X-Git-Url: https://sigrok.org/gitweb/?a=blobdiff_plain;f=src%2Finput%2Fcsv.c;h=000599a8d305bf9140fa78679aefb1b5cad72174;hb=2355d2291925ace0bd9c85b21f2ce7e2c6606c7f;hp=5399027bf823305ab0a33c79b02ec9783e1ca5c3;hpb=4555d3bda00ed6f12016d4aca28bbef3cb459988;p=libsigrok.git diff --git a/src/input/csv.c b/src/input/csv.c index 5399027b..000599a8 100644 --- a/src/input/csv.c +++ b/src/input/csv.c @@ -26,6 +26,8 @@ #define LOG_PREFIX "input/csv" +#define DATAFEED_MAX_SAMPLES (128 * 1024) + /* * The CSV input module has the following options: * @@ -67,6 +69,39 @@ * than 0. The default line number to start processing is 1. */ +/* + * TODO + * + * - Determine how the text line handling can get improved, regarding + * all of robustness and flexibility and correctness. + * - The current implementation splits on "any run of CR and LF". Which + * translates to: Line numbers are wrong in the presence of empty + * lines in the input stream. See below for an (expensive) fix. + * - Dropping support for CR style end-of-line markers could improve + * the situation a lot. Code could search for and split on LF, and + * trim optional trailing CR. This would result in proper support + * for CRLF (Windows) as well as LF (Unix), and allow for correct + * line number counts. + * - When support for CR-only line termination cannot get dropped, + * then the current implementation is inappropriate. Currently the + * input stream is scanned for the first occurance of either of the + * supported termination styles (which is good). For the remaining + * session a consistent encoding of the text lines is assumed (which + * is acceptable). + * - When line numbers need to be correct and reliable, _and_ the full + * set of previously supported line termination sequences are required, + * and potentially more are to get added for improved compatibility + * with more platforms or generators, then the current approach of + * splitting on runs of termination characters needs to get replaced, + * by the more expensive approach to scan for and count the initially + * determined termination sequence. + * + * - Add support for analog input data? (optional) + * - Needs a syntax first for user specs which channels (columns) are + * logic and which are analog. May need heuristics(?) to guess from + * input data in the absence of user provided specs. + */ + /* Single column formats. */ enum { FORMAT_BIN, @@ -123,11 +158,12 @@ struct context { /* Format sample data is stored in single column mode. */ int format; - /* Size of the sample buffer. */ - size_t sample_buffer_size; + size_t sample_unit_size; /**!< Byte count for a single sample. */ + uint8_t *sample_buffer; /**!< Buffer for a single sample. */ - /* Buffer to store sample data. */ - uint8_t *sample_buffer; + uint8_t *datafeed_buffer; /**!< Queue for datafeed submission. */ + size_t datafeed_buf_size; + size_t datafeed_buf_fill; /* Current line number. */ size_t line_number; @@ -157,7 +193,7 @@ static int parse_binstr(const char *str, struct context *inc) } /* Clear buffer in order to set bits only. */ - memset(inc->sample_buffer, 0, (inc->num_channels + 7) >> 3); + memset(inc->sample_buffer, 0, inc->sample_unit_size); i = inc->first_channel; @@ -189,7 +225,7 @@ static int parse_hexstr(const char *str, struct context *inc) } /* Clear buffer in order to set bits only. */ - memset(inc->sample_buffer, 0, (inc->num_channels + 7) >> 3); + memset(inc->sample_buffer, 0, inc->sample_unit_size); /* Calculate the position of the first hexadecimal digit. */ i = inc->first_channel / 4; @@ -233,7 +269,7 @@ static int parse_octstr(const char *str, struct context *inc) } /* Clear buffer in order to set bits only. */ - memset(inc->sample_buffer, 0, (inc->num_channels + 7) >> 3); + memset(inc->sample_buffer, 0, inc->sample_unit_size); /* Calculate the position of the first octal digit. */ i = inc->first_channel / 3; @@ -316,7 +352,7 @@ static int parse_multi_columns(char **columns, struct context *inc) char *column; /* Clear buffer in order to set bits only. */ - memset(inc->sample_buffer, 0, (inc->num_channels + 7) >> 3); + memset(inc->sample_buffer, 0, inc->sample_unit_size); for (i = 0; i < inc->num_channels; i++) { column = columns[i]; @@ -358,26 +394,47 @@ static int parse_single_column(const char *column, struct context *inc) return res; } -static int send_samples(const struct sr_dev_inst *sdi, uint8_t *buffer, - gsize buffer_size, gsize count) +static int flush_samples(const struct sr_input *in) { + struct context *inc; struct sr_datafeed_packet packet; struct sr_datafeed_logic logic; - int res; - gsize i; + int rc; + + inc = in->priv; + if (!inc->datafeed_buf_fill) + return SR_OK; + memset(&packet, 0, sizeof(packet)); + memset(&logic, 0, sizeof(logic)); packet.type = SR_DF_LOGIC; packet.payload = &logic; - logic.unitsize = buffer_size; - logic.length = buffer_size; - logic.data = buffer; - - for (i = 0; i < count; i++) { - res = sr_session_send(sdi, &packet); - if (res != SR_OK) - return res; - } + logic.unitsize = inc->sample_unit_size; + logic.length = inc->datafeed_buf_fill; + logic.data = inc->datafeed_buffer; + + rc = sr_session_send(in->sdi, &packet); + if (rc != SR_OK) + return rc; + + inc->datafeed_buf_fill = 0; + return SR_OK; +} + +static int queue_samples(const struct sr_input *in) +{ + struct context *inc; + int rc; + inc = in->priv; + + inc->datafeed_buf_fill += inc->sample_unit_size; + if (inc->datafeed_buf_fill == inc->datafeed_buf_size) { + rc = flush_samples(in); + if (rc != SR_OK) + return rc; + } + inc->sample_buffer = &inc->datafeed_buffer[inc->datafeed_buf_fill]; return SR_OK; } @@ -560,11 +617,18 @@ static int initial_parse(const struct sr_input *in, GString *buf) g_string_free(channel_name, TRUE); /* - * Calculate the minimum buffer size to store the sample data of the - * channels. + * Calculate the minimum buffer size to store the set of samples + * of all channels (unit size). Determine a larger buffer size + * for datafeed submission that is a multiple of the unit size. + * Allocate the larger buffer, and have the "sample buffer" point + * to a location within that large buffer. */ - inc->sample_buffer_size = (inc->num_channels + 7) >> 3; - inc->sample_buffer = g_malloc(inc->sample_buffer_size); + inc->sample_unit_size = (inc->num_channels + 7) / 8; + inc->datafeed_buf_size = DATAFEED_MAX_SAMPLES; + inc->datafeed_buf_size *= inc->sample_unit_size; + inc->datafeed_buffer = g_malloc(inc->datafeed_buf_size); + inc->datafeed_buf_fill = 0; + inc->sample_buffer = &inc->datafeed_buffer[inc->datafeed_buf_fill]; out: if (columns) @@ -574,6 +638,27 @@ out: return ret; } +/* + * Gets called from initial_receive(), which runs until the end-of-line + * encoding of the input stream could get determined. Assumes that this + * routine receives enough buffered initial input data to either see the + * BOM when there is one, or that no BOM will follow when a text line + * termination sequence was seen. Silently drops the UTF-8 BOM sequence + * from the input buffer if one was seen. Does not care to protect + * against multiple execution or dropping the BOM multiple times -- + * there should be at most one in the input stream. + */ +static void initial_bom_check(const struct sr_input *in) +{ + static const char *utf8_bom = "\xef\xbb\xbf"; + + if (in->buf->len < strlen(utf8_bom)) + return; + if (strncmp(in->buf->str, utf8_bom, strlen(utf8_bom)) != 0) + return; + g_string_erase(in->buf, 0, strlen(utf8_bom)); +} + static int initial_receive(const struct sr_input *in) { struct context *inc; @@ -582,6 +667,8 @@ static int initial_receive(const struct sr_input *in) char *p; const char *termination; + initial_bom_check(in); + inc = in->priv; termination = get_line_termination(in->buf); @@ -609,7 +696,7 @@ static int initial_receive(const struct sr_input *in) return ret; } -static int process_buffer(struct sr_input *in) +static int process_buffer(struct sr_input *in, gboolean is_eof) { struct sr_datafeed_packet packet; struct sr_datafeed_meta meta; @@ -650,13 +737,24 @@ static int process_buffer(struct sr_input *in) * maximum amount of accumulated data that consists of full text * lines, and process what has been received so far, leaving not * yet complete lines for the next invocation. + * + * Enforce that all previously buffered data gets processed in + * the "EOF" condition. Do not insist in the presence of the + * termination sequence for the last line (may often be missing + * on Windows). A present termination sequence will just result + * in the "execution of an empty line", and does not harm. */ if (!in->buf->len) return SR_OK; - p = g_strrstr_len(in->buf->str, in->buf->len, inc->termination); - if (!p) - return SR_ERR; - *p = '\0'; + if (is_eof) { + p = in->buf->str + in->buf->len; + } else { + p = g_strrstr_len(in->buf->str, in->buf->len, inc->termination); + if (!p) + return SR_ERR; + *p = '\0'; + p += strlen(inc->termination); + } g_strstrip(in->buf->str); ret = SR_OK; @@ -686,6 +784,7 @@ static int process_buffer(struct sr_input *in) columns = parse_line(line, inc, max_columns); if (!columns) { sr_err("Error while parsing line %zu.", inc->line_number); + g_strfreev(lines); return SR_ERR; } num_columns = g_strv_length(columns); @@ -693,6 +792,7 @@ static int process_buffer(struct sr_input *in) sr_err("Column %u in line %zu is out of bounds.", inc->first_column, inc->line_number); g_strfreev(columns); + g_strfreev(lines); return SR_ERR; } /* @@ -703,6 +803,7 @@ static int process_buffer(struct sr_input *in) sr_err("Not enough columns for desired number of channels in line %zu.", inc->line_number); g_strfreev(columns); + g_strfreev(lines); return SR_ERR; } @@ -712,20 +813,23 @@ static int process_buffer(struct sr_input *in) ret = parse_single_column(columns[0], inc); if (ret != SR_OK) { g_strfreev(columns); + g_strfreev(lines); return SR_ERR; } /* Send sample data to the session bus. */ - ret = send_samples(in->sdi, inc->sample_buffer, - inc->sample_buffer_size, 1); + ret = queue_samples(in); if (ret != SR_OK) { sr_err("Sending samples failed."); + g_strfreev(columns); + g_strfreev(lines); return SR_ERR; } + g_strfreev(columns); } g_strfreev(lines); - g_string_erase(in->buf, 0, p - in->buf->str + 1); + g_string_erase(in->buf, 0, p - in->buf->str); return ret; } @@ -751,7 +855,7 @@ static int receive(struct sr_input *in, GString *buf) return SR_OK; } - ret = process_buffer(in); + ret = process_buffer(in, FALSE); return ret; } @@ -762,9 +866,15 @@ static int end(struct sr_input *in) int ret; if (in->sdi_ready) - ret = process_buffer(in); + ret = process_buffer(in, TRUE); else ret = SR_OK; + if (ret != SR_OK) + return ret; + + ret = flush_samples(in); + if (ret != SR_OK) + return ret; inc = in->priv; if (inc->started) @@ -786,7 +896,7 @@ static void cleanup(struct sr_input *in) g_string_free(inc->comment, TRUE); g_free(inc->termination); - g_free(inc->sample_buffer); + g_free(inc->datafeed_buffer); } static int reset(struct sr_input *in)