File: ggcm_read_mhd.C

package info (click to toggle)
paraview 4.0.1-1~bpo70%2B1
  • links: PTS, VCS
  • area: main
  • in suites: wheezy-backports
  • size: 526,572 kB
  • sloc: cpp: 2,284,430; ansic: 816,374; python: 239,936; xml: 70,162; tcl: 48,295; fortran: 39,116; yacc: 5,466; java: 3,518; perl: 3,107; lex: 1,620; sh: 1,555; makefile: 932; asm: 471; pascal: 228
file content (691 lines) | stat: -rw-r--r-- 27,581 bytes parent folder | download | duplicates (6)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
/** Module to read run length encoded ASCII MHD files generated by
 * the GGCM code.  These subroutines were taken from Bill Sherman's
 * "convertmhd.c" code, which was designed to create VTK formatted
 * output files.
 *
 * NOTE:
 *   Compiling works under both Linux and IRIX.  However, the Linux output
 *   seems to round up a bit higher than the IRIX.  Ie. the last significant
 *   digit in the Linux output may be one numeral higher than IRIX.  Since
 *   we're dealing with values that have some loss from the MHD compression
 *   it probably doesn't matter much -- but it makes it difficult to compare
 *   outputs with "diff".
 *
 * HISTORY:
 *   This code was initially created by Bill Sherman in late 2003.  John C.
 *   Dorelli cleaned things up a bit and made it easier to use for us GGCM
 *   people.  Maintainence has since been handed to Tom Fogal, who has turned
 *   the program into a library and done other various fixes.
 *   The old history that was here has been removed; most of it didn't make
 *   sense anymore anyway.  You can grep it out of the subversion logs / old
 *   versions if you really want it.
 ****/

//   Jeremy Meredith, Wed Jan  6 17:46:31 EST 2010
//   Added VisIt exceptions in place of aborts.
//

#include "libggcm.h"

#include <stdio.h>
#include <string.h>
#include <math.h>
#include <stdlib.h>
#include <assert.h>

#include <InvalidFilesException.h>

/** fscanf that is expected to work; aborts on error */
#define SCAN(fp, match, v1)                   \
    {                                         \
        int err;                              \
        err = fscanf(fp, match, v1);          \
        if(err == EOF) {                      \
            perror("scanning in libggcm");    \
            EXCEPTION1(InvalidFilesException, mhd->filename); \
        }                                     \
    }
/** more stringent; give the number of arguments which must match. */
#define NSCAN(fp, num, match, v1, v2, v3, v4, v5, v6)     \
    {                                                     \
        int err;                                          \
        err = fscanf(fp, match, v1, v2, v3, v4, v5, v6);  \
        if(err == EOF || err != num) {                    \
            perror("scanning in libggcm");                \
            fprintf(stderr, "%d matches\n", err);         \
            EXCEPTION1(InvalidFilesException, mhd->filename); \
        }                                                 \
    }

static int wrndec(unsigned char *in_string, int *out_array);
static int readRLEData(MHDdata *data);
static int field_match(const char * const match_list[],
                       const char * const field_name);
static int readMHDField(MHDdata *data, char *field_info,
                        const char * const match_list[]);
static int read_time_line(MHDdata *);
static int ggcm_read_field_metadata(MHDdata *);

/** alias for ggcm_init_mhd */
void initMHD(MHDdata *data) { ggcm_init_mhd(data); }

/** alias for ggcm_read_mhd */
MHDdata *readMHDFile(const char *file_name, const char *field_list[], long file_start)
{
    return ggcm_read_mhd(file_name, field_list);
}

/** alias for ggcm_free_data */
void freeMHDData(MHDdata *md) { ggcm_free_data(md); }

/** initialize an MHDdata record */
void ggcm_init_mhd(MHDdata *data)
{
    data->next = NULL;
    data->field_name[0] = '\0';

    data->regular = 1;
    data->subsample[0] = 1;
    data->subsample[1] = 1;
    data->subsample[2] = 1;
    data->subsamp_dim[0] = 1;
    data->subsamp_dim[1] = 1;
    data->subsamp_dim[2] = 1;

    data->line3_info[0] = '\0';
    data->line4_value = 0;

    data->filename[0] = '\0';
    data->mhd_fp = NULL;
    data->file_position = -1;

    data->dim_rank = 0;
    data->dim_size[0] = 1;
    data->dim_size[1] = 1;
    data->dim_size[2] = 1;
    data->value_rank = 1;
    data->number = 0;
    data->write_data = 1;
    data->num_covecs = 0;
    data->data = NULL;
    data->grid[0] = NULL;
    data->grid[1] = NULL;
    data->grid[2] = NULL;
    data->co_vector[0] = NULL;
    data->co_vector[1] = NULL;
    data->co_vector[2] = NULL;
}

/** creates a list which contains all of the metadata for a given GGCM file,
 * such as the fields contained within it, data points / field, time code
 * information, etc. */
MHDdata *ggcm_read_metadata(const char *filename)
{
    MHDdata *head, *cur, *prev;   /* linked list of field information */
    unsigned int lines, i;        /* how many lines to skip, current line */
    size_t len;                   /* length of temp buffer */
    char *line;                   /* temp buffer for junking lines */
    int fields = 0;               /* more fields to read? */

    head = prev = cur = (MHDdata *)malloc(sizeof(MHDdata));
    head->mhd_fp = fopen(filename, "r");
    strcpy(head->filename,filename); // for error reporting

    len = 1024;
    line = (char *)malloc(len);
    while((fields = ggcm_read_field_metadata(cur)) != 0) {
        prev = cur;
        /* how many lines should we skip? */
        lines = ((cur->number % 64) != 0) ?
                ((cur->number/64)+1) : (cur->number/64);
        lines *= 2;

        /* 'seek' to the next field in the file.  We'd like to be able to use
         * fseek here, but we only know how many *lines* the data are, not how
         * many *bytes* they are.  So we have to scan every line... */
        for(i=0; i < lines; ++i) {
            /* fgets works pretty well too; getline is slightly faster.
               fscanf is extremely slow.  Using fgets since it is portable,
               whereas getline is only on linux. */
            fgets(line, len, cur->mhd_fp);
        }

        cur->next = (MHDdata *)malloc(sizeof(MHDdata)); /* need another elem in the list */
        cur->next->mhd_fp = cur->mhd_fp;     /* field_metadata needs a good FP */

        cur = cur->next;
        cur->next = NULL; /* NULL for the sentinel */
    }

    /* bogus extra element */
    free(cur);
    prev->next = NULL;

    free(line); /* buffer for getline */
    fclose(head->mhd_fp);
    return head;
}

/** reads the metadata for a given field.
 * @todo add a field to MHDdata records which includes the file offset (via
 *       ftell(3)) of the field in question, and set that here right after
 *       reading the last bit of information ("line 5").  This will allow a
 *       more efficient reader when I rewrite that in the future, since it can
 *       know exactly where to seek to. */
static int ggcm_read_field_metadata(MHDdata *mhd)
{
    char junk[512];

    /* read the first field header, describing the field type (2D? 3D? etc.).
     * If this read fails, we're probably at EOF, indicating there are no more
     * fields. */
    fscanf(mhd->mhd_fp, "FIELD-%[^\n ]\n", junk);
    if(ferror(mhd->mhd_fp)) {
        perror("Unrecoverable reading field metadata");
        EXCEPTION1(InvalidFilesException, mhd->filename);
    }
    if(feof(mhd->mhd_fp)) {
        return 0;
    }

    SCAN(mhd->mhd_fp, "%s\n", mhd->field_name);
    read_time_line(mhd);

    /* line 4 has info on dimensionality, etc; ignored for now. */
    SCAN(mhd->mhd_fp, "%[^\n]\n", junk);

    /* the important thing on line 5 is the 'number' -- how many data points are
     * in this field. */
    NSCAN(mhd->mhd_fp, 6, "WRN2 %d%f %f %f %d%[^\n]\n", &mhd->number,
                          &mhd->wrn2_z1, &mhd->wrn2_z2, &mhd->wrn2_rid,
                          &mhd->wrn2_it, mhd->wrn2_cid);
    return 1;
}

/** wrndec(): decodes a line of run-length encoded ascii data into an
 *   integer array ("out_array") which is returned via call-by-reference.
 *   The length of the output array is returned normally.  The input to
 *   this function is a character array of the string to decode.
 * @param in_string characters to decode (nee "c" & replacement for "iu")
 * @param out_array integer array number 1 -- return values (nee "i1")
 * @return the number of values (nee "n+1") */
int wrndec(unsigned char *in_string, int *out_array)
{
    int    index_out = 0;    /* index into the outgoing "out_array" array (nee "i")  */
    int    index_in = 0;    /* index into the "in_string" character array (nee "j") */
    int    input_checksum;    /* checksum of the incoming character string (nee i2[0])*/
    int    checksum;    /* checksum value (nee "ick") */
    int    run_value;    /* running value (nee "ir") */
    int    run_count;    /* ren-length count (nee "ic") */
    int    count;

    /* set some initial values in case of error */
    for (count = 0; count < 64; count++) {
        out_array[count] = 34;        /* NOTE: why 34?  seems like "33" would be better */
    }

    input_checksum = in_string[0];
    checksum = 0;

    index_in = 1; /* because C is 0 based, and perhaps the original code is 1-based */
    while ((run_value = (int)(in_string[index_in])) != 0) {
        /* NOTE: in Fortran code the value "32" is the terminal value.  Need  */
        /*   to find out what significance that value has related to strings. */

        if (run_value < 0) {
            printf("adjusting sign -- preferable to deal with unsigned ints: run_value = %d\n", run_value);
            run_value += 256; /* shouldn't this be an abs() call? */
        }

        if (run_value > 127) {
            /* this means the value is actually a run-count, so get the   */
            /*   count, and the next number is the actual repeating value */
            run_count = run_value-170;
            run_value = (int)(in_string[++index_in]);
            if (run_value < 0) {
                printf("adjusting sign (2) -- preferable to deal with unsigned ints: run_value = %d\n", run_value);
                run_value += 256;
            }

            for (count = 0; count < run_count; count++) {
                if (index_out < 67)
                    out_array[index_out++] = run_value;
                else    printf("wrndec(): catastrophic error -- trying to write outside the output array\n");
                checksum += run_value;
            }
        } else {
            if (index_out < 67)
                out_array[index_out++] = run_value;
            else    printf("wrndec(): catastrophic error -- trying to write outside the output array\n");
            checksum += run_value;
        }

        index_in++;
        if (index_in > 67) {
            printf("error:wrndec: can't find end of encoded record -- index_in = %d\n", index_in);
            return index_out;
        }
    }

    if (index_out > 64) {
        printf("error: wrndec: range error -- index_out(%d) > 64\n", index_out);
        return 64;
    }
    if (input_checksum != 33 + (checksum % 92)) {
        printf("error: wrndec: checksum error, mismatch: %d vs %d -- index_out = %d\n", input_checksum, 33 + (checksum % 92), index_out);
        return 0;
    } else {
#if 0
        printf("checksum match: %d\n", input_checksum);
#endif
    }

    return index_out;
}

/** readRLEData(): Read a run-length ascii encoding with N=2 bytes
 * Reads ascii-encoded run-length data from the file pointed to by
 * data->mhd_fp.  The "data" structure is also used to pass in values
 * of the "wrn" encoding, such as the the min/max range of the absolute
 * values, and the number and organization of the data (though this
 * routine could really care less about the organization).
 *
 * In the original (util1.for) version, this routine is responsible for
 *   reading "line 5", however in my version, I've already read and
 *   parsed that info.
 * returns number of elements (aka "n" in the fortran code)
 *
 * NOTE: in rmhd06/util.for, the rdn2() procedure takes six arguments:
 *   iu  -- file id (called "data->mhd_fp" here)
 *   a   -- data array (floats) (called "data->data" here)
 *   n   -- number of elements [returned value?] (called "data->number")
 *   cid -- type of data (called "data->wrn2_cid" here)
 *   it  -- [no longer used] (was timecode)
 *   rid -- [not used]
 *
 * Modifications:
 *   Mark C. Miller, Thu Nov  6 11:12:56 PST 2008
 *   Fixed compilation warning on line ~389 converting double to int
 **/
int readRLEData(MHDdata *data)
{
    unsigned char buffer[128];    /* buffer for input string -- from wrndec()         */
    int    count;
    int    block64;    /* for counting through the elements 64 numbers at a time (nee "k") */
    int    block_size;    /* number of expected values to be returned by wrndec() -- (nee "nk+1") */
    int    msbits[72];    /* integer array of most significant bits decoded by wrndec() (nee "i1") */
    int    lsbits[72];    /* integer array of least significant bits decoded by wrndec() (nee "i2") */
    int    int_value;    /* reconstructed integer value with all the bits (nee "i3" array) */
                /*   NOTE: this doesn't need to be an array.                */
#if 0 /* fltarray_1 isn't necessary because it gets immediately stored in the actual data array */
    float    fltarray_1[64];    /* float array that decodes one block of data (nee "a1")*/
#endif
    int    num_decoded;    /* value of "n" as returned by wrndec() (nee "nn") */
    float    dzi;        /* some sort of dynamic range value */
    float    sign;        /* the sign of the result (nee "sig") */
#if 0 /* data_index is no longer necessary, because it is equivalent to "count+block64" */
    int    data_index = 0;    /* index into the full data array of where to write next (nee "l") */
#endif
    int    scan;        /* return value from call to fscanf */
    int    bytes_read;    /* number of bytes read by call to fscanf */

    /* Free the data array memory if it exists */
    if (data->data != NULL)
        free(data->data);

    /* Need to allocate the memory for the data array */
    data->data = (float *)calloc(data->number, sizeof(float));

    /* If the min and the max are the same, then all elements are that value. */
    if(data->wrn2_z1 == data->wrn2_z2) {
        for (count = 0; count < data->number; count++)
            data->data[count] = data->wrn2_z1;
        return data->number;
    }

    /* The decoding is done in 64 element chunks (each stored in two separate byte arrays) */
    for(block64 = 0; block64 < data->number; block64 += 64) {
        /* set block_size to the number of elements expected from the decoder */
        /*   (either the number of remaining elements, or a cap of 64).       */
        block_size = data->number - block64;
        if (block_size > 64)
            block_size = 64;

        /* decode one block of input (if all is well, this will match the expected block_size) */

        /* read one line from input file and then decode it */
        scan = fscanf(data->mhd_fp, "%[^\n]%n\n", buffer, &bytes_read);
        if(scan < 0) {
            fprintf(stderr, "Difficulting reading from MHD file.\n");
        }
        num_decoded = wrndec(buffer, msbits);
        if(num_decoded != block_size) {
#ifdef DEBUG_READ_VERBOSE
            printf("readRLEData[1]: num_decoded (%d) != block_size (%d) [number = %d, block64 = %d]\n",
                   num_decoded, block_size, data->number, block64);
#endif
            data->number = -2;    /* return an error code */
            return data->number;
        }

        /* read one line from input file and then decode it */
        scan = fscanf(data->mhd_fp, "%[^\n]%n\n", buffer, &bytes_read);
        if(scan < 0) {
            fprintf(stderr, "Difficulting reading from MHD file.\n");
        }
        num_decoded = wrndec(buffer, lsbits);
        if(num_decoded != block_size) {
#ifdef DEBUG_READ_VERBOSE
            printf("readRLEData[2]: num_decoded (%d) != block_size (%d) [number = %d, block64 = %d]\n",
                   num_decoded, block_size, data->number, block64);
#endif
            data->number = -2;    /* return an error code */
            return data->number;
        }

        /* Okay, we've decoded the MSbits and the LSbits of */
        /* data now convert them into floating point values */
        dzi = (data->wrn2_z2 - data->wrn2_z1) / 4410.0;
        for (count = 0; count < block_size; count++) {
            /* subtract the 33 that was added to put it above the ascii control-characters */
            msbits[count] -= 33;
            lsbits[count] -= 33;

            /* the sign is stored in msbits[] as the 47-or-greater-bit */
            sign = 1.0;
            if (msbits[count] >= 47) {
                sign = -1.0;
                msbits[count] -= 47;
            }

            /* reconstruct the unsigned integer value */
            int_value = (int) (lsbits[count] + 94.0 * msbits[count]);

            /* reconstruct the original floating point value from the range information & sign */
            /* Note that the following expression uses float(exp()) instead */
            /* of expf(). This is because we are using gcc 3.2 on a solaris */
            /* machine, which doesn't support expf. We should change this */
            /* when we no longer support gcc 3.2. */
           
            data->data[block64+count] = sign * float(exp(dzi * int_value + data->wrn2_z1));
        }
    }

    return data->number;
}

/** field_match(): returns a boolean value of whether the given field
 * name is in the list of requested fields.  If found, the name is
 * removed from the list, and a 1 is returned.  Otherwise 0 is
 * returned.  If the value of the match_list is NULL, then it is
 * assumed that all fields should be read, and therefore 1 is
 * returned */
int field_match(const char * const match_list[], const char * const field_name)
{
    int    count;
    int    match = 0;    /* no match found so far */
    if (match_list == NULL)
        return 1;

#if 0
    char *check;
    for(check = match_list[count = 0]; check != NULL; check = match_list[++count]) {
        if(strcmp(check, field_name) == 0) {
            match = 1;
        }
    }
#else
    for(count = 0; match_list[count] != NULL; ++count) {
        if(strcmp(match_list[count], field_name) == 0) {
            match = 1;
            break;
        }
    }
#endif
    return match;
}

/** readMHDField() reads the next wrn ASCII-encoded MHD data field into
 * the MHDdata structure
 *
 * NOTE:  the first line of the field is given by the "field_info"
 *  argument.  The rest of the data is from the "mhd_fp" field of
 *  MHDField -- though that may change in the (near) future.
 *  The header information and a pointer to the data array are
 *  returned via the MHDdata structure
 *
 * The format of the ASCII field representation is 5 lines of header
 *  info, followed by approximately (number-of-elements/32) lines
 *  of WRN ASCII encoded data.
 *    Line 1: ...
 *    Line 2: ...
 *    Line 3: ...
 *    Line 4: ...
 *    Line 5: ... */
int readMHDField(MHDdata *data, char *field_info,
                 const char * const match_list[])
{
    char    junk[256];
    int    scan;          /* return value from a scanf() call */
    int    rank;
    int    num_elements;  /* return value of this function */

    /* determine the beginning position of the file */
    data->file_position = ftell(data->mhd_fp);
    data->file_position -= (7 + strlen(field_info));

    /* Line 1: the number of dimensions, and rank of each element */
    scan = sscanf(field_info, "%dD-%d%[^\n]\n", &data->dim_rank, &data->value_rank, junk);
    if (scan != 2) {
        printf("readMHDField() trouble scanning line 1, got %d values instead of 2\n", scan);
    }

    /* Line 2: the name of this field */
    scan = fscanf(data->mhd_fp, "%[^\n ]\n", &data->field_name[0]);
    /* scan = fscanf(data->mhd_fp, "%[a-z]\n", &data->field_name[0]) */;
    if (scan != 1) {
        printf("readMHDField() trouble scanning line 2, got %d values instead of 1\n", scan);
    }
#ifdef DEBUG_READ_VERBOSE
    printf("reading Field '%s' of type '%s'\n", data->field_name, field_info);
#endif

    if(strcmp(data->field_name,"bx") == 0 ||
       strcmp(data->field_name,"by") == 0 ||
       strcmp(data->field_name,"bz") == 0 ||
       strcmp(data->field_name,"vx") == 0 ||
       strcmp(data->field_name,"vy") == 0 ||
       strcmp(data->field_name,"vz") == 0 ||
       strcmp(data->field_name,"xjx") == 0 ||
       strcmp(data->field_name,"xjy") == 0 ||
       strcmp(data->field_name,"xjz") == 0 ||
       strcmp(data->field_name,"rr") == 0 ||
       strcmp(data->field_name,"pp") == 0 ||
       strcmp(data->field_name,"resis") == 0) {
        /* then Line 3 is time data. (TJF) */
        read_time_line(data);
#if DEBUG
        fprintf(stderr, "read time data: %lf, %lf\n\t %u:%u:%u:%u:%u:%lf\n",
                        data->time_since_model_start,
                        data->some_time,
                        data->year, data->month, data->day,
                        data->hour, data->minute, data->second);
#endif
    } else {
        /* then Line 3 is the name of the grid file, or something else we don't
         * care about. */
        scan = fscanf(data->mhd_fp, "%[^\n]\n", &data->line3_info[0]);
        if(scan != 1) {
            printf("readMHDField() trouble scanning line 3, got ");
            printf("%d values instead of 1\n", scan);
        }
    }

    /* Line 4: info about the dimensionality of the data */
    scan = fscanf(data->mhd_fp, " %d", &data->line4_value);
    if (scan != 1) {
        printf("readMHDField() trouble scanning line 4a, got %d values instead of 1\n", scan);
    }
    for (rank = 0; rank < data->dim_rank; rank++) {
      scan = fscanf(data->mhd_fp, " %d", &data->dim_size[rank]);
        if (scan != 1) {
            printf("readMHDField() trouble scanning line 4b, got %d values instead of 1\n", scan);
        }
    }
    scan = fscanf(data->mhd_fp, "%[^\n]\n", junk);  /* go to beginning of next line */
    /* Line 5: The WRN2 line */
    /* NOTE: perhaps the "2" in WRN2 is an indicator of the number of bytes */
    /*   into which the values are stored.                                  */
    /* Sometimes "WRN2" (data->encoding) is not separated from the # of values
     *  (data->number) via a space. not sure why, but the original code gets
     *  screwed up because it tries to read everything in one go. This is a
     * slightly modified fscanf that limits the first field to 4 chars */
    scan = fscanf(data->mhd_fp, "%4s %d%f %f %f %d%[^\n]\n", data->encoding,
                  &data->number, &data->wrn2_z1, &data->wrn2_z2,
                  &data->wrn2_rid, &data->wrn2_it, &data->wrn2_cid[0]);
    if(scan != 7) {
        printf("readMHDField() trouble scanning line 5, got %d values instead of 7\n", scan);
    }
    /* Lines 6 to ~(data->number / 32): the ascii encoded data */
    /*   Only read the data if this field matches a requested field, otherwise skip */
    if(!field_match(match_list, data->field_name)) {
#ifdef DEBUG_READ
      printf("readMHDField() No match.\n");
#endif
        int    count;    /* count through the lines */
        int    skip;    /* number of lines to skip */

        skip = data->number / 64;
        if(data->number % 64 != 0)
            skip++;
        skip *= 2;
        for(count = 0; count < skip; count++)
            fscanf(data->mhd_fp, "%*[^\n]\n");    /* read one line */

        num_elements = 0;
    } else {
#ifdef DEBUG_READ
      printf("readMHDField() reading data\n");
#endif
        num_elements = readRLEData(data);
    }
    return num_elements;
}

/* returns the mhd file in a linked list sort of structure called 'MHDdata'. As
 * far as I can tell, the structure itself is dynamically allocated, as well as
 * the 'data' pointer within each element. The calling program will need to
 * deallocate these.
 * 'field_list' also seems to be modified by this function or another it calls.
 * At best, this causes a memory leak. Worse, usually field_list ends up being
 * allocated on the stack... this might cause stack corruption!
 ****
 * Tom Fogal, Thu Aug 19 15:42:47 EST 2004
 *    Comments, fixes. I removed the ability to read a file from stdin. Seems
 *    an unlikely use anyway. */
MHDdata *ggcm_read_mhd(const char *file_name, const char *field_list[])
{
    FILE    *mhd_fp;   /* pointer to the MHD file to read from */
    MHDdata *data;     /* pointer to the current MHD dataset */
    MHDdata *first;    /* pointer to the first MHD dataset */
    MHDdata *last;     /* pointer to the last MHD dataset */
    char    info[256];
    int     scan;      /* return value from a scanf */

    mhd_fp = fopen(file_name, "r");
    if (mhd_fp == NULL) {
        fprintf(stderr, "Error, unable to open '%s'\n", file_name);
        exit(1);
    }

    /* first scan the preliminary listing -- I guess that's what it is */
    while (fscanf(mhd_fp, "FIELD-%[^\n]\n", info)) {
#ifdef DEBUG_READ
        fprintf(stderr,"prelim: expect field of '%s'\n", info);
#endif
    }

    first = (MHDdata *)malloc(sizeof(MHDdata));
    last = first;
    data = first;
    ggcm_init_mhd(data);

    /* Now scan the actual data listings */
    /*   We can stop either when there are no more FIELDs to read, or */
    /*   when all the fields in the field_list have been read.        */
    while (!feof(mhd_fp) && field_list[0] != NULL) {
        scan = fscanf(mhd_fp, "FIELD-%[^\n]\n", info);
        if (scan < 0) {
            fprintf(stderr, "Difficulty reading from MHD file.\n");
        }
        data->mhd_fp = mhd_fp;
        strcpy(data->filename, (char* const)file_name);
        if(readMHDField(data, info, field_list) > 0) {
#ifdef DEBUG_READ
            printf("Just read data '%s'\n", data->field_name);
#endif
            assert(data->data); /* must have read in some data */
            /* we read some data, so there was a match with the field_list */
            data->next = (MHDdata *)malloc(sizeof(MHDdata));
            ggcm_init_mhd(data->next); /* initialize space for the next field */
            last = data;
            data = data->next;
        } else {
#ifdef DEBUG_READ
            printf("Just ignored data '%s'\n", data->field_name);
#endif
            /* if the field didn't match, then reset the values in "data" */
            ggcm_init_mhd(data);
        }
        info[0] = 'y'; info[1] = 'o'; info[2] = 0;
    }

    /* remove the link to the empty data set from the last read MHD strcuture */
    ggcm_free_data(last->next);
    last->next = NULL;

    fclose(mhd_fp);
    return first;
}

void ggcm_free_data(MHDdata *mhd)
{
    MHDdata *p, *prev;

    prev = p = mhd;
    while(p) {
        p = p->next;
        free(prev->data);
        free(prev);
        prev = p;
    };
}

static int read_time_line(MHDdata *data)
{
    char junk[256]; /* for throwing away some filename data */
    int matches;

    /* file pointer must be valid */
    assert(data != NULL);
    assert(data->mhd_fp != NULL);
    /* and have data ready for us. */
    assert(feof(data->mhd_fp) == 0 && ferror(data->mhd_fp) == 0);

    matches = fscanf(data->mhd_fp,
                     "time= %lf %lf %u:%u:%u:%u:%u:%lf %[^\n ]\n",
                     &data->time_since_model_start,  /* %lf */
                     &data->some_time,               /* %lf */
                     &data->year,                    /* %u */
                     &data->month,                   /* %u */
                     &data->day,                     /* %u */
                     &data->hour,                    /* %u */
                     &data->minute,                  /* %u */
                     &data->second,                  /* %lf */
                     junk);                          /* %[\n ] */
    if(matches != 9) {
        fputs("Error reading time code information!", stderr);
        return -1;
    }
    return 0;
}