1 /* Copyright (C) 2002-2016 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
6 This file is part of the GNU Fortran runtime library (libgfortran).
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)
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.
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.
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/>. */
36 #define star_fill(p, n) memset(p, '*', n)
38 typedef unsigned char uchar
;
40 /* Helper functions for character(kind=4) internal units. These are needed
41 by write_float.def. */
44 memcpy4 (gfc_char4_t
*dest
, const char *source
, int k
)
48 const char *p
= source
;
49 for (j
= 0; j
< k
; j
++)
50 *dest
++ = (gfc_char4_t
) *p
++;
53 /* This include contains the heart and soul of formatted floating point. */
54 #include "write_float.def"
56 /* Write out default char4. */
59 write_default_char4 (st_parameter_dt
*dtp
, const gfc_char4_t
*source
,
60 int src_len
, int w_len
)
67 /* Take care of preceding blanks. */
71 p
= write_block (dtp
, k
);
74 if (is_char4_unit (dtp
))
76 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
83 /* Get ready to handle delimiters if needed. */
84 switch (dtp
->u
.p
.current_unit
->delim_status
)
86 case DELIM_APOSTROPHE
:
97 /* Now process the remaining characters, one at a time. */
98 for (j
= 0; j
< src_len
; j
++)
101 if (is_char4_unit (dtp
))
104 /* Handle delimiters if any. */
105 if (c
== d
&& d
!= ' ')
107 p
= write_block (dtp
, 2);
110 q
= (gfc_char4_t
*) p
;
115 p
= write_block (dtp
, 1);
118 q
= (gfc_char4_t
*) p
;
124 /* Handle delimiters if any. */
125 if (c
== d
&& d
!= ' ')
127 p
= write_block (dtp
, 2);
134 p
= write_block (dtp
, 1);
138 *p
= c
> 255 ? '?' : (uchar
) c
;
144 /* Write out UTF-8 converted from char4. */
147 write_utf8_char4 (st_parameter_dt
*dtp
, gfc_char4_t
*source
,
148 int src_len
, int w_len
)
153 static const uchar masks
[6] = { 0x00, 0xC0, 0xE0, 0xF0, 0xF8, 0xFC };
154 static const uchar limits
[6] = { 0x80, 0xE0, 0xF0, 0xF8, 0xFC, 0xFE };
158 /* Take care of preceding blanks. */
162 p
= write_block (dtp
, k
);
168 /* Get ready to handle delimiters if needed. */
169 switch (dtp
->u
.p
.current_unit
->delim_status
)
171 case DELIM_APOSTROPHE
:
182 /* Now process the remaining characters, one at a time. */
183 for (j
= k
; j
< src_len
; j
++)
188 /* Handle the delimiters if any. */
189 if (c
== d
&& d
!= ' ')
191 p
= write_block (dtp
, 2);
198 p
= write_block (dtp
, 1);
206 /* Convert to UTF-8 sequence. */
212 *--q
= ((c
& 0x3F) | 0x80);
216 while (c
>= 0x3F || (c
& limits
[nbytes
-1]));
218 *--q
= (c
| masks
[nbytes
-1]);
220 p
= write_block (dtp
, nbytes
);
232 write_a (st_parameter_dt
*dtp
, const fnode
*f
, const char *source
, int len
)
237 wlen
= f
->u
.string
.length
< 0
238 || (f
->format
== FMT_G
&& f
->u
.string
.length
== 0)
239 ? len
: f
->u
.string
.length
;
242 /* If this is formatted STREAM IO convert any embedded line feed characters
243 to CR_LF on systems that use that sequence for newlines. See F2003
244 Standard sections 10.6.3 and 9.9 for further information. */
245 if (is_stream_io (dtp
))
247 const char crlf
[] = "\r\n";
251 /* Write out any padding if needed. */
254 p
= write_block (dtp
, wlen
- len
);
257 memset (p
, ' ', wlen
- len
);
260 /* Scan the source string looking for '\n' and convert it if found. */
261 for (i
= 0; i
< wlen
; i
++)
263 if (source
[i
] == '\n')
265 /* Write out the previously scanned characters in the string. */
268 p
= write_block (dtp
, bytes
);
271 memcpy (p
, &source
[q
], bytes
);
276 /* Write out the CR_LF sequence. */
278 p
= write_block (dtp
, 2);
287 /* Write out any remaining bytes if no LF was found. */
290 p
= write_block (dtp
, bytes
);
293 memcpy (p
, &source
[q
], bytes
);
299 p
= write_block (dtp
, wlen
);
303 if (unlikely (is_char4_unit (dtp
)))
305 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
307 memcpy4 (p4
, source
, wlen
);
310 memset4 (p4
, ' ', wlen
- len
);
311 memcpy4 (p4
+ wlen
- len
, source
, len
);
317 memcpy (p
, source
, wlen
);
320 memset (p
, ' ', wlen
- len
);
321 memcpy (p
+ wlen
- len
, source
, len
);
329 /* The primary difference between write_a_char4 and write_a is that we have to
330 deal with writing from the first byte of the 4-byte character and pay
331 attention to the most significant bytes. For ENCODING="default" write the
332 lowest significant byte. If the 3 most significant bytes contain
333 non-zero values, emit a '?'. For ENCODING="utf-8", convert the UCS-32 value
334 to the UTF-8 encoded string before writing out. */
337 write_a_char4 (st_parameter_dt
*dtp
, const fnode
*f
, const char *source
, int len
)
342 wlen
= f
->u
.string
.length
< 0
343 || (f
->format
== FMT_G
&& f
->u
.string
.length
== 0)
344 ? len
: f
->u
.string
.length
;
346 q
= (gfc_char4_t
*) source
;
348 /* If this is formatted STREAM IO convert any embedded line feed characters
349 to CR_LF on systems that use that sequence for newlines. See F2003
350 Standard sections 10.6.3 and 9.9 for further information. */
351 if (is_stream_io (dtp
))
353 const gfc_char4_t crlf
[] = {0x000d,0x000a};
358 /* Write out any padding if needed. */
362 p
= write_block (dtp
, wlen
- len
);
365 memset (p
, ' ', wlen
- len
);
368 /* Scan the source string looking for '\n' and convert it if found. */
369 qq
= (gfc_char4_t
*) source
;
370 for (i
= 0; i
< wlen
; i
++)
374 /* Write out the previously scanned characters in the string. */
377 if (dtp
->u
.p
.current_unit
->flags
.encoding
== ENCODING_UTF8
)
378 write_utf8_char4 (dtp
, q
, bytes
, 0);
380 write_default_char4 (dtp
, q
, bytes
, 0);
384 /* Write out the CR_LF sequence. */
385 write_default_char4 (dtp
, crlf
, 2, 0);
391 /* Write out any remaining bytes if no LF was found. */
394 if (dtp
->u
.p
.current_unit
->flags
.encoding
== ENCODING_UTF8
)
395 write_utf8_char4 (dtp
, q
, bytes
, 0);
397 write_default_char4 (dtp
, q
, bytes
, 0);
403 if (dtp
->u
.p
.current_unit
->flags
.encoding
== ENCODING_UTF8
)
404 write_utf8_char4 (dtp
, q
, len
, wlen
);
406 write_default_char4 (dtp
, q
, len
, wlen
);
413 static GFC_INTEGER_LARGEST
414 extract_int (const void *p
, int len
)
416 GFC_INTEGER_LARGEST i
= 0;
426 memcpy ((void *) &tmp
, p
, len
);
433 memcpy ((void *) &tmp
, p
, len
);
440 memcpy ((void *) &tmp
, p
, len
);
447 memcpy ((void *) &tmp
, p
, len
);
451 #ifdef HAVE_GFC_INTEGER_16
455 memcpy ((void *) &tmp
, p
, len
);
461 internal_error (NULL
, "bad integer kind");
467 static GFC_UINTEGER_LARGEST
468 extract_uint (const void *p
, int len
)
470 GFC_UINTEGER_LARGEST i
= 0;
480 memcpy ((void *) &tmp
, p
, len
);
481 i
= (GFC_UINTEGER_1
) tmp
;
487 memcpy ((void *) &tmp
, p
, len
);
488 i
= (GFC_UINTEGER_2
) tmp
;
494 memcpy ((void *) &tmp
, p
, len
);
495 i
= (GFC_UINTEGER_4
) tmp
;
501 memcpy ((void *) &tmp
, p
, len
);
502 i
= (GFC_UINTEGER_8
) tmp
;
505 #ifdef HAVE_GFC_INTEGER_16
509 GFC_INTEGER_16 tmp
= 0;
510 memcpy ((void *) &tmp
, p
, len
);
511 i
= (GFC_UINTEGER_16
) tmp
;
516 internal_error (NULL
, "bad integer kind");
524 write_l (st_parameter_dt
*dtp
, const fnode
*f
, char *source
, int len
)
528 GFC_INTEGER_LARGEST n
;
530 wlen
= (f
->format
== FMT_G
&& f
->u
.w
== 0) ? 1 : f
->u
.w
;
532 p
= write_block (dtp
, wlen
);
536 n
= extract_int (source
, len
);
538 if (unlikely (is_char4_unit (dtp
)))
540 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
541 memset4 (p4
, ' ', wlen
-1);
542 p4
[wlen
- 1] = (n
) ? 'T' : 'F';
546 memset (p
, ' ', wlen
-1);
547 p
[wlen
- 1] = (n
) ? 'T' : 'F';
552 write_boz (st_parameter_dt
*dtp
, const fnode
*f
, const char *q
, int n
)
554 int w
, m
, digits
, nzero
, nblank
;
562 if (m
== 0 && n
== 0)
567 p
= write_block (dtp
, w
);
570 if (unlikely (is_char4_unit (dtp
)))
572 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
573 memset4 (p4
, ' ', w
);
582 /* Select a width if none was specified. The idea here is to always
586 w
= ((digits
< m
) ? m
: digits
);
588 p
= write_block (dtp
, w
);
596 /* See if things will work. */
598 nblank
= w
- (nzero
+ digits
);
600 if (unlikely (is_char4_unit (dtp
)))
602 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
605 memset4 (p4
, '*', w
);
609 if (!dtp
->u
.p
.no_leading_blank
)
611 memset4 (p4
, ' ', nblank
);
613 memset4 (p4
, '0', nzero
);
615 memcpy4 (p4
, q
, digits
);
619 memset4 (p4
, '0', nzero
);
621 memcpy4 (p4
, q
, digits
);
623 memset4 (p4
, ' ', nblank
);
624 dtp
->u
.p
.no_leading_blank
= 0;
635 if (!dtp
->u
.p
.no_leading_blank
)
637 memset (p
, ' ', nblank
);
639 memset (p
, '0', nzero
);
641 memcpy (p
, q
, digits
);
645 memset (p
, '0', nzero
);
647 memcpy (p
, q
, digits
);
649 memset (p
, ' ', nblank
);
650 dtp
->u
.p
.no_leading_blank
= 0;
658 write_decimal (st_parameter_dt
*dtp
, const fnode
*f
, const char *source
,
660 const char *(*conv
) (GFC_INTEGER_LARGEST
, char *, size_t))
662 GFC_INTEGER_LARGEST n
= 0;
663 int w
, m
, digits
, nsign
, nzero
, nblank
;
667 char itoa_buf
[GFC_BTOA_BUF_SIZE
];
670 m
= f
->format
== FMT_G
? -1 : f
->u
.integer
.m
;
672 n
= extract_int (source
, len
);
675 if (m
== 0 && n
== 0)
680 p
= write_block (dtp
, w
);
683 if (unlikely (is_char4_unit (dtp
)))
685 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
686 memset4 (p4
, ' ', w
);
693 sign
= calculate_sign (dtp
, n
< 0);
696 nsign
= sign
== S_NONE
? 0 : 1;
698 /* conv calls itoa which sets the negative sign needed
699 by write_integer. The sign '+' or '-' is set below based on sign
700 calculated above, so we just point past the sign in the string
701 before proceeding to avoid double signs in corner cases.
703 q
= conv (n
, itoa_buf
, sizeof (itoa_buf
));
709 /* Select a width if none was specified. The idea here is to always
713 w
= ((digits
< m
) ? m
: digits
) + nsign
;
715 p
= write_block (dtp
, w
);
723 /* See if things will work. */
725 nblank
= w
- (nsign
+ nzero
+ digits
);
727 if (unlikely (is_char4_unit (dtp
)))
729 gfc_char4_t
* p4
= (gfc_char4_t
*) p
;
732 memset4 (p4
, '*', w
);
736 memset4 (p4
, ' ', nblank
);
751 memset4 (p4
, '0', nzero
);
754 memcpy4 (p4
, q
, digits
);
764 memset (p
, ' ', nblank
);
779 memset (p
, '0', nzero
);
782 memcpy (p
, q
, digits
);
789 /* Convert unsigned octal to ascii. */
792 otoa (GFC_UINTEGER_LARGEST n
, char *buffer
, size_t len
)
796 assert (len
>= GFC_OTOA_BUF_SIZE
);
801 p
= buffer
+ GFC_OTOA_BUF_SIZE
- 1;
806 *--p
= '0' + (n
& 7);
814 /* Convert unsigned binary to ascii. */
817 btoa (GFC_UINTEGER_LARGEST n
, char *buffer
, size_t len
)
821 assert (len
>= GFC_BTOA_BUF_SIZE
);
826 p
= buffer
+ GFC_BTOA_BUF_SIZE
- 1;
831 *--p
= '0' + (n
& 1);
838 /* The following three functions, btoa_big, otoa_big, and ztoa_big, are needed
839 to convert large reals with kind sizes that exceed the largest integer type
840 available on certain platforms. In these cases, byte by byte conversion is
841 performed. Endianess is taken into account. */
843 /* Conversion to binary. */
846 btoa_big (const char *s
, char *buffer
, int len
, GFC_UINTEGER_LARGEST
*n
)
855 for (i
= 0; i
< len
; i
++)
859 /* Test for zero. Needed by write_boz later. */
863 for (j
= 0; j
< 8; j
++)
865 *q
++ = (c
& 128) ? '1' : '0';
873 const char *p
= s
+ len
- 1;
874 for (i
= 0; i
< len
; i
++)
878 /* Test for zero. Needed by write_boz later. */
882 for (j
= 0; j
< 8; j
++)
884 *q
++ = (c
& 128) ? '1' : '0';
896 /* Move past any leading zeros. */
897 while (*buffer
== '0')
904 /* Conversion to octal. */
907 otoa_big (const char *s
, char *buffer
, int len
, GFC_UINTEGER_LARGEST
*n
)
913 q
= buffer
+ GFC_OTOA_BUF_SIZE
- 1;
919 const char *p
= s
+ len
- 1;
923 /* Test for zero. Needed by write_boz later. */
927 for (j
= 0; j
< 3 && i
< len
; j
++)
929 octet
|= (c
& 1) << j
;
948 /* Test for zero. Needed by write_boz later. */
952 for (j
= 0; j
< 3 && i
< len
; j
++)
954 octet
|= (c
& 1) << j
;
971 /* Move past any leading zeros. */
978 /* Conversion to hexidecimal. */
981 ztoa_big (const char *s
, char *buffer
, int len
, GFC_UINTEGER_LARGEST
*n
)
983 static char a
[16] = {'0', '1', '2', '3', '4', '5', '6', '7',
984 '8', '9', 'A', 'B', 'C', 'D', 'E', 'F'};
995 for (i
= 0; i
< len
; i
++)
997 /* Test for zero. Needed by write_boz later. */
1001 h
= (*p
>> 4) & 0x0F;
1009 const char *p
= s
+ len
- 1;
1010 for (i
= 0; i
< len
; i
++)
1012 /* Test for zero. Needed by write_boz later. */
1016 h
= (*p
>> 4) & 0x0F;
1028 /* Move past any leading zeros. */
1029 while (*buffer
== '0')
1037 write_i (st_parameter_dt
*dtp
, const fnode
*f
, const char *p
, int len
)
1039 write_decimal (dtp
, f
, p
, len
, (void *) gfc_itoa
);
1044 write_b (st_parameter_dt
*dtp
, const fnode
*f
, const char *source
, int len
)
1047 char itoa_buf
[GFC_BTOA_BUF_SIZE
];
1048 GFC_UINTEGER_LARGEST n
= 0;
1050 if (len
> (int) sizeof (GFC_UINTEGER_LARGEST
))
1052 p
= btoa_big (source
, itoa_buf
, len
, &n
);
1053 write_boz (dtp
, f
, p
, n
);
1057 n
= extract_uint (source
, len
);
1058 p
= btoa (n
, itoa_buf
, sizeof (itoa_buf
));
1059 write_boz (dtp
, f
, p
, n
);
1065 write_o (st_parameter_dt
*dtp
, const fnode
*f
, const char *source
, int len
)
1068 char itoa_buf
[GFC_OTOA_BUF_SIZE
];
1069 GFC_UINTEGER_LARGEST n
= 0;
1071 if (len
> (int) sizeof (GFC_UINTEGER_LARGEST
))
1073 p
= otoa_big (source
, itoa_buf
, len
, &n
);
1074 write_boz (dtp
, f
, p
, n
);
1078 n
= extract_uint (source
, len
);
1079 p
= otoa (n
, itoa_buf
, sizeof (itoa_buf
));
1080 write_boz (dtp
, f
, p
, n
);
1085 write_z (st_parameter_dt
*dtp
, const fnode
*f
, const char *source
, int len
)
1088 char itoa_buf
[GFC_XTOA_BUF_SIZE
];
1089 GFC_UINTEGER_LARGEST n
= 0;
1091 if (len
> (int) sizeof (GFC_UINTEGER_LARGEST
))
1093 p
= ztoa_big (source
, itoa_buf
, len
, &n
);
1094 write_boz (dtp
, f
, p
, n
);
1098 n
= extract_uint (source
, len
);
1099 p
= gfc_xtoa (n
, itoa_buf
, sizeof (itoa_buf
));
1100 write_boz (dtp
, f
, p
, n
);
1104 /* Take care of the X/TR descriptor. */
1107 write_x (st_parameter_dt
*dtp
, int len
, int nspaces
)
1111 p
= write_block (dtp
, len
);
1114 if (nspaces
> 0 && len
- nspaces
>= 0)
1116 if (unlikely (is_char4_unit (dtp
)))
1118 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
1119 memset4 (&p4
[len
- nspaces
], ' ', nspaces
);
1122 memset (&p
[len
- nspaces
], ' ', nspaces
);
1127 /* List-directed writing. */
1130 /* Write a single character to the output. Returns nonzero if
1131 something goes wrong. */
1134 write_char (st_parameter_dt
*dtp
, int c
)
1138 p
= write_block (dtp
, 1);
1141 if (unlikely (is_char4_unit (dtp
)))
1143 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
1154 /* Write a list-directed logical value. */
1157 write_logical (st_parameter_dt
*dtp
, const char *source
, int length
)
1159 write_char (dtp
, extract_int (source
, length
) ? 'T' : 'F');
1163 /* Write a list-directed integer value. */
1166 write_integer (st_parameter_dt
*dtp
, const char *source
, int length
)
1172 char itoa_buf
[GFC_ITOA_BUF_SIZE
];
1174 q
= gfc_itoa (extract_int (source
, length
), itoa_buf
, sizeof (itoa_buf
));
1199 digits
= strlen (q
);
1203 p
= write_block (dtp
, width
);
1207 if (unlikely (is_char4_unit (dtp
)))
1209 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
1210 if (dtp
->u
.p
.no_leading_blank
)
1212 memcpy4 (p4
, q
, digits
);
1213 memset4 (p4
+ digits
, ' ', width
- digits
);
1217 memset4 (p4
, ' ', width
- digits
);
1218 memcpy4 (p4
+ width
- digits
, q
, digits
);
1223 if (dtp
->u
.p
.no_leading_blank
)
1225 memcpy (p
, q
, digits
);
1226 memset (p
+ digits
, ' ', width
- digits
);
1230 memset (p
, ' ', width
- digits
);
1231 memcpy (p
+ width
- digits
, q
, digits
);
1236 /* Write a list-directed string. We have to worry about delimiting
1237 the strings if the file has been opened in that mode. */
1243 write_character (st_parameter_dt
*dtp
, const char *source
, int kind
, int length
, int mode
)
1250 switch (dtp
->u
.p
.current_unit
->delim_status
)
1252 case DELIM_APOSTROPHE
:
1274 for (i
= 0; i
< length
; i
++)
1279 p
= write_block (dtp
, length
+ extra
);
1283 if (unlikely (is_char4_unit (dtp
)))
1285 gfc_char4_t d4
= (gfc_char4_t
) d
;
1286 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
1289 memcpy4 (p4
, source
, length
);
1294 for (i
= 0; i
< length
; i
++)
1296 *p4
++ = (gfc_char4_t
) source
[i
];
1307 memcpy (p
, source
, length
);
1312 for (i
= 0; i
< length
; i
++)
1326 if (dtp
->u
.p
.current_unit
->flags
.encoding
== ENCODING_UTF8
)
1327 write_utf8_char4 (dtp
, (gfc_char4_t
*) source
, length
, 0);
1329 write_default_char4 (dtp
, (gfc_char4_t
*) source
, length
, 0);
1333 p
= write_block (dtp
, 1);
1336 if (dtp
->u
.p
.current_unit
->flags
.encoding
== ENCODING_UTF8
)
1337 write_utf8_char4 (dtp
, (gfc_char4_t
*) source
, length
, 0);
1339 write_default_char4 (dtp
, (gfc_char4_t
*) source
, length
, 0);
1341 p
= write_block (dtp
, 1);
1347 /* Floating point helper functions. */
1349 #define BUF_STACK_SZ 256
1352 get_precision (st_parameter_dt
*dtp
, const fnode
*f
, const char *source
, int kind
)
1354 if (f
->format
!= FMT_EN
)
1355 return determine_precision (dtp
, f
, kind
);
1357 return determine_en_precision (dtp
, f
, source
, kind
);
1361 select_buffer (int precision
, char *buf
, size_t *size
)
1364 *size
= BUF_STACK_SZ
/ 2 + precision
;
1365 if (*size
> BUF_STACK_SZ
)
1366 result
= xmalloc (*size
);
1373 select_string (const fnode
*f
, char *buf
, size_t *size
)
1376 *size
= f
->u
.real
.w
+ 1;
1377 if (*size
> BUF_STACK_SZ
)
1378 result
= xmalloc (*size
);
1385 write_float_string (st_parameter_dt
*dtp
, char *fstr
, size_t len
)
1387 char *p
= write_block (dtp
, len
);
1391 if (unlikely (is_char4_unit (dtp
)))
1393 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
1394 memcpy4 (p4
, fstr
, len
);
1397 memcpy (p
, fstr
, len
);
1401 write_float_0 (st_parameter_dt
*dtp
, const fnode
*f
, const char *source
, int kind
)
1403 char buf_stack
[BUF_STACK_SZ
];
1404 char str_buf
[BUF_STACK_SZ
];
1405 char *buffer
, *result
;
1406 size_t buf_size
, res_len
;
1408 /* Precision for snprintf call. */
1409 int precision
= get_precision (dtp
, f
, source
, kind
);
1411 /* String buffer to hold final result. */
1412 result
= select_string (f
, str_buf
, &res_len
);
1414 buffer
= select_buffer (precision
, buf_stack
, &buf_size
);
1416 get_float_string (dtp
, f
, source
, kind
, 0, buffer
,
1417 precision
, buf_size
, result
, &res_len
);
1418 write_float_string (dtp
, result
, res_len
);
1420 if (buf_size
> BUF_STACK_SZ
)
1422 if (res_len
> BUF_STACK_SZ
)
1427 write_d (st_parameter_dt
*dtp
, const fnode
*f
, const char *p
, int len
)
1429 write_float_0 (dtp
, f
, p
, len
);
1434 write_e (st_parameter_dt
*dtp
, const fnode
*f
, const char *p
, int len
)
1436 write_float_0 (dtp
, f
, p
, len
);
1441 write_f (st_parameter_dt
*dtp
, const fnode
*f
, const char *p
, int len
)
1443 write_float_0 (dtp
, f
, p
, len
);
1448 write_en (st_parameter_dt
*dtp
, const fnode
*f
, const char *p
, int len
)
1450 write_float_0 (dtp
, f
, p
, len
);
1455 write_es (st_parameter_dt
*dtp
, const fnode
*f
, const char *p
, int len
)
1457 write_float_0 (dtp
, f
, p
, len
);
1461 /* Set an fnode to default format. */
1464 set_fnode_default (st_parameter_dt
*dtp
, fnode
*f
, int length
)
1485 /* Adjust decimal precision depending on binary precision, 106 or 113. */
1486 #if GFC_REAL_16_DIGITS == 113
1497 internal_error (&dtp
->common
, "bad real kind");
1502 /* Output a real number with default format.
1503 To guarantee that a binary -> decimal -> binary roundtrip conversion
1504 recovers the original value, IEEE 754-2008 requires 9, 17, 21 and 36
1505 significant digits for REAL kinds 4, 8, 10, and 16, respectively.
1506 Thus, we use 1PG16.9E2 for REAL(4), 1PG25.17E3 for REAL(8), 1PG30.21E4
1507 for REAL(10) and 1PG45.36E4 for REAL(16). The exception is that the
1508 Fortran standard requires outputting an extra digit when the scale
1509 factor is 1 and when the magnitude of the value is such that E
1510 editing is used. However, gfortran compensates for this, and thus
1511 for list formatted the same number of significant digits is
1512 generated both when using F and E editing. */
1515 write_real (st_parameter_dt
*dtp
, const char *source
, int kind
)
1518 char buf_stack
[BUF_STACK_SZ
];
1519 char str_buf
[BUF_STACK_SZ
];
1520 char *buffer
, *result
;
1521 size_t buf_size
, res_len
;
1522 int orig_scale
= dtp
->u
.p
.scale_factor
;
1523 dtp
->u
.p
.scale_factor
= 1;
1524 set_fnode_default (dtp
, &f
, kind
);
1526 /* Precision for snprintf call. */
1527 int precision
= get_precision (dtp
, &f
, source
, kind
);
1529 /* String buffer to hold final result. */
1530 result
= select_string (&f
, str_buf
, &res_len
);
1532 /* scratch buffer to hold final result. */
1533 buffer
= select_buffer (precision
, buf_stack
, &buf_size
);
1535 get_float_string (dtp
, &f
, source
, kind
, 1, buffer
,
1536 precision
, buf_size
, result
, &res_len
);
1537 write_float_string (dtp
, result
, res_len
);
1539 dtp
->u
.p
.scale_factor
= orig_scale
;
1540 if (buf_size
> BUF_STACK_SZ
)
1542 if (res_len
> BUF_STACK_SZ
)
1546 /* Similar to list formatted REAL output, for kPG0 where k > 0 we
1547 compensate for the extra digit. */
1550 write_real_g0 (st_parameter_dt
*dtp
, const char *source
, int kind
, int d
)
1553 char buf_stack
[BUF_STACK_SZ
];
1554 char str_buf
[BUF_STACK_SZ
];
1555 char *buffer
, *result
;
1556 size_t buf_size
, res_len
;
1558 set_fnode_default (dtp
, &f
, kind
);
1563 /* Compensate for extra digits when using scale factor, d is not
1564 specified, and the magnitude is such that E editing is used. */
1565 if (dtp
->u
.p
.scale_factor
> 0 && d
== 0)
1569 dtp
->u
.p
.g0_no_blanks
= 1;
1571 /* Precision for snprintf call. */
1572 int precision
= get_precision (dtp
, &f
, source
, kind
);
1574 /* String buffer to hold final result. */
1575 result
= select_string (&f
, str_buf
, &res_len
);
1577 buffer
= select_buffer (precision
, buf_stack
, &buf_size
);
1579 get_float_string (dtp
, &f
, source
, kind
, comp_d
, buffer
,
1580 precision
, buf_size
, result
, &res_len
);
1581 write_float_string (dtp
, result
, res_len
);
1583 dtp
->u
.p
.g0_no_blanks
= 0;
1584 if (buf_size
> BUF_STACK_SZ
)
1586 if (res_len
> BUF_STACK_SZ
)
1592 write_complex (st_parameter_dt
*dtp
, const char *source
, int kind
, size_t size
)
1595 dtp
->u
.p
.current_unit
->decimal_status
== DECIMAL_POINT
? ',' : ';';
1597 /* Set for no blanks so we get a string result with no leading
1598 blanks. We will pad left later. */
1599 dtp
->u
.p
.g0_no_blanks
= 1;
1602 char buf_stack
[BUF_STACK_SZ
];
1603 char str1_buf
[BUF_STACK_SZ
];
1604 char str2_buf
[BUF_STACK_SZ
];
1605 char *buffer
, *result1
, *result2
;
1606 size_t buf_size
, res_len1
, res_len2
;
1607 int width
, lblanks
, orig_scale
= dtp
->u
.p
.scale_factor
;
1609 dtp
->u
.p
.scale_factor
= 1;
1610 set_fnode_default (dtp
, &f
, kind
);
1612 /* Set width for two values, parenthesis, and comma. */
1613 width
= 2 * f
.u
.real
.w
+ 3;
1615 /* Set for no blanks so we get a string result with no leading
1616 blanks. We will pad left later. */
1617 dtp
->u
.p
.g0_no_blanks
= 1;
1619 /* Precision for snprintf call. */
1620 int precision
= get_precision (dtp
, &f
, source
, kind
);
1622 /* String buffers to hold final result. */
1623 result1
= select_string (&f
, str1_buf
, &res_len1
);
1624 result2
= select_string (&f
, str2_buf
, &res_len2
);
1626 buffer
= select_buffer (precision
, buf_stack
, &buf_size
);
1628 get_float_string (dtp
, &f
, source
, kind
, 0, buffer
,
1629 precision
, buf_size
, result1
, &res_len1
);
1630 get_float_string (dtp
, &f
, source
+ size
/ 2 , kind
, 0, buffer
,
1631 precision
, buf_size
, result2
, &res_len2
);
1632 lblanks
= width
- res_len1
- res_len2
- 3;
1634 write_x (dtp
, lblanks
, lblanks
);
1635 write_char (dtp
, '(');
1636 write_float_string (dtp
, result1
, res_len1
);
1637 write_char (dtp
, semi_comma
);
1638 write_float_string (dtp
, result2
, res_len2
);
1639 write_char (dtp
, ')');
1641 dtp
->u
.p
.scale_factor
= orig_scale
;
1642 dtp
->u
.p
.g0_no_blanks
= 0;
1643 if (buf_size
> BUF_STACK_SZ
)
1645 if (res_len1
> BUF_STACK_SZ
)
1647 if (res_len2
> BUF_STACK_SZ
)
1652 /* Write the separator between items. */
1655 write_separator (st_parameter_dt
*dtp
)
1659 p
= write_block (dtp
, options
.separator_len
);
1662 if (unlikely (is_char4_unit (dtp
)))
1664 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
1665 memcpy4 (p4
, options
.separator
, options
.separator_len
);
1668 memcpy (p
, options
.separator
, options
.separator_len
);
1672 /* Write an item with list formatting.
1673 TODO: handle skipping to the next record correctly, particularly
1677 list_formatted_write_scalar (st_parameter_dt
*dtp
, bt type
, void *p
, int kind
,
1680 if (dtp
->u
.p
.current_unit
== NULL
)
1683 if (dtp
->u
.p
.first_item
)
1685 dtp
->u
.p
.first_item
= 0;
1686 write_char (dtp
, ' ');
1690 if (type
!= BT_CHARACTER
|| !dtp
->u
.p
.char_flag
||
1691 (dtp
->u
.p
.current_unit
->delim_status
!= DELIM_NONE
1692 && dtp
->u
.p
.current_unit
->delim_status
!= DELIM_UNSPECIFIED
))
1693 write_separator (dtp
);
1699 write_integer (dtp
, p
, kind
);
1702 write_logical (dtp
, p
, kind
);
1705 write_character (dtp
, p
, kind
, size
, DELIM
);
1708 write_real (dtp
, p
, kind
);
1711 write_complex (dtp
, p
, kind
, size
);
1714 internal_error (&dtp
->common
, "list_formatted_write(): Bad type");
1717 fbuf_flush_list (dtp
->u
.p
.current_unit
, LIST_WRITING
);
1718 dtp
->u
.p
.char_flag
= (type
== BT_CHARACTER
);
1723 list_formatted_write (st_parameter_dt
*dtp
, bt type
, void *p
, int kind
,
1724 size_t size
, size_t nelems
)
1728 size_t stride
= type
== BT_CHARACTER
?
1729 size
* GFC_SIZE_OF_CHAR_KIND(kind
) : size
;
1733 /* Big loop over all the elements. */
1734 for (elem
= 0; elem
< nelems
; elem
++)
1736 dtp
->u
.p
.item_count
++;
1737 list_formatted_write_scalar (dtp
, type
, tmp
+ elem
* stride
, kind
, size
);
1743 nml_write_obj writes a namelist object to the output stream. It is called
1744 recursively for derived type components:
1745 obj = is the namelist_info for the current object.
1746 offset = the offset relative to the address held by the object for
1747 derived type arrays.
1748 base = is the namelist_info of the derived type, when obj is a
1750 base_name = the full name for a derived type, including qualifiers
1752 The returned value is a pointer to the object beyond the last one
1753 accessed, including nested derived types. Notice that the namelist is
1754 a linear linked list of objects, including derived types and their
1755 components. A tree, of sorts, is implied by the compound names of
1756 the derived type components and this is how this function recurses through
1759 /* A generous estimate of the number of characters needed to print
1760 repeat counts and indices, including commas, asterices and brackets. */
1762 #define NML_DIGITS 20
1765 namelist_write_newline (st_parameter_dt
*dtp
)
1767 if (!is_internal_unit (dtp
))
1770 write_character (dtp
, "\r\n", 1, 2, NODELIM
);
1772 write_character (dtp
, "\n", 1, 1, NODELIM
);
1777 if (is_array_io (dtp
))
1782 int length
= dtp
->u
.p
.current_unit
->bytes_left
;
1784 p
= write_block (dtp
, length
);
1788 if (unlikely (is_char4_unit (dtp
)))
1790 gfc_char4_t
*p4
= (gfc_char4_t
*) p
;
1791 memset4 (p4
, ' ', length
);
1794 memset (p
, ' ', length
);
1796 /* Now that the current record has been padded out,
1797 determine where the next record in the array is. */
1798 record
= next_array_record (dtp
, dtp
->u
.p
.current_unit
->ls
,
1801 dtp
->u
.p
.current_unit
->endfile
= AT_ENDFILE
;
1804 /* Now seek to this record */
1805 record
= record
* dtp
->u
.p
.current_unit
->recl
;
1807 if (sseek (dtp
->u
.p
.current_unit
->s
, record
, SEEK_SET
) < 0)
1809 generate_error (&dtp
->common
, LIBERROR_INTERNAL_UNIT
, NULL
);
1813 dtp
->u
.p
.current_unit
->bytes_left
= dtp
->u
.p
.current_unit
->recl
;
1817 write_character (dtp
, " ", 1, 1, NODELIM
);
1821 static namelist_info
*
1822 nml_write_obj (st_parameter_dt
*dtp
, namelist_info
* obj
, index_type offset
,
1823 namelist_info
* base
, char * base_name
)
1829 index_type obj_size
;
1833 index_type elem_ctr
;
1834 size_t obj_name_len
;
1840 size_t ext_name_len
;
1841 char rep_buff
[NML_DIGITS
];
1842 namelist_info
* cmp
;
1843 namelist_info
* retval
= obj
->next
;
1844 size_t base_name_len
;
1845 size_t base_var_name_len
;
1848 /* Set the character to be used to separate values
1849 to a comma or semi-colon. */
1852 dtp
->u
.p
.current_unit
->decimal_status
== DECIMAL_POINT
? ',' : ';';
1854 /* Write namelist variable names in upper case. If a derived type,
1855 nothing is output. If a component, base and base_name are set. */
1857 if (obj
->type
!= BT_DERIVED
)
1859 namelist_write_newline (dtp
);
1860 write_character (dtp
, " ", 1, 1, NODELIM
);
1865 len
= strlen (base
->var_name
);
1866 base_name_len
= strlen (base_name
);
1867 for (dim_i
= 0; dim_i
< base_name_len
; dim_i
++)
1869 cup
= toupper ((int) base_name
[dim_i
]);
1870 write_character (dtp
, &cup
, 1, 1, NODELIM
);
1873 clen
= strlen (obj
->var_name
);
1874 for (dim_i
= len
; dim_i
< clen
; dim_i
++)
1876 cup
= toupper ((int) obj
->var_name
[dim_i
]);
1879 write_character (dtp
, &cup
, 1, 1, NODELIM
);
1881 write_character (dtp
, "=", 1, 1, NODELIM
);
1884 /* Counts the number of data output on a line, including names. */
1894 obj_size
= size_from_real_kind (len
);
1898 obj_size
= size_from_complex_kind (len
);
1902 obj_size
= obj
->string_length
;
1910 obj_size
= obj
->size
;
1912 /* Set the index vector and count the number of elements. */
1915 for (dim_i
= 0; dim_i
< (size_t) obj
->var_rank
; dim_i
++)
1917 obj
->ls
[dim_i
].idx
= GFC_DESCRIPTOR_LBOUND(obj
, dim_i
);
1918 nelem
= nelem
* GFC_DESCRIPTOR_EXTENT (obj
, dim_i
);
1921 /* Main loop to output the data held in the object. */
1924 for (elem_ctr
= 0; elem_ctr
< nelem
; elem_ctr
++)
1927 /* Build the pointer to the data value. The offset is passed by
1928 recursive calls to this function for arrays of derived types.
1929 Is NULL otherwise. */
1931 p
= (void *)(obj
->mem_pos
+ elem_ctr
* obj_size
);
1934 /* Check for repeat counts of intrinsic types. */
1936 if ((elem_ctr
< (nelem
- 1)) &&
1937 (obj
->type
!= BT_DERIVED
) &&
1938 !memcmp (p
, (void*)(p
+ obj_size
), obj_size
))
1943 /* Execute a repeated output. Note the flag no_leading_blank that
1944 is used in the functions used to output the intrinsic types. */
1950 snprintf(rep_buff
, NML_DIGITS
, " %d*", rep_ctr
);
1951 write_character (dtp
, rep_buff
, 1, strlen (rep_buff
), NODELIM
);
1952 dtp
->u
.p
.no_leading_blank
= 1;
1956 /* Output the data, if an intrinsic type, or recurse into this
1957 routine to treat derived types. */
1963 write_integer (dtp
, p
, len
);
1967 write_logical (dtp
, p
, len
);
1971 if (dtp
->u
.p
.current_unit
->flags
.encoding
== ENCODING_UTF8
)
1972 write_character (dtp
, p
, 4, obj
->string_length
, DELIM
);
1974 write_character (dtp
, p
, 1, obj
->string_length
, DELIM
);
1978 write_real (dtp
, p
, len
);
1982 dtp
->u
.p
.no_leading_blank
= 0;
1984 write_complex (dtp
, p
, len
, obj_size
);
1989 /* To treat a derived type, we need to build two strings:
1990 ext_name = the name, including qualifiers that prepends
1991 component names in the output - passed to
1993 obj_name = the derived type name with no qualifiers but %
1994 appended. This is used to identify the
1997 /* First ext_name => get length of all possible components */
1999 base_name_len
= base_name
? strlen (base_name
) : 0;
2000 base_var_name_len
= base
? strlen (base
->var_name
) : 0;
2001 ext_name_len
= base_name_len
+ base_var_name_len
2002 + strlen (obj
->var_name
) + obj
->var_rank
* NML_DIGITS
+ 1;
2003 ext_name
= xmalloc (ext_name_len
);
2006 memcpy (ext_name
, base_name
, base_name_len
);
2007 clen
= strlen (obj
->var_name
+ base_var_name_len
);
2008 memcpy (ext_name
+ base_name_len
,
2009 obj
->var_name
+ base_var_name_len
, clen
);
2011 /* Append the qualifier. */
2013 tot_len
= base_name_len
+ clen
;
2014 for (dim_i
= 0; dim_i
< (size_t) obj
->var_rank
; dim_i
++)
2018 ext_name
[tot_len
] = '(';
2021 snprintf (ext_name
+ tot_len
, ext_name_len
- tot_len
, "%d",
2022 (int) obj
->ls
[dim_i
].idx
);
2023 tot_len
+= strlen (ext_name
+ tot_len
);
2024 ext_name
[tot_len
] = ((int) dim_i
== obj
->var_rank
- 1) ? ')' : ',';
2028 ext_name
[tot_len
] = '\0';
2029 for (q
= ext_name
; *q
; q
++)
2035 obj_name_len
= strlen (obj
->var_name
) + 1;
2036 obj_name
= xmalloc (obj_name_len
+ 1);
2037 memcpy (obj_name
, obj
->var_name
, obj_name_len
-1);
2038 memcpy (obj_name
+ obj_name_len
-1, "%", 2);
2040 /* Now loop over the components. Update the component pointer
2041 with the return value from nml_write_obj => this loop jumps
2042 past nested derived types. */
2044 for (cmp
= obj
->next
;
2045 cmp
&& !strncmp (cmp
->var_name
, obj_name
, obj_name_len
);
2048 retval
= nml_write_obj (dtp
, cmp
,
2049 (index_type
)(p
- obj
->mem_pos
),
2058 internal_error (&dtp
->common
, "Bad type for namelist write");
2061 /* Reset the leading blank suppression, write a comma (or semi-colon)
2062 and, if 5 values have been output, write a newline and advance
2063 to column 2. Reset the repeat counter. */
2065 dtp
->u
.p
.no_leading_blank
= 0;
2066 if (obj
->type
== BT_CHARACTER
)
2068 if (dtp
->u
.p
.nml_delim
!= '\0')
2069 write_character (dtp
, &semi_comma
, 1, 1, NODELIM
);
2072 write_character (dtp
, &semi_comma
, 1, 1, NODELIM
);
2076 if (dtp
->u
.p
.nml_delim
== '\0')
2077 write_character (dtp
, &semi_comma
, 1, 1, NODELIM
);
2078 namelist_write_newline (dtp
);
2079 write_character (dtp
, " ", 1, 1, NODELIM
);
2084 /* Cycle through and increment the index vector. */
2089 for (dim_i
= 0; nml_carry
&& (dim_i
< (size_t) obj
->var_rank
); dim_i
++)
2091 obj
->ls
[dim_i
].idx
+= nml_carry
;
2093 if (obj
->ls
[dim_i
].idx
> GFC_DESCRIPTOR_UBOUND(obj
,dim_i
))
2095 obj
->ls
[dim_i
].idx
= GFC_DESCRIPTOR_LBOUND(obj
,dim_i
);
2101 /* Return a pointer beyond the furthest object accessed. */
2107 /* This is the entry function for namelist writes. It outputs the name
2108 of the namelist and iterates through the namelist by calls to
2109 nml_write_obj. The call below has dummys in the arguments used in
2110 the treatment of derived types. */
2113 namelist_write (st_parameter_dt
*dtp
)
2115 namelist_info
* t1
, *t2
, *dummy
= NULL
;
2117 index_type dummy_offset
= 0;
2119 char * dummy_name
= NULL
;
2121 /* Set the delimiter for namelist output. */
2122 switch (dtp
->u
.p
.current_unit
->delim_status
)
2124 case DELIM_APOSTROPHE
:
2125 dtp
->u
.p
.nml_delim
= '\'';
2128 case DELIM_UNSPECIFIED
:
2129 dtp
->u
.p
.nml_delim
= '"';
2132 dtp
->u
.p
.nml_delim
= '\0';
2135 write_character (dtp
, "&", 1, 1, NODELIM
);
2137 /* Write namelist name in upper case - f95 std. */
2138 for (i
= 0 ;i
< dtp
->namelist_name_len
;i
++ )
2140 c
= toupper ((int) dtp
->namelist_name
[i
]);
2141 write_character (dtp
, &c
, 1 ,1, NODELIM
);
2144 if (dtp
->u
.p
.ionml
!= NULL
)
2146 t1
= dtp
->u
.p
.ionml
;
2150 t1
= nml_write_obj (dtp
, t2
, dummy_offset
, dummy
, dummy_name
);
2154 namelist_write_newline (dtp
);
2155 write_character (dtp
, " /", 1, 2, NODELIM
);