]> git.ipfire.org Git - thirdparty/gcc.git/blob - libgfortran/io/write.c
re PR fortran/60148 (strings in NAMELIST do not honor DELIM= in open statement)
[thirdparty/gcc.git] / libgfortran / io / write.c
1 /* Copyright (C) 2002-2014 Free Software Foundation, Inc.
2 Contributed by Andy Vaught
3 Namelist output contributed by Paul Thomas
4 F2003 I/O support contributed by Jerry DeLisle
5
6 This file is part of the GNU Fortran runtime library (libgfortran).
7
8 Libgfortran is free software; you can redistribute it and/or modify
9 it under the terms of the GNU General Public License as published by
10 the Free Software Foundation; either version 3, or (at your option)
11 any later version.
12
13 Libgfortran is distributed in the hope that it will be useful,
14 but WITHOUT ANY WARRANTY; without even the implied warranty of
15 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 GNU General Public License for more details.
17
18 Under Section 7 of GPL version 3, you are granted additional
19 permissions described in the GCC Runtime Library Exception, version
20 3.1, as published by the Free Software Foundation.
21
22 You should have received a copy of the GNU General Public License and
23 a copy of the GCC Runtime Library Exception along with this program;
24 see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
25 <http://www.gnu.org/licenses/>. */
26
27 #include "io.h"
28 #include "format.h"
29 #include "unix.h"
30 #include <assert.h>
31 #include <string.h>
32 #include <ctype.h>
33 #include <stdlib.h>
34 #include <errno.h>
35 #define star_fill(p, n) memset(p, '*', n)
36
37 typedef unsigned char uchar;
38
39 /* Helper functions for character(kind=4) internal units. These are needed
40 by write_float.def. */
41
42 static void
43 memcpy4 (gfc_char4_t *dest, const char *source, int k)
44 {
45 int j;
46
47 const char *p = source;
48 for (j = 0; j < k; j++)
49 *dest++ = (gfc_char4_t) *p++;
50 }
51
52 /* This include contains the heart and soul of formatted floating point. */
53 #include "write_float.def"
54
55 /* Write out default char4. */
56
57 static void
58 write_default_char4 (st_parameter_dt *dtp, const gfc_char4_t *source,
59 int src_len, int w_len)
60 {
61 char *p;
62 int j, k = 0;
63 gfc_char4_t c;
64 uchar d;
65
66 /* Take care of preceding blanks. */
67 if (w_len > src_len)
68 {
69 k = w_len - src_len;
70 p = write_block (dtp, k);
71 if (p == NULL)
72 return;
73 if (is_char4_unit (dtp))
74 {
75 gfc_char4_t *p4 = (gfc_char4_t *) p;
76 memset4 (p4, ' ', k);
77 }
78 else
79 memset (p, ' ', k);
80 }
81
82 /* Get ready to handle delimiters if needed. */
83 switch (dtp->u.p.current_unit->delim_status)
84 {
85 case DELIM_APOSTROPHE:
86 d = '\'';
87 break;
88 case DELIM_QUOTE:
89 d = '"';
90 break;
91 default:
92 d = ' ';
93 break;
94 }
95
96 /* Now process the remaining characters, one at a time. */
97 for (j = 0; j < src_len; j++)
98 {
99 c = source[j];
100 if (is_char4_unit (dtp))
101 {
102 gfc_char4_t *q;
103 /* Handle delimiters if any. */
104 if (c == d && d != ' ')
105 {
106 p = write_block (dtp, 2);
107 if (p == NULL)
108 return;
109 q = (gfc_char4_t *) p;
110 *q++ = c;
111 }
112 else
113 {
114 p = write_block (dtp, 1);
115 if (p == NULL)
116 return;
117 q = (gfc_char4_t *) p;
118 }
119 *q = c;
120 }
121 else
122 {
123 /* Handle delimiters if any. */
124 if (c == d && d != ' ')
125 {
126 p = write_block (dtp, 2);
127 if (p == NULL)
128 return;
129 *p++ = (uchar) c;
130 }
131 else
132 {
133 p = write_block (dtp, 1);
134 if (p == NULL)
135 return;
136 }
137 *p = c > 255 ? '?' : (uchar) c;
138 }
139 }
140 }
141
142
143 /* Write out UTF-8 converted from char4. */
144
145 static void
146 write_utf8_char4 (st_parameter_dt *dtp, gfc_char4_t *source,
147 int src_len, int w_len)
148 {
149 char *p;
150 int j, k = 0;
151 gfc_char4_t c;
152 static const uchar masks[6] = { 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
153 static const uchar limits[6] = { 0x80, 0xE0, 0xF0, 0xF8, 0xFC, 0xFE };
154 int nbytes;
155 uchar buf[6], d, *q;
156
157 /* Take care of preceding blanks. */
158 if (w_len > src_len)
159 {
160 k = w_len - src_len;
161 p = write_block (dtp, k);
162 if (p == NULL)
163 return;
164 memset (p, ' ', k);
165 }
166
167 /* Get ready to handle delimiters if needed. */
168 switch (dtp->u.p.current_unit->delim_status)
169 {
170 case DELIM_APOSTROPHE:
171 d = '\'';
172 break;
173 case DELIM_QUOTE:
174 d = '"';
175 break;
176 default:
177 d = ' ';
178 break;
179 }
180
181 /* Now process the remaining characters, one at a time. */
182 for (j = k; j < src_len; j++)
183 {
184 c = source[j];
185 if (c < 0x80)
186 {
187 /* Handle the delimiters if any. */
188 if (c == d && d != ' ')
189 {
190 p = write_block (dtp, 2);
191 if (p == NULL)
192 return;
193 *p++ = (uchar) c;
194 }
195 else
196 {
197 p = write_block (dtp, 1);
198 if (p == NULL)
199 return;
200 }
201 *p = (uchar) c;
202 }
203 else
204 {
205 /* Convert to UTF-8 sequence. */
206 nbytes = 1;
207 q = &buf[6];
208
209 do
210 {
211 *--q = ((c & 0x3F) | 0x80);
212 c >>= 6;
213 nbytes++;
214 }
215 while (c >= 0x3F || (c & limits[nbytes-1]));
216
217 *--q = (c | masks[nbytes-1]);
218
219 p = write_block (dtp, nbytes);
220 if (p == NULL)
221 return;
222
223 while (q < &buf[6])
224 *p++ = *q++;
225 }
226 }
227 }
228
229
230 void
231 write_a (st_parameter_dt *dtp, const fnode *f, const char *source, int len)
232 {
233 int wlen;
234 char *p;
235
236 wlen = f->u.string.length < 0
237 || (f->format == FMT_G && f->u.string.length == 0)
238 ? len : f->u.string.length;
239
240 #ifdef HAVE_CRLF
241 /* If this is formatted STREAM IO convert any embedded line feed characters
242 to CR_LF on systems that use that sequence for newlines. See F2003
243 Standard sections 10.6.3 and 9.9 for further information. */
244 if (is_stream_io (dtp))
245 {
246 const char crlf[] = "\r\n";
247 int i, q, bytes;
248 q = bytes = 0;
249
250 /* Write out any padding if needed. */
251 if (len < wlen)
252 {
253 p = write_block (dtp, wlen - len);
254 if (p == NULL)
255 return;
256 memset (p, ' ', wlen - len);
257 }
258
259 /* Scan the source string looking for '\n' and convert it if found. */
260 for (i = 0; i < wlen; i++)
261 {
262 if (source[i] == '\n')
263 {
264 /* Write out the previously scanned characters in the string. */
265 if (bytes > 0)
266 {
267 p = write_block (dtp, bytes);
268 if (p == NULL)
269 return;
270 memcpy (p, &source[q], bytes);
271 q += bytes;
272 bytes = 0;
273 }
274
275 /* Write out the CR_LF sequence. */
276 q++;
277 p = write_block (dtp, 2);
278 if (p == NULL)
279 return;
280 memcpy (p, crlf, 2);
281 }
282 else
283 bytes++;
284 }
285
286 /* Write out any remaining bytes if no LF was found. */
287 if (bytes > 0)
288 {
289 p = write_block (dtp, bytes);
290 if (p == NULL)
291 return;
292 memcpy (p, &source[q], bytes);
293 }
294 }
295 else
296 {
297 #endif
298 p = write_block (dtp, wlen);
299 if (p == NULL)
300 return;
301
302 if (unlikely (is_char4_unit (dtp)))
303 {
304 gfc_char4_t *p4 = (gfc_char4_t *) p;
305 if (wlen < len)
306 memcpy4 (p4, source, wlen);
307 else
308 {
309 memset4 (p4, ' ', wlen - len);
310 memcpy4 (p4 + wlen - len, source, len);
311 }
312 return;
313 }
314
315 if (wlen < len)
316 memcpy (p, source, wlen);
317 else
318 {
319 memset (p, ' ', wlen - len);
320 memcpy (p + wlen - len, source, len);
321 }
322 #ifdef HAVE_CRLF
323 }
324 #endif
325 }
326
327
328 /* The primary difference between write_a_char4 and write_a is that we have to
329 deal with writing from the first byte of the 4-byte character and pay
330 attention to the most significant bytes. For ENCODING="default" write the
331 lowest significant byte. If the 3 most significant bytes contain
332 non-zero values, emit a '?'. For ENCODING="utf-8", convert the UCS-32 value
333 to the UTF-8 encoded string before writing out. */
334
335 void
336 write_a_char4 (st_parameter_dt *dtp, const fnode *f, const char *source, int len)
337 {
338 int wlen;
339 gfc_char4_t *q;
340
341 wlen = f->u.string.length < 0
342 || (f->format == FMT_G && f->u.string.length == 0)
343 ? len : f->u.string.length;
344
345 q = (gfc_char4_t *) source;
346 #ifdef HAVE_CRLF
347 /* If this is formatted STREAM IO convert any embedded line feed characters
348 to CR_LF on systems that use that sequence for newlines. See F2003
349 Standard sections 10.6.3 and 9.9 for further information. */
350 if (is_stream_io (dtp))
351 {
352 const gfc_char4_t crlf[] = {0x000d,0x000a};
353 int i, bytes;
354 gfc_char4_t *qq;
355 bytes = 0;
356
357 /* Write out any padding if needed. */
358 if (len < wlen)
359 {
360 char *p;
361 p = write_block (dtp, wlen - len);
362 if (p == NULL)
363 return;
364 memset (p, ' ', wlen - len);
365 }
366
367 /* Scan the source string looking for '\n' and convert it if found. */
368 qq = (gfc_char4_t *) source;
369 for (i = 0; i < wlen; i++)
370 {
371 if (qq[i] == '\n')
372 {
373 /* Write out the previously scanned characters in the string. */
374 if (bytes > 0)
375 {
376 if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8)
377 write_utf8_char4 (dtp, q, bytes, 0);
378 else
379 write_default_char4 (dtp, q, bytes, 0);
380 bytes = 0;
381 }
382
383 /* Write out the CR_LF sequence. */
384 write_default_char4 (dtp, crlf, 2, 0);
385 }
386 else
387 bytes++;
388 }
389
390 /* Write out any remaining bytes if no LF was found. */
391 if (bytes > 0)
392 {
393 if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8)
394 write_utf8_char4 (dtp, q, bytes, 0);
395 else
396 write_default_char4 (dtp, q, bytes, 0);
397 }
398 }
399 else
400 {
401 #endif
402 if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8)
403 write_utf8_char4 (dtp, q, len, wlen);
404 else
405 write_default_char4 (dtp, q, len, wlen);
406 #ifdef HAVE_CRLF
407 }
408 #endif
409 }
410
411
412 static GFC_INTEGER_LARGEST
413 extract_int (const void *p, int len)
414 {
415 GFC_INTEGER_LARGEST i = 0;
416
417 if (p == NULL)
418 return i;
419
420 switch (len)
421 {
422 case 1:
423 {
424 GFC_INTEGER_1 tmp;
425 memcpy ((void *) &tmp, p, len);
426 i = tmp;
427 }
428 break;
429 case 2:
430 {
431 GFC_INTEGER_2 tmp;
432 memcpy ((void *) &tmp, p, len);
433 i = tmp;
434 }
435 break;
436 case 4:
437 {
438 GFC_INTEGER_4 tmp;
439 memcpy ((void *) &tmp, p, len);
440 i = tmp;
441 }
442 break;
443 case 8:
444 {
445 GFC_INTEGER_8 tmp;
446 memcpy ((void *) &tmp, p, len);
447 i = tmp;
448 }
449 break;
450 #ifdef HAVE_GFC_INTEGER_16
451 case 16:
452 {
453 GFC_INTEGER_16 tmp;
454 memcpy ((void *) &tmp, p, len);
455 i = tmp;
456 }
457 break;
458 #endif
459 default:
460 internal_error (NULL, "bad integer kind");
461 }
462
463 return i;
464 }
465
466 static GFC_UINTEGER_LARGEST
467 extract_uint (const void *p, int len)
468 {
469 GFC_UINTEGER_LARGEST i = 0;
470
471 if (p == NULL)
472 return i;
473
474 switch (len)
475 {
476 case 1:
477 {
478 GFC_INTEGER_1 tmp;
479 memcpy ((void *) &tmp, p, len);
480 i = (GFC_UINTEGER_1) tmp;
481 }
482 break;
483 case 2:
484 {
485 GFC_INTEGER_2 tmp;
486 memcpy ((void *) &tmp, p, len);
487 i = (GFC_UINTEGER_2) tmp;
488 }
489 break;
490 case 4:
491 {
492 GFC_INTEGER_4 tmp;
493 memcpy ((void *) &tmp, p, len);
494 i = (GFC_UINTEGER_4) tmp;
495 }
496 break;
497 case 8:
498 {
499 GFC_INTEGER_8 tmp;
500 memcpy ((void *) &tmp, p, len);
501 i = (GFC_UINTEGER_8) tmp;
502 }
503 break;
504 #ifdef HAVE_GFC_INTEGER_16
505 case 10:
506 case 16:
507 {
508 GFC_INTEGER_16 tmp = 0;
509 memcpy ((void *) &tmp, p, len);
510 i = (GFC_UINTEGER_16) tmp;
511 }
512 break;
513 #endif
514 default:
515 internal_error (NULL, "bad integer kind");
516 }
517
518 return i;
519 }
520
521
522 void
523 write_l (st_parameter_dt *dtp, const fnode *f, char *source, int len)
524 {
525 char *p;
526 int wlen;
527 GFC_INTEGER_LARGEST n;
528
529 wlen = (f->format == FMT_G && f->u.w == 0) ? 1 : f->u.w;
530
531 p = write_block (dtp, wlen);
532 if (p == NULL)
533 return;
534
535 n = extract_int (source, len);
536
537 if (unlikely (is_char4_unit (dtp)))
538 {
539 gfc_char4_t *p4 = (gfc_char4_t *) p;
540 memset4 (p4, ' ', wlen -1);
541 p4[wlen - 1] = (n) ? 'T' : 'F';
542 return;
543 }
544
545 memset (p, ' ', wlen -1);
546 p[wlen - 1] = (n) ? 'T' : 'F';
547 }
548
549
550 static void
551 write_boz (st_parameter_dt *dtp, const fnode *f, const char *q, int n)
552 {
553 int w, m, digits, nzero, nblank;
554 char *p;
555
556 w = f->u.integer.w;
557 m = f->u.integer.m;
558
559 /* Special case: */
560
561 if (m == 0 && n == 0)
562 {
563 if (w == 0)
564 w = 1;
565
566 p = write_block (dtp, w);
567 if (p == NULL)
568 return;
569 if (unlikely (is_char4_unit (dtp)))
570 {
571 gfc_char4_t *p4 = (gfc_char4_t *) p;
572 memset4 (p4, ' ', w);
573 }
574 else
575 memset (p, ' ', w);
576 goto done;
577 }
578
579 digits = strlen (q);
580
581 /* Select a width if none was specified. The idea here is to always
582 print something. */
583
584 if (w == 0)
585 w = ((digits < m) ? m : digits);
586
587 p = write_block (dtp, w);
588 if (p == NULL)
589 return;
590
591 nzero = 0;
592 if (digits < m)
593 nzero = m - digits;
594
595 /* See if things will work. */
596
597 nblank = w - (nzero + digits);
598
599 if (unlikely (is_char4_unit (dtp)))
600 {
601 gfc_char4_t *p4 = (gfc_char4_t *) p;
602 if (nblank < 0)
603 {
604 memset4 (p4, '*', w);
605 return;
606 }
607
608 if (!dtp->u.p.no_leading_blank)
609 {
610 memset4 (p4, ' ', nblank);
611 q += nblank;
612 memset4 (p4, '0', nzero);
613 q += nzero;
614 memcpy4 (p4, q, digits);
615 }
616 else
617 {
618 memset4 (p4, '0', nzero);
619 q += nzero;
620 memcpy4 (p4, q, digits);
621 q += digits;
622 memset4 (p4, ' ', nblank);
623 dtp->u.p.no_leading_blank = 0;
624 }
625 return;
626 }
627
628 if (nblank < 0)
629 {
630 star_fill (p, w);
631 goto done;
632 }
633
634 if (!dtp->u.p.no_leading_blank)
635 {
636 memset (p, ' ', nblank);
637 p += nblank;
638 memset (p, '0', nzero);
639 p += nzero;
640 memcpy (p, q, digits);
641 }
642 else
643 {
644 memset (p, '0', nzero);
645 p += nzero;
646 memcpy (p, q, digits);
647 p += digits;
648 memset (p, ' ', nblank);
649 dtp->u.p.no_leading_blank = 0;
650 }
651
652 done:
653 return;
654 }
655
656 static void
657 write_decimal (st_parameter_dt *dtp, const fnode *f, const char *source,
658 int len,
659 const char *(*conv) (GFC_INTEGER_LARGEST, char *, size_t))
660 {
661 GFC_INTEGER_LARGEST n = 0;
662 int w, m, digits, nsign, nzero, nblank;
663 char *p;
664 const char *q;
665 sign_t sign;
666 char itoa_buf[GFC_BTOA_BUF_SIZE];
667
668 w = f->u.integer.w;
669 m = f->format == FMT_G ? -1 : f->u.integer.m;
670
671 n = extract_int (source, len);
672
673 /* Special case: */
674 if (m == 0 && n == 0)
675 {
676 if (w == 0)
677 w = 1;
678
679 p = write_block (dtp, w);
680 if (p == NULL)
681 return;
682 if (unlikely (is_char4_unit (dtp)))
683 {
684 gfc_char4_t *p4 = (gfc_char4_t *) p;
685 memset4 (p4, ' ', w);
686 }
687 else
688 memset (p, ' ', w);
689 goto done;
690 }
691
692 sign = calculate_sign (dtp, n < 0);
693 if (n < 0)
694 n = -n;
695 nsign = sign == S_NONE ? 0 : 1;
696
697 /* conv calls itoa which sets the negative sign needed
698 by write_integer. The sign '+' or '-' is set below based on sign
699 calculated above, so we just point past the sign in the string
700 before proceeding to avoid double signs in corner cases.
701 (see PR38504) */
702 q = conv (n, itoa_buf, sizeof (itoa_buf));
703 if (*q == '-')
704 q++;
705
706 digits = strlen (q);
707
708 /* Select a width if none was specified. The idea here is to always
709 print something. */
710
711 if (w == 0)
712 w = ((digits < m) ? m : digits) + nsign;
713
714 p = write_block (dtp, w);
715 if (p == NULL)
716 return;
717
718 nzero = 0;
719 if (digits < m)
720 nzero = m - digits;
721
722 /* See if things will work. */
723
724 nblank = w - (nsign + nzero + digits);
725
726 if (unlikely (is_char4_unit (dtp)))
727 {
728 gfc_char4_t * p4 = (gfc_char4_t *) p;
729 if (nblank < 0)
730 {
731 memset4 (p4, '*', w);
732 goto done;
733 }
734
735 memset4 (p4, ' ', nblank);
736 p4 += nblank;
737
738 switch (sign)
739 {
740 case S_PLUS:
741 *p4++ = '+';
742 break;
743 case S_MINUS:
744 *p4++ = '-';
745 break;
746 case S_NONE:
747 break;
748 }
749
750 memset4 (p4, '0', nzero);
751 p4 += nzero;
752
753 memcpy4 (p4, q, digits);
754 return;
755 }
756
757 if (nblank < 0)
758 {
759 star_fill (p, w);
760 goto done;
761 }
762
763 memset (p, ' ', nblank);
764 p += nblank;
765
766 switch (sign)
767 {
768 case S_PLUS:
769 *p++ = '+';
770 break;
771 case S_MINUS:
772 *p++ = '-';
773 break;
774 case S_NONE:
775 break;
776 }
777
778 memset (p, '0', nzero);
779 p += nzero;
780
781 memcpy (p, q, digits);
782
783 done:
784 return;
785 }
786
787
788 /* Convert unsigned octal to ascii. */
789
790 static const char *
791 otoa (GFC_UINTEGER_LARGEST n, char *buffer, size_t len)
792 {
793 char *p;
794
795 assert (len >= GFC_OTOA_BUF_SIZE);
796
797 if (n == 0)
798 return "0";
799
800 p = buffer + GFC_OTOA_BUF_SIZE - 1;
801 *p = '\0';
802
803 while (n != 0)
804 {
805 *--p = '0' + (n & 7);
806 n >>= 3;
807 }
808
809 return p;
810 }
811
812
813 /* Convert unsigned binary to ascii. */
814
815 static const char *
816 btoa (GFC_UINTEGER_LARGEST n, char *buffer, size_t len)
817 {
818 char *p;
819
820 assert (len >= GFC_BTOA_BUF_SIZE);
821
822 if (n == 0)
823 return "0";
824
825 p = buffer + GFC_BTOA_BUF_SIZE - 1;
826 *p = '\0';
827
828 while (n != 0)
829 {
830 *--p = '0' + (n & 1);
831 n >>= 1;
832 }
833
834 return p;
835 }
836
837 /* The following three functions, btoa_big, otoa_big, and ztoa_big, are needed
838 to convert large reals with kind sizes that exceed the largest integer type
839 available on certain platforms. In these cases, byte by byte conversion is
840 performed. Endianess is taken into account. */
841
842 /* Conversion to binary. */
843
844 static const char *
845 btoa_big (const char *s, char *buffer, int len, GFC_UINTEGER_LARGEST *n)
846 {
847 char *q;
848 int i, j;
849
850 q = buffer;
851 if (big_endian)
852 {
853 const char *p = s;
854 for (i = 0; i < len; i++)
855 {
856 char c = *p;
857
858 /* Test for zero. Needed by write_boz later. */
859 if (*p != 0)
860 *n = 1;
861
862 for (j = 0; j < 8; j++)
863 {
864 *q++ = (c & 128) ? '1' : '0';
865 c <<= 1;
866 }
867 p++;
868 }
869 }
870 else
871 {
872 const char *p = s + len - 1;
873 for (i = 0; i < len; i++)
874 {
875 char c = *p;
876
877 /* Test for zero. Needed by write_boz later. */
878 if (*p != 0)
879 *n = 1;
880
881 for (j = 0; j < 8; j++)
882 {
883 *q++ = (c & 128) ? '1' : '0';
884 c <<= 1;
885 }
886 p--;
887 }
888 }
889
890 *q = '\0';
891
892 if (*n == 0)
893 return "0";
894
895 /* Move past any leading zeros. */
896 while (*buffer == '0')
897 buffer++;
898
899 return buffer;
900
901 }
902
903 /* Conversion to octal. */
904
905 static const char *
906 otoa_big (const char *s, char *buffer, int len, GFC_UINTEGER_LARGEST *n)
907 {
908 char *q;
909 int i, j, k;
910 uint8_t octet;
911
912 q = buffer + GFC_OTOA_BUF_SIZE - 1;
913 *q = '\0';
914 i = k = octet = 0;
915
916 if (big_endian)
917 {
918 const char *p = s + len - 1;
919 char c = *p;
920 while (i < len)
921 {
922 /* Test for zero. Needed by write_boz later. */
923 if (*p != 0)
924 *n = 1;
925
926 for (j = 0; j < 3 && i < len; j++)
927 {
928 octet |= (c & 1) << j;
929 c >>= 1;
930 if (++k > 7)
931 {
932 i++;
933 k = 0;
934 c = *--p;
935 }
936 }
937 *--q = '0' + octet;
938 octet = 0;
939 }
940 }
941 else
942 {
943 const char *p = s;
944 char c = *p;
945 while (i < len)
946 {
947 /* Test for zero. Needed by write_boz later. */
948 if (*p != 0)
949 *n = 1;
950
951 for (j = 0; j < 3 && i < len; j++)
952 {
953 octet |= (c & 1) << j;
954 c >>= 1;
955 if (++k > 7)
956 {
957 i++;
958 k = 0;
959 c = *++p;
960 }
961 }
962 *--q = '0' + octet;
963 octet = 0;
964 }
965 }
966
967 if (*n == 0)
968 return "0";
969
970 /* Move past any leading zeros. */
971 while (*q == '0')
972 q++;
973
974 return q;
975 }
976
977 /* Conversion to hexidecimal. */
978
979 static const char *
980 ztoa_big (const char *s, char *buffer, int len, GFC_UINTEGER_LARGEST *n)
981 {
982 static char a[16] = {'0', '1', '2', '3', '4', '5', '6', '7',
983 '8', '9', 'A', 'B', 'C', 'D', 'E', 'F'};
984
985 char *q;
986 uint8_t h, l;
987 int i;
988
989 q = buffer;
990
991 if (big_endian)
992 {
993 const char *p = s;
994 for (i = 0; i < len; i++)
995 {
996 /* Test for zero. Needed by write_boz later. */
997 if (*p != 0)
998 *n = 1;
999
1000 h = (*p >> 4) & 0x0F;
1001 l = *p++ & 0x0F;
1002 *q++ = a[h];
1003 *q++ = a[l];
1004 }
1005 }
1006 else
1007 {
1008 const char *p = s + len - 1;
1009 for (i = 0; i < len; i++)
1010 {
1011 /* Test for zero. Needed by write_boz later. */
1012 if (*p != 0)
1013 *n = 1;
1014
1015 h = (*p >> 4) & 0x0F;
1016 l = *p-- & 0x0F;
1017 *q++ = a[h];
1018 *q++ = a[l];
1019 }
1020 }
1021
1022 *q = '\0';
1023
1024 if (*n == 0)
1025 return "0";
1026
1027 /* Move past any leading zeros. */
1028 while (*buffer == '0')
1029 buffer++;
1030
1031 return buffer;
1032 }
1033
1034 /* gfc_itoa()-- Integer to decimal conversion.
1035 The itoa function is a widespread non-standard extension to standard
1036 C, often declared in <stdlib.h>. Even though the itoa defined here
1037 is a static function we take care not to conflict with any prior
1038 non-static declaration. Hence the 'gfc_' prefix, which is normally
1039 reserved for functions with external linkage. */
1040
1041 static const char *
1042 gfc_itoa (GFC_INTEGER_LARGEST n, char *buffer, size_t len)
1043 {
1044 int negative;
1045 char *p;
1046 GFC_UINTEGER_LARGEST t;
1047
1048 assert (len >= GFC_ITOA_BUF_SIZE);
1049
1050 if (n == 0)
1051 return "0";
1052
1053 negative = 0;
1054 t = n;
1055 if (n < 0)
1056 {
1057 negative = 1;
1058 t = -n; /*must use unsigned to protect from overflow*/
1059 }
1060
1061 p = buffer + GFC_ITOA_BUF_SIZE - 1;
1062 *p = '\0';
1063
1064 while (t != 0)
1065 {
1066 *--p = '0' + (t % 10);
1067 t /= 10;
1068 }
1069
1070 if (negative)
1071 *--p = '-';
1072 return p;
1073 }
1074
1075
1076 void
1077 write_i (st_parameter_dt *dtp, const fnode *f, const char *p, int len)
1078 {
1079 write_decimal (dtp, f, p, len, (void *) gfc_itoa);
1080 }
1081
1082
1083 void
1084 write_b (st_parameter_dt *dtp, const fnode *f, const char *source, int len)
1085 {
1086 const char *p;
1087 char itoa_buf[GFC_BTOA_BUF_SIZE];
1088 GFC_UINTEGER_LARGEST n = 0;
1089
1090 if (len > (int) sizeof (GFC_UINTEGER_LARGEST))
1091 {
1092 p = btoa_big (source, itoa_buf, len, &n);
1093 write_boz (dtp, f, p, n);
1094 }
1095 else
1096 {
1097 n = extract_uint (source, len);
1098 p = btoa (n, itoa_buf, sizeof (itoa_buf));
1099 write_boz (dtp, f, p, n);
1100 }
1101 }
1102
1103
1104 void
1105 write_o (st_parameter_dt *dtp, const fnode *f, const char *source, int len)
1106 {
1107 const char *p;
1108 char itoa_buf[GFC_OTOA_BUF_SIZE];
1109 GFC_UINTEGER_LARGEST n = 0;
1110
1111 if (len > (int) sizeof (GFC_UINTEGER_LARGEST))
1112 {
1113 p = otoa_big (source, itoa_buf, len, &n);
1114 write_boz (dtp, f, p, n);
1115 }
1116 else
1117 {
1118 n = extract_uint (source, len);
1119 p = otoa (n, itoa_buf, sizeof (itoa_buf));
1120 write_boz (dtp, f, p, n);
1121 }
1122 }
1123
1124 void
1125 write_z (st_parameter_dt *dtp, const fnode *f, const char *source, int len)
1126 {
1127 const char *p;
1128 char itoa_buf[GFC_XTOA_BUF_SIZE];
1129 GFC_UINTEGER_LARGEST n = 0;
1130
1131 if (len > (int) sizeof (GFC_UINTEGER_LARGEST))
1132 {
1133 p = ztoa_big (source, itoa_buf, len, &n);
1134 write_boz (dtp, f, p, n);
1135 }
1136 else
1137 {
1138 n = extract_uint (source, len);
1139 p = gfc_xtoa (n, itoa_buf, sizeof (itoa_buf));
1140 write_boz (dtp, f, p, n);
1141 }
1142 }
1143
1144
1145 void
1146 write_d (st_parameter_dt *dtp, const fnode *f, const char *p, int len)
1147 {
1148 write_float (dtp, f, p, len, 0);
1149 }
1150
1151
1152 void
1153 write_e (st_parameter_dt *dtp, const fnode *f, const char *p, int len)
1154 {
1155 write_float (dtp, f, p, len, 0);
1156 }
1157
1158
1159 void
1160 write_f (st_parameter_dt *dtp, const fnode *f, const char *p, int len)
1161 {
1162 write_float (dtp, f, p, len, 0);
1163 }
1164
1165
1166 void
1167 write_en (st_parameter_dt *dtp, const fnode *f, const char *p, int len)
1168 {
1169 write_float (dtp, f, p, len, 0);
1170 }
1171
1172
1173 void
1174 write_es (st_parameter_dt *dtp, const fnode *f, const char *p, int len)
1175 {
1176 write_float (dtp, f, p, len, 0);
1177 }
1178
1179
1180 /* Take care of the X/TR descriptor. */
1181
1182 void
1183 write_x (st_parameter_dt *dtp, int len, int nspaces)
1184 {
1185 char *p;
1186
1187 p = write_block (dtp, len);
1188 if (p == NULL)
1189 return;
1190 if (nspaces > 0 && len - nspaces >= 0)
1191 {
1192 if (unlikely (is_char4_unit (dtp)))
1193 {
1194 gfc_char4_t *p4 = (gfc_char4_t *) p;
1195 memset4 (&p4[len - nspaces], ' ', nspaces);
1196 }
1197 else
1198 memset (&p[len - nspaces], ' ', nspaces);
1199 }
1200 }
1201
1202
1203 /* List-directed writing. */
1204
1205
1206 /* Write a single character to the output. Returns nonzero if
1207 something goes wrong. */
1208
1209 static int
1210 write_char (st_parameter_dt *dtp, int c)
1211 {
1212 char *p;
1213
1214 p = write_block (dtp, 1);
1215 if (p == NULL)
1216 return 1;
1217 if (unlikely (is_char4_unit (dtp)))
1218 {
1219 gfc_char4_t *p4 = (gfc_char4_t *) p;
1220 *p4 = c;
1221 return 0;
1222 }
1223
1224 *p = (uchar) c;
1225
1226 return 0;
1227 }
1228
1229
1230 /* Write a list-directed logical value. */
1231
1232 static void
1233 write_logical (st_parameter_dt *dtp, const char *source, int length)
1234 {
1235 write_char (dtp, extract_int (source, length) ? 'T' : 'F');
1236 }
1237
1238
1239 /* Write a list-directed integer value. */
1240
1241 static void
1242 write_integer (st_parameter_dt *dtp, const char *source, int length)
1243 {
1244 char *p;
1245 const char *q;
1246 int digits;
1247 int width;
1248 char itoa_buf[GFC_ITOA_BUF_SIZE];
1249
1250 q = gfc_itoa (extract_int (source, length), itoa_buf, sizeof (itoa_buf));
1251
1252 switch (length)
1253 {
1254 case 1:
1255 width = 4;
1256 break;
1257
1258 case 2:
1259 width = 6;
1260 break;
1261
1262 case 4:
1263 width = 11;
1264 break;
1265
1266 case 8:
1267 width = 20;
1268 break;
1269
1270 default:
1271 width = 0;
1272 break;
1273 }
1274
1275 digits = strlen (q);
1276
1277 if (width < digits)
1278 width = digits;
1279 p = write_block (dtp, width);
1280 if (p == NULL)
1281 return;
1282
1283 if (unlikely (is_char4_unit (dtp)))
1284 {
1285 gfc_char4_t *p4 = (gfc_char4_t *) p;
1286 if (dtp->u.p.no_leading_blank)
1287 {
1288 memcpy4 (p4, q, digits);
1289 memset4 (p4 + digits, ' ', width - digits);
1290 }
1291 else
1292 {
1293 memset4 (p4, ' ', width - digits);
1294 memcpy4 (p4 + width - digits, q, digits);
1295 }
1296 return;
1297 }
1298
1299 if (dtp->u.p.no_leading_blank)
1300 {
1301 memcpy (p, q, digits);
1302 memset (p + digits, ' ', width - digits);
1303 }
1304 else
1305 {
1306 memset (p, ' ', width - digits);
1307 memcpy (p + width - digits, q, digits);
1308 }
1309 }
1310
1311
1312 /* Write a list-directed string. We have to worry about delimiting
1313 the strings if the file has been opened in that mode. */
1314
1315 #define DELIM 1
1316 #define NODELIM 0
1317
1318 static void
1319 write_character (st_parameter_dt *dtp, const char *source, int kind, int length, int mode)
1320 {
1321 int i, extra;
1322 char *p, d;
1323
1324 if (mode == DELIM)
1325 {
1326 switch (dtp->u.p.current_unit->delim_status)
1327 {
1328 case DELIM_APOSTROPHE:
1329 d = '\'';
1330 break;
1331 case DELIM_QUOTE:
1332 d = '"';
1333 break;
1334 default:
1335 d = ' ';
1336 break;
1337 }
1338 }
1339 else
1340 d = ' ';
1341
1342 if (kind == 1)
1343 {
1344 if (d == ' ')
1345 extra = 0;
1346 else
1347 {
1348 extra = 2;
1349
1350 for (i = 0; i < length; i++)
1351 if (source[i] == d)
1352 extra++;
1353 }
1354
1355 p = write_block (dtp, length + extra);
1356 if (p == NULL)
1357 return;
1358
1359 if (unlikely (is_char4_unit (dtp)))
1360 {
1361 gfc_char4_t d4 = (gfc_char4_t) d;
1362 gfc_char4_t *p4 = (gfc_char4_t *) p;
1363
1364 if (d4 == ' ')
1365 memcpy4 (p4, source, length);
1366 else
1367 {
1368 *p4++ = d4;
1369
1370 for (i = 0; i < length; i++)
1371 {
1372 *p4++ = (gfc_char4_t) source[i];
1373 if (source[i] == d)
1374 *p4++ = d4;
1375 }
1376
1377 *p4 = d4;
1378 }
1379 return;
1380 }
1381
1382 if (d == ' ')
1383 memcpy (p, source, length);
1384 else
1385 {
1386 *p++ = d;
1387
1388 for (i = 0; i < length; i++)
1389 {
1390 *p++ = source[i];
1391 if (source[i] == d)
1392 *p++ = d;
1393 }
1394
1395 *p = d;
1396 }
1397 }
1398 else
1399 {
1400 if (d == ' ')
1401 {
1402 if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8)
1403 write_utf8_char4 (dtp, (gfc_char4_t *) source, length, 0);
1404 else
1405 write_default_char4 (dtp, (gfc_char4_t *) source, length, 0);
1406 }
1407 else
1408 {
1409 p = write_block (dtp, 1);
1410 *p = d;
1411
1412 if (dtp->u.p.current_unit->flags.encoding == ENCODING_UTF8)
1413 write_utf8_char4 (dtp, (gfc_char4_t *) source, length, 0);
1414 else
1415 write_default_char4 (dtp, (gfc_char4_t *) source, length, 0);
1416
1417 p = write_block (dtp, 1);
1418 *p = d;
1419 }
1420 }
1421 }
1422
1423
1424 /* Set an fnode to default format. */
1425
1426 static void
1427 set_fnode_default (st_parameter_dt *dtp, fnode *f, int length)
1428 {
1429 f->format = FMT_G;
1430 switch (length)
1431 {
1432 case 4:
1433 f->u.real.w = 16;
1434 f->u.real.d = 9;
1435 f->u.real.e = 2;
1436 break;
1437 case 8:
1438 f->u.real.w = 25;
1439 f->u.real.d = 17;
1440 f->u.real.e = 3;
1441 break;
1442 case 10:
1443 f->u.real.w = 30;
1444 f->u.real.d = 21;
1445 f->u.real.e = 4;
1446 break;
1447 case 16:
1448 f->u.real.w = 45;
1449 f->u.real.d = 36;
1450 f->u.real.e = 4;
1451 break;
1452 default:
1453 internal_error (&dtp->common, "bad real kind");
1454 break;
1455 }
1456 }
1457
1458 /* Output a real number with default format. To guarantee that a
1459 binary -> decimal -> binary roundtrip conversion recovers the
1460 original value, IEEE 754-2008 requires 9, 17, 21 and 36 significant
1461 digits for REAL kinds 4, 8, 10, and 16, respectively. Thus, we use
1462 1PG16.9E2 for REAL(4), 1PG25.17E3 for REAL(8), 1PG30.21E4 for
1463 REAL(10) and 1PG45.36E4 for REAL(16). The exception is that the
1464 Fortran standard requires outputting an extra digit when the scale
1465 factor is 1 and when the magnitude of the value is such that E
1466 editing is used. However, gfortran compensates for this, and thus
1467 for list formatted the same number of significant digits is
1468 generated both when using F and E editing. */
1469
1470 void
1471 write_real (st_parameter_dt *dtp, const char *source, int length)
1472 {
1473 fnode f ;
1474 int org_scale = dtp->u.p.scale_factor;
1475 dtp->u.p.scale_factor = 1;
1476 set_fnode_default (dtp, &f, length);
1477 write_float (dtp, &f, source , length, 1);
1478 dtp->u.p.scale_factor = org_scale;
1479 }
1480
1481 /* Similar to list formatted REAL output, for kPG0 where k > 0 we
1482 compensate for the extra digit. */
1483
1484 void
1485 write_real_g0 (st_parameter_dt *dtp, const char *source, int length, int d)
1486 {
1487 fnode f;
1488 int comp_d;
1489 set_fnode_default (dtp, &f, length);
1490 if (d > 0)
1491 f.u.real.d = d;
1492
1493 /* Compensate for extra digits when using scale factor, d is not
1494 specified, and the magnitude is such that E editing is used. */
1495 if (dtp->u.p.scale_factor > 0 && d == 0)
1496 comp_d = 1;
1497 else
1498 comp_d = 0;
1499 dtp->u.p.g0_no_blanks = 1;
1500 write_float (dtp, &f, source , length, comp_d);
1501 dtp->u.p.g0_no_blanks = 0;
1502 }
1503
1504
1505 static void
1506 write_complex (st_parameter_dt *dtp, const char *source, int kind, size_t size)
1507 {
1508 char semi_comma =
1509 dtp->u.p.current_unit->decimal_status == DECIMAL_POINT ? ',' : ';';
1510
1511 if (write_char (dtp, '('))
1512 return;
1513 write_real (dtp, source, kind);
1514
1515 if (write_char (dtp, semi_comma))
1516 return;
1517 write_real (dtp, source + size / 2, kind);
1518
1519 write_char (dtp, ')');
1520 }
1521
1522
1523 /* Write the separator between items. */
1524
1525 static void
1526 write_separator (st_parameter_dt *dtp)
1527 {
1528 char *p;
1529
1530 p = write_block (dtp, options.separator_len);
1531 if (p == NULL)
1532 return;
1533 if (unlikely (is_char4_unit (dtp)))
1534 {
1535 gfc_char4_t *p4 = (gfc_char4_t *) p;
1536 memcpy4 (p4, options.separator, options.separator_len);
1537 }
1538 else
1539 memcpy (p, options.separator, options.separator_len);
1540 }
1541
1542
1543 /* Write an item with list formatting.
1544 TODO: handle skipping to the next record correctly, particularly
1545 with strings. */
1546
1547 static void
1548 list_formatted_write_scalar (st_parameter_dt *dtp, bt type, void *p, int kind,
1549 size_t size)
1550 {
1551 if (dtp->u.p.current_unit == NULL)
1552 return;
1553
1554 if (dtp->u.p.first_item)
1555 {
1556 dtp->u.p.first_item = 0;
1557 write_char (dtp, ' ');
1558 }
1559 else
1560 {
1561 if (type != BT_CHARACTER || !dtp->u.p.char_flag ||
1562 (dtp->u.p.current_unit->delim_status != DELIM_NONE
1563 && dtp->u.p.current_unit->delim_status != DELIM_UNSPECIFIED))
1564 write_separator (dtp);
1565 }
1566
1567 switch (type)
1568 {
1569 case BT_INTEGER:
1570 write_integer (dtp, p, kind);
1571 break;
1572 case BT_LOGICAL:
1573 write_logical (dtp, p, kind);
1574 break;
1575 case BT_CHARACTER:
1576 write_character (dtp, p, kind, size, DELIM);
1577 break;
1578 case BT_REAL:
1579 write_real (dtp, p, kind);
1580 break;
1581 case BT_COMPLEX:
1582 write_complex (dtp, p, kind, size);
1583 break;
1584 default:
1585 internal_error (&dtp->common, "list_formatted_write(): Bad type");
1586 }
1587
1588 dtp->u.p.char_flag = (type == BT_CHARACTER);
1589 }
1590
1591
1592 void
1593 list_formatted_write (st_parameter_dt *dtp, bt type, void *p, int kind,
1594 size_t size, size_t nelems)
1595 {
1596 size_t elem;
1597 char *tmp;
1598 size_t stride = type == BT_CHARACTER ?
1599 size * GFC_SIZE_OF_CHAR_KIND(kind) : size;
1600
1601 tmp = (char *) p;
1602
1603 /* Big loop over all the elements. */
1604 for (elem = 0; elem < nelems; elem++)
1605 {
1606 dtp->u.p.item_count++;
1607 list_formatted_write_scalar (dtp, type, tmp + elem * stride, kind, size);
1608 }
1609 }
1610
1611 /* NAMELIST OUTPUT
1612
1613 nml_write_obj writes a namelist object to the output stream. It is called
1614 recursively for derived type components:
1615 obj = is the namelist_info for the current object.
1616 offset = the offset relative to the address held by the object for
1617 derived type arrays.
1618 base = is the namelist_info of the derived type, when obj is a
1619 component.
1620 base_name = the full name for a derived type, including qualifiers
1621 if any.
1622 The returned value is a pointer to the object beyond the last one
1623 accessed, including nested derived types. Notice that the namelist is
1624 a linear linked list of objects, including derived types and their
1625 components. A tree, of sorts, is implied by the compound names of
1626 the derived type components and this is how this function recurses through
1627 the list. */
1628
1629 /* A generous estimate of the number of characters needed to print
1630 repeat counts and indices, including commas, asterices and brackets. */
1631
1632 #define NML_DIGITS 20
1633
1634 static void
1635 namelist_write_newline (st_parameter_dt *dtp)
1636 {
1637 if (!is_internal_unit (dtp))
1638 {
1639 #ifdef HAVE_CRLF
1640 write_character (dtp, "\r\n", 1, 2, NODELIM);
1641 #else
1642 write_character (dtp, "\n", 1, 1, NODELIM);
1643 #endif
1644 return;
1645 }
1646
1647 if (is_array_io (dtp))
1648 {
1649 gfc_offset record;
1650 int finished;
1651 char *p;
1652 int length = dtp->u.p.current_unit->bytes_left;
1653
1654 p = write_block (dtp, length);
1655 if (p == NULL)
1656 return;
1657
1658 if (unlikely (is_char4_unit (dtp)))
1659 {
1660 gfc_char4_t *p4 = (gfc_char4_t *) p;
1661 memset4 (p4, ' ', length);
1662 }
1663 else
1664 memset (p, ' ', length);
1665
1666 /* Now that the current record has been padded out,
1667 determine where the next record in the array is. */
1668 record = next_array_record (dtp, dtp->u.p.current_unit->ls,
1669 &finished);
1670 if (finished)
1671 dtp->u.p.current_unit->endfile = AT_ENDFILE;
1672 else
1673 {
1674 /* Now seek to this record */
1675 record = record * dtp->u.p.current_unit->recl;
1676
1677 if (sseek (dtp->u.p.current_unit->s, record, SEEK_SET) < 0)
1678 {
1679 generate_error (&dtp->common, LIBERROR_INTERNAL_UNIT, NULL);
1680 return;
1681 }
1682
1683 dtp->u.p.current_unit->bytes_left = dtp->u.p.current_unit->recl;
1684 }
1685 }
1686 else
1687 write_character (dtp, " ", 1, 1, NODELIM);
1688 }
1689
1690
1691 static namelist_info *
1692 nml_write_obj (st_parameter_dt *dtp, namelist_info * obj, index_type offset,
1693 namelist_info * base, char * base_name)
1694 {
1695 int rep_ctr;
1696 int num;
1697 int nml_carry;
1698 int len;
1699 index_type obj_size;
1700 index_type nelem;
1701 size_t dim_i;
1702 size_t clen;
1703 index_type elem_ctr;
1704 size_t obj_name_len;
1705 void * p ;
1706 char cup;
1707 char * obj_name;
1708 char * ext_name;
1709 size_t ext_name_len;
1710 char rep_buff[NML_DIGITS];
1711 namelist_info * cmp;
1712 namelist_info * retval = obj->next;
1713 size_t base_name_len;
1714 size_t base_var_name_len;
1715 size_t tot_len;
1716
1717 /* Set the character to be used to separate values
1718 to a comma or semi-colon. */
1719
1720 char semi_comma =
1721 dtp->u.p.current_unit->decimal_status == DECIMAL_POINT ? ',' : ';';
1722
1723 /* Write namelist variable names in upper case. If a derived type,
1724 nothing is output. If a component, base and base_name are set. */
1725
1726 if (obj->type != BT_DERIVED)
1727 {
1728 namelist_write_newline (dtp);
1729 write_character (dtp, " ", 1, 1, NODELIM);
1730
1731 len = 0;
1732 if (base)
1733 {
1734 len = strlen (base->var_name);
1735 base_name_len = strlen (base_name);
1736 for (dim_i = 0; dim_i < base_name_len; dim_i++)
1737 {
1738 cup = toupper ((int) base_name[dim_i]);
1739 write_character (dtp, &cup, 1, 1, NODELIM);
1740 }
1741 }
1742 clen = strlen (obj->var_name);
1743 for (dim_i = len; dim_i < clen; dim_i++)
1744 {
1745 cup = toupper ((int) obj->var_name[dim_i]);
1746 write_character (dtp, &cup, 1, 1, NODELIM);
1747 }
1748 write_character (dtp, "=", 1, 1, NODELIM);
1749 }
1750
1751 /* Counts the number of data output on a line, including names. */
1752
1753 num = 1;
1754
1755 len = obj->len;
1756
1757 switch (obj->type)
1758 {
1759
1760 case BT_REAL:
1761 obj_size = size_from_real_kind (len);
1762 break;
1763
1764 case BT_COMPLEX:
1765 obj_size = size_from_complex_kind (len);
1766 break;
1767
1768 case BT_CHARACTER:
1769 obj_size = obj->string_length;
1770 break;
1771
1772 default:
1773 obj_size = len;
1774 }
1775
1776 if (obj->var_rank)
1777 obj_size = obj->size;
1778
1779 /* Set the index vector and count the number of elements. */
1780
1781 nelem = 1;
1782 for (dim_i = 0; dim_i < (size_t) obj->var_rank; dim_i++)
1783 {
1784 obj->ls[dim_i].idx = GFC_DESCRIPTOR_LBOUND(obj, dim_i);
1785 nelem = nelem * GFC_DESCRIPTOR_EXTENT (obj, dim_i);
1786 }
1787
1788 /* Main loop to output the data held in the object. */
1789
1790 rep_ctr = 1;
1791 for (elem_ctr = 0; elem_ctr < nelem; elem_ctr++)
1792 {
1793
1794 /* Build the pointer to the data value. The offset is passed by
1795 recursive calls to this function for arrays of derived types.
1796 Is NULL otherwise. */
1797
1798 p = (void *)(obj->mem_pos + elem_ctr * obj_size);
1799 p += offset;
1800
1801 /* Check for repeat counts of intrinsic types. */
1802
1803 if ((elem_ctr < (nelem - 1)) &&
1804 (obj->type != BT_DERIVED) &&
1805 !memcmp (p, (void*)(p + obj_size ), obj_size ))
1806 {
1807 rep_ctr++;
1808 }
1809
1810 /* Execute a repeated output. Note the flag no_leading_blank that
1811 is used in the functions used to output the intrinsic types. */
1812
1813 else
1814 {
1815 if (rep_ctr > 1)
1816 {
1817 snprintf(rep_buff, NML_DIGITS, " %d*", rep_ctr);
1818 write_character (dtp, rep_buff, 1, strlen (rep_buff), NODELIM);
1819 dtp->u.p.no_leading_blank = 1;
1820 }
1821 num++;
1822
1823 /* Output the data, if an intrinsic type, or recurse into this
1824 routine to treat derived types. */
1825
1826 switch (obj->type)
1827 {
1828
1829 case BT_INTEGER:
1830 write_integer (dtp, p, len);
1831 break;
1832
1833 case BT_LOGICAL:
1834 write_logical (dtp, p, len);
1835 break;
1836
1837 case BT_CHARACTER:
1838 write_character (dtp, p, 1, obj->string_length, DELIM);
1839 break;
1840
1841 case BT_REAL:
1842 write_real (dtp, p, len);
1843 break;
1844
1845 case BT_COMPLEX:
1846 dtp->u.p.no_leading_blank = 0;
1847 num++;
1848 write_complex (dtp, p, len, obj_size);
1849 break;
1850
1851 case BT_DERIVED:
1852
1853 /* To treat a derived type, we need to build two strings:
1854 ext_name = the name, including qualifiers that prepends
1855 component names in the output - passed to
1856 nml_write_obj.
1857 obj_name = the derived type name with no qualifiers but %
1858 appended. This is used to identify the
1859 components. */
1860
1861 /* First ext_name => get length of all possible components */
1862
1863 base_name_len = base_name ? strlen (base_name) : 0;
1864 base_var_name_len = base ? strlen (base->var_name) : 0;
1865 ext_name_len = base_name_len + base_var_name_len
1866 + strlen (obj->var_name) + obj->var_rank * NML_DIGITS + 1;
1867 ext_name = (char*)xmalloc (ext_name_len);
1868
1869 memcpy (ext_name, base_name, base_name_len);
1870 clen = strlen (obj->var_name + base_var_name_len);
1871 memcpy (ext_name + base_name_len,
1872 obj->var_name + base_var_name_len, clen);
1873
1874 /* Append the qualifier. */
1875
1876 tot_len = base_name_len + clen;
1877 for (dim_i = 0; dim_i < (size_t) obj->var_rank; dim_i++)
1878 {
1879 if (!dim_i)
1880 {
1881 ext_name[tot_len] = '(';
1882 tot_len++;
1883 }
1884 snprintf (ext_name + tot_len, ext_name_len - tot_len, "%d",
1885 (int) obj->ls[dim_i].idx);
1886 tot_len += strlen (ext_name + tot_len);
1887 ext_name[tot_len] = ((int) dim_i == obj->var_rank - 1) ? ')' : ',';
1888 tot_len++;
1889 }
1890
1891 ext_name[tot_len] = '\0';
1892
1893 /* Now obj_name. */
1894
1895 obj_name_len = strlen (obj->var_name) + 1;
1896 obj_name = xmalloc (obj_name_len+1);
1897 memcpy (obj_name, obj->var_name, obj_name_len-1);
1898 memcpy (obj_name + obj_name_len-1, "%", 2);
1899
1900 /* Now loop over the components. Update the component pointer
1901 with the return value from nml_write_obj => this loop jumps
1902 past nested derived types. */
1903
1904 for (cmp = obj->next;
1905 cmp && !strncmp (cmp->var_name, obj_name, obj_name_len);
1906 cmp = retval)
1907 {
1908 retval = nml_write_obj (dtp, cmp,
1909 (index_type)(p - obj->mem_pos),
1910 obj, ext_name);
1911 }
1912
1913 free (obj_name);
1914 free (ext_name);
1915 goto obj_loop;
1916
1917 default:
1918 internal_error (&dtp->common, "Bad type for namelist write");
1919 }
1920
1921 /* Reset the leading blank suppression, write a comma (or semi-colon)
1922 and, if 5 values have been output, write a newline and advance
1923 to column 2. Reset the repeat counter. */
1924
1925 dtp->u.p.no_leading_blank = 0;
1926 if (obj->type == BT_CHARACTER)
1927 {
1928 if (dtp->u.p.nml_delim != '\0')
1929 write_character (dtp, &semi_comma, 1, 1, NODELIM);
1930 }
1931 else
1932 write_character (dtp, &semi_comma, 1, 1, NODELIM);
1933 if (num > 5)
1934 {
1935 num = 0;
1936 if (dtp->u.p.nml_delim == '\0')
1937 write_character (dtp, &semi_comma, 1, 1, NODELIM);
1938 namelist_write_newline (dtp);
1939 write_character (dtp, " ", 1, 1, NODELIM);
1940 }
1941 rep_ctr = 1;
1942 }
1943
1944 /* Cycle through and increment the index vector. */
1945
1946 obj_loop:
1947
1948 nml_carry = 1;
1949 for (dim_i = 0; nml_carry && (dim_i < (size_t) obj->var_rank); dim_i++)
1950 {
1951 obj->ls[dim_i].idx += nml_carry ;
1952 nml_carry = 0;
1953 if (obj->ls[dim_i].idx > GFC_DESCRIPTOR_UBOUND(obj,dim_i))
1954 {
1955 obj->ls[dim_i].idx = GFC_DESCRIPTOR_LBOUND(obj,dim_i);
1956 nml_carry = 1;
1957 }
1958 }
1959 }
1960
1961 /* Return a pointer beyond the furthest object accessed. */
1962
1963 return retval;
1964 }
1965
1966
1967 /* This is the entry function for namelist writes. It outputs the name
1968 of the namelist and iterates through the namelist by calls to
1969 nml_write_obj. The call below has dummys in the arguments used in
1970 the treatment of derived types. */
1971
1972 void
1973 namelist_write (st_parameter_dt *dtp)
1974 {
1975 namelist_info * t1, *t2, *dummy = NULL;
1976 index_type i;
1977 index_type dummy_offset = 0;
1978 char c;
1979 char * dummy_name = NULL;
1980
1981 /* Set the delimiter for namelist output. */
1982 switch (dtp->u.p.current_unit->delim_status)
1983 {
1984 case DELIM_APOSTROPHE:
1985 dtp->u.p.nml_delim = '\'';
1986 break;
1987 case DELIM_QUOTE:
1988 case DELIM_UNSPECIFIED:
1989 dtp->u.p.nml_delim = '"';
1990 break;
1991 default:
1992 dtp->u.p.nml_delim = '\0';
1993 }
1994
1995 write_character (dtp, "&", 1, 1, NODELIM);
1996
1997 /* Write namelist name in upper case - f95 std. */
1998 for (i = 0 ;i < dtp->namelist_name_len ;i++ )
1999 {
2000 c = toupper ((int) dtp->namelist_name[i]);
2001 write_character (dtp, &c, 1 ,1, NODELIM);
2002 }
2003
2004 if (dtp->u.p.ionml != NULL)
2005 {
2006 t1 = dtp->u.p.ionml;
2007 while (t1 != NULL)
2008 {
2009 t2 = t1;
2010 t1 = nml_write_obj (dtp, t2, dummy_offset, dummy, dummy_name);
2011 }
2012 }
2013
2014 namelist_write_newline (dtp);
2015 write_character (dtp, " /", 1, 2, NODELIM);
2016 }
2017
2018 #undef NML_DIGITS