.TH FMTINSTALL 2.SH NAMEfmtinstall, dofmt, dorfmt, fmtprint, fmtvprint, fmtrune, fmtstrcpy, fmtrunestrcpy, fmtfdinit, fmtfdflush, fmtstrinit, fmtstrflush, runefmtstrinit, runefmtstrflush, errfmt \- support for user-defined print formats and output routines.SH SYNOPSIS.B #include <u.h>.br.B #include <libc.h>.PP.ft L.nf.ta \w' 'u +\w' 'u +\w' 'u +\w' 'u +\w' 'utypedef struct Fmt Fmt;struct Fmt{ uchar runes; /* output buffer is runes or chars? */ void *start; /* of buffer */ void *to; /* current place in the buffer */ void *stop; /* end of the buffer; overwritten if flush fails */ int (*flush)(Fmt*); /* called when to == stop */ void *farg; /* to make flush a closure */ int nfmt; /* num chars formatted so far */ va_list args; /* args passed to dofmt */ int r; /* % format Rune */ int width; int prec; ulong flags;};.sp 0.3venum{ FmtWidth = 1, FmtLeft = FmtWidth << 1, FmtPrec = FmtLeft << 1, FmtSharp = FmtPrec << 1, FmtSpace = FmtSharp << 1, FmtSign = FmtSpace << 1, FmtZero = FmtSign << 1, FmtUnsigned = FmtZero << 1, FmtShort = FmtUnsigned << 1, FmtLong = FmtShort << 1, FmtVLong = FmtLong << 1, FmtComma = FmtVLong << 1,.sp 0.3v FmtFlag = FmtComma << 1};.fi.PP.B.ta \w'\fLchar* 'u.sp 0.3v.PP.Bint fmtfdinit(Fmt *f, int fd, char *buf, int nbuf);.PP.Bint fmtfdflush(Fmt *f);.PP.Bint fmtstrinit(Fmt *f);.PP.Bchar* fmtstrflush(Fmt *f);.PP.Bint runefmtstrinit(Fmt *f);.PP.BRune* runefmtstrflush(Fmt *f);.sp 0.3v.PP.Bint fmtinstall(int c, int (*fn)(Fmt*));.PP.Bint dofmt(Fmt *f, char *fmt);.PP.Bint dorfmt(Fmt*, Rune *fmt);.PP.Bint fmtprint(Fmt *f, char *fmt, ...);.PP.Bint fmtvprint(Fmt *f, char *fmt, va_list v);.PP.Bint fmtrune(Fmt *f, int r);.PP.Bint fmtstrcpy(Fmt *f, char *s);.PP.Bint fmtrunestrcpy(Fmt *f, Rune *s);.PP.Bint errfmt(Fmt *f);.SH DESCRIPTIONThe interface described here allows the construction of custom.IR print (2)verbs and output routines.In essence, they provide access to the workings of the formatted print code..PPThe.IR print (2)suite maintains its state with a data structure called.BR Fmt .A typical call to.IR print (2)or its relatives initializes a.B Fmtstructure, passes it to subsidiary routines to process the output,and finishes by emitting any saved state recorded in the.BR Fmt .The details of the.B Fmtare unimportant to outside users, except insofar as the generaldesign influences the interface.The.B Fmtrecords whether the output is in runes or bytes,the verb being processed, its precision and width,and buffering parameters.Most important, it also records a.I flushroutine that the library will call if a buffer overflows.When printing to a file descriptor, the flush routine willemit saved characters and reset the buffer; when printingto an allocated string, it will resize the string to receive more output.The flush routine is nil when printing to fixed-size buffers.User code need never provide a flush routine; this is done internallyby the library..SS Custom output routinesTo write a custom output routine, such as an error handler thatformats and prints custom error messages, the output sequence can be runfrom outside the library using the routines described here.There are two main cases: output to an open file descriptorand output to a string..PPTo write to a file descriptor, call.I fmtfdinitto initialize the local.B Fmtstructure.IR f ,giving the file descriptor.IR fd ,the buffer.IR buf ,and its size.IR nbuf .Then call.IR fmtprintor.IR fmtvprintto generate the output.These behave like.B fprint(see.IR print (2))or.B vfprintexcept that the characters are buffered until.I fmtfdflushis called and the return value is either 0 or \-1.A typical example of this sequence appears in the Examples section..PPThe same basic sequence applies when outputting to an allocated string:call.I fmtstrinitto initialize the.BR Fmt ,then call.I fmtprintand.I fmtvprintto generate the output.Finally,.I fmtstrflushwill return the allocated string, which should be freed after use.To output to a rune string, use.I runefmtstrinitand.IR runefmtstrflush .Regardless of the output style or type,.I fmtprintor.I fmtvprintgenerates the characters..SS Custom format verbs.I Fmtinstallis used to install custom verbs and flags labeled by character.IR c ,which may be any non-zero Unicode character..I Fnshould be declared as.IP.EXint fn(Fmt*).EE.PP.IB Fp ->ris the flag or verb character to cause.I fnto be called.In.IR fn ,.IB fp ->width ,.IB fp ->precare the width and precision, and.IB fp ->flagsthe decoded flags for the verb (see.IR print (2)for a description of these items).The standard flag values are:.B FmtSign.RB ( + ),.B FmtLeft.RB ( - ),.B FmtSpace.RB ( '\ ' ),.B FmtSharp.RB ( # ),.B FmtComma.RB ( , ),.B FmtLong.RB ( l ),.B FmtShort.RB ( h ),.B FmtUnsigned.RB ( u ),and.B FmtVLong.RB ( ll ).The flag bits.B FmtWidthand.B FmtPrecidentify whether a width and precision were specified..PP.I Fnis passed a pointer to the.B Fmtstructure recording the state of the output.If.IB fp ->ris a verb (rather than a flag),.I fnshould use .B Fmt->argsto fetch its argument from the list,then format it, and return zero.If.IB fp ->ris a flag,.I fnshould return one.All interpretation of.IB fp ->width\f1,.IB fp ->prec\f1,and.IB fp-> flagsis left up to the conversion routine..I Fmtinstallreturns 0 if the installation succeeds, \-1 if it fails..PP.IR Fmtprintand.IR fmtvprintmay be called tohelp prepare output in custom conversion routines.However, these functions clear the width, precision, and flags.Both functions return 0 for success and \-1 for failure..PPThe functions.I dofmtand.I dorfmtare the underlying formatters; theyuse the existing contents of.B Fmtand should be called only by sophisticated conversion routines.These routines return the number of characters (bytes of UTF or runes)produced..PPSome internal functions may be useful to format primitive types.They honor the width, precision and flags as described in.IR print (2)..I Fmtruneformats a single character.BR r ..I Fmtstrcpyformats a string.BR s ;.I fmtrunestrcpyformats a rune string.BR s ..I Errfmtformats the system error string.All these routines return zero for successful execution.Conversion routines that call these functions will work properlyregardless of whether the output is bytes or runes..PP.IR 2c (1)describes the C directive.B #pragma.B varargckthat can be used to provide type-checking for custom print verbs and output routines..SH EXAMPLESThis function prints an error message with a variablenumber of arguments and then quits.Compared to the corresponding example in.IR print (2),this version uses a smaller buffer, will never truncatethe output message, but might generate multiple.B writesystem calls to produce its output..IP.EX.ta 6n +6n +6n +6n +6n +6n +6n +6n +6n#pragma varargck argpos fatal 1.sp 0.3vvoidfatal(char *fmt, ...){ Fmt f; char buf[64]; va_list arg;.sp 0.3v fmtfdinit(&f, 1, buf, sizeof buf); fmtprint(&f, "fatal: "); va_start(arg, fmt); fmtvprint(&f, fmt, arg); va_end(arg); fmtprint(&f, "\en"); fmtfdflush(&f); exits("fatal error");}.EE.PPThis example adds a verb to print complex numbers..IP.EXtypedef struct { double r, i;} Complex;.sp 0.3v#pragma varargck type "X" Complex.sp 0.3vintXfmt(Fmt *f){ Complex c;.sp 0.3v c = va_arg(f->args, Complex); return fmtprint(f, "(%g,%g)", c.r, c.i);}.sp 0.3vmain(...){ Complex x = (Complex){ 1.5, -2.3 };.sp 0.3v fmtinstall('X', Xfmt); print("x = %X\en", x);}.EE.SH SOURCE.B /sys/src/libc/fmt.SH SEE ALSO.IR print (2),.IR utf (6),.IR errstr (2).SH DIAGNOSTICSThese routines return negative numbers or nil for errors and set.IR errstr .
.TH FMTINSTALL 2
.SH NAME
fmtinstall, dofmt, dorfmt, fmtprint, fmtvprint, fmtrune, fmtstrcpy, fmtrunestrcpy, fmtfdinit, fmtfdflush, fmtstrinit, fmtstrflush, runefmtstrinit, runefmtstrflush, errfmt \- support for user-defined print formats and output routines
.SH SYNOPSIS
.B #include <u.h>
.br
.B #include <libc.h>
.PP
.ft L
.nf
.ta \w' 'u +\w' 'u +\w' 'u +\w' 'u +\w' 'u
typedef struct Fmt Fmt;
struct Fmt{
uchar runes; /* output buffer is runes or chars? */
void *start; /* of buffer */
void *to; /* current place in the buffer */
void *stop; /* end of the buffer; overwritten if flush fails */
int (*flush)(Fmt*); /* called when to == stop */
void *farg; /* to make flush a closure */
int nfmt; /* num chars formatted so far */
va_list args; /* args passed to dofmt */
int r; /* % format Rune */
int width;
int prec;
ulong flags;
};
.sp 0.3v
enum{
FmtWidth = 1,
FmtLeft = FmtWidth << 1,
FmtPrec = FmtLeft << 1,
FmtSharp = FmtPrec << 1,
FmtSpace = FmtSharp << 1,
FmtSign = FmtSpace << 1,
FmtZero = FmtSign << 1,
FmtUnsigned = FmtZero << 1,
FmtShort = FmtUnsigned << 1,
FmtLong = FmtShort << 1,
FmtVLong = FmtLong << 1,
FmtComma = FmtVLong << 1,
.sp 0.3v
FmtFlag = FmtComma << 1
};
.fi
.PP
.B
.ta \w'\fLchar* 'u
.sp 0.3v
.PP
.B
int fmtfdinit(Fmt *f, int fd, char *buf, int nbuf);
.PP
.B
int fmtfdflush(Fmt *f);
.PP
.B
int fmtstrinit(Fmt *f);
.PP
.B
char* fmtstrflush(Fmt *f);
.PP
.B
int runefmtstrinit(Fmt *f);
.PP
.B
Rune* runefmtstrflush(Fmt *f);
.sp 0.3v
.PP
.B
int fmtinstall(int c, int (*fn)(Fmt*));
.PP
.B
int dofmt(Fmt *f, char *fmt);
.PP
.B
int dorfmt(Fmt*, Rune *fmt);
.PP
.B
int fmtprint(Fmt *f, char *fmt, ...);
.PP
.B
int fmtvprint(Fmt *f, char *fmt, va_list v);
.PP
.B
int fmtrune(Fmt *f, int r);
.PP
.B
int fmtstrcpy(Fmt *f, char *s);
.PP
.B
int fmtrunestrcpy(Fmt *f, Rune *s);
.PP
.B
int errfmt(Fmt *f);
.SH DESCRIPTION
The interface described here allows the construction of custom
.IR print (2)
verbs and output routines.
In essence, they provide access to the workings of the formatted print code.
.PP
The
.IR print (2)
suite maintains its state with a data structure called
.BR Fmt .
A typical call to
.IR print (2)
or its relatives initializes a
.B Fmt
structure, passes it to subsidiary routines to process the output,
and finishes by emitting any saved state recorded in the
.BR Fmt .
The details of the
.B Fmt
are unimportant to outside users, except insofar as the general
design influences the interface.
The
.B Fmt
records whether the output is in runes or bytes,
the verb being processed, its precision and width,
and buffering parameters.
Most important, it also records a
.I flush
routine that the library will call if a buffer overflows.
When printing to a file descriptor, the flush routine will
emit saved characters and reset the buffer; when printing
to an allocated string, it will resize the string to receive more output.
The flush routine is nil when printing to fixed-size buffers.
User code need never provide a flush routine; this is done internally
by the library.
.SS Custom output routines
To write a custom output routine, such as an error handler that
formats and prints custom error messages, the output sequence can be run
from outside the library using the routines described here.
There are two main cases: output to an open file descriptor
and output to a string.
.PP
To write to a file descriptor, call
.I fmtfdinit
to initialize the local
.B Fmt
structure
.IR f ,
giving the file descriptor
.IR fd ,
the buffer
.IR buf ,
and its size
.IR nbuf .
Then call
.IR fmtprint
or
.IR fmtvprint
to generate the output.
These behave like
.B fprint
(see
.IR print (2))
or
.B vfprint
except that the characters are buffered until
.I fmtfdflush
is called and the return value is either 0 or \-1.
A typical example of this sequence appears in the Examples section.
.PP
The same basic sequence applies when outputting to an allocated string:
call
.I fmtstrinit
to initialize the
.BR Fmt ,
then call
.I fmtprint
and
.I fmtvprint
to generate the output.
Finally,
.I fmtstrflush
will return the allocated string, which should be freed after use.
To output to a rune string, use
.I runefmtstrinit
and
.IR runefmtstrflush .
Regardless of the output style or type,
.I fmtprint
or
.I fmtvprint
generates the characters.
.SS Custom format verbs
.I Fmtinstall
is used to install custom verbs and flags labeled by character
.IR c ,
which may be any non-zero Unicode character.
.I Fn
should be declared as
.IP
.EX
int fn(Fmt*)
.EE
.PP
.IB Fp ->r
is the flag or verb character to cause
.I fn
to be called.
In
.IR fn ,
.IB fp ->width ,
.IB fp ->prec
are the width and precision, and
.IB fp ->flags
the decoded flags for the verb (see
.IR print (2)
for a description of these items).
The standard flag values are:
.B FmtSign
.RB ( + ),
.B FmtLeft
.RB ( - ),
.B FmtSpace
.RB ( '\ ' ),
.B FmtSharp
.RB ( # ),
.B FmtComma
.RB ( , ),
.B FmtLong
.RB ( l ),
.B FmtShort
.RB ( h ),
.B FmtUnsigned
.RB ( u ),
and
.B FmtVLong
.RB ( ll ).
The flag bits
.B FmtWidth
and
.B FmtPrec
identify whether a width and precision were specified.
.PP
.I Fn
is passed a pointer to the
.B Fmt
structure recording the state of the output.
If
.IB fp ->r
is a verb (rather than a flag),
.I fn
should use
.B Fmt->args
to fetch its argument from the list,
then format it, and return zero.
If
.IB fp ->r
is a flag,
.I fn
should return one.
All interpretation of
.IB fp ->width\f1,
.IB fp ->prec\f1,
and
.IB fp-> flags
is left up to the conversion routine.
.I Fmtinstall
returns 0 if the installation succeeds, \-1 if it fails.
.PP
.IR Fmtprint
and
.IR fmtvprint
may be called to
help prepare output in custom conversion routines.
However, these functions clear the width, precision, and flags.
Both functions return 0 for success and \-1 for failure.
.PP
The functions
.I dofmt
and
.I dorfmt
are the underlying formatters; they
use the existing contents of
.B Fmt
and should be called only by sophisticated conversion routines.
These routines return the number of characters (bytes of UTF or runes)
produced.
.PP
Some internal functions may be useful to format primitive types.
They honor the width, precision and flags as described in
.IR print (2).
.I Fmtrune
formats a single character
.BR r .
.I Fmtstrcpy
formats a string
.BR s ;
.I fmtrunestrcpy
formats a rune string
.BR s .
.I Errfmt
formats the system error string.
All these routines return zero for successful execution.
Conversion routines that call these functions will work properly
regardless of whether the output is bytes or runes.
.PP
.IR 2c (1)
describes the C directive
.B #pragma
.B varargck
that can be used to provide type-checking for custom print verbs and output routines.
.SH EXAMPLES
This function prints an error message with a variable
number of arguments and then quits.
Compared to the corresponding example in
.IR print (2),
this version uses a smaller buffer, will never truncate
the output message, but might generate multiple
.B write
system calls to produce its output.
.IP
.EX
.ta 6n +6n +6n +6n +6n +6n +6n +6n +6n
#pragma varargck argpos fatal 1
.sp 0.3v
void
fatal(char *fmt, ...)
{
Fmt f;
char buf[64];
va_list arg;
.sp 0.3v
fmtfdinit(&f, 1, buf, sizeof buf);
fmtprint(&f, "fatal: ");
va_start(arg, fmt);
fmtvprint(&f, fmt, arg);
va_end(arg);
fmtprint(&f, "\en");
fmtfdflush(&f);
exits("fatal error");
}
.EE
.PP
This example adds a verb to print complex numbers.
.IP
.EX
typedef struct {
double r, i;
} Complex;
.sp 0.3v
#pragma varargck type "X" Complex
.sp 0.3v
int
Xfmt(Fmt *f)
{
Complex c;
.sp 0.3v
c = va_arg(f->args, Complex);
return fmtprint(f, "(%g,%g)", c.r, c.i);
}
.sp 0.3v
main(...)
{
Complex x = (Complex){ 1.5, -2.3 };
.sp 0.3v
fmtinstall('X', Xfmt);
print("x = %X\en", x);
}
.EE
.SH SOURCE
.B /sys/src/libc/fmt
.SH SEE ALSO
.IR print (2),
.IR utf (6),
.IR errstr (2)
.SH DIAGNOSTICS
These routines return negative numbers or nil for errors and set
.IR errstr .