As a namespace convention, everything (files, typedefs, functions,
macros) not defined in ISO/IEC 9899 is prefixed with _PDCLIB_*.
-As identifiers starting with '_' and a capital letter are reserved
-for the implementation, and the chances of you compiler using an
-identifier in the _PDCLIB_* range are slim, any strictly conforming
-application should work with PDCLib.
+The standard defines any identifiers starting with '_' and a capital
+letter as reserved for the implementation, and since the chances of
+your compiler using an identifier in the _PDCLIB_* range are slim,
+any strictly conforming application should work with PDCLib.
PDCLib consists of several parts:
3) internal header files keeping complex stuff out of the standard
headers;
4) the central, platform-specific file _PDCLIB_config.h;
-5) optimization overlay implementation files (optional).
-
-The standard headers only contain what they are defined to contain.
-Where additional logic or macro magic is necessary, that is deferred
-to the internal files. This has been done so that the headers are
-actually educational as to what they provide (as opposed to how the
-library does it).
-
-There is a seperate implementation file for every function defined
-by the standard, named {function}.c. Not only does this avoid linking
-in huge amounts of unused code when you use but a single function,
-it also allows the optimization overlay to work (see below).
-
-Then there are internal header files, which contain all the "black
-magic" and "code fu" that were kept out of the standard headers. You
-should not have to touch them if you want to adapt PDCLib to a new
-platform. If you do, note that the PDCLib author would consider it
-a serious design flaw, and would be happy to fix it in the next PDCLib
-release. Any adaption work should be covered by the config header
-(and, possibly, the optimization overlay).
+5) platform-specific implementation files;
+6) platform-specific, optimized "overlay" implementations (optional).
+
+The standard headers (in ./includes/) only contain what they are
+defined to contain. Where additional logic or macro magic is
+necessary, that is deferred to the internal files. This has been done
+so that the headers are actually educational as to what they provide
+(as opposed to how the library does it).
+
+Note that there *might* be some feature to remove this additional
+level of indirection for a production release, to ease the workload
+put on the preprocessor.
+
+There is a seperate implementation file (in ./function/{header}/) for
+every function defined by the standard, named {function}.c. Not only
+does this avoid linking in huge amounts of unused code when you use
+but a single function, it also allows the optimization overlay to work
+(see below).
+
+(The directory ./functions/_PDCLIB/ contains internal and helper
+functions that are not part of the standard.)
+
+Then there are internal header files (in ./internal/), which contain
+all the "black magic" and "code fu" that was kept out of the standard
+headers. You should not have to touch them if you want to adapt PDCLib
+to a new platform. If you *do* have to touch them, note that the PDCLib
+author would consider it a serious design flaw, and would be happy to
+fix it in the next PDCLib release. Any adaption work should be covered
+by the steps detailed below.
For adapting PDCLib to a new platform (the trinity of CPU, operating
-system, and compiler), open _PDCLIB_config.h in your favourite text
-editor, have a look at the comments, and modify it as appropriate for
-your platform. That should be all that is actually required for such
-an adaption (see previous paragraph).
+system, and compiler), make a copy of the ./platform/example/ directory
+(named after your target platform), and modify the files of your copy
+to suit the constraints of your platform. When you are done, copy the
+contents of your platform directory over the source directory structure
+of PDCLib (or link them into the appropriate places). That should be
+all that is actually required to make PDCLib work for your platform.
Of course, your platform might provide more efficient replacements
for the generic implementations offered by PDCLib. The math functions
crafted assembler or compiler build-ins. So the author wanted to
provide a means to modify PDCLib to run more efficiently on a given
platform, without cluttering the main branch with tons of #ifdef
-statements and "featureset defines" that grow stale quickly.
+statements and "featureset #defines" that grow stale quickly.
The solution is the "optimization overlay". Every function has its
-own implementation file, and _PDCLIB_config.h should be the only
-header that must be modified. So, a platform-specific overlay is
-copied over the main PDCLib branch - replacing _PDCLIB_config.h and
-any number of implementation files - to create a PDCLib adapted /
-optimized for the platform in question. That overlay could be part
-of the PDCLib source tree (for established platforms where maintainers
-won't bother with PDCLib), or part of that platform's source tree
-(for under-development platforms PDCLib maintainers won't bother with).
+own implementation file, which makes it possible to replace them
+piecemeal by copying a platform-specific overlay over the main PDCLib
+branch to create a PDCLib adapted / optimized for the platform in
+question. That overlay could be part of the PDCLib source tree (for
+established platforms where maintainers won't bother with PDCLib), or
+part of that platform's source tree (for under-development platforms
+PDCLib maintainers won't bother with).
So, to use PDCLib on your given platform, you unpack PDCLib (as you
obviously have done already since you are reading this), and copy
point conversions, and the wide-/multibyte-character functions.
v0.5 - unreleased
-Implementations for parts of <stdio.h>. Bug fixes. Still no locale /
-wide-char support. Enabled all GCC compiler warnings I could find, and
-fixed everything that threw a warning. Fixed all open bugs in the v0.4
-release.
-
+Implementations for parts of <stdio.h>. Still no locale / wide-char
+support. Enabled all GCC compiler warnings I could find, and fixed
+everything that threw a warning. (You see this, maintainers of Open
+Source software? No warnings whatsoever. Stop telling me it cannot
+be done.) Fixed all known bugs in the v0.4 release.
+++ /dev/null
-#include <stdarg.h>
-#include <stddef.h>
-#include <stdint.h>
-#include <stdio.h>
-#include <stdlib.h>
-
-/* These can be removed once integrated into PDCLIB make procedure */
-#undef TEST
-#include </home/solar/src/pdclib/functions/_PDCLIB/digits.c>
-#include </home/solar/src/pdclib/functions/_PDCLIB/Xdigits.c>
-
-/* Using an integer's bits as flags for both the conversion flags and length
- modifiers.
-*/
-#define E_minus 1<<0
-#define E_plus 1<<1
-#define E_alt 1<<2
-#define E_space 1<<3
-#define E_zero 1<<4
-#define E_done 1<<5
-#define E_char 1<<6
-#define E_short 1<<7
-#define E_long 1<<8
-#define E_llong 1<<9
-#define E_intmax 1<<10
-#define E_size 1<<11
-#define E_ptrdiff 1<<12
-#define E_intptr 1<<13
-#define E_double 1<<14
-#define E_lower 1<<15
-#define E_unsigned 1<<16
-
-struct status_t
-{
- int base; /* base to which the value shall be converted */
- int_fast32_t flags; /* flags and length modifiers */
- size_t n; /* maximum number of characters to be written */
- size_t i; /* number of characters already written */
- size_t this; /* number of output chars in the current conversion */
- char * s; /* target buffer */
- size_t width; /* width of current field */
- size_t prec; /* precision of current field */
- FILE * stream;/* for to-stream output */
- va_list ap; /* the argument stack passed to the printf function */
-};
-
-const char * parse_out( const char * spec, struct status_t * status );
-inline void test( size_t n, const char * expect, ... );
-int _PDCLIB_vsnprintf( char * buffer, size_t n, const char * format, va_list ap );
-int _PDCLIB_snprintf( char * s, size_t n, const char * format, ... );
-
-/* The following only for testing. */
-#include <limits.h>
-#include <string.h>
-
-int main( void )
-{
- test( SIZE_MAX, "%hhd", CHAR_MIN );
- test( SIZE_MAX, "%hhd", CHAR_MAX );
- test( SIZE_MAX, "%hhd", 0 );
- test( SIZE_MAX, "%hd", SHRT_MIN );
- test( SIZE_MAX, "%hd", SHRT_MAX );
- test( SIZE_MAX, "%hd", 0 );
- test( SIZE_MAX, "%d", INT_MIN );
- test( SIZE_MAX, "%d", INT_MAX );
- test( SIZE_MAX, "%d", 0 );
- test( SIZE_MAX, "%ld", LONG_MIN );
- test( SIZE_MAX, "%ld", LONG_MAX );
- test( SIZE_MAX, "%ld", 0l );
- test( SIZE_MAX, "%lld", LLONG_MIN );
- test( SIZE_MAX, "%lld", LLONG_MAX );
- test( SIZE_MAX, "%lld", 0ll );
- test( SIZE_MAX, "%hhu", UCHAR_MAX );
- test( SIZE_MAX, "%hhu", (unsigned char)-1 );
- test( SIZE_MAX, "%hu", USHRT_MAX );
- test( SIZE_MAX, "%hu", (unsigned short)-1 );
- test( SIZE_MAX, "%u", UINT_MAX );
- test( SIZE_MAX, "%u", -1u );
- test( SIZE_MAX, "%lu", ULONG_MAX );
- test( SIZE_MAX, "%lu", -1ul );
- test( SIZE_MAX, "%llu", ULLONG_MAX );
- test( SIZE_MAX, "%llu", -1ull );
- test( SIZE_MAX, "%X", UINT_MAX );
- test( SIZE_MAX, "%#X", -1u );
- test( SIZE_MAX, "%x", UINT_MAX );
- test( SIZE_MAX, "%#x", -1u );
- test( SIZE_MAX, "%o", UINT_MAX );
- test( SIZE_MAX, "%#o", -1u );
- test( SIZE_MAX, "%.0#o", 0 );
- test( SIZE_MAX, "%+d", INT_MIN );
- test( SIZE_MAX, "%+d", INT_MAX );
- test( SIZE_MAX, "%+d", 0 );
- test( SIZE_MAX, "%+u", UINT_MAX );
- test( SIZE_MAX, "%+u", -1u );
- test( SIZE_MAX, "% d", INT_MIN );
- test( SIZE_MAX, "% d", INT_MAX );
- test( SIZE_MAX, "% d", 0 );
- test( SIZE_MAX, "% u", UINT_MAX );
- test( SIZE_MAX, "% u", -1u );
- test( SIZE_MAX, "%9d", INT_MIN );
- test( SIZE_MAX, "%9d", INT_MAX );
- test( SIZE_MAX, "%10d", INT_MIN );
- test( SIZE_MAX, "%10d", INT_MAX );
- test( SIZE_MAX, "%11d", INT_MIN );
- test( SIZE_MAX, "%11d", INT_MAX );
- test( SIZE_MAX, "%12d", INT_MIN );
- test( SIZE_MAX, "%12d", INT_MAX );
- test( SIZE_MAX, "%-9d", INT_MIN );
- test( SIZE_MAX, "%-9d", INT_MAX );
- test( SIZE_MAX, "%-10d", INT_MIN );
- test( SIZE_MAX, "%-10d", INT_MAX );
- test( SIZE_MAX, "%-11d", INT_MIN );
- test( SIZE_MAX, "%-11d", INT_MAX );
- test( SIZE_MAX, "%-12d", INT_MIN );
- test( SIZE_MAX, "%-12d", INT_MAX );
- test( SIZE_MAX, "%09d", INT_MIN );
- test( SIZE_MAX, "%09d", INT_MAX );
- test( SIZE_MAX, "%010d", INT_MIN );
- test( SIZE_MAX, "%010d", INT_MAX );
- test( SIZE_MAX, "%011d", INT_MIN );
- test( SIZE_MAX, "%011d", INT_MAX );
- test( SIZE_MAX, "%012d", INT_MIN );
- test( SIZE_MAX, "%012d", INT_MAX );
- test( SIZE_MAX, "%-09d", INT_MIN );
- test( SIZE_MAX, "%-09d", INT_MAX );
- test( SIZE_MAX, "%-010d", INT_MIN );
- test( SIZE_MAX, "%-010d", INT_MAX );
- test( SIZE_MAX, "%-011d", INT_MIN );
- test( SIZE_MAX, "%-011d", INT_MAX );
- test( SIZE_MAX, "%-012d", INT_MIN );
- test( SIZE_MAX, "%-012d", INT_MAX );
- test( 8, "%9d", INT_MAX );
- test( 8, "%9d", INT_MIN );
- test( 9, "%9d", INT_MAX );
- test( 9, "%9d", INT_MIN );
- test( 10, "%9d", INT_MAX );
- test( 10, "%9d", INT_MIN );
- test( 9, "%10d", INT_MAX );
- test( 9, "%10d", INT_MIN );
- test( 10, "%10d", INT_MAX );
- test( 10, "%10d", INT_MIN );
- test( 11, "%10d", INT_MAX );
- test( 11, "%10d", INT_MIN );
- test( 10, "%11d", INT_MAX );
- test( 10, "%11d", INT_MIN );
- test( 11, "%11d", INT_MAX );
- test( 11, "%11d", INT_MIN );
- test( 12, "%11d", INT_MAX );
- test( 12, "%11d", INT_MIN );
- test( 11, "%12d", INT_MAX );
- test( 11, "%12d", INT_MIN );
- test( 12, "%12d", INT_MAX );
- test( 12, "%12d", INT_MIN );
- test( 13, "%12d", INT_MAX );
- test( 13, "%12d", INT_MIN );
- test( SIZE_MAX, "%030.20d", INT_MAX );
- test( SIZE_MAX, "%.6x", UINT_MAX );
- test( SIZE_MAX, "%#6.3x", UINT_MAX );
- test( SIZE_MAX, "%#3.6x", UINT_MAX );
- test( SIZE_MAX, "%.6d", INT_MIN );
- test( SIZE_MAX, "%6.3d", INT_MIN );
- test( SIZE_MAX, "%3.6d", INT_MIN );
- test( SIZE_MAX, "%#0.6x", UINT_MAX );
- test( SIZE_MAX, "%#06.3x", UINT_MAX );
- test( SIZE_MAX, "%#03.6x", UINT_MAX );
- test( SIZE_MAX, "%#0.6d", INT_MAX );
- test( SIZE_MAX, "%#06.3d", INT_MAX );
- test( SIZE_MAX, "%#03.6d", INT_MAX );
- test( SIZE_MAX, "%#+.6d", INT_MAX );
- test( SIZE_MAX, "%#+6.3d", INT_MAX );
- test( SIZE_MAX, "%#+3.6d", INT_MAX );
- test( SIZE_MAX, "%+0.6d", INT_MAX );
- test( SIZE_MAX, "%+06.3d", INT_MAX );
- test( SIZE_MAX, "%+03.6d", INT_MAX );
- test( SIZE_MAX, "- %d", INT_MAX );
- test( SIZE_MAX, "- %d %% %d", INT_MAX, INT_MIN );
- test( SIZE_MAX, "%c", 'x' );
- test( SIZE_MAX, "%s", "abcdef" );
- test( SIZE_MAX, "%p", 0xdeadbeef );
- {
- char buffer[50];
- int val1, val2, val3, val4;
- snprintf( buffer, SIZE_MAX, "123456%n789%n", &val1, &val2 );
- _PDCLIB_snprintf( buffer, SIZE_MAX, "123456%n789%n", &val3, &val4 );
- if ( ( val1 != val3 ) || ( val2 != val4 ) )
- {
- printf( "Output %d/%d\nExpect %d/%d\n\n", val1, val2, val3, val4 );
- }
- }
- return 0;
-}
-
-/* This macro delivers a given character to either a memory buffer or a stream,
- depending on the contents of 'status' (struct status_t).
- x - the character to be delivered
- i - pointer to number of characters already delivered in this call
- n - pointer to maximum number of characters to be delivered in this call
- s - the buffer into which the character shall be delivered
-*/
-#define DELIVER( x ) do { if ( status->i < status->n ) { if ( status->stream != NULL ) putc( x, status->stream ); else status->s[status->i] = x; } ++(status->i); } while ( 0 )
-
-/* This function recursively converts a given integer value to a given base
- into a character string. Persistent information - like the number of digits
- parsed so far - is recorded in a struct status_t, which allows to avoid
- overwriting snprintf() limits, and enables the function to do the necessary
- padding / prefixing of the character string eventually printed.
-*/
-static void int2base( intmax_t value, struct status_t * status )
-{
- /* Registering the character being printed at the end of the function here
- already so it will be taken into account when the deepestmost recursion
- does the prefix / padding stuff.
- */
- ++(status->this);
- if ( ( value / status->base ) != 0 )
- {
- /* More digits to be done - recurse deeper */
- int2base( value / status->base, status );
- }
- else
- {
- /* We reached the last digit, the deepest point of our recursion, and
- only now know how long the number to be printed actually is. Now we
- have to do the sign, prefix, width, and precision padding stuff
- before printing the numbers while we resurface from the recursion.
- */
- /* At worst, we need two prefix characters (hex prefix). */
- char preface[3] = "\0";
- size_t preidx = 0;
- if ( ( status->flags & E_alt ) && ( status->base == 16 || status->base == 8 ) )
- {
- /* Octal / hexadecimal prefix for "%#" conversions */
- preface[ preidx++ ] = '0';
- if ( status->base == 16 )
- {
- preface[ preidx++ ] = ( status->flags & E_lower ) ? 'x' : 'X';
- }
- }
- if ( value < 0 )
- {
- /* Negative sign for negative values - at all times. */
- preface[ preidx++ ] = '-';
- }
- else if ( ! ( status->flags & E_unsigned ) )
- {
- /* plus sign / extra space are only for unsigned conversions */
- if ( status->flags & E_plus )
- {
- preface[ preidx++ ] = '+';
- }
- else if ( status->flags & E_space )
- {
- preface[ preidx++ ] = ' ';
- }
- }
- {
- size_t prec_pads = ( status->prec > status->this ) ? ( status->prec - status->this ) : 0;
- if ( ! ( status->flags & ( E_minus | E_zero ) ) )
- {
- /* Space padding is only done if no zero padding or left alignment
- is requested. Leave space for any prefixes determined above.
- */
- /* The number of characters to be printed, plus prefixes if any. */
- /* This line contained probably the most stupid, time-wasting bug
- I've ever perpetrated. Greetings to Samface, DevL, and all
- sceners at Breakpoint 2006.
- */
- size_t characters = preidx + ( ( status->this > status->prec ) ? status->this : status->prec );
- if ( status->width > characters )
- {
- for ( int i = 0; i < status->width - characters; ++i )
- {
- DELIVER( ' ' );
- ++(status->this);
- }
- }
- }
- /* Now we did the padding, do the prefixes (if any). */
- preidx = 0;
- while ( preface[ preidx ] != '\0' )
- {
- DELIVER( preface[ preidx++ ] );
- ++(status->this);
- }
- if ( ( ! ( status->flags & E_minus ) ) && ( status->flags & E_zero ) )
- {
- /* If field is not left aligned, and zero padding is requested, do
- so.
- */
- while ( status->this < status->width )
- {
- DELIVER( '0' );
- ++(status->this);
- }
- }
- /* Do the precision padding if necessary. */
- for ( int i = 0; i < prec_pads; ++i )
- {
- DELIVER( '0' );
- }
- }
- }
- /* Recursion tail - print the current digit. */
- {
- int digit = value % status->base;
- if ( digit < 0 )
- {
- digit *= -1;
- }
- if ( status->flags & E_lower )
- {
- /* Lowercase letters. Same array used for strto...(). */
- DELIVER( _PDCLIB_digits[ digit ] );
- }
- else
- {
- /* Uppercase letters. Array only used here, only 0-F. */
- DELIVER( _PDCLIB_Xdigits[ digit ] );
- }
- }
-}
-
-/* This function is to be called with spec pointing to the leading '%' of a
- printf() conversion specifier, with ap being
-*/
-const char * parse_out( const char * spec, struct status_t * status )
-{
- const char * orig_spec = spec;
- if ( *(++spec) == '%' )
- {
- DELIVER( *spec );
- return ++spec;
- }
- /* Initializing status structure */
- status->flags = 0;
- status->base = 0;
- status->this = 0;
- status->width = 0;
- status->prec = 0;
-
- /* First come 0..n flags */
- do
- {
- switch ( *spec )
- {
- case '-':
- status->flags |= E_minus;
- ++spec;
- break;
- case '+':
- status->flags |= E_plus;
- ++spec;
- break;
- case '#':
- status->flags |= E_alt;
- ++spec;
- break;
- case ' ':
- status->flags |= E_space;
- ++spec;
- break;
- case '0':
- status->flags |= E_zero;
- ++spec;
- break;
- default:
- status->flags |= E_done;
- break;
- }
- } while ( ! ( status->flags & E_done ) );
-
- /* Optional field width */
- if ( *spec == '*' )
- {
- /* Retrieve width value from argument stack */
- if ( ( status->width = va_arg( status->ap, int ) ) < 0 )
- {
- /* Negative value is '-' flag plus absolute value */
- status->flags |= E_minus;
- status->width *= -1;
- }
- ++spec;
- }
- else
- {
- /* If a width is given, strtol() will return its value. If not given,
- strtol() will return zero. In both cases, endptr will point to the
- rest of the conversion specifier - just what we need.
- */
- status->width = (int)strtol( spec, (char**)&spec, 10 );
- }
-
- /* Optional precision */
- if ( *spec == '.' )
- {
- ++spec;
- if ( *spec == '*' )
- {
- /* Retrieve precision value from argument stack. A negative value
- is as if no precision is given - as precision is initalized to
- EOF (negative), there is no need for testing for negative here.
- */
- status->prec = va_arg( status->ap, int );
- }
- else
- {
- char * endptr;
- status->prec = (int)strtol( spec, &endptr, 10 );
- if ( spec == endptr )
- {
- /* Decimal point but no number - bad conversion specifier. */
- return orig_spec;
- }
- spec = endptr;
- }
- /* Having a precision cancels out any zero flag. */
- status->flags ^= E_zero;
- }
-
- /* Optional length modifier
- We step one character ahead in any case, and step back only if we find
- there has been no length modifier (or step ahead another character if it
- has been "hh" or "ll").
- */
- switch ( *(spec++) )
- {
- case 'h':
- if ( *spec == 'h' )
- {
- status->flags |= E_char;
- ++spec;
- }
- else
- {
- status->flags |= E_short;
- }
- break;
- case 'l':
- if ( *spec == 'l' )
- {
- status->flags |= E_llong;
- ++spec;
- }
- else
- {
- status->flags |= E_long;
- }
- break;
- case 'j':
- status->flags |= E_intmax;
- break;
- case 'z':
- status->flags |= E_size;
- break;
- case 't':
- status->flags |= E_ptrdiff;
- break;
- case 'L':
- status->flags |= E_double;
- break;
- default:
- --spec;
- break;
- }
-
- /* Conversion specifier */
- switch ( *spec )
- {
- case 'd':
- /* FALLTHROUGH */
- case 'i':
- status->base = 10;
- break;
- case 'o':
- status->base = 8;
- status->flags |= E_unsigned;
- break;
- case 'u':
- status->base = 10;
- status->flags |= E_unsigned;
- break;
- case 'x':
- status->base = 16;
- status->flags |= ( E_lower | E_unsigned );
- break;
- case 'X':
- status->base = 16;
- status->flags |= E_unsigned;
- break;
- case 'f':
- case 'F':
- case 'e':
- case 'E':
- case 'g':
- case 'G':
- break;
- case 'a':
- case 'A':
- break;
- case 'c':
- /* TODO: Flags, wide chars. */
- DELIVER( va_arg( status->ap, int ) );
- return ++spec;
- case 's':
- /* TODO: Flags, wide chars. */
- {
- char * s = va_arg( status->ap, char * );
- while ( *s != '\0' )
- {
- DELIVER( *(s++) );
- }
- return ++spec;
- }
- case 'p':
- /* TODO: E_long -> E_intptr */
- status->base = 16;
- status->flags |= ( E_lower | E_unsigned | E_alt | E_long );
- break;
- case 'n':
- {
- int * val = va_arg( status->ap, int * );
- *val = status->i;
- return ++spec;
- }
- default:
- /* No conversion specifier. Bad conversion. */
- return orig_spec;
- }
-
- /* Do the actual output based on our findings */
- if ( status->base != 0 )
- {
- /* Integer conversions */
- /* TODO: Check for invalid flag combinations. */
- if ( status->flags & E_unsigned )
- {
- uintmax_t value;
- switch ( status->flags & ( E_char | E_short | E_long | E_llong | E_size ) )
- {
- case E_char:
- value = (uintmax_t)(unsigned char)va_arg( status->ap, int );
- break;
- case E_short:
- value = (uintmax_t)(unsigned short)va_arg( status->ap, int );
- break;
- case 0:
- value = (uintmax_t)va_arg( status->ap, unsigned int );
- break;
- case E_long:
- value = (uintmax_t)va_arg( status->ap, unsigned long );
- break;
- case E_llong:
- value = (uintmax_t)va_arg( status->ap, unsigned long long );
- break;
- case E_size:
- value = (uintmax_t)va_arg( status->ap, size_t );
- break;
- }
- ++(status->this);
- if ( ( value / status->base ) != 0 )
- {
- int2base( (intmax_t)(value / status->base), status );
- }
- int digit = value % status->base;
- if ( digit < 0 )
- {
- digit *= -1;
- }
- if ( status->flags & E_lower )
- {
- DELIVER( _PDCLIB_digits[ digit ] );
- }
- else
- {
- DELIVER( _PDCLIB_Xdigits[ digit ] );
- }
- }
- else
- {
- switch ( status->flags & ( E_char | E_short | E_long | E_llong | E_intmax ) )
- {
- case E_char:
- int2base( (intmax_t)(char)va_arg( status->ap, int ), status );
- break;
- case E_short:
- int2base( (intmax_t)(short)va_arg( status->ap, int ), status );
- break;
- case 0:
- int2base( (intmax_t)va_arg( status->ap, int ), status );
- break;
- case E_long:
- int2base( (intmax_t)va_arg( status->ap, long ), status );
- break;
- case E_llong:
- int2base( (intmax_t)va_arg( status->ap, long long ), status );
- break;
- case E_ptrdiff:
- int2base( (intmax_t)va_arg( status->ap, ptrdiff_t ), status );
- break;
- case E_intmax:
- int2base( va_arg( status->ap, intmax_t ), status );
- break;
- }
- }
- if ( status->flags & E_minus )
- {
- while ( status->this < status->width )
- {
- DELIVER( ' ' );
- ++(status->this);
- }
- }
- if ( status->i >= status->n )
- {
- status->s[status->n - 1] = '\0';
- }
- }
- return ++spec;
-}
-
-inline void test( size_t n, const char * expect, ... )
-{
- char * buffer1 = malloc( 50 );
- char * buffer2 = malloc( 50 );
- int myrc;
- int rc;
- va_list ap;
- va_start( ap, expect );
- myrc = _PDCLIB_vsnprintf( buffer1, n, expect, ap );
- rc = vsnprintf( buffer2, n, expect, ap );
- if ( ( strcmp( buffer1, buffer2 ) != 0 ) || ( myrc != rc ) )
- {
- printf( "Output '%s', RC %d\nExpect '%s', RC %d\n\n", buffer1, myrc, buffer2, rc );
- }
- free( buffer1 );
- free( buffer2 );
-}
-
-int _PDCLIB_vsnprintf( char * buffer, size_t n, const char * format, va_list ap )
-{
- struct status_t status = { 0, 0, n, 0, 0, buffer, 0, 0, NULL, ap };
- while ( *format != '\0' )
- {
- const char * rc;
- if ( ( *format != '%' ) || ( ( rc = parse_out( format, &status ) ) == format ) )
- {
- /* No conversion specifier, print verbatim */
- buffer[ status.i++ ] = *(format++);
- }
- else
- {
- /* Continue parsing after conversion specifier */
- format = rc;
- }
- }
- buffer[ status.i ] = '\0';
- return status.i;
-}
-
-int _PDCLIB_snprintf( char * s, size_t n, const char * format, ... )
-{
- va_list ap;
- va_start( ap, format );
- return _PDCLIB_vsnprintf( s, n, format, ap );
-}
-
-#if 0
-int _PDCLIB_fprintf( FILE * stream, const char * format, va_list ap )
-{
- struct status_t status = { 0, 0, SIZE_MAX, 0, 0, NULL, 0, 0, stream, ap };
- while ( *format != '\0' )
- {
- const char * rc;
- if ( ( *format != '%' ) || ( ( rc = parse_out( format, &status, ap ) ) == format ) )
- {
- /* No conversion specifier, print verbatim */
- putc( *(format++), stream );
- }
- else
- {
- /* Continue parsing after conversion specifier */
- format = rc;
- }
- }
- return status.i;
-}
-#endif