X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Fdata%2Fgnumeric-reader.c;h=0783e10f7c1b33452222b82238d95d51ae582004;hb=83239cbb53294662025d5ee957d7e24e5feb66a9;hp=454654c5443827a3b6b0db795cb25ad5b3cee4ef;hpb=0a881d629fb4ad0326796076d127c341c65832fe;p=pspp diff --git a/src/data/gnumeric-reader.c b/src/data/gnumeric-reader.c index 454654c544..0783e10f7c 100644 --- a/src/data/gnumeric-reader.c +++ b/src/data/gnumeric-reader.c @@ -28,7 +28,7 @@ #include "spreadsheet-reader.h" -#if !GNM_SUPPORT +#if !GNM_READ_SUPPORT struct casereader * gnumeric_open_reader (const struct spreadsheet_read_options *opts, struct dictionary **dict) @@ -48,6 +48,8 @@ gnumeric_open_reader (const struct spreadsheet_read_options *opts, struct dictio #include #include +#include "data/format.h" +#include "data/data-in.h" #include "data/case.h" #include "data/casereader-provider.h" #include "data/dictionary.h" @@ -59,6 +61,25 @@ gnumeric_open_reader (const struct spreadsheet_read_options *opts, struct dictio #include "gl/xalloc.h" + +/* Shamelessly lifted from the Gnumeric sources: + https://git.gnome.org/browse/gnumeric/tree/src/value.h + */ +enum gnm_value_type +{ + VALUE_EMPTY = 10, + VALUE_BOOLEAN = 20, + VALUE_INTEGER = 30, /* Note, this was removed from gnumeric in 2006 - old versions may of + course still be around. New ones are supposed to use float.*/ + VALUE_FLOAT = 40, + VALUE_ERROR = 50, + VALUE_STRING = 60, + VALUE_CELLRANGE = 70, + VALUE_ARRAY = 80 +}; + + + static void gnm_file_casereader_destroy (struct casereader *, void *); static struct ccase *gnm_file_casereader_read (struct casereader *, void *); @@ -128,7 +149,6 @@ state_data_destroy (struct state_data *sd) struct gnumeric_reader { struct spreadsheet spreadsheet; - int ref_cnt; struct state_data rsd; struct state_data msd; @@ -147,15 +167,17 @@ struct gnumeric_reader struct dictionary *dict; struct ccase *first_case; bool used_first_case; + + enum gnm_value_type vtype; }; void -gnumeric_destroy (struct spreadsheet *s) +gnumeric_unref (struct spreadsheet *s) { struct gnumeric_reader *r = (struct gnumeric_reader *) s; - if (0 == --r->ref_cnt) + if (0 == --s->ref_cnt) { int i; @@ -167,6 +189,8 @@ gnumeric_destroy (struct spreadsheet *s) free (r->sheets); state_data_destroy (&r->msd); + free (s->file_name); + free (r); } } @@ -203,7 +227,7 @@ gnumeric_get_sheet_range (struct spreadsheet *s, int n) process_node (gr, &gr->msd); } - return create_cell_ref ( + return create_cell_range ( gr->sheets[n].start_col, gr->sheets[n].start_row, gr->sheets[n].stop_col, @@ -221,12 +245,13 @@ gnm_file_casereader_destroy (struct casereader *reader UNUSED, void *r_) state_data_destroy (&r->rsd); - if ( ! r->used_first_case ) + if (r->first_case && ! r->used_first_case ) case_unref (r->first_case); - caseproto_unref (r->proto); + if (r->proto) + caseproto_unref (r->proto); - gnumeric_destroy (&r->spreadsheet); + gnumeric_unref (&r->spreadsheet); } @@ -428,7 +453,7 @@ process_node (struct gnumeric_reader *r, struct state_data *sd) */ static void convert_xml_string_to_value (struct ccase *c, const struct variable *var, - const xmlChar *xv) + const xmlChar *xv, enum gnm_value_type type, int col, int row) { union value *v = case_data_rw (c, var); @@ -436,7 +461,7 @@ convert_xml_string_to_value (struct ccase *c, const struct variable *var, value_set_missing (v, var_get_width (var)); else if ( var_is_alpha (var)) value_copy_str_rpad (v, var_get_width (var), xv, ' '); - else + else if (type == VALUE_FLOAT || type == VALUE_INTEGER) { const char *text = CHAR_CAST (const char *, xv); char *endptr; @@ -446,6 +471,29 @@ convert_xml_string_to_value (struct ccase *c, const struct variable *var, if ( errno != 0 || endptr == text) v->f = SYSMIS; } + else + { + const char *text = CHAR_CAST (const char *, xv); + + const struct fmt_spec *fmt = var_get_write_format (var); + + char *m = data_in (ss_cstr (text), "UTF-8", + fmt->type, + v, + var_get_width (var), + "UTF-8"); + + if (m) + { + char buf [FMT_STRING_LEN_MAX + 1]; + char *cell = create_cell_ref (col, row); + + msg (MW, _("Cannot convert the value in the spreadsheet cell %s to format (%s): %s"), + cell, fmt_to_string (fmt, buf), m); + free (cell); + } + free (m); + } } struct var_spec @@ -453,6 +501,7 @@ struct var_spec char *name; int width; xmlChar *first_value; + int first_type; }; @@ -472,7 +521,7 @@ gnumeric_error_handler (void *ctx, const char *mesg, static struct gnumeric_reader * gnumeric_reopen (struct gnumeric_reader *r, const char *filename, bool show_errors) { - int ret; + int ret = -1; struct state_data *sd; xmlTextReaderPtr xtr; @@ -508,7 +557,7 @@ gnumeric_reopen (struct gnumeric_reader *r, const char *filename, bool show_erro { r = xzalloc (sizeof *r); r->spreadsheet.n_sheets = -1; - r->spreadsheet.file_name = filename; + r->spreadsheet.file_name = strdup (filename); sd = &r->msd; } else @@ -525,7 +574,8 @@ gnumeric_reopen (struct gnumeric_reader *r, const char *filename, bool show_erro sd->row = sd->col = -1; sd->state = STATE_PRE_INIT; sd->xtr = xtr; - r->ref_cnt++; + r->spreadsheet.ref_cnt++; + /* Advance to the start of the workbook. This gives us some confidence that we are actually dealing with a gnumeric @@ -541,8 +591,7 @@ gnumeric_reopen (struct gnumeric_reader *r, const char *filename, bool show_erro if ( ret != 1) { /* Does not seem to be a gnumeric file */ - xmlFreeTextReader (sd->xtr); - free (r); + gnumeric_unref (&r->spreadsheet); return NULL; } @@ -581,6 +630,7 @@ struct casereader * gnumeric_make_reader (struct spreadsheet *spreadsheet, const struct spreadsheet_read_options *opts) { + int type = 0; int x = 0; struct gnumeric_reader *r = NULL; unsigned long int vstart = 0; @@ -592,8 +642,7 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet, r = (struct gnumeric_reader *) (spreadsheet); - if (r->rsd.row != -1) - r = gnumeric_reopen (r, NULL, true); + r = gnumeric_reopen (r, NULL, true); if ( opts->cell_range ) { @@ -618,6 +667,8 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet, r->target_sheet_index = opts->sheet_index; r->rsd.row = r->rsd.col = -1; r->rsd.current_sheet = -1; + r->first_case = NULL; + r->proto = NULL; /* Advance to the start of the cells for the target sheet */ while ( (r->rsd.state != STATE_CELL || r->rsd.row < r->start_row ) @@ -647,6 +698,7 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet, n_cases --; } + /* Read in the first row of cells, including the headers if read_names was set */ while ( @@ -655,9 +707,29 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet, ) { int idx; + + if (r->rsd.state == STATE_CELL && r->rsd.node_type == XML_READER_TYPE_TEXT) + { + xmlChar *attr = + xmlTextReaderGetAttribute (r->rsd.xtr, _xml ("ValueType")); + + type = _xmlchar_to_int (attr); + + xmlFree (attr); + } + process_node (r, &r->rsd); - if ( r->rsd.row > r->start_row ) break; + if ( r->rsd.row > r->start_row ) + { + xmlChar *attr = + xmlTextReaderGetAttribute (r->rsd.xtr, _xml ("ValueType")); + + r->vtype = _xmlchar_to_int (attr); + + xmlFree (attr); + break; + } if ( r->rsd.col < r->start_col || (r->stop_col != -1 && r->rsd.col > r->stop_col)) @@ -674,10 +746,13 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet, var_spec [i].name = NULL; var_spec [i].width = -1; var_spec [i].first_value = NULL; + var_spec [i].first_type = -1; } n_var_specs = idx + 1 ; } + var_spec [idx].first_type = type; + if ( r->rsd.node_type == XML_READER_TYPE_TEXT ) { xmlChar *value = xmlTextReaderValue (r->rsd.xtr); @@ -709,7 +784,7 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet, xmlChar *attr = xmlTextReaderGetAttribute (r->rsd.xtr, _xml ("ValueType")); - if ( NULL == attr || 60 != _xmlchar_to_int (attr)) + if ( NULL == attr || VALUE_STRING != _xmlchar_to_int (attr)) var_spec [idx].width = 0; free (attr); @@ -765,9 +840,12 @@ gnumeric_make_reader (struct spreadsheet *spreadsheet, continue; var = dict_get_var (r->dict, x++); - + convert_xml_string_to_value (r->first_case, var, - var_spec[i].first_value); + var_spec[i].first_value, + var_spec[i].first_type, + r->rsd.col + i - 1, + r->rsd.row - 1); } for ( i = 0 ; i < n_var_specs ; ++i ) @@ -826,11 +904,22 @@ gnm_file_casereader_read (struct casereader *reader UNUSED, void *r_) if (r->start_col == -1) r->start_col = r->rsd.min_col; + while ((r->rsd.state == STATE_CELL || r->rsd.state == STATE_CELLS_START ) && r->rsd.row == current_row && (ret = xmlTextReaderRead (r->rsd.xtr))) { process_node (r, &r->rsd); + if (r->rsd.state == STATE_CELL && r->rsd.node_type == XML_READER_TYPE_ELEMENT) + { + xmlChar *attr = + xmlTextReaderGetAttribute (r->rsd.xtr, _xml ("ValueType")); + + r->vtype = _xmlchar_to_int (attr); + + xmlFree (attr); + } + if ( r->rsd.col < r->start_col || (r->stop_col != -1 && r->rsd.col > r->stop_col)) continue; @@ -841,17 +930,17 @@ gnm_file_casereader_read (struct casereader *reader UNUSED, void *r_) if ( r->stop_row != -1 && r->rsd.row > r->stop_row) break; + if ( r->rsd.node_type == XML_READER_TYPE_TEXT ) { xmlChar *value = xmlTextReaderValue (r->rsd.xtr); - const int idx = r->rsd.col - r->start_col; - const struct variable *var = dict_get_var (r->dict, idx); - convert_xml_string_to_value (c, var, value); + convert_xml_string_to_value (c, var, value, r->vtype, + r->rsd.col, r->rsd.row); - free (value); + xmlFree (value); } } @@ -865,4 +954,4 @@ gnm_file_casereader_read (struct casereader *reader UNUSED, void *r_) } -#endif /* GNM_SUPPORT */ +#endif /* GNM_READ_SUPPORT */