X-Git-Url: https://sigrok.org/gitweb/?a=blobdiff_plain;f=src%2Finput%2Fcsv.c;h=4dcc77b8eaeb19e13999324e230786d8da19ef9f;hb=57140e5a804dd5b616120f05e38c3409241bb37c;hp=dabd52c86574a122d62f4159d737bd798614c83e;hpb=05719d75aa31627a25375567b3641d8787423567;p=libsigrok.git diff --git a/src/input/csv.c b/src/input/csv.c index dabd52c8..4dcc77b8 100644 --- a/src/input/csv.c +++ b/src/input/csv.c @@ -127,27 +127,15 @@ * and equidistant rows, there is no fancy support for textual unit * suffixes nor gaps in the stream of samples nor other non-linearity, * just '-' ignore the column if the format is not supported). + * + * IMPORTANT! Make sure the .format_match() logic matches the default + * values for the input module's options. Ideally the format match test + * shall pass for all input data that is supported by default. */ /* * TODO * - * - Extend support for analog input data. - * - Determine why analog samples of 'double' data type get scrambled - * in sigrok-cli screen output. Is analog.encoding->unitsize not - * handled properly? A sigrok-cli or libsigrok (src/output) issue? - * - Reconsider the channel creation after format processing. Current - * logic may "bleed" channel names into the analog group when logic - * channels' columns follow analog columns (seen with "-,2a,x8"). - * Trying to sort it out, a naive change used to map logic channels' - * data to incorrect bitmap positions. The whole channel numbering - * needs reconsideration. Probably it's easiest to first create _all_ - * logic channels so that they have adjacent numbers starting at 0 - * (addressing logic bits), then all analog channels (again adjacent) - * to simplify the calculation of their index in the sample set as - * well as their sdi channel index from the "analog column index". - * - Optionally get sample rate from timestamp column. Just best-effort - * approach, not necessarily reliable. Users can always specify rates. * - Add a test suite for input modules in general, and CSV in specific? * Becomes more important with the multitude of options and their * interaction. Could cover edge cases (BOM presence, line termination @@ -155,7 +143,7 @@ * samplerates, etc). */ -typedef float csv_analog_t; /* 'double' currently is flawed. */ +typedef double csv_analog_t; /* Single column formats. */ enum single_col_format { @@ -219,6 +207,7 @@ struct context { /* Current samplerate, optionally determined from input data. */ uint64_t samplerate; + uint64_t calc_samplerate; double prev_timestamp; gboolean samplerate_sent; @@ -266,6 +255,7 @@ struct context { /* List of previously created sigrok channels. */ GSList *prev_sr_channels; + GSList **prev_df_channels; }; /* @@ -284,19 +274,13 @@ struct context { static int flush_samplerate(const struct sr_input *in) { struct context *inc; - struct sr_datafeed_packet packet; - struct sr_datafeed_meta meta; - struct sr_config *src; inc = in->priv; - if (inc->samplerate && !inc->samplerate_sent) { - packet.type = SR_DF_META; - packet.payload = &meta; - src = sr_config_new(SR_CONF_SAMPLERATE, g_variant_new_uint64(inc->samplerate)); - meta.config = g_slist_append(NULL, src); - sr_session_send(in->sdi, &packet); - g_slist_free(meta.config); - sr_config_free(src); + if (!inc->calc_samplerate && inc->samplerate) + inc->calc_samplerate = inc->samplerate; + if (inc->calc_samplerate && !inc->samplerate_sent) { + (void)sr_session_send_meta(in->sdi, SR_CONF_SAMPLERATE, + g_variant_new_uint64(inc->calc_samplerate)); inc->samplerate_sent = TRUE; } @@ -355,6 +339,7 @@ static int flush_logic_samples(const struct sr_input *in) return rc; inc->datafeed_buf_fill = 0; + return SR_OK; } @@ -373,6 +358,7 @@ static int queue_logic_samples(const struct sr_input *in) if (rc != SR_OK) return rc; } + return SR_OK; } @@ -448,6 +434,7 @@ static int flush_analog_samples(const struct sr_input *in) } inc->analog_datafeed_buf_fill = 0; + return SR_OK; } @@ -466,6 +453,7 @@ static int queue_analog_samples(const struct sr_input *in) if (rc != SR_OK) return rc; } + return SR_OK; } @@ -669,10 +657,12 @@ static int make_column_details_from_format(const struct sr_input *in, * line won't get processed another time. */ column = column_texts[detail->col_nr - 1]; - if (inc->use_header && column && *column) + if (inc->use_header && column && *column) { + column = g_strstrip(column); caption = sr_scpi_unquote_string(column); - else + } else { caption = NULL; + } if (!caption || !*caption) caption = NULL; /* @@ -733,6 +723,7 @@ static const struct column_details *lookup_column_details(struct context *inc, s return NULL; if (!nr || nr > inc->column_want_count) return NULL; + return &inc->column_details[nr - 1]; } @@ -756,7 +747,7 @@ static void strip_comment(char *buf, const GString *prefix) } /** - * @brief Splits a text line into a set of columns. + * Splits a text line into a set of columns. * * @param[in] buf The input text line to split. * @param[in] inc The input module's context. @@ -767,11 +758,24 @@ static void strip_comment(char *buf, const GString *prefix) */ static char **split_line(char *buf, struct context *inc) { - return g_strsplit(buf, inc->delimiter->str, 0); + char **fields, *f; + size_t l; + + fields = g_strsplit(buf, inc->delimiter->str, 0); + if (!fields) + return NULL; + + l = g_strv_length(fields); + while (l--) { + f = fields[l]; + g_strchomp(f); + } + + return fields; } /** - * @brief Parse a multi-bit field into several logic channels. + * Parse a multi-bit field into several logic channels. * * @param[in] column The input text, a run of bin/hex/oct digits. * @param[in] inc The input module's context. @@ -881,7 +885,7 @@ static int parse_logic(const char *column, struct context *inc, } /** - * @brief Parse a floating point text into an analog value. + * Parse a floating point text into an analog value. * * @param[in] column The input text, a floating point number. * @param[in] inc The input module's context. @@ -930,7 +934,7 @@ static int parse_analog(const char *column, struct context *inc, } /** - * @brief Parse a timestamp text, auto-determine samplerate. + * Parse a timestamp text, auto-determine samplerate. * * @param[in] column The input text, a floating point number. * @param[in] inc The input module's context. @@ -981,13 +985,13 @@ static int parse_timestamp(const char *column, struct context *inc, * reduced rounding errors which result in odd rates. * - Support other formats ("2 ms" or similar)? */ - if (inc->samplerate) + if (inc->calc_samplerate) return SR_OK; ret = sr_atod_ascii(column, &ts); if (ret != SR_OK) ts = 0.0; if (!ts) { - sr_warn("Cannot convert timestamp text %s in line %zu (or zero value).", + sr_info("Cannot convert timestamp text %s in line %zu (or zero value).", column, inc->line_number); inc->prev_timestamp = 0.0; return SR_OK; @@ -1012,14 +1016,14 @@ static int parse_timestamp(const char *column, struct context *inc, rate += 0.5; rate = (uint64_t)rate; sr_dbg("Rate from timestamp %g in line %zu.", rate, inc->line_number); - inc->samplerate = rate; + inc->calc_samplerate = rate; inc->prev_timestamp = 0.0; return SR_OK; } /** - * @brief Parse routine which ignores the input text. + * Parse routine which ignores the input text. * * This routine exists to unify dispatch code paths, mapping input file * columns' data types to their respective parse routines. @@ -1030,6 +1034,7 @@ static int parse_ignore(const char *column, struct context *inc, (void)column; (void)inc; (void)details; + return SR_OK; } @@ -1161,6 +1166,7 @@ static int format_match(GHashTable *metadata, unsigned int *confidence) if (!status) return SR_ERR; + return SR_OK; } @@ -1266,15 +1272,30 @@ static int init(struct sr_input *in, GHashTable *options) * Check the channel list for consistency across file re-import. See * the VCD input module for more details and motivation. */ +static void release_df_channels(struct context *inc, GSList **l) +{ + size_t idx; + + if (!inc->analog_channels || !l) + return; + for (idx = 0; idx < inc->analog_channels; idx++) + g_slist_free(l[idx]); + g_free(l); +} static void keep_header_for_reread(const struct sr_input *in) { struct context *inc; inc = in->priv; + g_slist_free_full(inc->prev_sr_channels, sr_channel_free_cb); inc->prev_sr_channels = in->sdi->channels; in->sdi->channels = NULL; + + release_df_channels(inc, inc->prev_df_channels); + inc->prev_df_channels = inc->analog_datafeed_channels; + inc->analog_datafeed_channels = NULL; } static int check_header_in_reread(const struct sr_input *in) @@ -1293,10 +1314,15 @@ static int check_header_in_reread(const struct sr_input *in) sr_err("Channel list change not supported for file re-read."); return FALSE; } + g_slist_free_full(in->sdi->channels, sr_channel_free_cb); in->sdi->channels = inc->prev_sr_channels; inc->prev_sr_channels = NULL; + release_df_channels(inc, inc->analog_datafeed_channels); + inc->analog_datafeed_channels = inc->prev_df_channels; + inc->prev_df_channels = NULL; + return TRUE; } @@ -1373,7 +1399,7 @@ static int initial_parse(const struct sr_input *in, GString *buf) ret = SR_ERR; goto out; } - sr_dbg("DIAG Got %zu columns in text line: %s.", num_columns, line); + sr_dbg("Got %zu columns in text line: %s.", num_columns, line); /* * Interpret the user provided column format specs. This might @@ -1511,7 +1537,8 @@ static int process_buffer(struct sr_input *in, gboolean is_eof) const struct column_details *details; col_parse_cb parse_func; int ret; - char *p, **lines, *line, **columns, *column; + char *processed_up_to; + char **lines, *line, **columns, *column; inc = in->priv; if (!inc->started) { @@ -1535,16 +1562,17 @@ static int process_buffer(struct sr_input *in, gboolean is_eof) if (!in->buf->len) return SR_OK; if (is_eof) { - p = in->buf->str + in->buf->len; + processed_up_to = 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); + processed_up_to = g_strrstr_len(in->buf->str, in->buf->len, + inc->termination); + if (!processed_up_to) + return SR_OK; + *processed_up_to = '\0'; + processed_up_to += strlen(inc->termination); } - g_strstrip(in->buf->str); + /* Split input text lines and process their columns. */ ret = SR_OK; lines = g_strsplit(in->buf->str, inc->termination, 0); for (line_idx = 0; (line = lines[line_idx]); line_idx++) { @@ -1621,7 +1649,7 @@ static int process_buffer(struct sr_input *in, gboolean is_eof) g_strfreev(columns); } g_strfreev(lines); - g_string_erase(in->buf, 0, p - in->buf->str); + g_string_erase(in->buf, 0, processed_up_to - in->buf->str); return ret; } @@ -1678,10 +1706,12 @@ static int end(struct sr_input *in) static void cleanup(struct sr_input *in) { - struct context *inc; + struct context *inc, save_ctx; + /* Keep channel references between file re-imports. */ keep_header_for_reread(in); + /* Release dynamically allocated resources. */ inc = in->priv; g_free(inc->termination); @@ -1690,12 +1720,31 @@ static void cleanup(struct sr_input *in) inc->datafeed_buffer = NULL; g_free(inc->analog_datafeed_buffer); inc->analog_datafeed_buffer = NULL; + g_free(inc->analog_datafeed_digits); + inc->analog_datafeed_digits = NULL; + /* analog_datafeed_channels was released in keep_header_for_reread() */ + /* TODO Release channel names (before releasing details). */ + g_free(inc->column_details); + inc->column_details = NULL; + + /* Clear internal state, but keep what .init() has provided. */ + save_ctx = *inc; + memset(inc, 0, sizeof(*inc)); + inc->samplerate = save_ctx.samplerate; + inc->delimiter = save_ctx.delimiter; + inc->comment = save_ctx.comment; + inc->column_formats = save_ctx.column_formats; + inc->start_line = save_ctx.start_line; + inc->use_header = save_ctx.use_header; + inc->prev_sr_channels = save_ctx.prev_sr_channels; + inc->prev_df_channels = save_ctx.prev_df_channels; } static int reset(struct sr_input *in) { - struct context *inc = in->priv; + struct context *inc; + inc = in->priv; cleanup(in); inc->started = FALSE; g_string_truncate(in->buf, 0); @@ -1750,7 +1799,7 @@ static struct sr_option options[] = { }, [OPT_HEADER] = { "header", "Get channel names from first line.", - "Use the first processed line's column captions (when available) as channel names. Off by default", + "Use the first processed line's column captions (when available) as channel names. Enabled by default.", NULL, NULL, }, [OPT_SAMPLERATE] = { @@ -1787,7 +1836,7 @@ static const struct sr_option *get_options(void) l = g_slist_append(l, g_variant_ref_sink(g_variant_new_string("oct"))); options[OPT_SINGLE_FMT].values = l; options[OPT_START_LINE].def = g_variant_ref_sink(g_variant_new_uint32(1)); - options[OPT_HEADER].def = g_variant_ref_sink(g_variant_new_boolean(FALSE)); + options[OPT_HEADER].def = g_variant_ref_sink(g_variant_new_boolean(TRUE)); options[OPT_SAMPLERATE].def = g_variant_ref_sink(g_variant_new_uint64(0)); options[OPT_COL_SEP].def = g_variant_ref_sink(g_variant_new_string(",")); options[OPT_COMMENT].def = g_variant_ref_sink(g_variant_new_string(";"));