' ? 16 :
f.spec == 'o' ? 8 :
f.spec == 'b' ? 2 :
f.spec == 's' || f.spec == 'd' ? 10 :
0;
if (base == 0) throw new FormatError("integral");
// figure out sign and continue in unsigned mode
char forcedPrefix = void;
if (f.flPlus) forcedPrefix = '+';
else if (f.flSpace) forcedPrefix = ' ';
else forcedPrefix = 0;
if (base != 10)
{
// non-10 bases are always unsigned
forcedPrefix = 0;
}
else if (arg < 0)
{
// argument is signed
forcedPrefix = '-';
arg = -arg;
}
// fill the digits
char[] digits = void;
{
char buffer[64]; // 64 bits in base 2 at most
uint i = buffer.length;
auto n = cast(Unsigned!(Unqual!(D))) arg;
do
{
--i;
buffer[i] = cast(char) (n % base);
n /= base;
if (buffer[i] < 10) buffer[i] += '0';
else buffer[i] += (f.spec == 'x' ? 'a' : 'A') - 10;
} while (n);
digits = buffer[i .. $]; // got the digits without the sign
}
// adjust precision to print a '0' for octal if alternate format is on
if (base == 8 && f.flHash()
&& (f.precision <= digits.length)) // too low precision
{
//f.precision = digits.length + (arg != 0);
forcedPrefix = '0';
}
// write left pad; write sign; write 0x or 0X; write digits;
// write right pad
// Writing left pad
int spacesToPrint =
f.width // start with the minimum width
- digits.length // take away digits to print
- (forcedPrefix != 0) // take away the sign if any
- (base == 16 && f.flHash() && arg ? 2 : 0); // 0x or 0X
int delta = f.precision - digits.length;
if (delta > 0) spacesToPrint -= delta;
//writeln(spacesToPrint);
if (spacesToPrint > 0) // need to do some padding
{
if (leftPad == '0')
{
// pad with zeros
f.precision =
cast(typeof(f.precision)) (spacesToPrint + digits.length);
//to!(typeof(f.precision))(spacesToPrint + digits.length);
}
else if (leftPad) foreach (i ; 0 .. spacesToPrint) w.put(' ');
}
// write sign
if (forcedPrefix) w.put(forcedPrefix);
// write 0x or 0X
if (base == 16 && f.flHash() && arg) {
// @@@ overcome bug in dmd;
//w.write(f.spec == 'x' ? "0x" : "0X"); //crashes the compiler
w.put('0');
w.put(f.spec == 'x' ? 'x' : 'X'); // x or X
}
// write the digits
if (arg || f.precision)
{
int zerosToPrint = f.precision - digits.length;
foreach (i ; 0 .. zerosToPrint) w.put('0');
w.put(digits);
}
// write the spaces to the right if left-align
if (!leftPad) foreach (i ; 0 .. spacesToPrint) w.put(' ');
}
/*
* Formats a floating point number 'arg' according to 'f' and writes
* it to 'w'.
*/
private void formatImplFloat(Writer, D)(ref Writer w, D obj, FormatInfo f)
//if (isFloatingPoint!(D))
{
if (f.spec == 'r')
{
// raw write, skip all else and write the thing
auto begin = cast(const char*) &obj;
if (std.system.endian == Endian.LittleEndian && f.flPlus
|| std.system.endian == Endian.BigEndian && f.flDash)
{
// must swap bytes
foreach_reverse (i; 0 .. obj.sizeof)
w.put(begin[i]);
}
else
{
foreach (i; 0 .. obj.sizeof)
w.put(begin[i]);
}
return;
}
if (std.string.find("fgFGaAeEs", f.spec) < 0) {
throw new FormatError("floating");
}
if (f.spec == 's') f.spec = 'g';
char sprintfSpec[1 /*%*/ + 5 /*flags*/ + 3 /*width.prec*/ + 2 /*format*/
+ 1 /*\0*/] = void;
sprintfSpec[0] = '%';
uint i = 1;
if (f.flDash) sprintfSpec[i++] = '-';
if (f.flPlus) sprintfSpec[i++] = '+';
if (f.flZero) sprintfSpec[i++] = '0';
if (f.flSpace) sprintfSpec[i++] = ' ';
if (f.flHash) sprintfSpec[i++] = '#';
sprintfSpec[i .. i + 3] = "*.*";
i += 3;
if (is(const(D) == const(real))) sprintfSpec[i++] = 'L';
sprintfSpec[i++] = f.spec;
sprintfSpec[i] = 0;
//printf("format: '%s'; geeba: %g\n", sprintfSpec.ptr, obj);
char[512] buf;
invariant n = snprintf(buf.ptr, buf.length,
sprintfSpec.ptr,
f.width,
// negative precision is same as no precision specified
f.precision == FormatInfo.UNSPECIFIED ? -1 : f.precision,
obj);
if (n < 0) throw new FormatError("floating point formatting failure");
w.put(buf[0 .. strlen(buf.ptr)]);
}
/*
* Formats an object of type 'D' according to 'f' and writes it to
* 'w'. The pointer 'arg' is assumed to point to an object of type
* 'D'.
*/
private void formatGeneric(Writer, D)(ref Writer w, const(void)* arg,
FormatInfo f)
{
auto obj = *cast(D*) arg;
static if (is(const(D) == const(void[]))) {
auto s = cast(const char[]) obj;
w.put(s);
} else static if (is(D Original == enum)) {
formatGeneric!(Writer, Original)(w, arg, f);
} else static if (is(D Original == typedef)) {
formatGeneric!(Writer, Original)(w, arg, f);
} else static if (isFloatingPoint!(D)) {
formatImplFloat(w, obj, f);
} else static if (is(const(D) == const ifloat)) {
formatImplFloat(w, *cast(float*) &obj, f);
} else static if (is(const(D) == const idouble)) {
formatImplFloat(w, *cast(double*) &obj, f);
} else static if (is(const(D) == const ireal)) {
formatImplFloat(w, *cast(real*) &obj, f);
} else static if (is(const(D) == const cfloat)
|| is(const(D) == const cdouble)
|| is(const(D) == const creal)) {
formatImplFloat(w, obj.re, f);
// @@@BUG 2367@@@
//w.write("+");
w.put('+');
formatImplFloat(w, obj.im, f);
// @@@BUG 2367@@@
//w.put("i");
w.put('i');
} else static if (is(const D == const bool)) {
if (f.spec == 's') {
//@@@BUG 2606
//w.put(obj ? "true" : "false");
auto s = obj ? "true" : "false";
w.put(s);
} else {
formatImplInt(w, cast(int) obj, f);
}
} else static if (is(const(D) == const char)
|| is(const(D) == const wchar)
|| is(const(D) == const dchar)) {
if (f.spec == 's') {
w.put(obj);
} else {
formatImplInt(w, cast(uint) obj, f);
}
} else static if (isIntegral!(D)) {
formatImplInt(w, obj, f);
} else static if (is(D : const(char)[]) || is(D : const(wchar)[])
|| is(D : const(dchar)[])) {
auto s = obj[0 .. f.precision < $ ? f.precision : $];
if (!f.flDash)
{
// right align
if (f.width > s.length)
foreach (i ; 0 .. f.width - s.length) w.put(' ');
w.put(s);
}
else
{
// left align
w.put(s);
if (f.width > s.length)
foreach (i ; 0 .. f.width - s.length) w.put(' ');
}
} else static if (is(D == void[0])) {
w.put('[');
w.put(']');
} else static if (isArray!(D)) {
if (f.spec == 'r')
{
// raw writes
foreach (i, e; obj) formatGeneric!(Writer, typeof(e))(w, &e, f);
}
else
{
if (obj.length == 0) return;
// formatted writes
formatGeneric!(Writer, typeof(obj[0]))(w, &obj[0], f);
if (!f.innerTrailing)
{
foreach (i, e; obj[1 .. $])
{
w.put(' ');
formatGeneric!(Writer, typeof(e))(w, &e, f);
}
}
else
{
const hasEscapes =
std.algorithm.find(f.innerTrailing, '%').length > 0;
foreach (i, e; obj[1 .. $])
{
if (hasEscapes)
{
foreach (dchar c; f.innerTrailing)
{
if (c == '%') continue;
w.put(c);
}
}
else w.put(f.innerTrailing);
formatGeneric!(Writer, typeof(e))(w, &e, f);
}
}
}
} else static if (is(const(D) : const void*)) {
f.spec = 'X';
const fake = cast(ulong) obj;
formatGeneric!(Writer, ulong)(w, &fake, f);
} else static if (is(const(D) : const Object)) {
// @@@BUG 2367@@@
//if (obj is null) w.write("null");
if (obj is null) w.put("null"[]);
else w.put(obj.toString);
} else static if (isAssociativeArray!(D)) {
// somebody rid me of this hack
w.put(std.string.format("%s", obj));
} else static if (is(D : const(char))) {
// somebody rid me of this hack
w.put(obj);
} else static if (is(typeof(obj.toString))) {
auto s = obj.toString;
w.put(s);
} else {
// last resort: just print type name
w.put(D.stringof);
}
}
unittest
{
Appender!(char[]) w;
int[] a = [ 1, 3, 2 ];
formattedWrite(w, "testing %(s, ) embedded", a);
assert(w.data == "testing 1, 3, 2 embedded", w.data);
w.clear;
formattedWrite(w, "testing (%(s%) %()) embedded", a);
assert(w.data == "testing (1) (3) (2) embedded", w.data);
}
//------------------------------------------------------------------------------
// Fix for issue 1591
private int getNthInt(A...)(uint index, A args)
{
static if (A.length)
{
if (index)
{
return getNthInt(index - 1, args[1 .. $]);
}
static if (is(typeof(args[0]) : long) || is(typeof(arg) : ulong))
{
return to!(int)(args[0]);
}
else
{
throw new FormatError("int expected");
}
}
else
{
throw new FormatError("int expected");
}
}
/*
(Not public yet.) Formats arguments $(D args) according to the format
string $(D fmt) and writes the result to $(D w). $(D F) must be $(D
char), $(D wchar), or $(D dchar).
Example:
-------------------------
import std.c.stdio;
import std.format;
string myFormat(A...)(A args)
{
StringWriter!(char) writer;
std.format.formattedWrite(writer,
"%s et %s numeris romanis non sunt", args);
return writer.backend;
}
...
int x = 42;
assert(myFormat(x, 0) == "42 et 0 numeris romanis non sunt");
------------------------
$(D formattedWrite) supports positional parameter syntax in $(WEB
opengroup.org/onlinepubs/009695399/functions/printf.html, POSIX)
style. Example:
-------------------------
StringWriter!(char) writer;
std.format.formattedWrite(writer, "Date: %2$s %1$s", "October", 5);
assert(writer.backend == "Date: 5 October");
------------------------
The positional and non-positional styles can be mixed in the same
format string. (POSIX leaves this behavior undefined.) The internal
counter for non-positional parameters tracks the next parameter after
the largest positional parameter already used.
Warning:
This is the function internally used by writef* but it's still
undergoing active development. Do not rely on it.
*/
void formattedWrite(Writer, F, A...)(ref Writer w, const(F)[] fmt, A args)
{
enum len = args.length;
void function(ref Writer, const(void)*, FormatInfo) funs[len] = void;
const(void)* argsAddresses[len] = void;
foreach (i, arg; args)
{
funs[i] = &formatGeneric!(Writer, typeof(arg));
argsAddresses[i] = &arg;
}
// Are we already done with formats? Then just dump each parameter in turn
uint currentArg = 0;
for (;;)
{
writeUpToFormatSpec(w, fmt);
if (fmt.length == 0)
{
// New behavior: break if there are too few specifiers AND
// at least one specifier did exist
break;
}
auto spec = FormatInfo.parse(fmt);
if (currentArg == funs.length && !spec.index)
{
// leftover spec?
if (fmt.length)
{
throw new FormatError(
cast(string) ("Orphan format specifier: %" ~ fmt));
}
break;
}
if (spec.width == FormatInfo.DYNAMIC)
{
auto width = to!(typeof(spec.width))(getNthInt(currentArg, args));
if (width < 0)
{
spec.flDash = true;
width = cast(short)(0 - width);
}
spec.width = width;
++currentArg;
}
else if (spec.width < 0)
{
// means: get width as a positional parameter
auto index = cast(uint) -spec.width;
auto width = to!(typeof(spec.width))(getNthInt(index, args));
if (currentArg < index) currentArg = index;
if (width < 0)
{
spec.flDash = true;
width = cast(short)(0 - width);
}
spec.width = width;
}
if (spec.precision == FormatInfo.DYNAMIC)
{
auto precision = to!(typeof(spec.precision))(
getNthInt(currentArg, args));
if (precision >= 0) spec.precision = precision;
// else negative precision is same as no precision
else spec.precision = FormatInfo.UNSPECIFIED;
++currentArg;
}
else if (spec.precision < 0)
{
// means: get precision as a positional parameter
auto index = cast(uint) -spec.precision;
auto precision = to!(typeof(spec.precision))(
getNthInt(index, args));
if (currentArg < index) currentArg = index;
if (precision >= 0) spec.precision = precision;
// else negative precision is same as no precision
else spec.precision = FormatInfo.UNSPECIFIED;
}
// Format!
if (spec.index > 0)
{
// using positional parameters!
funs[spec.index - 1](w, argsAddresses[spec.index - 1], spec);
if (currentArg < spec.index) currentArg = spec.index;
}
else
{
funs[currentArg](w, argsAddresses[currentArg], spec);
++currentArg;
}
}
}
/* ======================== Unit Tests ====================================== */
unittest
{
Appender!(char[]) stream;
formattedWrite(stream, "%s", 1.1);
assert(stream.data == "1.1", stream.data);
}
unittest
{
// testing raw writes
Appender!(char[]) w;
uint a = 0x02030405;
formattedWrite(w, "%+r", a);
assert(w.data.length == 4 && w.data[0] == 2 && w.data[1] == 3
&& w.data[2] == 4 && w.data[3] == 5);
w.clear;
formattedWrite(w, "%-r", a);
assert(w.data.length == 4 && w.data[0] == 5 && w.data[1] == 4
&& w.data[2] == 3 && w.data[3] == 2);
}
unittest
{
// testing positional parameters
Appender!(char[]) w;
formattedWrite(w,
"Numbers %2$s and %1$s are reversed and %1$s%2$s repeated",
42, 0);
assert(w.data == "Numbers 0 and 42 are reversed and 420 repeated",
w.data);
w.clear;
formattedWrite(w, "asd%s", 23);
assert(w.data == "asd23", w.data);
w.clear;
formattedWrite(w, "%s%s", 23, 45);
assert(w.data == "2345", w.data);
}
unittest
{
debug(format) printf("std.format.format.unittest\n");
Appender!(char[]) stream;
//goto here;
formattedWrite(stream,
"hello world! %s %s ", true, 57, 1_000_000_000, 'x', " foo");
assert(stream.data == "hello world! true 57 ",
stream.data);
stream.clear;
formattedWrite(stream, "%g %A %s", 1.67, -1.28, float.nan);
//std.c.stdio.fwrite(stream.data.ptr, stream.data.length, 1, stderr);
/* The host C library is used to format floats.
* C99 doesn't specify what the hex digit before the decimal point
* is for %A.
*/
version (linux)
assert(stream.data == "1.67 -0X1.47AE147AE147BP+0 nan", stream.data);
else
assert(stream.data == "1.67 -0X1.47AE147AE147BP+0 nan");
stream.clear;
formattedWrite(stream, "%x %X", 0x1234AF, 0xAFAFAFAF);
assert(stream.data == "1234af AFAFAFAF");
stream.clear;
formattedWrite(stream, "%b %o", 0x1234AF, 0xAFAFAFAF);
assert(stream.data == "100100011010010101111 25753727657");
stream.clear;
formattedWrite(stream, "%d %s", 0x1234AF, 0xAFAFAFAF);
assert(stream.data == "1193135 2947526575");
stream.clear;
formattedWrite(stream, "%s", 1.2 + 3.4i);
assert(stream.data == "1.2+3.4i");
stream.clear;
formattedWrite(stream, "%a %A", 1.32, 6.78f);
//formattedWrite(stream, "%x %X", 1.32);
assert(stream.data == "0x1.51eb851eb851fp+0 0X1.B1EB86P+2");
stream.clear;
formattedWrite(stream, "%#06.*f",2,12.345);
assert(stream.data == "012.35");
stream.clear;
formattedWrite(stream, "%#0*.*f",6,2,12.345);
assert(stream.data == "012.35");
stream.clear;
const real constreal = 1;
formattedWrite(stream, "%g",constreal);
assert(stream.data == "1");
stream.clear;
formattedWrite(stream, "%7.4g:", 12.678);
assert(stream.data == " 12.68:");
stream.clear;
formattedWrite(stream, "%7.4g:", 12.678L);
assert(stream.data == " 12.68:");
stream.clear;
formattedWrite(stream, "%04f|%05d|%#05x|%#5x",-4.,-10,1,1);
assert(stream.data == "-4.000000|-0010|0x001| 0x1",
stream.data);
stream.clear;
int i;
string s;
i = -10;
formattedWrite(stream, "%d|%3d|%03d|%1d|%01.4f",i,i,i,i,cast(double) i);
assert(stream.data == "-10|-10|-10|-10|-10.0000");
stream.clear;
i = -5;
formattedWrite(stream, "%d|%3d|%03d|%1d|%01.4f",i,i,i,i,cast(double) i);
assert(stream.data == "-5| -5|-05|-5|-5.0000");
stream.clear;
i = 0;
formattedWrite(stream, "%d|%3d|%03d|%1d|%01.4f",i,i,i,i,cast(double) i);
assert(stream.data == "0| 0|000|0|0.0000");
stream.clear;
i = 5;
formattedWrite(stream, "%d|%3d|%03d|%1d|%01.4f",i,i,i,i,cast(double) i);
assert(stream.data == "5| 5|005|5|5.0000");
stream.clear;
i = 10;
formattedWrite(stream, "%d|%3d|%03d|%1d|%01.4f",i,i,i,i,cast(double) i);
assert(stream.data == "10| 10|010|10|10.0000");
stream.clear;
formattedWrite(stream, "%.0d", 0);
assert(stream.data == "");
stream.clear;
formattedWrite(stream, "%.g", .34);
assert(stream.data == "0.3");
stream.clear;
stream.clear; formattedWrite(stream, "%.0g", .34);
assert(stream.data == "0.3");
stream.clear; formattedWrite(stream, "%.2g", .34);
assert(stream.data == "0.34");
stream.clear; formattedWrite(stream, "%0.0008f", 1e-08);
assert(stream.data == "0.00000001");
stream.clear; formattedWrite(stream, "%0.0008f", 1e-05);
assert(stream.data == "0.00001000");
//return;
//std.c.stdio.fwrite(stream.data.ptr, stream.data.length, 1, stderr);
s = "helloworld";
string r;
stream.clear; formattedWrite(stream, "%.2s", s[0..5]);
assert(stream.data == "he");
stream.clear; formattedWrite(stream, "%.20s", s[0..5]);
assert(stream.data == "hello");
stream.clear; formattedWrite(stream, "%8s", s[0..5]);
assert(stream.data == " hello");
byte[] arrbyte = new byte[4];
arrbyte[0] = 100;
arrbyte[1] = -99;
arrbyte[3] = 0;
stream.clear; formattedWrite(stream, "%s", arrbyte);
assert(stream.data == "100 -99 0 0", stream.data);
ubyte[] arrubyte = new ubyte[4];
arrubyte[0] = 100;
arrubyte[1] = 200;
arrubyte[3] = 0;
stream.clear; formattedWrite(stream, "%s", arrubyte);
assert(stream.data == "100 200 0 0");
short[] arrshort = new short[4];
arrshort[0] = 100;
arrshort[1] = -999;
arrshort[3] = 0;
stream.clear; formattedWrite(stream, "%s", arrshort);
assert(stream.data == "100 -999 0 0");
stream.clear; formattedWrite(stream, "%s",arrshort);
assert(stream.data == "100 -999 0 0");
ushort[] arrushort = new ushort[4];
arrushort[0] = 100;
arrushort[1] = 20_000;
arrushort[3] = 0;
stream.clear; formattedWrite(stream, "%s", arrushort);
assert(stream.data == "100 20000 0 0");
int[] arrint = new int[4];
arrint[0] = 100;
arrint[1] = -999;
arrint[3] = 0;
stream.clear; formattedWrite(stream, "%s", arrint);
assert(stream.data == "100 -999 0 0");
stream.clear; formattedWrite(stream, "%s",arrint);
assert(stream.data == "100 -999 0 0");
long[] arrlong = new long[4];
arrlong[0] = 100;
arrlong[1] = -999;
arrlong[3] = 0;
stream.clear; formattedWrite(stream, "%s", arrlong);
assert(stream.data == "100 -999 0 0");
stream.clear; formattedWrite(stream, "%s",arrlong);
assert(stream.data == "100 -999 0 0");
ulong[] arrulong = new ulong[4];
arrulong[0] = 100;
arrulong[1] = 999;
arrulong[3] = 0;
stream.clear; formattedWrite(stream, "%s", arrulong);
assert(stream.data == "100 999 0 0");
string[] arr2 = new string[4];
arr2[0] = "hello";
arr2[1] = "world";
arr2[3] = "foo";
stream.clear; formattedWrite(stream, "%s", arr2);
assert(stream.data == "hello world foo");
stream.clear; formattedWrite(stream, "%.8d", 7);
assert(stream.data == "00000007");
stream.clear; formattedWrite(stream, "%.8x", 10);
assert(stream.data == "0000000a");
stream.clear; formattedWrite(stream, "%-3d", 7);
assert(stream.data == "7 ");
stream.clear; formattedWrite(stream, "%*d", -3, 7);
assert(stream.data == "7 ");
stream.clear; formattedWrite(stream, "%.*d", -3, 7);
//writeln(stream.data);
assert(stream.data == "7");
// assert(false);
// typedef int myint;
// myint m = -7;
// stream.clear; formattedWrite(stream, "", m);
// assert(stream.data == "-7");
stream.clear; formattedWrite(stream, "%s", "abc"c);
assert(stream.data == "abc");
stream.clear; formattedWrite(stream, "%s", "def"w);
assert(stream.data == "def", "["~stream.data~"]");
stream.clear; formattedWrite(stream, "%s", "ghi"d);
assert(stream.data == "ghi");
here:
void* p = cast(void*)0xDEADBEEF;
stream.clear; formattedWrite(stream, "%s", p);
assert(stream.data == "DEADBEEF");
stream.clear; formattedWrite(stream, "%#x", 0xabcd);
assert(stream.data == "0xabcd");
stream.clear; formattedWrite(stream, "%#X", 0xABCD);
assert(stream.data == "0XABCD");
stream.clear; formattedWrite(stream, "%#o", 012345);
assert(stream.data == "012345");
stream.clear; formattedWrite(stream, "%o", 9);
assert(stream.data == "11");
stream.clear; formattedWrite(stream, "%+d", 123);
assert(stream.data == "+123");
stream.clear; formattedWrite(stream, "%+d", -123);
assert(stream.data == "-123");
stream.clear; formattedWrite(stream, "% d", 123);
assert(stream.data == " 123");
stream.clear; formattedWrite(stream, "% d", -123);
assert(stream.data == "-123");
stream.clear; formattedWrite(stream, "%%");
assert(stream.data == "%");
stream.clear; formattedWrite(stream, "%d", true);
assert(stream.data == "1");
stream.clear; formattedWrite(stream, "%d", false);
assert(stream.data == "0");
stream.clear; formattedWrite(stream, "%d", 'a');
assert(stream.data == "97", stream.data);
wchar wc = 'a';
stream.clear; formattedWrite(stream, "%d", wc);
assert(stream.data == "97");
dchar dc = 'a';
stream.clear; formattedWrite(stream, "%d", dc);
assert(stream.data == "97");
byte b = byte.max;
stream.clear; formattedWrite(stream, "%x", b);
assert(stream.data == "7f");
stream.clear; formattedWrite(stream, "%x", ++b);
assert(stream.data == "80");
stream.clear; formattedWrite(stream, "%x", ++b);
assert(stream.data == "81");
short sh = short.max;
stream.clear; formattedWrite(stream, "%x", sh);
assert(stream.data == "7fff");
stream.clear; formattedWrite(stream, "%x", ++sh);
assert(stream.data == "8000");
stream.clear; formattedWrite(stream, "%x", ++sh);
assert(stream.data == "8001");
i = int.max;
stream.clear; formattedWrite(stream, "%x", i);
assert(stream.data == "7fffffff");
stream.clear; formattedWrite(stream, "%x", ++i);
assert(stream.data == "80000000");
stream.clear; formattedWrite(stream, "%x", ++i);
assert(stream.data == "80000001");
stream.clear; formattedWrite(stream, "%x", 10);
assert(stream.data == "a");
stream.clear; formattedWrite(stream, "%X", 10);
assert(stream.data == "A");
stream.clear; formattedWrite(stream, "%x", 15);
assert(stream.data == "f");
stream.clear; formattedWrite(stream, "%X", 15);
assert(stream.data == "F");
Object c = null;
stream.clear; formattedWrite(stream, "%s", c);
assert(stream.data == "null");
enum TestEnum
{
Value1, Value2
}
stream.clear; formattedWrite(stream, "%s", TestEnum.Value2);
assert(stream.data == "1", stream.data);
//invariant(char[5])[int] aa = ([3:"hello", 4:"betty"]);
//stream.clear; formattedWrite(stream, "%s", aa.values);
//std.c.stdio.fwrite(stream.data.ptr, stream.data.length, 1, stderr);
//assert(stream.data == "[[h,e,l,l,o],[b,e,t,t,y]]");
//stream.clear; formattedWrite(stream, "%s", aa);
//assert(stream.data == "[3:[h,e,l,l,o],4:[b,e,t,t,y]]");
static const dchar[] ds = ['a','b'];
for (int j = 0; j < ds.length; ++j)
{
stream.clear; formattedWrite(stream, " %d", ds[j]);
if (j == 0)
assert(stream.data == " 97");
else
assert(stream.data == " 98");
}
stream.clear; formattedWrite(stream, "%.-3d", 7);
assert(stream.data == "7", ">" ~ stream.data ~ "<");
// systematic test
const string[] flags = [ "-", "+", "#", "0", " ", "" ];
const string[] widths = [ "", "0", "4", "20" ];
const string[] precs = [ "", ".", ".0", ".4", ".20" ];
const string formats = "sdoxXeEfFgGaA";
/+
foreach (flag1; flags)
foreach (flag2; flags)
foreach (flag3; flags)
foreach (flag4; flags)
foreach (flag5; flags)
foreach (width; widths)
foreach (prec; precs)
foreach (format; formats)
{
stream.clear;
auto fmt = "%" ~ flag1 ~ flag2 ~ flag3
~ flag4 ~ flag5 ~ width ~ prec ~ format
~ '\0';
fmt = fmt[0 .. $ - 1]; // keep it zero-term
char buf[256];
buf[0] = 0;
switch (format)
{
case 's':
formattedWrite(stream, fmt, "wyda");
snprintf(buf.ptr, buf.length, fmt.ptr,
"wyda\0".ptr);
break;
case 'd':
formattedWrite(stream, fmt, 456);
snprintf(buf.ptr, buf.length, fmt.ptr,
456);
break;
case 'o':
formattedWrite(stream, fmt, 345);
snprintf(buf.ptr, buf.length, fmt.ptr,
345);
break;
case 'x':
formattedWrite(stream, fmt, 63546);
snprintf(buf.ptr, buf.length, fmt.ptr,
63546);
break;
case 'X':
formattedWrite(stream, fmt, 12566);
snprintf(buf.ptr, buf.length, fmt.ptr,
12566);
break;
case 'e':
formattedWrite(stream, fmt, 3245.345234);
snprintf(buf.ptr, buf.length, fmt.ptr,
3245.345234);
break;
case 'E':
formattedWrite(stream, fmt, 3245.2345234);
snprintf(buf.ptr, buf.length, fmt.ptr,
3245.2345234);
break;
case 'f':
formattedWrite(stream, fmt, 3245234.645675);
snprintf(buf.ptr, buf.length, fmt.ptr,
3245234.645675);
break;
case 'F':
formattedWrite(stream, fmt, 213412.43);
snprintf(buf.ptr, buf.length, fmt.ptr,
213412.43);
break;
case 'g':
formattedWrite(stream, fmt, 234134.34);
snprintf(buf.ptr, buf.length, fmt.ptr,
234134.34);
break;
case 'G':
formattedWrite(stream, fmt, 23141234.4321);
snprintf(buf.ptr, buf.length, fmt.ptr,
23141234.4321);
break;
case 'a':
formattedWrite(stream, fmt, 21341234.2134123);
snprintf(buf.ptr, buf.length, fmt.ptr,
21341234.2134123);
break;
case 'A':
formattedWrite(stream, fmt, 1092384098.45234);
snprintf(buf.ptr, buf.length, fmt.ptr,
1092384098.45234);
break;
default:
break;
}
auto exp = buf[0 .. strlen(buf.ptr)];
if (stream.data != exp)
{
writeln("Format: \"", fmt, '"');
writeln("Expected: >", exp, "<");
writeln("Actual: >", stream.data,
"<");
assert(false);
}
}+/
}
unittest
{
invariant(char[5])[int] aa = ([3:"hello", 4:"betty"]);
if (false) writeln(aa.keys);
assert(aa[3] == "hello");
assert(aa[4] == "betty");
if (false)
{
writeln(aa.values[0]);
writeln(aa.values[1]);
writefln("%s", typeid(typeof(aa.values)));
writefln("%s", aa[3]);
writefln("%s", aa[4]);
writefln("%s", aa.values);
//writefln("%s", aa);
wstring a = "abcd";
writefln(a);
dstring b = "abcd";
writefln(b);
}
Appender!(char[]) stream;
alias TypeTuple!(byte, ubyte, short, ushort, int, uint, long, ulong,
float, double, real,
ifloat, idouble, ireal, cfloat, cdouble, creal) AllNumerics;
foreach (T; AllNumerics)
{
static if (is(T : ireal))
T value = 1i;
else static if (is(T : creal))
T value = 1 + 1i;
else
T value = 1;
stream.clear; formattedWrite(stream, "%s", value);
static if (is(T : creal))
assert(stream.data == "1+1i");
else
assert(stream.data == "1");
// test typedefs too
typedef T Wyda;
Wyda another = 1;
stream.clear; formattedWrite(stream, "%s", another);
assert(stream.data == "1");
}
//auto r = std.string.format("%s", aa.values);
stream.clear; formattedWrite(stream, "%s", aa);
assert(stream.data == "[3:[h,e,l,l,o],4:[b,e,t,t,y]]", stream.data);
// r = std.string.format("%s", aa);
// assert(r == "[3:[h,e,l,l,o],4:[b,e,t,t,y]]");
}
//------------------------------------------------------------------------------
R formattedRead(R, S...)(R r, const(char)[] fmt, S args)
{
static if (!S.length)
{
r = parseToFormatSpec(r, fmt);
enforce(fmt.length == 0);
return r;
}
else
{
FormatInfo spec;
// The loop below accounts for '*' == fields meant to be read
// and skipped
for (;;)
{
r = parseToFormatSpec(r, fmt);
spec = FormatInfo.parse(fmt);
if (spec.width != spec.DYNAMIC) break;
// must skip this field
skipData(r, spec);
}
alias typeof(*args[0]) A;
//@@@BUG 2725
//static if (is(A X == Tuple!(T), T))
static if (is(A.Types[0]))
{
//@@@BUG
//args[0].field[0] = parse!(A.Types[0])(r);
// static if (A.Types.length > 1)
// return formattedRead(r, fmt,
// args[0].slice!(1, A.Types.length)(),
// args[1 .. $]);
// else
// return formattedRead(r, fmt, args[1 .. $]);
// assume it's a tuple
foreach (i, T; A.Types)
{
//writeln("Parsing ", r, " with format ", fmt);
args[0].field[i] = unformat!(T)(r, spec);
for (;;)
{
r = parseToFormatSpec(r, fmt);
spec = FormatInfo.parse(fmt);
if (spec.width != spec.DYNAMIC) break;
// must skip this guy
skipData(r, spec);
}
}
return formattedRead(r, fmt, args[1 .. $]);
}
else
{
*args[0] = unformat!(A)(r, spec);
return formattedRead(r, fmt, args[1 .. $]);
}
}
}
//------------------------------------------------------------------------------
void skipData(Range)(ref Range input, FormatInfo spec)
{
switch (spec.spec)
{
case 'c': input.popFront; break;
case 'd': assert(false, "Not implemented");
case 'u': while (!input.empty && isdigit(input.front)) input.popFront;
break;
default:
assert(false, text("Not understood: ", spec.spec));
}
}
//------------------------------------------------------------------------------
T unformat(T, Range)(ref Range input, FormatInfo spec) if (isArray!T)
{
Appender!(T) app;
for (;;)
{
auto e = parse!(ElementType!(T))(input);
app.put(e);
if (!std.string.startsWith(input, spec.innerTrailing)) break; // done
input = input[spec.innerTrailing.length .. $];
if (input.empty) break; // the trailing is terminator, not
// separator
}
return app.data;
}
private template acceptedSpecs(T)
{
static if (isIntegral!T) enum acceptedSpecs = "sdu";// + "coxX" (todo)
else static if (isFloatingPoint!T) enum acceptedSpecs = "seEfgG";
else enum acceptedSpecs = "";
}
//------------------------------------------------------------------------------
T unformat(T, Range)(ref Range input, FormatInfo spec) if (isIntegral!T)
{
enforce(std.algorithm.find("cdosuxX", spec.spec).length,
text("Wrong integral type specifier: `", spec.spec, "'"));
if (std.algorithm.find("dsu", spec.spec).length)
{
return parse!T(input);
}
assert(0, "Parsing spec '"~spec.spec~"' not implemented.");
}
//------------------------------------------------------------------------------
T unformat(T, Range)(ref Range input, FormatInfo spec) if (isFloatingPoint!T)
{
enforce(std.algorithm.find(acceptedSpecs!T, spec.spec).length,
text("Format specifier `", spec.spec,
"' not accepted for floating point types"));
return parse!T(input);
}
//------------------------------------------------------------------------------
private R parseToFormatSpec(R)(R r, ref const(char)[] fmt)
{
while (fmt.length)
{
if (fmt[0] == '%')
{
if (fmt.length > 1 && fmt[1] == '%')
{
assert(!r.empty);
// Require a '%'
if (r.front != '%') break;
fmt = fmt[1 .. $];
r.popFront();
}
else
{
fmt.popFront;
break;
}
}
else
{
if (fmt.front == ' ')
{
r = std.algorithm.find!(not!isspace)(r);
}
else
{
enforce(
!r.empty,
text("parseToFormatSpec: Cannot find character `",
fmt.front, "' in the input string `", r, "'"));
if (r.front != fmt.front) break;
r.popFront;
}
fmt.popFront;
}
}
return r;
}
unittest
{
char[] line = "1 2".dup;
string format = "%s %s";
int a, b;
formattedRead(line, format, &a, &b);
assert(a == 1 && b == 2);
Tuple!(int, float) t;
line = "1 2.125".dup;
formattedRead(line, format, &t);
assert(t.field[0] == 1 && t.field[1] == 2.125);
line = "1 7643 2.125".dup;
formattedRead(line, "%s %*u %s", &t);
assert(t.field[0] == 1 && t.field[1] == 2.125);
}
/*
* Copyright (C) 2004-2006 by Digital Mars, www.digitalmars.com
* Written by Walter Bright and Andrei Alexandrescu
*
* This software is provided 'as-is', without any express or implied
* warranty. In no event will the authors be held liable for any damages
* arising from the use of this software.
*
* Permission is granted to anyone to use this software for any purpose,
* including commercial applications, and to alter it and redistribute it
* freely, subject to the following restrictions:
*
* o The origin of this software must not be misrepresented; you must not
* claim that you wrote the original software. If you use this software
* in a product, an acknowledgment in the product documentation would be
* appreciated but is not required.
* o Altered source versions must be plainly marked as such, and must not
* be misrepresented as being the original software.
* o This notice may not be removed or altered from any source
* distribution.
*/