- Inkscape Devel - Inkscape Mailman

Formatting code
by Bob Jamison 13 Feb '04

13 Feb '04

Re: our conversation on Jabber about formatting data for the repr tree, here is some code I appropriated from Apache's APR lib, which can give us total control over formatting output strings. With Mental and Bryce's work in this area, these files might not be necessary, but then again they might be useful, so here they are. One very nice function for the repr tree would be: sp_repr_printf(repr, key, "format", args...) Would relieve the need for a lot of temporary buffers all over the place. Bob /*######################################################################## ## $Id$ ########################################################################*/ /* * Print formatter. This file's purpose is to * allow Inkscape developers to have complete * control over formatted strings. This can be * tweaked to suit. * * This file was basically copied from Apache's * apr_snprintf.c, but with extensive modifications * to ease embedding in a GLib project. Their license * is below. * * Copyright (C) 1999-2003 Authors * * Released under GNU GPL, read the file 'COPYING' for more information * */ /* ==================================================================== * The Apache Software License, Version 1.1 * * Copyright (c) 2000-2003 The Apache Software Foundation. All rights * reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in * the documentation and/or other materials provided with the * distribution. * * 3. The end-user documentation included with the redistribution, * if any, must include the following acknowledgment: * "This product includes software developed by the * Apache Software Foundation (http://www.apache.org/)." * Alternately, this acknowledgment may appear in the software itself, * if and wherever such third-party acknowledgments normally appear. * * 4. The names "Apache" and "Apache Software Foundation" must * not be used to endorse or promote products derived from this * software without prior written permission. For written * permission, please contact apache@...199... * * 5. Products derived from this software may not be called "Apache", * nor may "Apache" appear in their name, without prior written * permission of the Apache Software Foundation. * * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED OR IMPLIED * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE * DISCLAIMED. IN NO EVENT SHALL THE APACHE SOFTWARE FOUNDATION OR * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * ==================================================================== * * This software consists of voluntary contributions made by many * individuals on behalf of the Apache Software Foundation. For more * information on the Apache Software Foundation, please see * <http://www.apache.org/>. * * The ink_vsnprintf/ink_snprintf functions are based on, and used with the * permission of, the SIO stdio-replacement strx_* functions by Panos * Tsirigotis <panos@...200...> for xinetd. */ #include <math.h> #include <ctype.h> #include <string.h> #include <glib.h> #include "ink-snprintf.h" /*#################################### # D E F I N I T I O N S ####################################*/ /** * Structure used by the variable-formatter routines. */ struct ink_vformatter_buff_t { gchar *curpos;/** The current position */ gchar *endpos;/** The end position of the format string */ }; typedef struct ink_vformatter_buff_t ink_vformatter_buff_t; typedef enum { NO = 0, YES = 1 } boolean_e; #ifndef FALSE #define FALSE 0 #endif #ifndef TRUE #define TRUE 1 #endif #define NUL '\0' #define S_NULL "(null)" #define S_NULL_LEN 6 #define FLOAT_DIGITS 6 #define EXPONENT_LENGTH 10 /* * NUM_BUF_SIZE is the size of the buffer used for arithmetic conversions * * NOTICE: this is a magic number; do not decrease it */ #define NUM_BUF_SIZE 512 /* * cvt.c - IEEE floating point formatting routines for FreeBSD * from GNU libc-4.6.27. Modified to be thread safe. */ /* * ink_ecvt converts to decimal * the number of digits is specified by ndigit * decpt is set to the position of the decimal point * sign is set to 0 for positive, 1 for negative */ #define NDIG 80 /*#################################### # M A C R O S ####################################*/ /* * The INS_CHAR macro inserts a character in the buffer and writes * the buffer back to disk if necessary * It uses the char pointers sp and bep: * sp points to the next available character in the buffer * bep points to the end-of-buffer+1 * While using this macro, note that the nextb pointer is NOT updated. * * NOTE: Evaluation of the c argument should not have any side-effects */ #define INS_CHAR(c, sp, bep, cc) \ { \ if (sp) { \ if (sp >= bep) { \ vbuff->curpos = sp; \ if (flush_func(vbuff)) \ return -1; \ sp = vbuff->curpos; \ bep = vbuff->endpos; \ } \ *sp++ = (c); \ } \ cc++; \ } #define NUM(c) (c - '0') #define STR_TO_DEC(str, num) \ num = NUM(*str++); \ while (g_ascii_isdigit(*str)) \ { \ num *= 10 ; \ num += NUM(*str++); \ } /* * This macro does zero padding so that the precision * requirement is satisfied. The padding is done by * adding '0's to the left of the string that is going * to be printed. We don't allow precision to be large * enough that we continue past the start of s. * * NOTE: this makes use of the magic info that s is * always based on num_buf with a size of NUM_BUF_SIZE. */ #define FIX_PRECISION(adjust, precision, s, s_len) \ if (adjust) { \ gint p = precision < NUM_BUF_SIZE - 1 ? \ precision : NUM_BUF_SIZE - 1; \ while (s_len < p) { \ *--s = '0'; \ s_len++; \ } \ } /* * Macro that does padding. The padding is done by printing * the character ch. */ #define PAD(width, len, ch) \ do \ { \ INS_CHAR(ch, sp, bep, cc); \ width--; \ } \ while (width > len) /* * Prefix the character ch to the string str * Increase length * Set the has_prefix flag */ #define PREFIX(str, length, ch) \ *--str = ch; \ length++; \ has_prefix=YES; /*#################################### # CVT - float-to-ascii conversion ####################################*/ /* buf must have at least NDIG bytes */ static gchar * ink_cvt(gdouble arg, gint ndigits, gint *decpt, gint *sign, gint eflag, gchar *buf) { gint r2; gdouble fi, fj; gchar *p, *p1; if (ndigits >= NDIG - 1) ndigits = NDIG - 2; r2 = 0; *sign = 0; p = &buf[0]; if (arg < 0) { *sign = 1; arg = -arg; } arg = modf(arg, &fi); p1 = &buf[NDIG]; /* * Do integer part */ if (fi != 0) { p1 = &buf[NDIG]; while (p1 > &buf[0] && fi != 0) { fj = modf(fi / 10, &fi); *--p1 = (int) ((fj + .03) * 10) + '0'; r2++; } while (p1 < &buf[NDIG]) *p++ = *p1++; } else if (arg > 0) { while ((fj = arg * 10) < 1) { arg = fj; r2--; } } p1 = &buf[ndigits]; if (eflag == 0) p1 += r2; *decpt = r2; if (p1 < &buf[0]) { buf[0] = '\0'; return (buf); } while (p <= p1 && p < &buf[NDIG]) { arg *= 10; arg = modf(arg, &fj); *p++ = (int) fj + '0'; } if (p1 >= &buf[NDIG]) { buf[NDIG - 1] = '\0'; return (buf); } p = p1; *p1 += 5; while (*p1 > '9') { *p1 = '0'; if (p1 > buf) ++ * --p1; else { *p1 = '1'; (*decpt)++; if (eflag == 0) { if (p > buf) *p = '0'; p++; } } } *p = '\0'; return (buf); } static gchar * ink_ecvt(gdouble arg, gint ndigits, gint *decpt, gint *sign, gchar *buf) { return (ink_cvt(arg, ndigits, decpt, sign, 1, buf)); } static gchar * ink_fcvt(gdouble arg, gint ndigits, gint *decpt, gint *sign, gchar *buf) { return (ink_cvt(arg, ndigits, decpt, sign, 0, buf)); } /* * ink_gcvt - Floating output conversion to * minimal length string */ static gchar * ink_gcvt(gdouble number, gint ndigit, gchar *buf, boolean_e altform) { gint sign, decpt; gchar *p1, *p2; gint i; gchar buf1[NDIG]; p1 = ink_ecvt(number, ndigit, &decpt, &sign, buf1); p2 = buf; if (sign) *p2++ = '-'; for (i = ndigit - 1; i > 0 && p1[i] == '0'; i--) ndigit--; if ((decpt >= 0 && decpt - ndigit > 4) || (decpt < 0 && decpt < -3)) { /* use E-style */ decpt--; *p2++ = *p1++; *p2++ = '.'; for (i = 1; i < ndigit; i++) *p2++ = *p1++; *p2++ = 'e'; if (decpt < 0) { decpt = -decpt; *p2++ = '-'; } else *p2++ = '+'; if (decpt / 100 > 0) *p2++ = decpt / 100 + '0'; if (decpt / 10 > 0) *p2++ = (decpt % 100) / 10 + '0'; *p2++ = decpt % 10 + '0'; } else { if (decpt <= 0) { if (*p1 != '0') *p2++ = '.'; while (decpt < 0) { decpt++; *p2++ = '0'; } } for (i = 1; i <= ndigit; i++) { *p2++ = *p1++; if (i == decpt) *p2++ = '.'; } if (ndigit < decpt) { while (ndigit++ < decpt) *p2++ = '0'; *p2++ = '.'; } } if (p2[-1] == '.' && !altform) p2--; *p2 = '\0'; return (buf); } /* * Convert num to its decimal format. * Return value: * - a pointer to a string containing the number (no sign) * - len contains the length of the string * - is_negative is set to TRUE or FALSE depending on the sign * of the number (always set to FALSE if is_unsigned is TRUE) * * The caller provides a buffer for the string: that is the buf_end argument * which is a pointer to the END of the buffer + 1 (i.e. if the buffer * is declared as buf[ 100 ], buf_end should be &buf[ 100 ]) * * Note: we have 2 versions. One is used when we need to use quads * (conv_10_quad), the other when we don't (conv_10). We're assuming the * latter is faster. */ static gchar * conv_10(gint32 num, gboolean is_unsigned, gboolean *is_negative, gchar *buf_end, gint *len) { gchar *p = buf_end; guint32 magnitude; if (is_unsigned) { magnitude = (guint32) num; *is_negative = FALSE; } else { *is_negative = (num < 0); /* * On a 2's complement machine, negating the most negative integer * results in a number that cannot be represented as a signed integer. * Here is what we do to obtain the number's magnitude: * a. add 1 to the number * b. negate it (becomes positive) * c. convert it to unsigned * d. add 1 */ if (*is_negative) { gint32 t = num + 1; magnitude = ((guint32) -t) + 1; } else magnitude = (guint32) num; } /* * We use a do-while loop so that we write at least 1 digit */ do { guint32 new_magnitude = magnitude / 10; *--p = (gchar) (magnitude - new_magnitude * 10 + '0'); magnitude = new_magnitude; } while (magnitude); *len = buf_end - p; return (p); } static gchar * conv_10_quad(gint64 num, gboolean is_unsigned, gboolean *is_negative, gchar *buf_end, gint *len) { gchar *p = buf_end; guint64 magnitude; /* * We see if we can use the faster non-quad version by checking the * number against the largest long value it can be. If <=, we * punt to the quicker version. */ if ( (num <= ULONG_MAX && is_unsigned) || (num <= LONG_MAX && !is_unsigned) ) return(conv_10( (gint32)num, is_unsigned, is_negative, buf_end, len)); if (is_unsigned) { magnitude = (guint64) num; *is_negative = FALSE; } else { *is_negative = (num < 0); /* * On a 2's complement machine, negating the most negative integer * results in a number that cannot be represented as a signed integer. * Here is what we do to obtain the number's magnitude: * a. add 1 to the number * b. negate it (becomes positive) * c. convert it to unsigned * d. add 1 */ if (*is_negative) { gint64 t = num + 1; magnitude = ((guint64) -t) + 1; } else magnitude = (guint64) num; } /* * We use a do-while loop so that we write at least 1 digit */ do { guint64 new_magnitude = magnitude / 10; *--p = (gchar) (magnitude - new_magnitude * 10 + '0'); magnitude = new_magnitude; } while (magnitude); *len = buf_end - p; return (p); } /* * Convert a floating point number to a string formats 'f', 'e' or 'E'. * The result is placed in buf, and len denotes the length of the string * The sign is returned in the is_negative argument (and is not placed * in buf). */ static gchar * conv_fp(gchar format, gdouble num, boolean_e add_dp, gint precision, gboolean *is_negative, gchar *buf, gint *len) { gchar *s = buf; gchar *p; gint decimal_point; gchar buf1[NDIG]; if (format == 'f') p = ink_fcvt(num, precision, &decimal_point, is_negative, buf1); else /* either e or E format */ p = ink_ecvt(num, precision + 1, &decimal_point, is_negative, buf1); /* * Check for Infinity and NaN */ if (g_ascii_isalpha(*p)) { *len = strlen(p); memcpy(buf, p, *len + 1); *is_negative = FALSE; return (buf); } if (format == 'f') { if (decimal_point <= 0) { *s++ = '0'; if (precision > 0) { *s++ = '.'; while (decimal_point++ < 0) *s++ = '0'; } else if (add_dp) *s++ = '.'; } else { while (decimal_point-- > 0) *s++ = *p++; if (precision > 0 || add_dp) *s++ = '.'; } } else { *s++ = *p++; if (precision > 0 || add_dp) *s++ = '.'; } /* * copy the rest of p, the NUL is NOT copied */ while (*p) *s++ = *p++; if (format != 'f') { gchar temp[EXPONENT_LENGTH]; /* for exponent conversion */ gint t_len; gboolean exponent_is_negative; *s++ = format; /* either e or E */ decimal_point--; if (decimal_point != 0) { p = conv_10((gint32) decimal_point, FALSE, &exponent_is_negative, &temp[EXPONENT_LENGTH], &t_len); *s++ = exponent_is_negative ? '-' : '+'; /* * Make sure the exponent has at least 2 digits */ if (t_len == 1) *s++ = '0'; while (t_len--) *s++ = *p++; } else { *s++ = '+'; *s++ = '0'; *s++ = '0'; } } *len = s - buf; return (buf); } /* * Convert num to a base X number where X is a power of 2. nbits determines X. * For example, if nbits is 3, we do base 8 conversion * Return value: * a pointer to a string containing the number * * The caller provides a buffer for the string: that is the buf_end argument * which is a pointer to the END of the buffer + 1 (i.e. if the buffer * is declared as buf[ 100 ], buf_end should be &buf[ 100 ]) * * As with conv_10, we have a faster version which is used when * the number isn't quad size. */ static gchar * conv_p2(guint32 num, gint nbits, gchar format, gchar *buf_end, gint *len) { gint mask = (1 << nbits) - 1; gchar *p = buf_end; const gchar low_digits[] = "0123456789abcdef"; const gchar upper_digits[] = "0123456789ABCDEF"; const gchar *digits = (format == 'X') ? upper_digits : low_digits; do { *--p = digits[num & mask]; num >>= nbits; } while (num); *len = buf_end - p; return (p); } static gchar * conv_p2_quad(guint64 num, gint nbits, gchar format, gchar *buf_end, gint *len) { gint mask = (1 << nbits) - 1; gchar *p = buf_end; const gchar low_digits[] = "0123456789abcdef"; const gchar upper_digits[] = "0123456789ABCDEF"; const gchar *digits = (format == 'X') ? upper_digits : low_digits; if (num <= ULONG_MAX) return(conv_p2((guint32)num, nbits, format, buf_end, len)); do { *--p = digits[num & mask]; num >>= nbits; } while (num); *len = buf_end - p; return (p); } /*#################################### # F O R M A T T I N G ####################################*/ /* * Do format conversion placing the output in buffer */ static gint ink_vformatter(gint (*flush_func)(ink_vformatter_buff_t *), ink_vformatter_buff_t *vbuff, const gchar *fmt, va_list ap) { gchar *sp; gchar *bep; gint cc = 0; gint i; gchar *s = NULL; gchar *q; gint s_len; gint min_width = 0; gint precision = 0; enum { LEFT, RIGHT } adjust; gchar pad_char; gchar prefix_char; gdouble fp_num; gint64 i_quad = (gint64) 0; guint64 ui_quad; gint32 i_num = (gint32) 0; guint32 ui_num; gchar num_buf[NUM_BUF_SIZE]; gchar char_buf[2]; /* for printing %% and %<unknown> */ enum var_type_enum { IS_QUAD, IS_LONG, IS_SHORT, IS_INT }; enum var_type_enum var_type = IS_INT; /* * Flag variables */ boolean_e alternate_form; boolean_e print_sign; boolean_e print_blank; boolean_e adjust_precision; boolean_e adjust_width; gboolean is_negative; sp = vbuff->curpos; bep = vbuff->endpos; while (*fmt) { if (*fmt != '%') { INS_CHAR(*fmt, sp, bep, cc); } else { /* * Default variable settings */ boolean_e print_something = YES; adjust = RIGHT; alternate_form = print_sign = print_blank = NO; pad_char = ' '; prefix_char = NUL; fmt++; /* * Try to avoid checking for flags, width or precision */ if (!g_ascii_islower(*fmt)) { /* * Recognize flags: -, #, BLANK, + */ for (;; fmt++) { if (*fmt == '-') adjust = LEFT; else if (*fmt == '+') print_sign = YES; else if (*fmt == '#') alternate_form = YES; else if (*fmt == ' ') print_blank = YES; else if (*fmt == '0') pad_char = '0'; else break; } /* * Check if a width was specified */ if (g_ascii_isdigit(*fmt)) { STR_TO_DEC(fmt, min_width); adjust_width = YES; } else if (*fmt == '*') { min_width = va_arg(ap, int); fmt++; adjust_width = YES; if (min_width < 0) { adjust = LEFT; min_width = -min_width; } } else adjust_width = NO; /* * Check if a precision was specified */ if (*fmt == '.') { adjust_precision = YES; fmt++; if ( g_ascii_isdigit (*fmt)) { STR_TO_DEC(fmt, precision); } else if (*fmt == '*') { precision = va_arg(ap, int); fmt++; if (precision < 0) precision = 0; } else precision = 0; } else adjust_precision = NO; } else adjust_precision = adjust_width = NO; /* * Modifier check */ if (*fmt == 'q') { var_type = IS_QUAD; fmt++; } else if (*fmt == 'l') { var_type = IS_LONG; fmt++; } else if (*fmt == 'h') { var_type = IS_SHORT; fmt++; } else { var_type = IS_INT; } /* * Argument extraction and printing. * First we determine the argument type. * Then, we convert the argument to a string. * On exit from the switch, s points to the string that * must be printed, s_len has the length of the string * The precision requirements, if any, are reflected in s_len. * * NOTE: pad_char may be set to '0' because of the 0 flag. * It is reset to ' ' by non-numeric formats */ switch (*fmt) { case 'u': if (var_type == IS_QUAD) { i_quad = va_arg(ap, guint64); s = conv_10_quad(i_quad, 1, &is_negative, &num_buf[NUM_BUF_SIZE], &s_len); } else { if (var_type == IS_LONG) i_num = (gint32) va_arg(ap, guint32); else if (var_type == IS_SHORT) i_num = (gint32) (unsigned short) va_arg(ap, unsigned int); else i_num = (gint32) va_arg(ap, unsigned int); s = conv_10(i_num, 1, &is_negative, &num_buf[NUM_BUF_SIZE], &s_len); } FIX_PRECISION(adjust_precision, precision, s, s_len); break; case 'd': case 'i': if (var_type == IS_QUAD) { i_quad = va_arg(ap, gint64); s = conv_10_quad(i_quad, 0, &is_negative, &num_buf[NUM_BUF_SIZE], &s_len); } else { if (var_type == IS_LONG) i_num = (gint32) va_arg(ap, gint32); else if (var_type == IS_SHORT) i_num = (gint32) (short) va_arg(ap, int); else i_num = (gint32) va_arg(ap, int); s = conv_10(i_num, 0, &is_negative, &num_buf[NUM_BUF_SIZE], &s_len); } FIX_PRECISION(adjust_precision, precision, s, s_len); if (is_negative) prefix_char = '-'; else if (print_sign) prefix_char = '+'; else if (print_blank) prefix_char = ' '; break; case 'o': if (var_type == IS_QUAD) { ui_quad = va_arg(ap, guint64); s = conv_p2_quad(ui_quad, 3, *fmt, &num_buf[NUM_BUF_SIZE], &s_len); } else { if (var_type == IS_LONG) ui_num = (guint32) va_arg(ap, guint32); else if (var_type == IS_SHORT) ui_num = (guint32) (unsigned short) va_arg(ap, unsigned int); else ui_num = (guint32) va_arg(ap, unsigned int); s = conv_p2(ui_num, 3, *fmt, &num_buf[NUM_BUF_SIZE], &s_len); } FIX_PRECISION(adjust_precision, precision, s, s_len); if (alternate_form && *s != '0') { *--s = '0'; s_len++; } break; case 'x': case 'X': if (var_type == IS_QUAD) { ui_quad = va_arg(ap, guint64); s = conv_p2_quad(ui_quad, 4, *fmt, &num_buf[NUM_BUF_SIZE], &s_len); } else { if (var_type == IS_LONG) ui_num = (guint32) va_arg(ap, guint32); else if (var_type == IS_SHORT) ui_num = (guint32) (unsigned short) va_arg(ap, unsigned int); else ui_num = (guint32) va_arg(ap, unsigned int); s = conv_p2(ui_num, 4, *fmt, &num_buf[NUM_BUF_SIZE], &s_len); } FIX_PRECISION(adjust_precision, precision, s, s_len); if (alternate_form && i_num != 0) { *--s = *fmt; /* 'x' or 'X' */ *--s = '0'; s_len += 2; } break; case 's': s = va_arg(ap, gchar *); if (s != NULL) { if (!adjust_precision) { s_len = strlen(s); } else { /* From the C library standard in section 7.9.6.1: * ...if the precision is specified, no more then * that many characters are written. If the * precision is not specified or is greater * than the size of the array, the array shall * contain a null character. * * My reading is is precision is specified and * is less then or equal to the size of the * array, no null character is required. So * we can't do a strlen. * * This figures out the length of the string * up to the precision. Once it's long enough * for the specified precision, we don't care * anymore. * * NOTE: you must do the length comparison * before the check for the null character. * Otherwise, you'll check one beyond the * last valid character. */ const gchar *walk; for (walk = s, s_len = 0; (s_len < precision) && (*walk != '\0'); ++walk, ++s_len); } } else { s = S_NULL; s_len = S_NULL_LEN; } pad_char = ' '; break; case 'f': case 'e': case 'E': fp_num = va_arg(ap, double); /* * We use &num_buf[ 1 ], so that we have room for the sign */ s = NULL; #ifdef HAVE_ISNAN if (isnan(fp_num)) { s = "nan"; s_len = 3; } #endif #ifdef HAVE_ISINF if (!s && isinf(fp_num)) { s = "inf"; s_len = 3; } #endif if (!s) { s = conv_fp(*fmt, fp_num, alternate_form, (adjust_precision == NO) ? FLOAT_DIGITS : precision, &is_negative, &num_buf[1], &s_len); if (is_negative) prefix_char = '-'; else if (print_sign) prefix_char = '+'; else if (print_blank) prefix_char = ' '; } break; case 'g': case 'G': if (adjust_precision == NO) precision = FLOAT_DIGITS; else if (precision == 0) precision = 1; /* * * We use &num_buf[ 1 ], so that we have room for the sign */ s = ink_gcvt(va_arg(ap, double), precision, &num_buf[1], alternate_form); if (*s == '-') prefix_char = *s++; else if (print_sign) prefix_char = '+'; else if (print_blank) prefix_char = ' '; s_len = strlen(s); if (alternate_form && (q = strchr(s, '.')) == NULL) { s[s_len++] = '.'; s[s_len] = '\0'; /* delimit for following strchr() */ } if (*fmt == 'G' && (q = strchr(s, 'e')) != NULL) *q = 'E'; break; case 'c': char_buf[0] = (char) (va_arg(ap, int)); s = &char_buf[0]; s_len = 1; pad_char = ' '; break; case '%': char_buf[0] = '%'; s = &char_buf[0]; s_len = 1; pad_char = ' '; break; case 'n': if (var_type == IS_QUAD) *(va_arg(ap, gint64 *)) = cc; else if (var_type == IS_LONG) *(va_arg(ap, long *)) = cc; else if (var_type == IS_SHORT) *(va_arg(ap, short *)) = cc; else *(va_arg(ap, int *)) = cc; print_something = NO; break; /* * This is where we extend the printf format, with a second * type specifier */ case 'p': switch(*++fmt) { /* * If the pointer size is equal to or smaller than the size * of the largest unsigned int, we convert the pointer to a * hex number, otherwise we print "%p" to indicate that we * don't handle "%p". */ case 'p': #ifdef VOID_P_IS_QUAD if (sizeof(void *) <= sizeof(guint64)) { ui_quad = (guint64) va_arg(ap, void *); s = conv_p2_quad(ui_quad, 4, 'x', &num_buf[NUM_BUF_SIZE], &s_len); } #else if (sizeof(void *) <= sizeof(guint32)) { ui_num = (guint32) va_arg(ap, void *); s = conv_p2(ui_num, 4, 'x', &num_buf[NUM_BUF_SIZE], &s_len); } #endif else { s = "%p"; s_len = 2; prefix_char = NUL; } pad_char = ' '; break; case 'T': char_buf[0] = '0'; s = &char_buf[0]; s_len = 1; pad_char = ' '; break; case NUL: /* if %p ends the string, oh well ignore it */ continue; default: s = "bogus %p"; s_len = 8; prefix_char = NUL; (void)va_arg(ap, void *); /* skip the bogus argument on the stack */ break; } break; case NUL: /* * The last character of the format string was %. * We ignore it. */ continue; /* * The default case is for unrecognized %'s. * We print %<char> to help the user identify what * option is not understood. * This is also useful in case the user wants to pass * the output of format_converter to another function * that understands some other %<char> (like syslog). * Note that we can't point s inside fmt because the * unknown <char> could be preceded by width etc. */ default: char_buf[0] = '%'; char_buf[1] = *fmt; s = char_buf; s_len = 2; pad_char = ' '; break; } if (prefix_char != NUL && s != S_NULL && s != char_buf) { *--s = prefix_char; s_len++; } if (adjust_width && adjust == RIGHT && min_width > s_len) { if (pad_char == '0' && prefix_char != NUL) { INS_CHAR(*s, sp, bep, cc); s++; s_len--; min_width--; } PAD(min_width, s_len, pad_char); } /* * Print the string s. */ if (print_something == YES) { for (i = s_len; i != 0; i--) { INS_CHAR(*s, sp, bep, cc); s++; } } if (adjust_width && adjust == LEFT && min_width > s_len) PAD(min_width, s_len, pad_char); } fmt++; } vbuff->curpos = sp; return cc; } static int snprintf_flush(ink_vformatter_buff_t *vbuff) { /* if the buffer fills we have to abort immediately, there is no way * to "flush" an ink_snprintf... there's nowhere to flush it to. */ return -1; } gint ink_snprintf(gchar *buf, gint len, const gchar *format, ...) { gint cc; va_list ap; ink_vformatter_buff_t vbuff; if (len == 0) { /* NOTE: This is a special case; we just want to return the number * of chars that would be written (minus \0) if the buffer * size was infinite. We leverage the fact that INS_CHAR * just does actual inserts iff the buffer pointer is non-NULL. * In this case, we don't care what buf is; it can be NULL, since * we don't touch it at all. */ vbuff.curpos = NULL; vbuff.endpos = NULL; } else { /* save one byte for nul terminator */ vbuff.curpos = buf; vbuff.endpos = buf + len - 1; } va_start(ap, format); cc = ink_vformatter(snprintf_flush, &vbuff, format, ap); va_end(ap); if (len != 0) { *vbuff.curpos = '\0'; } return (cc == -1) ? (gint)len : cc; } gint ink_vsnprintf(gchar *buf, gint len, const gchar *format, va_list ap) { gint cc; ink_vformatter_buff_t vbuff; if (len == 0) { /* See above note */ vbuff.curpos = NULL; vbuff.endpos = NULL; } else { /* save one byte for nul terminator */ vbuff.curpos = buf; vbuff.endpos = buf + len - 1; } cc = ink_vformatter(snprintf_flush, &vbuff, format, ap); if (len != 0) { *vbuff.curpos = '\0'; } return (cc == -1) ? (gint)len : cc; } #ifndef __INK_SNPRINTF_H__ #define __INK_SNPRINTF_H__ /*######################################################################## ## $Id$ ########################################################################*/ /* * Print formatter. This file's purpose is to * allow Inkscape developers to have complete * control over formatted strings. This can be * tweaked to suit. * * Copyright (C) 1999-2003 Authors * * Released under GNU GPL, read the file 'COPYING' for more information */ #include <glib.h> #include <stdarg.h> /** * Create a formatted string with a given format string * and a standard variable argument list. */ gint ink_snprintf(gchar *buf, gint len, const gchar *format, ...); /** * Create a formatted string with a given format string * and argument list. */ gint ink_vsnprintf(gchar *buf, gint len, const gchar *format, va_list ap); #endif /* __INK_SNPRINTF_H__ */

1 0

Re: [Inkscape-devel] Some fairly random feature reqs + bug reports.
by bulia byak 13 Feb '04

13 Feb '04

>>>What happened in the Fill/stroke dialog? the entry >>>boxs for RGB values on the wheel tab are gone >> >> >>That's temporary, we plan to restore them, although not RGB - the wheel is >>HSV and it will have HSV fields, if you need RGB switch to the RGB >>sliders. > >Well, that I'm not sure of. Jon, I do not mean restore the old system picker, but to add HSV fields to your new picker, right above the alpha spinbutton. _________________________________________________________________ Help STOP SPAM with the new MSN 8 and get 2 months FREE* http://join.msn.com/?page=dept/bcomm&pgmarket=en-ca&RU=http%3a%2f%2fjoin.ms…

2 1

RE: [Inkscape-devel] Some fairly random feature reqs + bug reports.
by bulia byak 13 Feb '04

13 Feb '04

Thanks for the feedback! > First, if I've got a new session of inkscape, and i >do file>open, why does it open it into a new window >rather than the blank one I've got in front of me? >every time it does it it strikes me as odd. Yes, it's in the plans. Maybe next version. > Second up, have you considered making use of the >boolean ops to add paintbrush/eraser tools ala flash? >the functionality is all there just spread across >several tools currently. The way I've been simulating >it currently is to use the calligraphy tool to 'paint' >an area, then whenever you paint additional >areas immediately switch to select and ctrl++ it to >the last section. Not sure about usefulness of this. The vector approach is to keep things separate and at a high level of abstraction for as long as possible, and only merge/flatten/convert to paths as the last stage if possible. So if I need this functionality, I would actually prefer to have shapes separate to I can move and adjust them, and do a union or difference on them if necessary (but I think in many cases, just grouping them together would be convenient enough). On the other hand, what you propose may be simple enough to implement, but we need to do a paint tool with user-definable strokes first, otherwise the booleans are of little use IMHO. > Another feature that I'd love to see would be >variable offsets, ie a dynamic offset with >node weighting allowing the offset to be modified at >each node so I could generate a line of varying >thickness (like the calligraphy tool) but still >editable by the basic centreline path. This would be a >godsend for doing comic style drawings where the line >weight is important. I know that ones a big task, just >putting the idea out there. Yes, useful, but perhaps more uesful in the short term and easier to implement is a variable-width stroke with some predefined shapes. I think our developers had some prototypes of this. > With regards to the 0.37 release, the builds i >downloaded from Bobs site (which is really useful, >both for builds and build instructions, cheers Bob) >seemed to have problems with the keyboard shortcuts >being messed up, the dialog ones were all muddled, >(for instance Ctrl+Shift+M was opening fill/stroke) >just checked the latest one there and its still the >case.(zip file Inkscape0402120319.zip) Very strange. Anyone else experiencing this? >What happened in the Fill/stroke dialog? the entry >boxs for RGB values on the wheel tab are gone That's temporary, we plan to restore them, although not RGB - the wheel is HSV and it will have HSV fields, if you need RGB switch to the RGB sliders. >as is >the dropper icon, both of which were really useful. Dropper on the dialog is redundant, you can use the dropper tool >(its a pain having to switch tool to do a dropper >action, What's the problem with the dropper? We discussed moving the dropper to the dialog and removing the tool, but then decided that the dropper will be very useful without the dialog, when we have a current color indicator and a palette. >btw, anyone got any suggestions as to the best place >to get started with getting familiar with the code? Read developer's howtos on Wiki - there's quite a lot of info there, although it's unorganized. Then look into the code and try to find what seems easiest - I don't know what it may be for you :) I started with adding new keyboard shortcuts, that was easy and fun, and gradually learned many parts of the codebase from there. _________________________________________________________________ Protect your PC - get McAfee.com VirusScan Online http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963

3 2

RE: [Inkscape-devel] openoffice fileformat
by bulia byak 13 Feb '04

13 Feb '04

Johan, >I've tested libinkjar and it can read openoffice fileformats now. >I'd like to know if I can hook this up and move it to cvs HEAD. Could you please add a paragraph describing what you did to http://inkscape.org/cgi-bin/wiki.pl?ReleaseNotes _________________________________________________________________ MSN 8 with e-mail virus protection service: 2 months FREE* http://join.msn.com/?page=features/virus&pgmarket=en-ca&RU=http%3a%2f%2fjoi…

1 0

RE: [Inkscape-devel] Some fairly random feature reqs + bug reports.
by bulia byak 12 Feb '04

12 Feb '04

>fair enough, however its a tool that flash users have >got used to, and is a good way of generating and >manipulating complex shapes in vector form without the >user needing to worry about bezier handles, adding >nodes etc. If you can use paint you can use that tool. You're right, this might be useful, after we have a "hose" tool that will draw by spitting out shapes. Could you please add a RFE to the tracker? >see >http://members.lycos.co.uk/simarilius/weirdness.jpg I'm stunned. How can this happen - anyone having any ideas? Mental, what do you think? > > HSV and it will have HSV fields, if you need RGB > > switch to the RGB sliders. > > > >they dont allow for 0-255 values like most art >programs tho, which is what was nice about the wheel >tab, it had commonality with a lot of other art >programs. I think you're right, we should switch all sliders on all color picker tabs from 0.00..1.00 to 0..255 (integer). It's more useful and gives more precision anyway. However, we must leave the possibility to use the old format as well, by typing e.g. "50%" or "0.5" which, after you press Enter, becomes 127. Again, could you please add an RFE? (Or if you wanted to try contributing, here's a nice little project for you to start :) >selecting the dropper tool switches you out of the >current context, hence you have to swap back. The one >in the wheel panel doesnt, it drops straight back to >whatever you were doing before, which is easier, and >less disruptive to the workflow. Anyway, i go into the >fill dialog to pick a color, isnt it logical to be >able to access the tool to do that from in there? Perhaps you are right, and we'll need to rethink the dropper role and position. I'd say let's do the more urgent color stuff first, which in my opition is a palette and a current color indicator, and then the perfect place of the dropper will be easier to find. _________________________________________________________________ MSN 8 helps eliminate e-mail viruses. Get 2 months FREE*. http://join.msn.com/?page=features/virus&pgmarket=en-ca&RU=http%3a%2f%2fjoi…

1 0

Re: [Inkscape-devel] Strange bug in Windows build?
by bulia byak 12 Feb '04

12 Feb '04

>). But you'll probably see it in the attached svg file that there are >freakish numbers showing... This used to be a small default values star... >:( Thanks, that helps. It would be even better if you could send a couple of "before" and "after" files that are only different by the offending convert to path operation, so we could see exactly what points were displaced and what were not. Please use a simplest possible shape for this (ellipse or square, not star). Being unable to reproduce, I cannot guarantee fixing this, but that would still be helpful. _________________________________________________________________ Tired of spam? Get advanced junk mail protection with MSN 8. http://join.msn.com/?page=dept/bcomm&pgmarket=en-ca&RU=http%3a%2f%2fjoin.ms…

3 4

About
by Daniel Díaz 12 Feb '04

12 Feb '04

I forgot to mention it again: The About Inkscape window is still showing "Frederico Mena", but it should read "Federico Mena-Quintero" (note only one R in Federico). Oh, by the way, when I right click in text written in the canvas, why should I care more about the Item Properties than Text and Font? I think it should be included in there. Greetings! Daniel Díaz yosoy@...31...

1 0

Fedora-1 rpms on sourceforge
by Peter Linnell 12 Feb '04

12 Feb '04

Hi all, Just a quick note to let yet know the Fedora-1 rpms are in the queue for loading on sourceforge. Peter aka mrdocs

2 1

Some fairly random feature reqs + bug reports.
by John Cliff 12 Feb '04

12 Feb '04

Hello all, been playing with both sodipodi and inkscape for a few weeks and have finally got round to looking at the source code. Have a couple of suggestions for features / usability issues. First, if I've got a new session of inkscape, and i do file>open, why does it open it into a new window rather than the blank one I've got in front of me? every time it does it it strikes me as odd. Second up, have you considered making use of the boolean ops to add paintbrush/eraser tools ala flash? the functionality is all there just spread across several tools currently. The way I've been simulating it currently is to use the calligraphy tool to 'paint' an area, then whenever you paint additional areas immediately switch to select and ctrl++ it to the last section. To use it as an eraser just paint over the bits you want to delete and do a subtract. What would be good would be if there was a tool that would automatically do the boolean to the current selection. (and possibly a simplify as well.) It could do with having constant stroke width rather than the calligriphic stuff tho, preferably of selectable width from the secondary toolbar, but still with the realtime preview of whats being drawn. I'd love to be able to try this myself, but am struggling even getting the codebase to compile atm, didnt help that the tar was lacking the mingw makefiles. Am a delphi person normally and havent touched C/C++ since uni, so am more than a little rusty. :) Another feature that I'd love to see would be variable offsets, ie a dynamic offset with node weighting allowing the offset to be modified at each node so I could generate a line of varying thickness (like the calligraphy tool) but still editable by the basic centreline path. This would be a godsend for doing comic style drawings where the line weight is important. I know that ones a big task, just putting the idea out there. With regards to the 0.37 release, the builds i downloaded from Bobs site (which is really useful, both for builds and build instructions, cheers Bob) seemed to have problems with the keyboard shortcuts being messed up, the dialog ones were all muddled, (for instance Ctrl+Shift+M was opening fill/stroke) just checked the latest one there and its still the case.(zip file Inkscape0402120319.zip) What happened in the Fill/stroke dialog? the entry boxs for RGB values on the wheel tab are gone as is the dropper icon, both of which were really useful. (its a pain having to switch tool to do a dropper action, and a lot of apps give their colors in RGB, so being able to enter the values from somewhere else was good.) It is a better size now, but that tab still has a fair bit of grey space in it which could be used. Also on the this build the icons are red crosses on the tabs and join&cap selectors in that dialog. Really good work on the tutorial btw, is really helpful, and looks great. On the website the mailing lists page doesnt link to the user list archive, which would be good. sorry to hit you with so much stuff at once, just been looking at it in increasing detail the last few days. keep up the good work John Cliff btw, anyone got any suggestions as to the best place to get started with getting familiar with the code? __________________________________ Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online. http://taxes.yahoo.com/filing.html

1 0

Inkscape 0.37 RPMS
by Bryce Harrington 12 Feb '04

12 Feb '04

I've uploaded i386 and i686 versions of the Mandrake RPM I mentioned yesterday. They're available from the Inkscape SourceForge download page (go to http://www.inkscape.org/ and click Download). The feedback I've received is that there may be some dependency issues, which I'm working on currently. I would appreciate additional feedback though, as that might help clarify what changes are needed. Bryce

1 0