/* Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010 Free Software Foundation, Inc. Contributed by Andy Vaught Namelist output contributed by Paul Thomas F2003 I/O support contributed by Jerry DeLisle This file is part of the GNU Fortran runtime library (libgfortran). Libgfortran is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 3, or (at your option) any later version. Libgfortran is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. Under Section 7 of GPL version 3, you are granted additional permissions described in the GCC Runtime Library Exception, version 3.1, as published by the Free Software Foundation. You should have received a copy of the GNU General Public License and a copy of the GCC Runtime Library Exception along with this program; see the files COPYING3 and COPYING.RUNTIME respectively. If not, see . */ #include "io.h" #include "format.h" #include "unix.h" #include #include #include #include #include #include #define star_fill(p, n) memset(p, '*', n) typedef unsigned char uchar; /* Helper functions for character(kind=4) internal units. These are needed by write_float.def. */ static inline void memset4 (gfc_char4_t *p, gfc_char4_t c, int k) { int j; for (j = 0; j < k; j++) *p++ = c; } static inline void memcpy4 (gfc_char4_t *dest, const char *source, int k) { int j; const char *p = source; for (j = 0; j < k; j++) *dest++ = (gfc_char4_t) *p++; } /* This include contains the heart and soul of formatted floating point. */ #include "write_float.def" /* Write out default char4. */ static void write_default_char4 (st_parameter_dt *dtp, const gfc_char4_t *source, int src_len, int w_len) { char *p; int j, k = 0; gfc_char4_t c; uchar d; /* Take care of preceding blanks. */ if (w_len > src_len) { k = w_len - src_len; p = write_block (dtp, k); if (p == NULL) return; if (is_char4_unit (dtp)) { gfc_char4_t *p4 = (gfc_char4_t *) p; memset4 (p4, ' ', k); } else memset (p, ' ', k); } /* Get ready to handle delimiters if needed. */ switch (dtp->u.p.current_unit->delim_status) { case DELIM_APOSTROPHE: d = '\''; break; case DELIM_QUOTE: d = '"'; break; default: d = ' '; break; } /* Now process the remaining characters, one at a time. */ for (j = 0; j < src_len; j++) { c = source[j]; if (is_char4_unit (dtp)) { gfc_char4_t *q; /* Handle delimiters if any. */ if (c == d && d != ' ') { p = write_block (dtp, 2); if (p == NULL) return; q = (gfc_char4_t *) p; *q++ = c; } else { p = write_block (dtp, 1); if (p == NULL) return; q = (gfc_char4_t *) p; } *q = c; } else { /* Handle delimiters if any. */ if (c == d && d != ' ') { p = write_block (dtp, 2); if (p == NULL) return; *p++ = (uchar) c; } else { p = write_block (dtp, 1); if (p == NULL) return; } *p = c > 255 ? '?' : (uchar) c; } } } /* Write out UTF-8 converted from char4. */ static void write_utf8_char4 (st_parameter_dt *dtp, gfc_char4_t *source, int src_len, int w_len) { char *p; int j, k = 0; gfc_char4_t c; static const uchar masks[6] = { 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC }; static const uchar limits[6] = { 0x80, 0xE0, 0xF0, 0xF8, 0xFC, 0xFE }; int nbytes; uchar buf[6], d, *q; /* Take care of preceding blanks. */ if (w_len > src_len) { k = w_len - src_len; p = write_block (dtp, k); if (p == NULL) return; memset (p, ' ', k); } /* Get ready to handle delimiters if needed. */ switch (dtp->u.p.current_unit->delim_status) { case DELIM_APOSTROPHE: d = '\''; break; case DELIM_QUOTE: d = '"'; break; default: d = ' '; break; } /* Now process the remaining characters, one at a time. */ for (j = k; j < src_len; j++) { c = source[j]; if (c < 0x80) { /* Handle the delimiters if any. */ if (c == d && d != ' ') { p = write_block (dtp, 2); if (p == NULL) return; *p++ = (uchar) c; } else { p = write_block (dtp, 1); if (p == NULL) return; } *p = (uchar) c; } else { /* Convert to UTF-8 sequence. */ nbytes = 1; q = &buf[6]; do { *--q = ((c & 0x3F) | 0x80); c >>= 6; nbytes++; } while (c >= 0x3F || (c & limits[nbytes-1])); *--q = (c | masks[nbytes-1]); p = write_block (dtp, nbytes); if (p == NULL) return; while (q < &buf[6]) *p++ = *q++; } } } void write_a (st_parameter_dt *dtp, const fnode *f, const char *source, int len) { int wlen; char *p; wlen = f->u.string.length < 0 || (f->format == FMT_G && f->u.string.length == 0) ? len : f->u.string.length; #ifdef HAVE_CRLF /* If this is formatted STREAM IO convert any embedded line feed characters to CR_LF on systems that use that sequence for newlines. See F2003 Standard sections 10.6.3 and 9.9 for further information. */ if (is_stream_io (dtp)) { const char crlf[] = "\r\n"; int i, q, bytes; q = bytes = 0; /* Write out any padding if needed. */ if (len < wlen) { p = write_block (dtp, wlen - len); if (p == NULL) return; memset (p, ' ', wlen - len); } /* Scan the source string looking for '\n' and convert it if found. */ for (i = 0; i < wlen; i++) { if (source[i] == '\n') { /* Write out the previously scanned characters in the string. */ if (bytes > 0) { p = write_block (dtp, bytes); if (p == NULL) return; memcpy (p, &source[q], bytes); q += bytes; bytes = 0; } /* Write out the CR_LF sequence. */ q++; p = write_block (dtp, 2); if (p == NULL) return; memcpy (p, crlf, 2); } else bytes++; } /* Write out any remaining bytes if no LF was found. */ if (bytes > 0) { p = write_block (dtp, bytes); if (p == NULL) return; memcpy (p, &source[q], bytes); } } else { #endif p = write_block (dtp, wlen); if (p == NULL) return; if (unlikely (is_char4_unit (dtp))) { gfc_char4_t *p4 = (gfc_char4_t *) p; if (wlen < len) memcpy4 (p4, source, wlen); else { memset4 (p4, ' ', wlen - len); memcpy4 (p4 + wlen - len, source, len); } return; } if (wlen < len) memcpy (p, source, wlen); else { memset (p, ' ', wlen - len); memcpy (p + wlen - len, source, len); } #ifdef HAVE_CRLF } #endif } /* The primary difference between write_a_char4 and write_a is that we have to deal with writing from the first byte of the 4-byte character and pay attention to the most significant bytes. For ENCODING="default" write the lowest significant byte. If the 3 most significant bytes contain non-zero values, emit a '?'. For ENCODING="utf-8", convert the UCS-32 value to the UTF-8 encoded string before writing out. */ void write_a_char4 (st_parameter_dt *dtp, const fnode *f, const char *source, int len) { int wlen; gfc_char4_t *q; wlen = f->u.string.length < 0 || (f->format == FMT_G && f->u.string.length == 0) ? len : f->u.string.length; q = (gfc_char4_t *) source; #ifdef HAVE_CRLF /* If this is formatted STREAM IO convert any embedded line feed characters to CR_LF on systems that use that sequence for newlines. See F2003 Standard sections 10.6.3 and 9.9 for further information. */ if (is_stream_io (dtp)) { const gfc_char4_t crlf[] = {0x000d,0x000a}; int i, bytes; gfc_char4_t *qq; bytes = 0; /* Write out any padding if needed. */ if (len < wlen) { char *p; p = write_block (dtp, wlen - len); if (p == NULL) return; memset (p, ' ', wlen - len); } /* Scan the source string looking for '\n' and convert it if found. */ qq = (gfc_char4_t *) source; for (i = 0; i < wlen; i++) { if (qq[i] == '\n') { /* Write out the previously scanned characters in the string. */ if (bytes > 0) { if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8) write_utf8_char4 (dtp, q, bytes, 0); else write_default_char4 (dtp, q, bytes, 0); bytes = 0; } /* Write out the CR_LF sequence. */ write_default_char4 (dtp, crlf, 2, 0); } else bytes++; } /* Write out any remaining bytes if no LF was found. */ if (bytes > 0) { if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8) write_utf8_char4 (dtp, q, bytes, 0); else write_default_char4 (dtp, q, bytes, 0); } } else { #endif if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8) write_utf8_char4 (dtp, q, len, wlen); else write_default_char4 (dtp, q, len, wlen); #ifdef HAVE_CRLF } #endif } static GFC_INTEGER_LARGEST extract_int (const void *p, int len) { GFC_INTEGER_LARGEST i = 0; if (p == NULL) return i; switch (len) { case 1: { GFC_INTEGER_1 tmp; memcpy ((void *) &tmp, p, len); i = tmp; } break; case 2: { GFC_INTEGER_2 tmp; memcpy ((void *) &tmp, p, len); i = tmp; } break; case 4: { GFC_INTEGER_4 tmp; memcpy ((void *) &tmp, p, len); i = tmp; } break; case 8: { GFC_INTEGER_8 tmp; memcpy ((void *) &tmp, p, len); i = tmp; } break; #ifdef HAVE_GFC_INTEGER_16 case 16: { GFC_INTEGER_16 tmp; memcpy ((void *) &tmp, p, len); i = tmp; } break; #endif default: internal_error (NULL, "bad integer kind"); } return i; } static GFC_UINTEGER_LARGEST extract_uint (const void *p, int len) { GFC_UINTEGER_LARGEST i = 0; if (p == NULL) return i; switch (len) { case 1: { GFC_INTEGER_1 tmp; memcpy ((void *) &tmp, p, len); i = (GFC_UINTEGER_1) tmp; } break; case 2: { GFC_INTEGER_2 tmp; memcpy ((void *) &tmp, p, len); i = (GFC_UINTEGER_2) tmp; } break; case 4: { GFC_INTEGER_4 tmp; memcpy ((void *) &tmp, p, len); i = (GFC_UINTEGER_4) tmp; } break; case 8: { GFC_INTEGER_8 tmp; memcpy ((void *) &tmp, p, len); i = (GFC_UINTEGER_8) tmp; } break; #ifdef HAVE_GFC_INTEGER_16 case 10: case 16: { GFC_INTEGER_16 tmp = 0; memcpy ((void *) &tmp, p, len); i = (GFC_UINTEGER_16) tmp; } break; #endif default: internal_error (NULL, "bad integer kind"); } return i; } void write_l (st_parameter_dt *dtp, const fnode *f, char *source, int len) { char *p; int wlen; GFC_INTEGER_LARGEST n; wlen = (f->format == FMT_G && f->u.w == 0) ? 1 : f->u.w; p = write_block (dtp, wlen); if (p == NULL) return; n = extract_int (source, len); if (unlikely (is_char4_unit (dtp))) { gfc_char4_t *p4 = (gfc_char4_t *) p; memset4 (p4, ' ', wlen -1); p4[wlen - 1] = (n) ? 'T' : 'F'; return; } memset (p, ' ', wlen -1); p[wlen - 1] = (n) ? 'T' : 'F'; } static void write_boz (st_parameter_dt *dtp, const fnode *f, const char *q, int n) { int w, m, digits, nzero, nblank; char *p; w = f->u.integer.w; m = f->u.integer.m; /* Special case: */ if (m == 0 && n == 0) { if (w == 0) w = 1; p = write_block (dtp, w); if (p == NULL) return; if (unlikely (is_char4_unit (dtp))) { gfc_char4_t *p4 = (gfc_char4_t *) p; memset4 (p4, ' ', w); } else memset (p, ' ', w); goto done; } digits = strlen (q); /* Select a width if none was specified. The idea here is to always print something. */ if (w == 0) w = ((digits < m) ? m : digits); p = write_block (dtp, w); if (p == NULL) return; nzero = 0; if (digits < m) nzero = m - digits; /* See if things will work. */ nblank = w - (nzero + digits); if (unlikely (is_char4_unit (dtp))) { gfc_char4_t *p4 = (gfc_char4_t *) p; if (nblank < 0) { memset4 (p4, '*', w); return; } if (!dtp->u.p.no_leading_blank) { memset4 (p4, ' ', nblank); q += nblank; memset4 (p4, '0', nzero); q += nzero; memcpy4 (p4, q, digits); } else { memset4 (p4, '0', nzero); q += nzero; memcpy4 (p4, q, digits); q += digits; memset4 (p4, ' ', nblank); dtp->u.p.no_leading_blank = 0; } return; } if (nblank < 0) { star_fill (p, w); goto done; } if (!dtp->u.p.no_leading_blank) { memset (p, ' ', nblank); p += nblank; memset (p, '0', nzero); p += nzero; memcpy (p, q, digits); } else { memset (p, '0', nzero); p += nzero; memcpy (p, q, digits); p += digits; memset (p, ' ', nblank); dtp->u.p.no_leading_blank = 0; } done: return; } static void write_decimal (st_parameter_dt *dtp, const fnode *f, const char *source, int len, const char *(*conv) (GFC_INTEGER_LARGEST, char *, size_t)) { GFC_INTEGER_LARGEST n = 0; int w, m, digits, nsign, nzero, nblank; char *p; const char *q; sign_t sign; char itoa_buf[GFC_BTOA_BUF_SIZE]; w = f->u.integer.w; m = f->format == FMT_G ? -1 : f->u.integer.m; n = extract_int (source, len); /* Special case: */ if (m == 0 && n == 0) { if (w == 0) w = 1; p = write_block (dtp, w); if (p == NULL) return; if (unlikely (is_char4_unit (dtp))) { gfc_char4_t *p4 = (gfc_char4_t *) p; memset4 (p4, ' ', w); } else memset (p, ' ', w); goto done; } sign = calculate_sign (dtp, n < 0); if (n < 0) n = -n; nsign = sign == S_NONE ? 0 : 1; /* conv calls itoa which sets the negative sign needed by write_integer. The sign '+' or '-' is set below based on sign calculated above, so we just point past the sign in the string before proceeding to avoid double signs in corner cases. (see PR38504) */ q = conv (n, itoa_buf, sizeof (itoa_buf)); if (*q == '-') q++; digits = strlen (q); /* Select a width if none was specified. The idea here is to always print something. */ if (w == 0) w = ((digits < m) ? m : digits) + nsign; p = write_block (dtp, w); if (p == NULL) return; nzero = 0; if (digits < m) nzero = m - digits; /* See if things will work. */ nblank = w - (nsign + nzero + digits); if (unlikely (is_char4_unit (dtp))) { gfc_char4_t * p4 = (gfc_char4_t *) p; if (nblank < 0) { memset4 (p4, '*', w); goto done; } memset4 (p4, ' ', nblank); p4 += nblank; switch (sign) { case S_PLUS: *p4++ = '+'; break; case S_MINUS: *p4++ = '-'; break; case S_NONE: break; } memset4 (p4, '0', nzero); p4 += nzero; memcpy4 (p4, q, digits); return; } if (nblank < 0) { star_fill (p, w); goto done; } memset (p, ' ', nblank); p += nblank; switch (sign) { case S_PLUS: *p++ = '+'; break; case S_MINUS: *p++ = '-'; break; case S_NONE: break; } memset (p, '0', nzero); p += nzero; memcpy (p, q, digits); done: return; } /* Convert unsigned octal to ascii. */ static const char * otoa (GFC_UINTEGER_LARGEST n, char *buffer, size_t len) { char *p; assert (len >= GFC_OTOA_BUF_SIZE); if (n == 0) return "0"; p = buffer + GFC_OTOA_BUF_SIZE - 1; *p = '\0'; while (n != 0) { *--p = '0' + (n & 7); n >>= 3; } return p; } /* Convert unsigned binary to ascii. */ static const char * btoa (GFC_UINTEGER_LARGEST n, char *buffer, size_t len) { char *p; assert (len >= GFC_BTOA_BUF_SIZE); if (n == 0) return "0"; p = buffer + GFC_BTOA_BUF_SIZE - 1; *p = '\0'; while (n != 0) { *--p = '0' + (n & 1); n >>= 1; } return p; } /* The following three functions, btoa_big, otoa_big, and ztoa_big, are needed to convert large reals with kind sizes that exceed the largest integer type available on certain platforms. In these cases, byte by byte conversion is performed. Endianess is taken into account. */ /* Conversion to binary. */ static const char * btoa_big (const char *s, char *buffer, int len, GFC_UINTEGER_LARGEST *n) { char *q; int i, j; q = buffer; if (big_endian) { const char *p = s; for (i = 0; i < len; i++) { char c = *p; /* Test for zero. Needed by write_boz later. */ if (*p != 0) *n = 1; for (j = 0; j < 8; j++) { *q++ = (c & 128) ? '1' : '0'; c <<= 1; } p++; } } else { const char *p = s + len - 1; for (i = 0; i < len; i++) { char c = *p; /* Test for zero. Needed by write_boz later. */ if (*p != 0) *n = 1; for (j = 0; j < 8; j++) { *q++ = (c & 128) ? '1' : '0'; c <<= 1; } p--; } } *q = '\0'; if (*n == 0) return "0"; /* Move past any leading zeros. */ while (*buffer == '0') buffer++; return buffer; } /* Conversion to octal. */ static const char * otoa_big (const char *s, char *buffer, int len, GFC_UINTEGER_LARGEST *n) { char *q; int i, j, k; uint8_t octet; q = buffer + GFC_OTOA_BUF_SIZE - 1; *q = '\0'; i = k = octet = 0; if (big_endian) { const char *p = s + len - 1; char c = *p; while (i < len) { /* Test for zero. Needed by write_boz later. */ if (*p != 0) *n = 1; for (j = 0; j < 3 && i < len; j++) { octet |= (c & 1) << j; c >>= 1; if (++k > 7) { i++; k = 0; c = *--p; } } *--q = '0' + octet; octet = 0; } } else { const char *p = s; char c = *p; while (i < len) { /* Test for zero. Needed by write_boz later. */ if (*p != 0) *n = 1; for (j = 0; j < 3 && i < len; j++) { octet |= (c & 1) << j; c >>= 1; if (++k > 7) { i++; k = 0; c = *++p; } } *--q = '0' + octet; octet = 0; } } if (*n == 0) return "0"; /* Move past any leading zeros. */ while (*q == '0') q++; return q; } /* Conversion to hexidecimal. */ static const char * ztoa_big (const char *s, char *buffer, int len, GFC_UINTEGER_LARGEST *n) { static char a[16] = {'0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'A', 'B', 'C', 'D', 'E', 'F'}; char *q; uint8_t h, l; int i; q = buffer; if (big_endian) { const char *p = s; for (i = 0; i < len; i++) { /* Test for zero. Needed by write_boz later. */ if (*p != 0) *n = 1; h = (*p >> 4) & 0x0F; l = *p++ & 0x0F; *q++ = a[h]; *q++ = a[l]; } } else { const char *p = s + len - 1; for (i = 0; i < len; i++) { /* Test for zero. Needed by write_boz later. */ if (*p != 0) *n = 1; h = (*p >> 4) & 0x0F; l = *p-- & 0x0F; *q++ = a[h]; *q++ = a[l]; } } *q = '\0'; if (*n == 0) return "0"; /* Move past any leading zeros. */ while (*buffer == '0') buffer++; return buffer; } /* gfc_itoa()-- Integer to decimal conversion. The itoa function is a widespread non-standard extension to standard C, often declared in . Even though the itoa defined here is a static function we take care not to conflict with any prior non-static declaration. Hence the 'gfc_' prefix, which is normally reserved for functions with external linkage. */ static const char * gfc_itoa (GFC_INTEGER_LARGEST n, char *buffer, size_t len) { int negative; char *p; GFC_UINTEGER_LARGEST t; assert (len >= GFC_ITOA_BUF_SIZE); if (n == 0) return "0"; negative = 0; t = n; if (n < 0) { negative = 1; t = -n; /*must use unsigned to protect from overflow*/ } p = buffer + GFC_ITOA_BUF_SIZE - 1; *p = '\0'; while (t != 0) { *--p = '0' + (t % 10); t /= 10; } if (negative) *--p = '-'; return p; } void write_i (st_parameter_dt *dtp, const fnode *f, const char *p, int len) { write_decimal (dtp, f, p, len, (void *) gfc_itoa); } void write_b (st_parameter_dt *dtp, const fnode *f, const char *source, int len) { const char *p; char itoa_buf[GFC_BTOA_BUF_SIZE]; GFC_UINTEGER_LARGEST n = 0; if (len > (int) sizeof (GFC_UINTEGER_LARGEST)) { p = btoa_big (source, itoa_buf, len, &n); write_boz (dtp, f, p, n); } else { n = extract_uint (source, len); p = btoa (n, itoa_buf, sizeof (itoa_buf)); write_boz (dtp, f, p, n); } } void write_o (st_parameter_dt *dtp, const fnode *f, const char *source, int len) { const char *p; char itoa_buf[GFC_OTOA_BUF_SIZE]; GFC_UINTEGER_LARGEST n = 0; if (len > (int) sizeof (GFC_UINTEGER_LARGEST)) { p = otoa_big (source, itoa_buf, len, &n); write_boz (dtp, f, p, n); } else { n = extract_uint (source, len); p = otoa (n, itoa_buf, sizeof (itoa_buf)); write_boz (dtp, f, p, n); } } void write_z (st_parameter_dt *dtp, const fnode *f, const char *source, int len) { const char *p; char itoa_buf[GFC_XTOA_BUF_SIZE]; GFC_UINTEGER_LARGEST n = 0; if (len > (int) sizeof (GFC_UINTEGER_LARGEST)) { p = ztoa_big (source, itoa_buf, len, &n); write_boz (dtp, f, p, n); } else { n = extract_uint (source, len); p = gfc_xtoa (n, itoa_buf, sizeof (itoa_buf)); write_boz (dtp, f, p, n); } } void write_d (st_parameter_dt *dtp, const fnode *f, const char *p, int len) { write_float (dtp, f, p, len); } void write_e (st_parameter_dt *dtp, const fnode *f, const char *p, int len) { write_float (dtp, f, p, len); } void write_f (st_parameter_dt *dtp, const fnode *f, const char *p, int len) { write_float (dtp, f, p, len); } void write_en (st_parameter_dt *dtp, const fnode *f, const char *p, int len) { write_float (dtp, f, p, len); } void write_es (st_parameter_dt *dtp, const fnode *f, const char *p, int len) { write_float (dtp, f, p, len); } /* Take care of the X/TR descriptor. */ void write_x (st_parameter_dt *dtp, int len, int nspaces) { char *p; p = write_block (dtp, len); if (p == NULL) return; if (nspaces > 0 && len - nspaces >= 0) { if (unlikely (is_char4_unit (dtp))) { gfc_char4_t *p4 = (gfc_char4_t *) p; memset4 (&p4[len - nspaces], ' ', nspaces); } else memset (&p[len - nspaces], ' ', nspaces); } } /* List-directed writing. */ /* Write a single character to the output. Returns nonzero if something goes wrong. */ static int write_char (st_parameter_dt *dtp, int c) { char *p; p = write_block (dtp, 1); if (p == NULL) return 1; if (unlikely (is_char4_unit (dtp))) { gfc_char4_t *p4 = (gfc_char4_t *) p; *p4 = c; return 0; } *p = (uchar) c; return 0; } /* Write a list-directed logical value. */ static void write_logical (st_parameter_dt *dtp, const char *source, int length) { write_char (dtp, extract_int (source, length) ? 'T' : 'F'); } /* Write a list-directed integer value. */ static void write_integer (st_parameter_dt *dtp, const char *source, int length) { char *p; const char *q; int digits; int width; char itoa_buf[GFC_ITOA_BUF_SIZE]; q = gfc_itoa (extract_int (source, length), itoa_buf, sizeof (itoa_buf)); switch (length) { case 1: width = 4; break; case 2: width = 6; break; case 4: width = 11; break; case 8: width = 20; break; default: width = 0; break; } digits = strlen (q); if (width < digits) width = digits; p = write_block (dtp, width); if (p == NULL) return; if (unlikely (is_char4_unit (dtp))) { gfc_char4_t *p4 = (gfc_char4_t *) p; if (dtp->u.p.no_leading_blank) { memcpy4 (p4, q, digits); memset4 (p4 + digits, ' ', width - digits); } else { memset4 (p4, ' ', width - digits); memcpy4 (p4 + width - digits, q, digits); } return; } if (dtp->u.p.no_leading_blank) { memcpy (p, q, digits); memset (p + digits, ' ', width - digits); } else { memset (p, ' ', width - digits); memcpy (p + width - digits, q, digits); } } /* Write a list-directed string. We have to worry about delimiting the strings if the file has been opened in that mode. */ static void write_character (st_parameter_dt *dtp, const char *source, int kind, int length) { int i, extra; char *p, d; switch (dtp->u.p.current_unit->delim_status) { case DELIM_APOSTROPHE: d = '\''; break; case DELIM_QUOTE: d = '"'; break; default: d = ' '; break; } if (kind == 1) { if (d == ' ') extra = 0; else { extra = 2; for (i = 0; i < length; i++) if (source[i] == d) extra++; } p = write_block (dtp, length + extra); if (p == NULL) return; if (unlikely (is_char4_unit (dtp))) { gfc_char4_t d4 = (gfc_char4_t) d; gfc_char4_t *p4 = (gfc_char4_t *) p; if (d4 == ' ') memcpy4 (p4, source, length); else { *p4++ = d4; for (i = 0; i < length; i++) { *p4++ = (gfc_char4_t) source[i]; if (source[i] == d) *p4++ = d4; } *p4 = d4; } return; } if (d == ' ') memcpy (p, source, length); else { *p++ = d; for (i = 0; i < length; i++) { *p++ = source[i]; if (source[i] == d) *p++ = d; } *p = d; } } else { if (d == ' ') { if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8) write_utf8_char4 (dtp, (gfc_char4_t *) source, length, 0); else write_default_char4 (dtp, (gfc_char4_t *) source, length, 0); } else { p = write_block (dtp, 1); *p = d; if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8) write_utf8_char4 (dtp, (gfc_char4_t *) source, length, 0); else write_default_char4 (dtp, (gfc_char4_t *) source, length, 0); p = write_block (dtp, 1); *p = d; } } } /* Set an fnode to default format. */ static void set_fnode_default (st_parameter_dt *dtp, fnode *f, int length) { f->format = FMT_G; switch (length) { case 4: f->u.real.w = 15; f->u.real.d = 8; f->u.real.e = 2; break; case 8: f->u.real.w = 25; f->u.real.d = 17; f->u.real.e = 3; break; case 10: f->u.real.w = 29; f->u.real.d = 20; f->u.real.e = 4; break; case 16: f->u.real.w = 44; f->u.real.d = 35; f->u.real.e = 4; break; default: internal_error (&dtp->common, "bad real kind"); break; } } /* Output a real number with default format. This is 1PG14.7E2 for REAL(4), 1PG23.15E3 for REAL(8), 1PG28.19E4 for REAL(10) and 1PG43.34E4 for REAL(16). */ // FX -- FIXME: should we change the default format for __float128-real(16)? void write_real (st_parameter_dt *dtp, const char *source, int length) { fnode f ; int org_scale = dtp->u.p.scale_factor; dtp->u.p.scale_factor = 1; set_fnode_default (dtp, &f, length); write_float (dtp, &f, source , length); dtp->u.p.scale_factor = org_scale; } void write_real_g0 (st_parameter_dt *dtp, const char *source, int length, int d) { fnode f ; set_fnode_default (dtp, &f, length); if (d > 0) f.u.real.d = d; dtp->u.p.g0_no_blanks = 1; write_float (dtp, &f, source , length); dtp->u.p.g0_no_blanks = 0; } static void write_complex (st_parameter_dt *dtp, const char *source, int kind, size_t size) { char semi_comma = dtp->u.p.current_unit->decimal_status == DECIMAL_POINT ? ',' : ';'; if (write_char (dtp, '(')) return; write_real (dtp, source, kind); if (write_char (dtp, semi_comma)) return; write_real (dtp, source + size / 2, kind); write_char (dtp, ')'); } /* Write the separator between items. */ static void write_separator (st_parameter_dt *dtp) { char *p; p = write_block (dtp, options.separator_len); if (p == NULL) return; if (unlikely (is_char4_unit (dtp))) { gfc_char4_t *p4 = (gfc_char4_t *) p; memcpy4 (p4, options.separator, options.separator_len); } else memcpy (p, options.separator, options.separator_len); } /* Write an item with list formatting. TODO: handle skipping to the next record correctly, particularly with strings. */ static void list_formatted_write_scalar (st_parameter_dt *dtp, bt type, void *p, int kind, size_t size) { if (dtp->u.p.current_unit == NULL) return; if (dtp->u.p.first_item) { dtp->u.p.first_item = 0; write_char (dtp, ' '); } else { if (type != BT_CHARACTER || !dtp->u.p.char_flag || dtp->u.p.current_unit->delim_status != DELIM_NONE) write_separator (dtp); } switch (type) { case BT_INTEGER: write_integer (dtp, p, kind); break; case BT_LOGICAL: write_logical (dtp, p, kind); break; case BT_CHARACTER: write_character (dtp, p, kind, size); break; case BT_REAL: write_real (dtp, p, kind); break; case BT_COMPLEX: write_complex (dtp, p, kind, size); break; default: internal_error (&dtp->common, "list_formatted_write(): Bad type"); } dtp->u.p.char_flag = (type == BT_CHARACTER); } void list_formatted_write (st_parameter_dt *dtp, bt type, void *p, int kind, size_t size, size_t nelems) { size_t elem; char *tmp; size_t stride = type == BT_CHARACTER ? size * GFC_SIZE_OF_CHAR_KIND(kind) : size; tmp = (char *) p; /* Big loop over all the elements. */ for (elem = 0; elem < nelems; elem++) { dtp->u.p.item_count++; list_formatted_write_scalar (dtp, type, tmp + elem * stride, kind, size); } } /* NAMELIST OUTPUT nml_write_obj writes a namelist object to the output stream. It is called recursively for derived type components: obj = is the namelist_info for the current object. offset = the offset relative to the address held by the object for derived type arrays. base = is the namelist_info of the derived type, when obj is a component. base_name = the full name for a derived type, including qualifiers if any. The returned value is a pointer to the object beyond the last one accessed, including nested derived types. Notice that the namelist is a linear linked list of objects, including derived types and their components. A tree, of sorts, is implied by the compound names of the derived type components and this is how this function recurses through the list. */ /* A generous estimate of the number of characters needed to print repeat counts and indices, including commas, asterices and brackets. */ #define NML_DIGITS 20 static void namelist_write_newline (st_parameter_dt *dtp) { if (!is_internal_unit (dtp)) { #ifdef HAVE_CRLF write_character (dtp, "\r\n", 1, 2); #else write_character (dtp, "\n", 1, 1); #endif return; } if (is_array_io (dtp)) { gfc_offset record; int finished; char *p; int length = dtp->u.p.current_unit->bytes_left; p = write_block (dtp, length); if (p == NULL) return; if (unlikely (is_char4_unit (dtp))) { gfc_char4_t *p4 = (gfc_char4_t *) p; memset4 (p4, ' ', length); } else memset (p, ' ', length); /* Now that the current record has been padded out, determine where the next record in the array is. */ record = next_array_record (dtp, dtp->u.p.current_unit->ls, &finished); if (finished) dtp->u.p.current_unit->endfile = AT_ENDFILE; else { /* Now seek to this record */ record = record * dtp->u.p.current_unit->recl; if (sseek (dtp->u.p.current_unit->s, record, SEEK_SET) < 0) { generate_error (&dtp->common, LIBERROR_INTERNAL_UNIT, NULL); return; } dtp->u.p.current_unit->bytes_left = dtp->u.p.current_unit->recl; } } else write_character (dtp, " ", 1, 1); } static namelist_info * nml_write_obj (st_parameter_dt *dtp, namelist_info * obj, index_type offset, namelist_info * base, char * base_name) { int rep_ctr; int num; int nml_carry; int len; index_type obj_size; index_type nelem; size_t dim_i; size_t clen; index_type elem_ctr; size_t obj_name_len; void * p ; char cup; char * obj_name; char * ext_name; char rep_buff[NML_DIGITS]; namelist_info * cmp; namelist_info * retval = obj->next; size_t base_name_len; size_t base_var_name_len; size_t tot_len; unit_delim tmp_delim; /* Set the character to be used to separate values to a comma or semi-colon. */ char semi_comma = dtp->u.p.current_unit->decimal_status == DECIMAL_POINT ? ',' : ';'; /* Write namelist variable names in upper case. If a derived type, nothing is output. If a component, base and base_name are set. */ if (obj->type != BT_DERIVED) { namelist_write_newline (dtp); write_character (dtp, " ", 1, 1); len = 0; if (base) { len = strlen (base->var_name); base_name_len = strlen (base_name); for (dim_i = 0; dim_i < base_name_len; dim_i++) { cup = toupper ((int) base_name[dim_i]); write_character (dtp, &cup, 1, 1); } } clen = strlen (obj->var_name); for (dim_i = len; dim_i < clen; dim_i++) { cup = toupper ((int) obj->var_name[dim_i]); write_character (dtp, &cup, 1, 1); } write_character (dtp, "=", 1, 1); } /* Counts the number of data output on a line, including names. */ num = 1; len = obj->len; switch (obj->type) { case BT_REAL: obj_size = size_from_real_kind (len); break; case BT_COMPLEX: obj_size = size_from_complex_kind (len); break; case BT_CHARACTER: obj_size = obj->string_length; break; default: obj_size = len; } if (obj->var_rank) obj_size = obj->size; /* Set the index vector and count the number of elements. */ nelem = 1; for (dim_i = 0; dim_i < (size_t) obj->var_rank; dim_i++) { obj->ls[dim_i].idx = GFC_DESCRIPTOR_LBOUND(obj, dim_i); nelem = nelem * GFC_DESCRIPTOR_EXTENT (obj, dim_i); } /* Main loop to output the data held in the object. */ rep_ctr = 1; for (elem_ctr = 0; elem_ctr < nelem; elem_ctr++) { /* Build the pointer to the data value. The offset is passed by recursive calls to this function for arrays of derived types. Is NULL otherwise. */ p = (void *)(obj->mem_pos + elem_ctr * obj_size); p += offset; /* Check for repeat counts of intrinsic types. */ if ((elem_ctr < (nelem - 1)) && (obj->type != BT_DERIVED) && !memcmp (p, (void*)(p + obj_size ), obj_size )) { rep_ctr++; } /* Execute a repeated output. Note the flag no_leading_blank that is used in the functions used to output the intrinsic types. */ else { if (rep_ctr > 1) { sprintf(rep_buff, " %d*", rep_ctr); write_character (dtp, rep_buff, 1, strlen (rep_buff)); dtp->u.p.no_leading_blank = 1; } num++; /* Output the data, if an intrinsic type, or recurse into this routine to treat derived types. */ switch (obj->type) { case BT_INTEGER: write_integer (dtp, p, len); break; case BT_LOGICAL: write_logical (dtp, p, len); break; case BT_CHARACTER: tmp_delim = dtp->u.p.current_unit->delim_status; if (dtp->u.p.nml_delim == '"') dtp->u.p.current_unit->delim_status = DELIM_QUOTE; if (dtp->u.p.nml_delim == '\'') dtp->u.p.current_unit->delim_status = DELIM_APOSTROPHE; write_character (dtp, p, 1, obj->string_length); dtp->u.p.current_unit->delim_status = tmp_delim; break; case BT_REAL: write_real (dtp, p, len); break; case BT_COMPLEX: dtp->u.p.no_leading_blank = 0; num++; write_complex (dtp, p, len, obj_size); break; case BT_DERIVED: /* To treat a derived type, we need to build two strings: ext_name = the name, including qualifiers that prepends component names in the output - passed to nml_write_obj. obj_name = the derived type name with no qualifiers but % appended. This is used to identify the components. */ /* First ext_name => get length of all possible components */ base_name_len = base_name ? strlen (base_name) : 0; base_var_name_len = base ? strlen (base->var_name) : 0; ext_name = (char*)get_mem ( base_name_len + base_var_name_len + strlen (obj->var_name) + obj->var_rank * NML_DIGITS + 1); memcpy (ext_name, base_name, base_name_len); clen = strlen (obj->var_name + base_var_name_len); memcpy (ext_name + base_name_len, obj->var_name + base_var_name_len, clen); /* Append the qualifier. */ tot_len = base_name_len + clen; for (dim_i = 0; dim_i < (size_t) obj->var_rank; dim_i++) { if (!dim_i) { ext_name[tot_len] = '('; tot_len++; } sprintf (ext_name + tot_len, "%d", (int) obj->ls[dim_i].idx); tot_len += strlen (ext_name + tot_len); ext_name[tot_len] = ((int) dim_i == obj->var_rank - 1) ? ')' : ','; tot_len++; } ext_name[tot_len] = '\0'; /* Now obj_name. */ obj_name_len = strlen (obj->var_name) + 1; obj_name = get_mem (obj_name_len+1); memcpy (obj_name, obj->var_name, obj_name_len-1); memcpy (obj_name + obj_name_len-1, "%", 2); /* Now loop over the components. Update the component pointer with the return value from nml_write_obj => this loop jumps past nested derived types. */ for (cmp = obj->next; cmp && !strncmp (cmp->var_name, obj_name, obj_name_len); cmp = retval) { retval = nml_write_obj (dtp, cmp, (index_type)(p - obj->mem_pos), obj, ext_name); } free (obj_name); free (ext_name); goto obj_loop; default: internal_error (&dtp->common, "Bad type for namelist write"); } /* Reset the leading blank suppression, write a comma (or semi-colon) and, if 5 values have been output, write a newline and advance to column 2. Reset the repeat counter. */ dtp->u.p.no_leading_blank = 0; write_character (dtp, &semi_comma, 1, 1); if (num > 5) { num = 0; namelist_write_newline (dtp); write_character (dtp, " ", 1, 1); } rep_ctr = 1; } /* Cycle through and increment the index vector. */ obj_loop: nml_carry = 1; for (dim_i = 0; nml_carry && (dim_i < (size_t) obj->var_rank); dim_i++) { obj->ls[dim_i].idx += nml_carry ; nml_carry = 0; if (obj->ls[dim_i].idx > (ssize_t) GFC_DESCRIPTOR_UBOUND(obj,dim_i)) { obj->ls[dim_i].idx = GFC_DESCRIPTOR_LBOUND(obj,dim_i); nml_carry = 1; } } } /* Return a pointer beyond the furthest object accessed. */ return retval; } /* This is the entry function for namelist writes. It outputs the name of the namelist and iterates through the namelist by calls to nml_write_obj. The call below has dummys in the arguments used in the treatment of derived types. */ void namelist_write (st_parameter_dt *dtp) { namelist_info * t1, *t2, *dummy = NULL; index_type i; index_type dummy_offset = 0; char c; char * dummy_name = NULL; unit_delim tmp_delim = DELIM_UNSPECIFIED; /* Set the delimiter for namelist output. */ tmp_delim = dtp->u.p.current_unit->delim_status; dtp->u.p.nml_delim = tmp_delim == DELIM_APOSTROPHE ? '\'' : '"'; /* Temporarily disable namelist delimters. */ dtp->u.p.current_unit->delim_status = DELIM_NONE; write_character (dtp, "&", 1, 1); /* Write namelist name in upper case - f95 std. */ for (i = 0 ;i < dtp->namelist_name_len ;i++ ) { c = toupper ((int) dtp->namelist_name[i]); write_character (dtp, &c, 1 ,1); } if (dtp->u.p.ionml != NULL) { t1 = dtp->u.p.ionml; while (t1 != NULL) { t2 = t1; t1 = nml_write_obj (dtp, t2, dummy_offset, dummy, dummy_name); } } namelist_write_newline (dtp); write_character (dtp, " /", 1, 2); /* Restore the original delimiter. */ dtp->u.p.current_unit->delim_status = tmp_delim; } #undef NML_DIGITS