]>
Commit | Line | Data |
---|---|---|
4a35548b MS |
1 | /* |
2 | * This file is part of the libsigrok project. | |
3 | * | |
4 | * Copyright (C) 2013 Marc Schink <sigrok-dev@marcschink.de> | |
5 | * | |
6 | * This program is free software: you can redistribute it and/or modify | |
7 | * it under the terms of the GNU General Public License as published by | |
8 | * the Free Software Foundation, either version 3 of the License, or | |
9 | * (at your option) any later version. | |
10 | * | |
11 | * This program is distributed in the hope that it will be useful, | |
12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
14 | * GNU General Public License for more details. | |
15 | * | |
16 | * You should have received a copy of the GNU General Public License | |
17 | * along with this program. If not, see <http://www.gnu.org/licenses/>. | |
18 | */ | |
19 | ||
e05f1827 GS |
20 | #include "config.h" |
21 | ||
22 | #include <glib.h> | |
4a35548b MS |
23 | #include <stdlib.h> |
24 | #include <string.h> | |
e05f1827 | 25 | |
c1aae900 | 26 | #include <libsigrok/libsigrok.h> |
4a35548b MS |
27 | #include "libsigrok-internal.h" |
28 | ||
3544f848 | 29 | #define LOG_PREFIX "input/csv" |
4a35548b | 30 | |
9a4fd01a | 31 | #define CHUNK_SIZE (4 * 1024 * 1024) |
cd59e6ec | 32 | |
4a35548b MS |
33 | /* |
34 | * The CSV input module has the following options: | |
35 | * | |
36 | * single-column: Specifies the column number which stores the sample data for | |
37 | * single column mode and enables single column mode. Multi | |
38 | * column mode is used if this parameter is omitted. | |
39 | * | |
ba7dd8bb UH |
40 | * numchannels: Specifies the number of channels to use. In multi column mode |
41 | * the number of channels are the number of columns and in single | |
4a35548b | 42 | * column mode the number of bits (LSB first) beginning at |
ba7dd8bb | 43 | * 'first-channel'. |
4a35548b MS |
44 | * |
45 | * delimiter: Specifies the delimiter for columns. Must be at least one | |
46 | * character. Comma is used as default delimiter. | |
47 | * | |
48 | * format: Specifies the format of the sample data in single column mode. | |
49 | * Available formats are: 'bin', 'hex' and 'oct'. The binary | |
50 | * format is used by default. This option has no effect in multi | |
51 | * column mode. | |
52 | * | |
53 | * comment: Specifies the prefix character(s) for comments. No prefix | |
54 | * characters are used by default which disables removing of | |
55 | * comments. | |
56 | * | |
57 | * samplerate: Samplerate which the sample data was captured with. Default | |
58 | * value is 0. | |
59 | * | |
ba7dd8bb UH |
60 | * first-channel: Column number of the first channel in multi column mode and |
61 | * position of the bit for the first channel in single column mode. | |
4a35548b MS |
62 | * Default value is 0. |
63 | * | |
64 | * header: Determines if the first line should be treated as header | |
ba7dd8bb UH |
65 | * and used for channel names in multi column mode. Empty header |
66 | * names will be replaced by the channel number. If enabled in | |
4a35548b MS |
67 | * single column mode the first line will be skipped. Usage of |
68 | * header is disabled by default. | |
69 | * | |
70 | * startline: Line number to start processing sample data. Must be greater | |
71 | * than 0. The default line number to start processing is 1. | |
72 | */ | |
73 | ||
ccff468b GS |
74 | /* |
75 | * TODO | |
76 | * | |
77 | * - Determine how the text line handling can get improved, regarding | |
78 | * all of robustness and flexibility and correctness. | |
79 | * - The current implementation splits on "any run of CR and LF". Which | |
80 | * translates to: Line numbers are wrong in the presence of empty | |
de788af4 | 81 | * lines in the input stream. See below for an (expensive) fix. |
ccff468b GS |
82 | * - Dropping support for CR style end-of-line markers could improve |
83 | * the situation a lot. Code could search for and split on LF, and | |
84 | * trim optional trailing CR. This would result in proper support | |
85 | * for CRLF (Windows) as well as LF (Unix), and allow for correct | |
86 | * line number counts. | |
87 | * - When support for CR-only line termination cannot get dropped, | |
88 | * then the current implementation is inappropriate. Currently the | |
89 | * input stream is scanned for the first occurance of either of the | |
90 | * supported termination styles (which is good). For the remaining | |
91 | * session a consistent encoding of the text lines is assumed (which | |
de788af4 | 92 | * is acceptable). |
ccff468b GS |
93 | * - When line numbers need to be correct and reliable, _and_ the full |
94 | * set of previously supported line termination sequences are required, | |
95 | * and potentially more are to get added for improved compatibility | |
96 | * with more platforms or generators, then the current approach of | |
97 | * splitting on runs of termination characters needs to get replaced, | |
98 | * by the more expensive approach to scan for and count the initially | |
99 | * determined termination sequence. | |
100 | * | |
101 | * - Add support for analog input data? (optional) | |
102 | * - Needs a syntax first for user specs which channels (columns) are | |
103 | * logic and which are analog. May need heuristics(?) to guess from | |
104 | * input data in the absence of user provided specs. | |
105 | */ | |
106 | ||
4a35548b MS |
107 | /* Single column formats. */ |
108 | enum { | |
109 | FORMAT_BIN, | |
110 | FORMAT_HEX, | |
111 | FORMAT_OCT | |
112 | }; | |
113 | ||
114 | struct context { | |
41d214f6 BV |
115 | gboolean started; |
116 | ||
4a35548b MS |
117 | /* Current selected samplerate. */ |
118 | uint64_t samplerate; | |
119 | ||
ba7dd8bb | 120 | /* Number of channels. */ |
6433156c | 121 | unsigned int num_channels; |
4a35548b MS |
122 | |
123 | /* Column delimiter character(s). */ | |
124 | GString *delimiter; | |
125 | ||
126 | /* Comment prefix character(s). */ | |
127 | GString *comment; | |
128 | ||
d9251a2c | 129 | /* Termination character(s) used in current stream. */ |
41d214f6 BV |
130 | char *termination; |
131 | ||
4a35548b MS |
132 | /* Determines if sample data is stored in multiple columns. */ |
133 | gboolean multi_column_mode; | |
134 | ||
135 | /* Column number of the sample data in single column mode. */ | |
6433156c | 136 | unsigned int single_column; |
4a35548b MS |
137 | |
138 | /* | |
139 | * Number of the first column to parse. Equivalent to the number of the | |
ba7dd8bb | 140 | * first channel in multi column mode and the single column number in |
4a35548b MS |
141 | * single column mode. |
142 | */ | |
6433156c | 143 | unsigned int first_column; |
4a35548b MS |
144 | |
145 | /* | |
ba7dd8bb UH |
146 | * Column number of the first channel in multi column mode and position of |
147 | * the bit for the first channel in single column mode. | |
4a35548b | 148 | */ |
6433156c | 149 | unsigned int first_channel; |
4a35548b MS |
150 | |
151 | /* Line number to start processing. */ | |
6433156c | 152 | size_t start_line; |
4a35548b MS |
153 | |
154 | /* | |
155 | * Determines if the first line should be treated as header and used for | |
ba7dd8bb | 156 | * channel names in multi column mode. |
4a35548b MS |
157 | */ |
158 | gboolean header; | |
159 | ||
160 | /* Format sample data is stored in single column mode. */ | |
161 | int format; | |
162 | ||
cd59e6ec GS |
163 | size_t sample_unit_size; /**!< Byte count for a single sample. */ |
164 | uint8_t *sample_buffer; /**!< Buffer for a single sample. */ | |
4a35548b | 165 | |
cd59e6ec GS |
166 | uint8_t *datafeed_buffer; /**!< Queue for datafeed submission. */ |
167 | size_t datafeed_buf_size; | |
168 | size_t datafeed_buf_fill; | |
4a35548b | 169 | |
4a35548b | 170 | /* Current line number. */ |
6433156c | 171 | size_t line_number; |
affaf540 GS |
172 | |
173 | /* List of previously created sigrok channels. */ | |
174 | GSList *prev_sr_channels; | |
4a35548b MS |
175 | }; |
176 | ||
41d214f6 | 177 | static void strip_comment(char *buf, const GString *prefix) |
4a35548b MS |
178 | { |
179 | char *ptr; | |
180 | ||
181 | if (!prefix->len) | |
182 | return; | |
183 | ||
41d214f6 BV |
184 | if ((ptr = strstr(buf, prefix->str))) |
185 | *ptr = '\0'; | |
4a35548b MS |
186 | } |
187 | ||
41d214f6 | 188 | static int parse_binstr(const char *str, struct context *inc) |
4a35548b MS |
189 | { |
190 | gsize i, j, length; | |
191 | ||
192 | length = strlen(str); | |
193 | ||
194 | if (!length) { | |
6433156c | 195 | sr_err("Column %u in line %zu is empty.", inc->single_column, |
41d214f6 | 196 | inc->line_number); |
4a35548b MS |
197 | return SR_ERR; |
198 | } | |
199 | ||
200 | /* Clear buffer in order to set bits only. */ | |
cd59e6ec | 201 | memset(inc->sample_buffer, 0, inc->sample_unit_size); |
4a35548b | 202 | |
41d214f6 | 203 | i = inc->first_channel; |
4a35548b | 204 | |
41d214f6 | 205 | for (j = 0; i < length && j < inc->num_channels; i++, j++) { |
4a35548b | 206 | if (str[length - i - 1] == '1') { |
41d214f6 | 207 | inc->sample_buffer[j / 8] |= (1 << (j % 8)); |
4a35548b | 208 | } else if (str[length - i - 1] != '0') { |
6433156c | 209 | sr_err("Invalid value '%s' in column %u in line %zu.", |
41d214f6 | 210 | str, inc->single_column, inc->line_number); |
4a35548b MS |
211 | return SR_ERR; |
212 | } | |
213 | } | |
214 | ||
215 | return SR_OK; | |
216 | } | |
217 | ||
41d214f6 | 218 | static int parse_hexstr(const char *str, struct context *inc) |
4a35548b MS |
219 | { |
220 | gsize i, j, k, length; | |
221 | uint8_t value; | |
222 | char c; | |
223 | ||
224 | length = strlen(str); | |
225 | ||
226 | if (!length) { | |
6433156c | 227 | sr_err("Column %u in line %zu is empty.", inc->single_column, |
41d214f6 | 228 | inc->line_number); |
4a35548b MS |
229 | return SR_ERR; |
230 | } | |
231 | ||
232 | /* Clear buffer in order to set bits only. */ | |
cd59e6ec | 233 | memset(inc->sample_buffer, 0, inc->sample_unit_size); |
4a35548b MS |
234 | |
235 | /* Calculate the position of the first hexadecimal digit. */ | |
41d214f6 | 236 | i = inc->first_channel / 4; |
4a35548b | 237 | |
41d214f6 | 238 | for (j = 0; i < length && j < inc->num_channels; i++) { |
4a35548b MS |
239 | c = str[length - i - 1]; |
240 | ||
241 | if (!g_ascii_isxdigit(c)) { | |
6433156c | 242 | sr_err("Invalid value '%s' in column %u in line %zu.", |
41d214f6 | 243 | str, inc->single_column, inc->line_number); |
4a35548b MS |
244 | return SR_ERR; |
245 | } | |
246 | ||
247 | value = g_ascii_xdigit_value(c); | |
248 | ||
41d214f6 | 249 | k = (inc->first_channel + j) % 4; |
4a35548b | 250 | |
41d214f6 | 251 | for (; j < inc->num_channels && k < 4; k++) { |
4a35548b | 252 | if (value & (1 << k)) |
41d214f6 | 253 | inc->sample_buffer[j / 8] |= (1 << (j % 8)); |
4a35548b MS |
254 | |
255 | j++; | |
256 | } | |
257 | } | |
258 | ||
259 | return SR_OK; | |
260 | } | |
261 | ||
41d214f6 | 262 | static int parse_octstr(const char *str, struct context *inc) |
4a35548b MS |
263 | { |
264 | gsize i, j, k, length; | |
265 | uint8_t value; | |
266 | char c; | |
267 | ||
268 | length = strlen(str); | |
269 | ||
270 | if (!length) { | |
6433156c | 271 | sr_err("Column %u in line %zu is empty.", inc->single_column, |
41d214f6 | 272 | inc->line_number); |
4a35548b MS |
273 | return SR_ERR; |
274 | } | |
275 | ||
276 | /* Clear buffer in order to set bits only. */ | |
cd59e6ec | 277 | memset(inc->sample_buffer, 0, inc->sample_unit_size); |
4a35548b MS |
278 | |
279 | /* Calculate the position of the first octal digit. */ | |
41d214f6 | 280 | i = inc->first_channel / 3; |
4a35548b | 281 | |
41d214f6 | 282 | for (j = 0; i < length && j < inc->num_channels; i++) { |
4a35548b MS |
283 | c = str[length - i - 1]; |
284 | ||
285 | if (c < '0' || c > '7') { | |
6433156c | 286 | sr_err("Invalid value '%s' in column %u in line %zu.", |
41d214f6 | 287 | str, inc->single_column, inc->line_number); |
4a35548b MS |
288 | return SR_ERR; |
289 | } | |
290 | ||
291 | value = g_ascii_xdigit_value(c); | |
292 | ||
41d214f6 | 293 | k = (inc->first_channel + j) % 3; |
4a35548b | 294 | |
41d214f6 | 295 | for (; j < inc->num_channels && k < 3; k++) { |
4a35548b | 296 | if (value & (1 << k)) |
41d214f6 | 297 | inc->sample_buffer[j / 8] |= (1 << (j % 8)); |
4a35548b MS |
298 | |
299 | j++; | |
300 | } | |
301 | } | |
302 | ||
303 | return SR_OK; | |
304 | } | |
305 | ||
41d214f6 | 306 | static char **parse_line(char *buf, struct context *inc, int max_columns) |
4a35548b MS |
307 | { |
308 | const char *str, *remainder; | |
309 | GSList *list, *l; | |
310 | char **columns; | |
311 | char *column; | |
312 | gsize n, k; | |
313 | ||
314 | n = 0; | |
315 | k = 0; | |
316 | list = NULL; | |
317 | ||
41d214f6 BV |
318 | remainder = buf; |
319 | str = strstr(remainder, inc->delimiter->str); | |
4a35548b MS |
320 | |
321 | while (str && max_columns) { | |
41d214f6 | 322 | if (n >= inc->first_column) { |
4a35548b MS |
323 | column = g_strndup(remainder, str - remainder); |
324 | list = g_slist_prepend(list, g_strstrip(column)); | |
325 | ||
326 | max_columns--; | |
327 | k++; | |
328 | } | |
329 | ||
41d214f6 BV |
330 | remainder = str + inc->delimiter->len; |
331 | str = strstr(remainder, inc->delimiter->str); | |
4a35548b MS |
332 | n++; |
333 | } | |
334 | ||
41d214f6 | 335 | if (buf[0] && max_columns && n >= inc->first_column) { |
4a35548b MS |
336 | column = g_strdup(remainder); |
337 | list = g_slist_prepend(list, g_strstrip(column)); | |
338 | k++; | |
339 | } | |
340 | ||
341 | if (!(columns = g_try_new(char *, k + 1))) | |
342 | return NULL; | |
343 | ||
344 | columns[k--] = NULL; | |
345 | ||
346 | for (l = list; l; l = l->next) | |
347 | columns[k--] = l->data; | |
348 | ||
349 | g_slist_free(list); | |
350 | ||
351 | return columns; | |
352 | } | |
353 | ||
41d214f6 | 354 | static int parse_multi_columns(char **columns, struct context *inc) |
4a35548b MS |
355 | { |
356 | gsize i; | |
df0db9fd | 357 | char *column; |
4a35548b MS |
358 | |
359 | /* Clear buffer in order to set bits only. */ | |
cd59e6ec | 360 | memset(inc->sample_buffer, 0, inc->sample_unit_size); |
4a35548b | 361 | |
41d214f6 | 362 | for (i = 0; i < inc->num_channels; i++) { |
df0db9fd GS |
363 | column = columns[i]; |
364 | if (column[0] == '1') { | |
41d214f6 | 365 | inc->sample_buffer[i / 8] |= (1 << (i % 8)); |
df0db9fd | 366 | } else if (!strlen(column)) { |
4a35548b | 367 | sr_err("Column %zu in line %zu is empty.", |
41d214f6 | 368 | inc->first_channel + i, inc->line_number); |
4a35548b | 369 | return SR_ERR; |
df0db9fd | 370 | } else if (column[0] != '0') { |
4a35548b | 371 | sr_err("Invalid value '%s' in column %zu in line %zu.", |
df0db9fd | 372 | column, inc->first_channel + i, |
41d214f6 | 373 | inc->line_number); |
4a35548b MS |
374 | return SR_ERR; |
375 | } | |
376 | } | |
377 | ||
378 | return SR_OK; | |
379 | } | |
380 | ||
41d214f6 | 381 | static int parse_single_column(const char *column, struct context *inc) |
4a35548b MS |
382 | { |
383 | int res; | |
384 | ||
385 | res = SR_ERR; | |
386 | ||
0c5f2abc | 387 | switch (inc->format) { |
4a35548b | 388 | case FORMAT_BIN: |
41d214f6 | 389 | res = parse_binstr(column, inc); |
4a35548b MS |
390 | break; |
391 | case FORMAT_HEX: | |
41d214f6 | 392 | res = parse_hexstr(column, inc); |
4a35548b MS |
393 | break; |
394 | case FORMAT_OCT: | |
41d214f6 | 395 | res = parse_octstr(column, inc); |
4a35548b MS |
396 | break; |
397 | } | |
398 | ||
399 | return res; | |
400 | } | |
401 | ||
cd59e6ec | 402 | static int flush_samples(const struct sr_input *in) |
4a35548b | 403 | { |
cd59e6ec | 404 | struct context *inc; |
4a35548b MS |
405 | struct sr_datafeed_packet packet; |
406 | struct sr_datafeed_logic logic; | |
cd59e6ec | 407 | int rc; |
4a35548b | 408 | |
cd59e6ec GS |
409 | inc = in->priv; |
410 | if (!inc->datafeed_buf_fill) | |
411 | return SR_OK; | |
412 | ||
413 | memset(&packet, 0, sizeof(packet)); | |
414 | memset(&logic, 0, sizeof(logic)); | |
4a35548b MS |
415 | packet.type = SR_DF_LOGIC; |
416 | packet.payload = &logic; | |
cd59e6ec GS |
417 | logic.unitsize = inc->sample_unit_size; |
418 | logic.length = inc->datafeed_buf_fill; | |
419 | logic.data = inc->datafeed_buffer; | |
420 | ||
421 | rc = sr_session_send(in->sdi, &packet); | |
422 | if (rc != SR_OK) | |
423 | return rc; | |
4a35548b | 424 | |
cd59e6ec GS |
425 | inc->datafeed_buf_fill = 0; |
426 | return SR_OK; | |
427 | } | |
428 | ||
429 | static int queue_samples(const struct sr_input *in) | |
430 | { | |
431 | struct context *inc; | |
432 | int rc; | |
433 | ||
434 | inc = in->priv; | |
435 | ||
436 | inc->datafeed_buf_fill += inc->sample_unit_size; | |
437 | if (inc->datafeed_buf_fill == inc->datafeed_buf_size) { | |
438 | rc = flush_samples(in); | |
439 | if (rc != SR_OK) | |
440 | return rc; | |
441 | } | |
442 | inc->sample_buffer = &inc->datafeed_buffer[inc->datafeed_buf_fill]; | |
4a35548b MS |
443 | return SR_OK; |
444 | } | |
445 | ||
41d214f6 | 446 | static int init(struct sr_input *in, GHashTable *options) |
4a35548b | 447 | { |
41d214f6 BV |
448 | struct context *inc; |
449 | const char *s; | |
4a35548b | 450 | |
aac29cc1 | 451 | in->sdi = g_malloc0(sizeof(struct sr_dev_inst)); |
41d214f6 | 452 | in->priv = inc = g_malloc0(sizeof(struct context)); |
4a35548b | 453 | |
41d214f6 BV |
454 | inc->single_column = g_variant_get_int32(g_hash_table_lookup(options, "single-column")); |
455 | inc->multi_column_mode = inc->single_column == 0; | |
4a35548b | 456 | |
41d214f6 | 457 | inc->num_channels = g_variant_get_int32(g_hash_table_lookup(options, "numchannels")); |
4a35548b | 458 | |
41d214f6 BV |
459 | inc->delimiter = g_string_new(g_variant_get_string( |
460 | g_hash_table_lookup(options, "delimiter"), NULL)); | |
461 | if (inc->delimiter->len == 0) { | |
462 | sr_err("Delimiter must be at least one character."); | |
463 | return SR_ERR_ARG; | |
4a35548b MS |
464 | } |
465 | ||
41d214f6 BV |
466 | s = g_variant_get_string(g_hash_table_lookup(options, "format"), NULL); |
467 | if (!g_ascii_strncasecmp(s, "bin", 3)) { | |
468 | inc->format = FORMAT_BIN; | |
469 | } else if (!g_ascii_strncasecmp(s, "hex", 3)) { | |
470 | inc->format = FORMAT_HEX; | |
471 | } else if (!g_ascii_strncasecmp(s, "oct", 3)) { | |
472 | inc->format = FORMAT_OCT; | |
473 | } else { | |
474 | sr_err("Invalid format: '%s'", s); | |
475 | return SR_ERR_ARG; | |
4a35548b MS |
476 | } |
477 | ||
41d214f6 BV |
478 | inc->comment = g_string_new(g_variant_get_string( |
479 | g_hash_table_lookup(options, "comment"), NULL)); | |
480 | if (g_string_equal(inc->comment, inc->delimiter)) { | |
481 | /* That's never going to work. Likely the result of the user | |
482 | * setting the delimiter to ; -- the default comment. Clearing | |
483 | * the comment setting will work in that case. */ | |
484 | g_string_truncate(inc->comment, 0); | |
4a35548b MS |
485 | } |
486 | ||
6e8d95a5 | 487 | inc->samplerate = g_variant_get_uint64(g_hash_table_lookup(options, "samplerate")); |
4a35548b | 488 | |
41d214f6 | 489 | inc->first_channel = g_variant_get_int32(g_hash_table_lookup(options, "first-channel")); |
4a35548b | 490 | |
41d214f6 | 491 | inc->header = g_variant_get_boolean(g_hash_table_lookup(options, "header")); |
4a35548b | 492 | |
41d214f6 BV |
493 | inc->start_line = g_variant_get_int32(g_hash_table_lookup(options, "startline")); |
494 | if (inc->start_line < 1) { | |
6433156c | 495 | sr_err("Invalid start line %zu.", inc->start_line); |
41d214f6 | 496 | return SR_ERR_ARG; |
4a35548b MS |
497 | } |
498 | ||
41d214f6 BV |
499 | if (inc->multi_column_mode) |
500 | inc->first_column = inc->first_channel; | |
4a35548b | 501 | else |
41d214f6 | 502 | inc->first_column = inc->single_column; |
4a35548b | 503 | |
41d214f6 | 504 | if (!inc->multi_column_mode && !inc->num_channels) { |
ba7dd8bb | 505 | sr_err("Number of channels needs to be specified in single column mode."); |
41d214f6 | 506 | return SR_ERR_ARG; |
4a35548b MS |
507 | } |
508 | ||
41d214f6 BV |
509 | return SR_OK; |
510 | } | |
4a35548b | 511 | |
affaf540 GS |
512 | /* |
513 | * Check the channel list for consistency across file re-import. See | |
514 | * the VCD input module for more details and motivation. | |
515 | */ | |
516 | ||
517 | static void keep_header_for_reread(const struct sr_input *in) | |
518 | { | |
519 | struct context *inc; | |
520 | ||
521 | inc = in->priv; | |
522 | g_slist_free_full(inc->prev_sr_channels, sr_channel_free_cb); | |
523 | inc->prev_sr_channels = in->sdi->channels; | |
524 | in->sdi->channels = NULL; | |
525 | } | |
526 | ||
527 | static int check_header_in_reread(const struct sr_input *in) | |
528 | { | |
529 | struct context *inc; | |
530 | ||
531 | if (!in) | |
532 | return FALSE; | |
533 | inc = in->priv; | |
534 | if (!inc) | |
535 | return FALSE; | |
536 | if (!inc->prev_sr_channels) | |
537 | return TRUE; | |
538 | ||
539 | if (sr_channel_lists_differ(inc->prev_sr_channels, in->sdi->channels)) { | |
540 | sr_err("Channel list change not supported for file re-read."); | |
541 | return FALSE; | |
542 | } | |
543 | g_slist_free_full(in->sdi->channels, sr_channel_free_cb); | |
544 | in->sdi->channels = inc->prev_sr_channels; | |
545 | inc->prev_sr_channels = NULL; | |
546 | ||
547 | return TRUE; | |
548 | } | |
549 | ||
492dfa90 GS |
550 | static const char *delim_set = "\r\n"; |
551 | ||
329733d9 | 552 | static const char *get_line_termination(GString *buf) |
41d214f6 | 553 | { |
329733d9 | 554 | const char *term; |
4a35548b | 555 | |
41d214f6 BV |
556 | term = NULL; |
557 | if (g_strstr_len(buf->str, buf->len, "\r\n")) | |
558 | term = "\r\n"; | |
559 | else if (memchr(buf->str, '\n', buf->len)) | |
560 | term = "\n"; | |
561 | else if (memchr(buf->str, '\r', buf->len)) | |
562 | term = "\r"; | |
4a35548b | 563 | |
41d214f6 BV |
564 | return term; |
565 | } | |
4a35548b | 566 | |
41d214f6 BV |
567 | static int initial_parse(const struct sr_input *in, GString *buf) |
568 | { | |
569 | struct context *inc; | |
41d214f6 | 570 | GString *channel_name; |
6433156c DE |
571 | unsigned int num_columns, i; |
572 | size_t line_number, l; | |
41d214f6 | 573 | int ret; |
df0db9fd | 574 | char **lines, *line, **columns, *column; |
41d214f6 BV |
575 | |
576 | ret = SR_OK; | |
577 | inc = in->priv; | |
578 | columns = NULL; | |
579 | ||
580 | line_number = 0; | |
492dfa90 | 581 | lines = g_strsplit_set(buf->str, delim_set, 0); |
41d214f6 BV |
582 | for (l = 0; lines[l]; l++) { |
583 | line_number++; | |
df0db9fd | 584 | line = lines[l]; |
41d214f6 BV |
585 | if (inc->start_line > line_number) { |
586 | sr_spew("Line %zu skipped.", line_number); | |
4a35548b MS |
587 | continue; |
588 | } | |
df0db9fd | 589 | if (line[0] == '\0') { |
41d214f6 BV |
590 | sr_spew("Blank line %zu skipped.", line_number); |
591 | continue; | |
592 | } | |
df0db9fd GS |
593 | strip_comment(line, inc->comment); |
594 | if (line[0] == '\0') { | |
41d214f6 | 595 | sr_spew("Comment-only line %zu skipped.", line_number); |
4a35548b MS |
596 | continue; |
597 | } | |
598 | ||
41d214f6 BV |
599 | /* Reached first proper line. */ |
600 | break; | |
601 | } | |
602 | if (!lines[l]) { | |
603 | /* Not enough data for a proper line yet. */ | |
60107497 | 604 | ret = SR_ERR_NA; |
41d214f6 | 605 | goto out; |
4a35548b MS |
606 | } |
607 | ||
608 | /* | |
609 | * In order to determine the number of columns parse the current line | |
610 | * without limiting the number of columns. | |
611 | */ | |
df0db9fd GS |
612 | columns = parse_line(line, inc, -1); |
613 | if (!columns) { | |
41d214f6 BV |
614 | sr_err("Error while parsing line %zu.", line_number); |
615 | ret = SR_ERR; | |
616 | goto out; | |
4a35548b | 617 | } |
4a35548b MS |
618 | num_columns = g_strv_length(columns); |
619 | ||
620 | /* Ensure that the first column is not out of bounds. */ | |
621 | if (!num_columns) { | |
6433156c | 622 | sr_err("Column %u in line %zu is out of bounds.", |
41d214f6 BV |
623 | inc->first_column, line_number); |
624 | ret = SR_ERR; | |
625 | goto out; | |
4a35548b MS |
626 | } |
627 | ||
41d214f6 | 628 | if (inc->multi_column_mode) { |
4a35548b | 629 | /* |
ba7dd8bb | 630 | * Detect the number of channels in multi column mode |
4a35548b MS |
631 | * automatically if not specified. |
632 | */ | |
41d214f6 BV |
633 | if (!inc->num_channels) { |
634 | inc->num_channels = num_columns; | |
6433156c | 635 | sr_dbg("Number of auto-detected channels: %u.", |
41d214f6 | 636 | inc->num_channels); |
4a35548b MS |
637 | } |
638 | ||
639 | /* | |
ba7dd8bb | 640 | * Ensure that the number of channels does not exceed the number |
4a35548b MS |
641 | * of columns in multi column mode. |
642 | */ | |
41d214f6 | 643 | if (num_columns < inc->num_channels) { |
ba7dd8bb | 644 | sr_err("Not enough columns for desired number of channels in line %zu.", |
41d214f6 BV |
645 | line_number); |
646 | ret = SR_ERR; | |
647 | goto out; | |
4a35548b MS |
648 | } |
649 | } | |
650 | ||
41d214f6 BV |
651 | channel_name = g_string_sized_new(64); |
652 | for (i = 0; i < inc->num_channels; i++) { | |
df0db9fd GS |
653 | column = columns[i]; |
654 | if (inc->header && inc->multi_column_mode && column[0] != '\0') | |
655 | g_string_assign(channel_name, column); | |
4a35548b | 656 | else |
6433156c | 657 | g_string_printf(channel_name, "%u", i); |
5e23fcab | 658 | sr_channel_new(in->sdi, i, SR_CHANNEL_LOGIC, TRUE, channel_name->str); |
4a35548b | 659 | } |
41d214f6 | 660 | g_string_free(channel_name, TRUE); |
affaf540 GS |
661 | if (!check_header_in_reread(in)) { |
662 | ret = SR_ERR_DATA; | |
663 | goto out; | |
664 | } | |
4a35548b MS |
665 | |
666 | /* | |
cd59e6ec GS |
667 | * Calculate the minimum buffer size to store the set of samples |
668 | * of all channels (unit size). Determine a larger buffer size | |
669 | * for datafeed submission that is a multiple of the unit size. | |
670 | * Allocate the larger buffer, and have the "sample buffer" point | |
671 | * to a location within that large buffer. | |
4a35548b | 672 | */ |
cd59e6ec | 673 | inc->sample_unit_size = (inc->num_channels + 7) / 8; |
8bc2fa6d | 674 | inc->datafeed_buf_size = CHUNK_SIZE; |
cd59e6ec GS |
675 | inc->datafeed_buf_size *= inc->sample_unit_size; |
676 | inc->datafeed_buffer = g_malloc(inc->datafeed_buf_size); | |
677 | inc->datafeed_buf_fill = 0; | |
678 | inc->sample_buffer = &inc->datafeed_buffer[inc->datafeed_buf_fill]; | |
4a35548b | 679 | |
41d214f6 BV |
680 | out: |
681 | if (columns) | |
682 | g_strfreev(columns); | |
683 | g_strfreev(lines); | |
4a35548b | 684 | |
41d214f6 | 685 | return ret; |
4a35548b MS |
686 | } |
687 | ||
4439363a GS |
688 | /* |
689 | * Gets called from initial_receive(), which runs until the end-of-line | |
690 | * encoding of the input stream could get determined. Assumes that this | |
691 | * routine receives enough buffered initial input data to either see the | |
692 | * BOM when there is one, or that no BOM will follow when a text line | |
693 | * termination sequence was seen. Silently drops the UTF-8 BOM sequence | |
694 | * from the input buffer if one was seen. Does not care to protect | |
695 | * against multiple execution or dropping the BOM multiple times -- | |
696 | * there should be at most one in the input stream. | |
697 | */ | |
698 | static void initial_bom_check(const struct sr_input *in) | |
699 | { | |
700 | static const char *utf8_bom = "\xef\xbb\xbf"; | |
701 | ||
702 | if (in->buf->len < strlen(utf8_bom)) | |
703 | return; | |
704 | if (strncmp(in->buf->str, utf8_bom, strlen(utf8_bom)) != 0) | |
705 | return; | |
706 | g_string_erase(in->buf, 0, strlen(utf8_bom)); | |
707 | } | |
708 | ||
41d214f6 | 709 | static int initial_receive(const struct sr_input *in) |
4a35548b | 710 | { |
41d214f6 BV |
711 | struct context *inc; |
712 | GString *new_buf; | |
713 | int len, ret; | |
329733d9 UH |
714 | char *p; |
715 | const char *termination; | |
4a35548b | 716 | |
4439363a GS |
717 | initial_bom_check(in); |
718 | ||
41d214f6 | 719 | inc = in->priv; |
4a35548b | 720 | |
df0db9fd GS |
721 | termination = get_line_termination(in->buf); |
722 | if (!termination) | |
41d214f6 | 723 | /* Don't have a full line yet. */ |
d0181813 | 724 | return SR_ERR_NA; |
4a35548b | 725 | |
df0db9fd GS |
726 | p = g_strrstr_len(in->buf->str, in->buf->len, termination); |
727 | if (!p) | |
41d214f6 | 728 | /* Don't have a full line yet. */ |
d0181813 | 729 | return SR_ERR_NA; |
41d214f6 BV |
730 | len = p - in->buf->str - 1; |
731 | new_buf = g_string_new_len(in->buf->str, len); | |
732 | g_string_append_c(new_buf, '\0'); | |
4a35548b | 733 | |
41d214f6 BV |
734 | inc->termination = g_strdup(termination); |
735 | ||
736 | if (in->buf->str[0] != '\0') | |
737 | ret = initial_parse(in, new_buf); | |
738 | else | |
739 | ret = SR_OK; | |
740 | ||
741 | g_string_free(new_buf, TRUE); | |
742 | ||
743 | return ret; | |
744 | } | |
745 | ||
7f4c3a62 | 746 | static int process_buffer(struct sr_input *in, gboolean is_eof) |
41d214f6 BV |
747 | { |
748 | struct sr_datafeed_packet packet; | |
749 | struct sr_datafeed_meta meta; | |
750 | struct sr_config *src; | |
751 | struct context *inc; | |
752 | gsize num_columns; | |
753 | uint64_t samplerate; | |
754 | int max_columns, ret, l; | |
df0db9fd | 755 | char *p, **lines, *line, **columns; |
41d214f6 | 756 | |
41d214f6 | 757 | inc = in->priv; |
d0181813 | 758 | if (!inc->started) { |
bee2b016 | 759 | std_session_send_df_header(in->sdi); |
41d214f6 BV |
760 | |
761 | if (inc->samplerate) { | |
762 | packet.type = SR_DF_META; | |
763 | packet.payload = &meta; | |
764 | samplerate = inc->samplerate; | |
765 | src = sr_config_new(SR_CONF_SAMPLERATE, g_variant_new_uint64(samplerate)); | |
766 | meta.config = g_slist_append(NULL, src); | |
767 | sr_session_send(in->sdi, &packet); | |
c01378c9 | 768 | g_slist_free(meta.config); |
41d214f6 BV |
769 | sr_config_free(src); |
770 | } | |
d0181813 BV |
771 | |
772 | inc->started = TRUE; | |
4a35548b MS |
773 | } |
774 | ||
f9b74861 GS |
775 | /* Limit the number of columns to parse. */ |
776 | if (inc->multi_column_mode) | |
777 | max_columns = inc->num_channels; | |
778 | else | |
779 | max_columns = 1; | |
780 | ||
4555d3bd GS |
781 | /* |
782 | * Consider empty input non-fatal. Keep accumulating input until | |
783 | * at least one full text line has become available. Grab the | |
784 | * maximum amount of accumulated data that consists of full text | |
785 | * lines, and process what has been received so far, leaving not | |
786 | * yet complete lines for the next invocation. | |
7f4c3a62 GS |
787 | * |
788 | * Enforce that all previously buffered data gets processed in | |
789 | * the "EOF" condition. Do not insist in the presence of the | |
790 | * termination sequence for the last line (may often be missing | |
791 | * on Windows). A present termination sequence will just result | |
792 | * in the "execution of an empty line", and does not harm. | |
4555d3bd GS |
793 | */ |
794 | if (!in->buf->len) | |
795 | return SR_OK; | |
7f4c3a62 GS |
796 | if (is_eof) { |
797 | p = in->buf->str + in->buf->len; | |
798 | } else { | |
799 | p = g_strrstr_len(in->buf->str, in->buf->len, inc->termination); | |
800 | if (!p) | |
801 | return SR_ERR; | |
802 | *p = '\0'; | |
803 | p += strlen(inc->termination); | |
804 | } | |
41d214f6 | 805 | g_strstrip(in->buf->str); |
4a35548b | 806 | |
18078d05 | 807 | ret = SR_OK; |
492dfa90 | 808 | lines = g_strsplit_set(in->buf->str, delim_set, 0); |
41d214f6 BV |
809 | for (l = 0; lines[l]; l++) { |
810 | inc->line_number++; | |
df0db9fd GS |
811 | line = lines[l]; |
812 | if (line[0] == '\0') { | |
41d214f6 | 813 | sr_spew("Blank line %zu skipped.", inc->line_number); |
4a35548b MS |
814 | continue; |
815 | } | |
816 | ||
817 | /* Remove trailing comment. */ | |
df0db9fd GS |
818 | strip_comment(line, inc->comment); |
819 | if (line[0] == '\0') { | |
41d214f6 | 820 | sr_spew("Comment-only line %zu skipped.", inc->line_number); |
4a35548b MS |
821 | continue; |
822 | } | |
823 | ||
160691b9 JS |
824 | /* Skip the header line, its content was used as the channel names. */ |
825 | if (inc->header) { | |
826 | sr_spew("Header line %zu skipped.", inc->line_number); | |
827 | inc->header = FALSE; | |
828 | continue; | |
829 | } | |
830 | ||
df0db9fd GS |
831 | columns = parse_line(line, inc, max_columns); |
832 | if (!columns) { | |
41d214f6 | 833 | sr_err("Error while parsing line %zu.", inc->line_number); |
2355d229 | 834 | g_strfreev(lines); |
4a35548b MS |
835 | return SR_ERR; |
836 | } | |
4a35548b | 837 | num_columns = g_strv_length(columns); |
4a35548b | 838 | if (!num_columns) { |
6433156c | 839 | sr_err("Column %u in line %zu is out of bounds.", |
41d214f6 | 840 | inc->first_column, inc->line_number); |
4a35548b | 841 | g_strfreev(columns); |
2355d229 | 842 | g_strfreev(lines); |
4a35548b MS |
843 | return SR_ERR; |
844 | } | |
4a35548b | 845 | /* |
ba7dd8bb | 846 | * Ensure that the number of channels does not exceed the number |
4a35548b MS |
847 | * of columns in multi column mode. |
848 | */ | |
41d214f6 | 849 | if (inc->multi_column_mode && num_columns < inc->num_channels) { |
ba7dd8bb | 850 | sr_err("Not enough columns for desired number of channels in line %zu.", |
41d214f6 | 851 | inc->line_number); |
4a35548b | 852 | g_strfreev(columns); |
2355d229 | 853 | g_strfreev(lines); |
4a35548b MS |
854 | return SR_ERR; |
855 | } | |
856 | ||
41d214f6 BV |
857 | if (inc->multi_column_mode) |
858 | ret = parse_multi_columns(columns, inc); | |
4a35548b | 859 | else |
41d214f6 BV |
860 | ret = parse_single_column(columns[0], inc); |
861 | if (ret != SR_OK) { | |
4a35548b | 862 | g_strfreev(columns); |
2355d229 | 863 | g_strfreev(lines); |
4a35548b MS |
864 | return SR_ERR; |
865 | } | |
866 | ||
4a35548b | 867 | /* Send sample data to the session bus. */ |
cd59e6ec | 868 | ret = queue_samples(in); |
41d214f6 | 869 | if (ret != SR_OK) { |
4a35548b | 870 | sr_err("Sending samples failed."); |
cd59e6ec | 871 | g_strfreev(columns); |
2355d229 | 872 | g_strfreev(lines); |
4a35548b MS |
873 | return SR_ERR; |
874 | } | |
cd59e6ec | 875 | |
41d214f6 BV |
876 | g_strfreev(columns); |
877 | } | |
878 | g_strfreev(lines); | |
241c386a | 879 | g_string_erase(in->buf, 0, p - in->buf->str); |
41d214f6 | 880 | |
7066fd46 | 881 | return ret; |
41d214f6 BV |
882 | } |
883 | ||
7066fd46 | 884 | static int receive(struct sr_input *in, GString *buf) |
41d214f6 BV |
885 | { |
886 | struct context *inc; | |
7066fd46 BV |
887 | int ret; |
888 | ||
889 | g_string_append_len(in->buf, buf->str, buf->len); | |
41d214f6 BV |
890 | |
891 | inc = in->priv; | |
7066fd46 | 892 | if (!inc->termination) { |
df0db9fd GS |
893 | ret = initial_receive(in); |
894 | if (ret == SR_ERR_NA) | |
7066fd46 BV |
895 | /* Not enough data yet. */ |
896 | return SR_OK; | |
897 | else if (ret != SR_OK) | |
898 | return SR_ERR; | |
899 | ||
900 | /* sdi is ready, notify frontend. */ | |
901 | in->sdi_ready = TRUE; | |
41d214f6 | 902 | return SR_OK; |
7066fd46 BV |
903 | } |
904 | ||
7f4c3a62 | 905 | ret = process_buffer(in, FALSE); |
7066fd46 BV |
906 | |
907 | return ret; | |
908 | } | |
909 | ||
910 | static int end(struct sr_input *in) | |
911 | { | |
912 | struct context *inc; | |
7066fd46 | 913 | int ret; |
41d214f6 | 914 | |
7066fd46 | 915 | if (in->sdi_ready) |
7f4c3a62 | 916 | ret = process_buffer(in, TRUE); |
7066fd46 BV |
917 | else |
918 | ret = SR_OK; | |
cd59e6ec GS |
919 | if (ret != SR_OK) |
920 | return ret; | |
921 | ||
922 | ret = flush_samples(in); | |
923 | if (ret != SR_OK) | |
924 | return ret; | |
7066fd46 BV |
925 | |
926 | inc = in->priv; | |
3be42bc2 | 927 | if (inc->started) |
bee2b016 | 928 | std_session_send_df_end(in->sdi); |
4a35548b | 929 | |
7066fd46 BV |
930 | return ret; |
931 | } | |
932 | ||
d5cc282f | 933 | static void cleanup(struct sr_input *in) |
7066fd46 BV |
934 | { |
935 | struct context *inc; | |
936 | ||
affaf540 GS |
937 | keep_header_for_reread(in); |
938 | ||
7066fd46 BV |
939 | inc = in->priv; |
940 | ||
b1f83103 | 941 | g_free(inc->termination); |
539188e5 | 942 | inc->termination = NULL; |
cd59e6ec | 943 | g_free(inc->datafeed_buffer); |
539188e5 | 944 | inc->datafeed_buffer = NULL; |
4a35548b MS |
945 | } |
946 | ||
ad93bfb0 SA |
947 | static int reset(struct sr_input *in) |
948 | { | |
949 | struct context *inc = in->priv; | |
950 | ||
951 | cleanup(in); | |
952 | inc->started = FALSE; | |
953 | g_string_truncate(in->buf, 0); | |
954 | ||
955 | return SR_OK; | |
956 | } | |
957 | ||
41d214f6 | 958 | static struct sr_option options[] = { |
867293a1 UH |
959 | { "single-column", "Single column", "Enable single-column mode, using the specified column (>= 1); 0: multi-col. mode", NULL, NULL }, |
960 | { "numchannels", "Number of logic channels", "The number of (logic) channels (single-col. mode: number of bits beginning at 'first channel', LSB-first)", NULL, NULL }, | |
961 | { "delimiter", "Column delimiter", "The column delimiter (>= 1 characters)", NULL, NULL }, | |
962 | { "format", "Data format (single-col. mode)", "The numeric format of the data (single-col. mode): bin, hex, oct", NULL, NULL }, | |
963 | { "comment", "Comment character(s)", "The comment prefix character(s)", NULL, NULL }, | |
964 | { "samplerate", "Samplerate (Hz)", "The sample rate (used during capture) in Hz", NULL, NULL }, | |
965 | { "first-channel", "First channel", "The column number of the first channel (multi-col. mode); bit position for the first channel (single-col. mode)", NULL, NULL }, | |
966 | { "header", "Interpret first line as header (multi-col. mode)", "Treat the first line as header with channel names (multi-col. mode)", NULL, NULL }, | |
967 | { "startline", "Start line", "The line number at which to start processing samples (>= 1)", NULL, NULL }, | |
06ad20be | 968 | ALL_ZERO |
41d214f6 BV |
969 | }; |
970 | ||
2c240774 | 971 | static const struct sr_option *get_options(void) |
41d214f6 | 972 | { |
31c41782 UH |
973 | GSList *l; |
974 | ||
41d214f6 BV |
975 | if (!options[0].def) { |
976 | options[0].def = g_variant_ref_sink(g_variant_new_int32(0)); | |
977 | options[1].def = g_variant_ref_sink(g_variant_new_int32(0)); | |
978 | options[2].def = g_variant_ref_sink(g_variant_new_string(",")); | |
979 | options[3].def = g_variant_ref_sink(g_variant_new_string("bin")); | |
31c41782 UH |
980 | l = NULL; |
981 | l = g_slist_append(l, g_variant_ref_sink(g_variant_new_string("bin"))); | |
982 | l = g_slist_append(l, g_variant_ref_sink(g_variant_new_string("hex"))); | |
983 | l = g_slist_append(l, g_variant_ref_sink(g_variant_new_string("oct"))); | |
984 | options[3].values = l; | |
41d214f6 | 985 | options[4].def = g_variant_ref_sink(g_variant_new_string(";")); |
6e8d95a5 | 986 | options[5].def = g_variant_ref_sink(g_variant_new_uint64(0)); |
41d214f6 BV |
987 | options[6].def = g_variant_ref_sink(g_variant_new_int32(0)); |
988 | options[7].def = g_variant_ref_sink(g_variant_new_boolean(FALSE)); | |
989 | options[8].def = g_variant_ref_sink(g_variant_new_int32(1)); | |
990 | } | |
991 | ||
992 | return options; | |
993 | } | |
994 | ||
d4c93774 | 995 | SR_PRIV struct sr_input_module input_csv = { |
4a35548b | 996 | .id = "csv", |
41d214f6 BV |
997 | .name = "CSV", |
998 | .desc = "Comma-separated values", | |
c7bc82ff | 999 | .exts = (const char*[]){"csv", NULL}, |
41d214f6 | 1000 | .options = get_options, |
4a35548b | 1001 | .init = init, |
41d214f6 | 1002 | .receive = receive, |
7066fd46 | 1003 | .end = end, |
41d214f6 | 1004 | .cleanup = cleanup, |
ad93bfb0 | 1005 | .reset = reset, |
4a35548b | 1006 | }; |