1 2008-03-18 John Darrington <john@darrington.wattle.id.au>
3 * data-in.c: If category is custom currency, then use
4 numeric format for input.
6 2008-03-06 Ben Pfaff <blp@gnu.org>
8 * dictionary.c (var_name_is_insertable): New function.
9 (make_hinted_name): Don't accept variable names that match PSPP
10 keywords. Thanks to Jason Stover for reporting the problem.
12 2008-03-06 Ben Pfaff <blp@gnu.org>
14 * format-guesser.c (syntax): Require month names to be spelled out
15 as English words, so that single characters that happen to be
16 Roman numerals don't get detected as months. Thanks to John
17 Darrington for reporting this bug.
19 2008-03-04 Ben Pfaff <blp@gnu.org>
21 Patch #6441. Reviewed by John Darrington.
23 * format.c (fmt_fix): New function.
24 (fmt_fix_input): New function.
25 (fmt_fix_output): New function.
27 * format.def: Correct minimum width for DATETIME format. It was
28 7, should have been 17.
30 * automake.mk: Add new files.
32 * format-guesser.c: New file.
34 * format-guesser.h: New file.
36 2008-02-18 Ben Pfaff <blp@gnu.org>
38 Patch #6426. Reviewed by John Darrington.
40 * format.c (min_width): Renamed fmt_min_width and made public.
41 Updated all references.
42 (max_width): Renamed fmt_max_width and made public. Updated all
44 (max_decimals): Renamed fmt_max_decimals and made public. Updated
46 (var_create): Use the new functions for default variable
48 (var_default_formats): New function.
49 (var_default_measure): New function.
50 (var_default_alignment): New function.
52 * format.h (macro FMT_MAX_NUMERIC_WIDTH): New macro.
54 2008-02-09 Ben Pfaff <blp@gnu.org>
56 Add a couple of extensions to GET DATA TYPE=TXT. Patch #6412.
57 Thanks to John Darrington for review.
59 * data-in.c (data_in): Add new argument to designate the last
60 column of the data field being parsed, for use in error messages.
63 * data-parser (struct data_parser): New member `quote_escape'.
64 (data_parser_create): Initialize quote_escape.
65 (data_parser_set_quotes): New function.
66 (cut_field): Support escaped quotes.
67 (parse_delimited_span): Ditto.
68 (parse_delimited_no_span): Ditto.
70 * get-data.c (parse_get_txt): Support ESCAPE extension subcommand
71 in enhanced mode. Only support multiple quote characters in
74 2008-02-06 John Darrington <john@darrington.wattle.id.au>
76 psql-reader.c psql-reader.h: Read more than one tuple at
77 once. Fix bug reading a query which returns no data. Fix bug
78 when transformation followed a reader.
79 Ask the server for the number of records in the query, for the
82 2008-02-05 John Darrington <john@darrington.wattle.id.au>
84 psql-reader.c: So yesterday they release postgresql 8.3.0
85 which has money represented with 64 bits. They must get
88 2008-02-02 John Darrington <john@darrington.wattle.id.au>
90 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
91 for reviewing this code.
93 2008-02-02 Ben Pfaff <blp@gnu.org>
97 * sys-file-reader.c (read_variable_record): Allow missing values
98 to be specified on long string variables, but warn about them
99 because PSPP does not yet support them.
100 (read_extension_record): Ignore extension records 20 and 21, which
101 PSPP does not yet support.
102 (read_header): Fix error message when floating-point format cannot
105 2008-02-01 Ben Pfaff <blp@gnu.org>
107 Patch #6386. Thanks to John Darrington for review and for the
108 updates to gnumeric-reader.c.
110 * dictionary.c (make_hinted_name): New function.
111 (make_numeric_name): New function.
112 (dict_make_unique_var_name): New function.
114 * gnumeric-reader.c (devise_name): Removed.
115 (munge_name): Removed.
116 (gnumeric_open_reader): Use new function
117 dict_make_unique_var_name.
119 * short-names.c (set_var_short_name_suffix): Use new function
122 2008-01-19 John Darrington <john@darrington.wattle.id.au>
124 * settings.c settings.h: Moved static variables into a
125 single struct. Renamed functions to have a settings_ prefix.
127 2008-01-14 John Darrington <john@darrington.wattle.id.au>
129 * data-out.c (output_AHEX): Corrected number of bytes in
130 call to output_hex. Closes bug #22011
132 2008-01-02 John Darrington <john@darrington.wattle.id.au>
134 * variable.c variable.h: Replaced var_get_value_name with
135 var_append_value_name which doesn't use any static data.
136 Thanks to Ben for review.
138 2007-12-07 Ben Pfaff <blp@gnu.org>
142 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
143 made from an empty casereader produce a casegrouper with no
144 groups, instead of one with one group that has no cases.
146 * casereader.c (casereader_is_empty): New function.
148 2007-12-06 Ben Pfaff <blp@gnu.org>
152 * sys-file-reader.c (read_display_parameters): Handle variable
153 display parameters record with only 2 data items per variable.
154 Reported by Guido Gay <gay@irer.it>.
156 2007-12-04 Ben Pfaff <blp@gnu.org>
158 * identifier.c (lex_id_match_n): New function.
159 (lex_id_match): Reimplement in terms of lex_id_match_n.
161 2007-11-24 Ben Pfaff <blp@gnu.org>
163 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
166 2007-11-24 Ben Pfaff <blp@gnu.org>
168 Fix warning reported when reading back system files that include
169 very long string variables. Thanks to Guido Gay <gay@irer.it> for
172 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
173 was written where `j' was meant.
175 2007-11-18 Ben Pfaff <blp@gnu.org>
177 Properly write variables that include a range of missing values to
178 system files. Thanks to Guido Gay <gay@irer.it> for reporting
181 * sys-file-writer.c (write_variable): Correctly calculate
182 n_missing_values field when writing variables that include a range
185 2007-11-10 Ben Pfaff <blp@gnu.org>
187 Cleanups and bug fixes devised while writing up documentation.
190 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
192 * dict-class.c: New file.
193 (dict_class_from_id): Move here.
194 (dict_class_to_name): Move here.
196 * dict-class.h: New file.
197 (enum dict_class): Move here. Change from consecutive integers to
198 consecutive bits, to make testing for any of multiple values
199 easier. Add new DC_ALL constant.
201 * dictionary.c (struct dictionary): Change `case_limit' from
202 size_t to casenumber.
203 (dict_get_vars): Make final argument an enum dict_class.
204 (dict_get_vars_mutable): Ditto.
205 (dict_get_case_limit): Change return value to casenumber.
206 (dict_set_case_limit): Change final argument to a casenumber.
207 (dict_unset_split_var): Add assertion. Rephrase slightly.
208 (dict_set_label): Use xstrndup to simplify.
210 * format.c (fmt_step_width): AHEX format also needs 2-byte
212 (fmt_set_style): Simplify assertions.
214 * missing-values.c (mv_add_num_range): Rename mv_add_range.
215 Simplify implementation.
216 (mv_has_value): Simplify implementation.
217 (mv_pop_value): Remove the first value, not the last, to avoid
218 having GET followed by SAVE reverse the order of missing values.
219 (mv_peek_value): Rename mv_get_value. Simplify assertion.
220 (mv_has_range): Simplify implementation.
221 (mv_peek_range): Rename mv_get_range.
222 (can_resize_string): Removed.
223 (mv_is_resizable): Use value_is_resizable.
224 (mv_resize): Use value_resize.
226 * short-names.h (SHORT_NAME_LEN): Move here.
228 * val-type.h: New file, for definitions related to type and width
229 of abstract values. Before, these definitions were mixed among
230 those related to "union value" and those related to variables.
231 (macro SYSMIS): Move here.
232 (macro LOWEST): Move here.
233 (macro HIGHEST): Move here.
234 (macro MAX_STRING); Move here.
235 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
236 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
237 All references updated.
238 (val_type_is_valid): New function. Replaces var_type_is_valid.
239 All references updated.
240 (val_type_from_width): New function. Replaces
241 var_type_from_width. All references updated.
243 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
245 (val_labs_can_set_width): Use value_is_resizable.
246 (val_labs_add): Simply return false if the value labels set is too
247 wide, instead of having undefined behavior.
248 (val_labs_replace): Ditto.
249 (val_labs_replace): Ditto.
250 (val_labs_first): Set iterator to null if iteration is complete.
251 (val_labs_first_sorted): Ditto.
252 (val_labs_done): Become a no-op if the iterator is null.
254 * value.c (value_is_resizable): New function.
255 (value_resize): New function.
257 * variable.c (var_get_dict_class): New function.
259 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
262 2007-11-08 Ben Pfaff <blp@gnu.org>
264 * data-in.c: Make formatted data parsing locale-independent.
265 (parse_number): Use c_strtod instead of strtod, to avoid
266 locale-specific behavior.
269 2007-11-06 Ben Pfaff <blp@gnu.org>
271 Patch #6256: add support for binary, 360 file formats. Reviewed
274 * data-in.c (struct data_in): Add `encoding' member.
275 (data_in): Add `encoding' parameter, and re-encode the data passed
276 in where appropriate. Update all callers to pass it in.
277 (parse_A): Implement EBCDIC recoding wart described in manual.
278 (parse_AHEX): Implement EBCDIC recoding.
280 * data-out.c (data_out_legacy): New function.
281 (data_out): Make into a wrapper around data_out_legacy.
283 * file-handle-def.c (struct file_handle): New member `encoding'.
284 (fh_create_file): Set encoding.
285 (fh_default_properties): Set default encoding.
286 (fh_get_legacy_encoding): New function.
288 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
289 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
290 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
291 (struct fh_properties): New member `encoding'.
293 2007-11-05 Ben Pfaff <blp@gnu.org>
295 Patch #6258. Reviewed by John Darrington.
297 * file-handle-def.c (fh_lock): Add comment that TYPE should be
298 marked with N_() in the caller. Added these markings to each
299 caller too. Should make i18n easier.
300 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
302 2007-11-03 Ben Pfaff <blp@gnu.org>
304 Allow output files to overwrite input files (bug #21280).
306 * file-handle-def.c: Separate locking of files for input and for
307 output, to allow both to take place at once. Also, distinguish a
308 file handle from the identity of the underlying file, because the
309 identity of a file changes over time and the file handle can't
310 represent two different identities.
311 (struct file_handle): Remove `next', `open_cnt', `deleted',
312 `type', `open_mode', `aux', `identity' members. Change `id' from
313 char array to char *. Add `ref_cnt' member.
314 (file_handle_from_ll) New function.
315 (file_handles) Removed.
316 (named_handles) New variable.
317 (fh_init) Initialize named_handles.
318 (fh_done) Remove name from all named_handles.
319 (free_handle) Updated for modified struct file_handle.
320 (unname_handle) New function.
321 (fh_ref) New function.
322 (fh_from_file_name) Removed.
323 (fh_unref) New function.
324 (fh_unname) New function.
325 (fh_from_id) Rewritten.
326 (create_handle) Updated for modified struct file_handle.
327 (fh_inline_file) Increment the handle's ref count.
328 (fh_create_file) Updated for modified struct file_handle.
329 (fh_create_scratch) Ditto.
334 (fh_is_open) Removed.
335 (fh_get_id) Updated for modified struct file_handle.
336 (fh_get_default_handle) Increment the handle's ref count.
337 (fh_set_default_handle) Handle ref counts.
338 (struct fh_lock) New structure.
339 (locks) New static var.
340 (fh_lock) New function.
341 (fh_unlock) New function.
342 (fh_lock_get_aux) New function.
343 (fh_lock_set_aux) New function.
344 (fh_is_locked) New function.
345 (make_key) New function.
346 (free_key) New function.
347 (compare_fh_locks) New function.
348 (hash_fh_lock) New function.
350 * file-handle-def.h (enum fh_access) New enum.
352 * file-name.c: Made file_identity the same in all supported
354 (struct file_identity): New `name' member.
355 (fn_get_identity): For a file that doesn't exist, get the
356 dev/inode of its directory plus its name. If even the directory
357 doesn't exist, just use its name. Merge the Windows
358 implementation into the Unix one.
359 (fn_compare_file_identities): Rewritten. Merge the Windows
360 implementation into the Unix one.
361 (fn_hash_identity): New function.
363 * make-file.c (struct replace_file): New structure.
364 (all_files): New static var.
365 (replace_file_start): New function.
366 (replace_file_commit): New function.
367 (replace_file_abort): New function.
368 (free_replace_file): New function.
369 (unlink_replace_files): New function.
371 * por-file-reader.c (struct pfm_reader): Add `lock' member.
372 (close_reader): Unlock file.
373 (pfm_open_reader): Lock file.
375 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
377 (pfm_open_writer): Lock file and prepare for its replacement.
378 (close_writer): Unlock file.
380 * scratch-handle.h (struct scratch_handle): Add unique_id so that
381 different generations of a scratch file can be distinguished.
383 * scratch-reader.c (scratch_reader_open): Verify that the file is
386 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
387 remove scratch_handle member.
388 (scratch_writer_open): Lock handle. Prepare to replace handle
389 data, instead of doing it immediately.
390 (scratch_writer_casewriter_destroy): Replace handle data and
393 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
394 (sfm_open_reader): Lock file.
395 (close_reader): Unlock file.
397 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
399 (sfm_open_writer): Lock file and prepare for its replacement.
400 (close_writer): Unlock file.
402 2007-11-02 Ben Pfaff <blp@gnu.org>
404 * data-out.c (output_number): Use isfinite (from C99) instead of
405 gsl_isfinite, since we now have universal support for it in
407 (output_infinite): Ditto.
408 [!HAVE_ROUND] (round): Remove definition, since we now have a
409 replacement in gnulib.
411 2007-10-27 John Darrington <john@darrington.wattle.id.au>
413 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
414 wasn't initialised correctly. Closes bugs #21330 and 21397
417 2007-10-08 Ben Pfaff <blp@gnu.org>
419 Fix bug #21280. Thanks to John Darrington for review.
421 * file-name.c (create_stream): New function.
423 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
424 file handle before creating the file, to ensure that we don't
425 truncate a file that we're reading. Make code easier to read by
428 * sys-file-write.c (sfm_open_writer): Ditto.
430 2007-10-01 Ben Pfaff <blp@gnu.org>
432 Fix bug #21192. Thanks to John Darrington for review.
434 * casereader.c (casereader_read): Decrement case_cnt before
435 calling the casereader's "read" member function, so that we
436 interact properly with lazy_casereader.
438 * datasheet.c: Add regression test for above bug fix.
439 (clone_datasheet): New function.
440 (lazy_callback): New function.
441 (check_datasheet_casereader): New function.
442 (check_datasheet): Check datasheet contents are reported correctly
443 through an ordinary casereader and a lazy casereader.
444 (clone_model): Use clone_datasheet.
446 2007-09-24 Ben Pfaff <blp@gnu.org>
448 Patch #6210. Reviewed by John Darrington.
449 * settings.c: Drop "long view", which is not needed any longer.
450 (static var long_view): Removed.
451 (force_long_view): Removed.
452 (get_viewwidth): Removed.
453 (init_viewport): Removed long_view reference.
455 2007-09-19 John Darrington <john@darrington.wattle.id.au>
457 * settings.c settings.h: Changed viewport's length and width to be
458 owned by the user interface which uses the data library. This allows
459 better abstraction, and makes dynamically adjustable dimensions easier.
461 2007-09-18 Ben Pfaff <blp@gnu.org>
463 * procedure.c (proc_extract_active_file_data): New function.
465 * lazy-casereader.h: New file.
467 * lazy-casereader.c: New file.
469 * casereader.c (casereader_dynamic_cast): New function.
471 2007-09-14 Ben Pfaff <blp@gnu.org>
473 * dictionary.c (dict_clone): Copy case indexes from cloned
474 dictionary. Fixes bug #21061. Reviewed by John Darrington.
476 2007-09-13 John Darrington <john@darrington.wattle.id.au>
478 * value.c value.h (create_value): New function.
480 2007-09-12 Ben Pfaff <blp@gnu.org>
482 Make it clear that translator casereader and casewriter translate
483 functions are supposed to destroy their input case.
485 * casereader-translator.c (struct casereader_translator): Change
486 input case parameter of translate member function from const
487 struct ccase * to struct ccase *.
488 (casereader_create_translator): Ditto, for translate parameter.
490 * casewriter-translator.c (struct casewriter_translator): Ditto.
491 (casewriter_create_translator): Ditto.
493 2007-08-27 John Darrington <john@darrington.wattle.id.au>
495 * sys-file-reader.c (read_display_parameters): Force display width
496 to 8, if the sys file says 0 (like SPSS does).
498 2007-08-12 Ben Pfaff <blp@gnu.org>
500 * dictionary.c (dict_dump): New function.
502 2007-08-12 Ben Pfaff <blp@gnu.org>
504 Drop dict_compactor in favor of using the new struct case_map.
506 * dictionary.c (struct copy_map): Removed.
507 (struct dict_compactor): Removed.
508 (dict_make_compactor): Removed.
509 (dict_compactor_compact): Removed.
510 (dict_compactor_destroy): Removed.
512 * procedure.c (struct data_set): Change `compactor' member to be a
514 (proc_open): Use case_map_to_compact_dict instead of
516 (proc_casereader_read): Use case_map_execute instead of
517 dict_compactor_compact.
518 (proc_commit): Use case_map_destroy instead of
519 dict_compactor_destroy.
521 * scratch-writer.c (struct scratch_writer): Change `compactor'
522 member to be a struct case_map *.
523 (scratch_writer_open): Use case_map_to_compact_dict instead of
525 (scratch_writer_casewriter_write): Use case_map_execute instead of
526 dict_compactor_compact.
528 2007-08-12 Ben Pfaff <blp@gnu.org>
530 * automake.mk: Add case-map.c, case-map.h.
532 * case-map.c: New file.
534 * case-map.h: New file.
536 2007-08-12 Ben Pfaff <blp@gnu.org>
538 * dictionary.c (dict_compact_values): Don't delete scratch
539 variables as well as compacting case indexes. Update all callers.
540 (dict_get_compacted_value_cnt): Rename dict_count_values and
541 change interface. Update all callers.
542 (dict_get_compacted_value_cnt): Remove.
543 (dict_compacting_would_shrink): Remove.
544 (dict_compacting_would_change): Remove.
545 (dict_make_compactor): Add new parameter. Update all callers.
547 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
548 instead of dict_count_values, changing an O(N) operation into
551 2007-08-12 Ben Pfaff <blp@gnu.org>
553 * casereader.c (casereader_read): Don't require cases read by a
554 casereader to be exactly the expected size: as long as they're big
557 2007-08-12 Ben Pfaff <blp@gnu.org>
559 Make casewriters keep track of the number of `union value's in
560 each case. This is useful for two reasons: casewriter_write can
561 then check that the case being written is large enough, and later
562 recipients of the casewriter can determine the size of the case.
564 * casewriter-translator.c (casewriter_create_translator): Add
567 * casewriter.c (struct casewriter): Add value_cnt member.
568 (casewriter_write): Check that the case passed in is big enough.
569 (casewriter_get_value_cnt): New function.
570 (casewriter_create): Add value_cnt parameter.
572 2007-08-09 Ben Pfaff <blp@gnu.org>
574 Fix bug reported by Jason Stover.
575 * settings.c: Move get_termcap_viewport and in particular the
576 #include for <curses.h> to the end of the file. curses.h
577 redefines bool on some systems (e.g. OpenBSD), which causes
578 disagreement between uses of bool before its inclusion and
580 Tested by Jason Stover.
582 2007-07-29 Ben Pfaff <blp@gnu.org>
584 Provisional fix for bug #18692 and bug #20161. Reviewed by John
587 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
588 argument (never "rb" or "wb") because SUSv3 says that only those
589 modes are defined, and glibc in fact rejects other modes.
591 Open portable files with fn_open so that they can be read from
592 pipes. Fix missing fh_close call to go along with fh_open.
593 Report an error if the file close reports an error.
594 * por-file-reader.c (close_reader): New function.
595 (por_file_casereader_destroy): Use close_reader.
596 (pfm_open_reader): Open file with fn_open.
598 2007-07-28 Ben Pfaff <blp@gnu.org>
600 Make PSPP able to read all the portable files I could find on the
601 web. Thanks to John Darrington for review. Bug #17620.
602 * por-file-reader.c (struct pfm_reader): New member `line_length'.
603 (error): Print file offset in hexadecimal.
604 (warning): New function.
605 (advance): Treat lines less than 80 bytes long as padded to 80
607 (pfm_open_reader): Call read_documents if we find an "E" record.
608 (convert_format): Convert invalid formats to the default format
609 instead of aborting reading the file.
610 (read_variables): Rename duplicate variable names instead of
611 aborting reading the file.
612 (read_value_label): Allow string variables of different widths to
613 be assigned value labels in the same record. Replace duplicate
614 value labels instead of aborting.
615 (read_documents): New function.
617 * por-file-writer.c (pfm_open_writer): Call write_documents if the
618 dictionary has documents.
619 (write_documents): New function.
621 2007-07-25 Ben Pfaff <blp@gnu.org>
623 Fix bugs related to bug #17213.
625 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
626 since the former is what config.h has. Include the needed ncurses
628 (static var echo) Rename to `do_echo' because the original name is
629 the same as an ncurses identifier.
630 (get_termcap_viewport) Use error instead of msg.
632 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
633 (fn_close): Don't close stdin, stdout, stderr.
635 2007-07-26 John Darrington <john@darrington.wattle.id.au>
637 * procedure.c procedure.h: Added callbacks which get invoked whenever
638 a dataset's transformation chain changes.
640 2007-07-24 Ben Pfaff <blp@gnu.org>
643 * sys-file-writer.c (write_variable_display_parameters): Use new
644 var_default_display_width function to choose display width of
645 segments after the first one in a given variable.
646 * variable.c (var_create): Use var_default_display_width to pick
647 new variable's display width.
648 (var_default_display_width): New function.
649 Reviewed by John Darrington.
651 2007-07-24 Ben Pfaff <blp@gnu.org>
654 * por-file-writer.c (write_variables): Write weight variable.
655 Reviewed by John Darrington.
657 2007-07-23 Ben Pfaff <blp@gnu.org>
659 Improvements to system file reader and writer.
661 First, move all detailed knowledge of very long strings into
662 sys-file-private.[ch], so that this nasty stuff can be isolated.
664 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
665 (EFFECTIVE_VLS_CHUNK): New macro.
666 (min_int): New function.
667 (max_int): New function.
668 (sfm_width_to_bytes): Rewrite.
669 (sfm_width_to_octs): New function.
670 (sfm_segment_alloc_width): New function.
671 (sfm_segment_alloc_bytes): New function.
672 (sfm_segment_used_bytes): New function.
673 (sfm_segment_offset): New function.
674 (sfm_segment_effective_offset): New function.
675 (sfm_dictionary_to_sfm_vars): New function.
677 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
678 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
679 (struct sfm_var): New structure.
681 Next, improvements to the system file reader.
683 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
684 type casenumber. Added `version_major', `version_minor',
687 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
688 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
689 Change `case_cnt' to type casenumber. Removed `has_vls'.
690 (struct sfm_var): Removed.
691 (sfm_open_reader): Don't warn on wrong case size if the file was
692 written by SPSS 13, which tends to get it wrong. Use
693 sfm_dictionary_to_sfm_vars.
694 (read_header): Always output system file info.
695 (read_variable_record): Simplify code for reading missing values.
696 (read_machine_int32_info): Save version numbers from system file
697 into info struct passed as new argument.
698 (read_long_string_map): Restructured to use new sys-file-private
700 (read_value_labels): Use size_overflow_p.
701 (sys_file_casereader_read): Get rid of distinction between fast
702 and slow paths. Use information provided by sys-file-primate's
703 struct sfm_var to simplify code.
704 (skip_whole_strings): New function.
705 (read_int32): Renamed read_int. Changed return value to int.
707 (read_flt64): Renamed read_float. Changed return value to
708 double. Updated all callers.
709 (int32_to_native): Removed. Changed callers to use
711 (flt64_to_double): Removed. Changed callers to use float_convert.
713 Finally, get rid of int32, flt64 terminology and types in system
714 file writer. The former wasn't very useful since a POSIX "int"
715 can hold the whole range of int32 and we generally didn't have a
716 need for it to be exactly-32-bits, just at-least-32-bits. The
717 latter was inconvenient because we had to assume that it could be
718 different from double and thereby convert special values SYSMIS,
719 HIGHEST, LOWEST to and from it in multiple places. Instead, now
720 we just use "int" and "double" in most places, and do conversions,
721 if necessary, very close to where we do I/O. This change meant
722 that the writer code couldn't represent records in the file as C
723 structs any longer, but that's no great loss. The code actually
724 seems to be more readable without them.
726 Simplify the compression buffering code: only buffer as much as
727 necessary, which is no more than eight 8-byte units at any given
730 * sys-file-writer.c (typedef flt64): Removed.
731 (macro second_lowest_flt64): Removed.
732 (struct sysfile_header): Removed.
733 (struct sysfile_variable): Removed.
734 (struct sfm_writer): Removed `needs_translation', `has_vls',
735 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
736 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
737 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
738 compression buffering by `opcodes', `opcode_cnt', `data',
739 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
740 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
741 data writer functions instead of structures.
742 (calc_oct_idx): New function.
743 (write_header): Use simple data writer functions instead of
745 (write_format_spec): Renamed write_format. New argument.
746 (write_variable_continuation_records): New function.
747 (write_variable): Use simple data writer functions instead of
748 structures. Use write_variable_continuation_records. Write
749 entire very long string instead of requiring caller to understand
751 (write_value_labels): Use simple data writer functions instead of
753 (write_documents): Ditto.
754 (write_variable_display_parameters): Use sys-file-private
755 functions to simplify. Use simple data writer functions instead
757 (write_vls_length_table): Use simple data writer functions instead
759 (write_longvar_table): Ditto.
760 (write_rec_7_34): Break into new functions
761 write_integer_info_record, write_float_info_record. Use simple
762 data writer functions instead of structures.
763 (buf_write): Removed.
764 (append_string_max): Removed.
765 (ensure_buf_space): Removed.
766 (sys_file_casewriter_write): Get rid of the distinction between
767 fast and slow paths, which didn't seem to be too useful. Use new
768 functions write_case_uncompressed, write_case_compressed.
769 (put_instruction): Removed.
770 (put_element): Removed.
771 (write_compressed_data): Removed.
772 (close_writer): Use flush_compressed. Only write case count to
773 system file if it will fit in the field.
774 (write_case_compressed): New function.
775 (write_case_uncompressed): New function.
776 (flush_compressed): New function.
777 (put_cmp_opcode): New function.
778 (put_cmp_number): New function.
779 (write_int): New function.
780 (convert_double_to_output_format): New function.
781 (write_float): New function.
782 (write_value): New function.
783 (write_string): New function.
784 (write_bytes): New function.
785 (write_zeros): New function.
786 (write_spaces): New function.
788 Reviewed by John Darrington.
790 2007-07-22 Ben Pfaff <blp@gnu.org>
792 Don't try to write very long strings to portable files. The
793 format does not support it.
795 * por-file-writer.c (MAX_POR_WIDTH): New macro.
796 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
797 (write_format): Add arg to take width to resize format to.
798 (write_value): Limit width of value written to MAX_POR_WIDTH.
799 (write_variables): Limit width of variable and its output formats
801 Reviewed by John Darrington.
803 2007-07-22 Ben Pfaff <blp@gnu.org>
805 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
806 local variable instead of literal 5.
807 Reviewed by John Darrington.
809 2007-07-22 Ben Pfaff <blp@gnu.org>
811 Fix problems with uniqueness of short names in system files with
812 very long string variables. Now a variable may have multiple
815 * automake.mk (src_data_libdata_a_SOURCES): Add new files
816 short-names.c, short-names.h.
818 * dictionary.c (dict_clone): Clone all the short names.
819 (compare_strings): Move into short-names.c.
820 (hash_strings): Ditto.
821 (set_var_short_name_suffix): Ditto.
822 (dict_assign_short_names): Ditto, rename short_names_assign,
823 change to assign all short names.
825 * por-file-writer.c (write_variables): Use short_names_assign
826 instead of dict_assign_short_names.
828 * short-names.c: New file.
830 * short-names.h: New file.
832 * sys-file-private.c (sfm_width_to_segments): New function.
834 * sys-file-reader.c (read_long_var_name_map): Save and restore all
835 the short names, not just the first one.
837 * sys-file-writer.c (cont_var_name): Removed.
838 (sfm_open_writer): Use short_names_assign instead of
839 dict_assign_short_names. Use unique short names assigned by
840 short_names_assign instead of those generated by cont_var_name.
842 * variable.c (struct variable): Remove `short_name' member,
843 replace by `short_names' and `short_name_cnt'.
844 (var_create) Initialize new members.
845 (var_get_short_name_cnt): New function.
846 (var_get_short_name): Now takes an index argument. Changed most
848 (var_set_short_name): Ditto.
849 (var_clear_short_name): Renamed var_clear_short_names, changed to
850 clear all short names.
852 Reviewed by John Darrington.
854 2007-07-22 Ben Pfaff <blp@gnu.org>
856 * variable.c (var_set_width): Use new var_set_width function.
858 * missing-values.c (mv_n_values): Drop assertion, which was not
861 * format.c (fmt_default_for_width): New function.
862 (fmt_resize): New function.
864 Reviewed by John Darrington.
866 2007-07-18 John Darrington <john@darrington.wattle.id.au>
868 * datasheet.c (datasheet_delete_columns): Added assertion to check
869 we're not deleting outside the range of the sheet.
872 * dictionary.c dictionary.h variable.c: Added the ability for string
873 variables to be resized.
875 * vardict.h: Added some prototypes (moved from dictionary.h) as
876 these should only be called by variable.c
879 2007-07-14 John Darrington <john@darrington.wattle.id.au>
881 * sfm-reader.c: Respect case_cnt field in file header.
883 2007-07-01 John Darrington <john@darrington.wattle.id.au>
885 * transformation.c transformation.h (trns_chain_execute): Changed the
886 signature (Patch #6057)
888 2007-06-10 Ben Pfaff <blp@gnu.org>
890 * casereader-filter.c (casereader_filter_destroy): Make sure to
891 write all the remaining excluded cases to the casewriter, if any.
893 * caseinit.c (init_list_destroy): Rewrite.
894 (init_list_clear): Ditto.
896 * casegrouper.c (casegrouper_get_next_group): Always set *reader
897 to null when returning false.
899 2007-06-06 Ben Pfaff <blp@gnu.org>
901 Actually implement the new procedure code and adapt all of its
902 clients to match. Also adapt all of the other case sources and
903 sinks in the tree and their clients to use the
904 casereader/casewriter infrastructure.
906 * automake.mk: Add and remove files.
908 * any-reader.c: Change into a casereader.
909 * por-file-reader.c: Ditto.
910 * scratch-reader.c: Ditto.
911 * sys-file-reader.c: Ditto.
913 * any-writer.c: Change into a casewriter.
914 * por-file-writer.c: Ditto.
915 * scratch-writer.c: Ditto.
916 * sys-file-writer.c: Ditto.
918 * procedure.c: Change to use casereader, casewriter, caseinit, and
919 other new infrastructure.
921 * scratch-handle.c: Adapt to new infrastructure.
923 * case-sink.c: Removed, now dead code.
924 * case-sink.h: Ditto.
925 * case-source.c: Ditto.
926 * case-source.h: Ditto.
927 * casefile-factory.c: Ditto.
928 * casefile-private.h: Ditto.
931 * casefilter.c: Ditto.
932 * casefilter.h: Ditto.
935 * fastfile-factory.c: Ditto.
936 * fastfile-factory.h: Ditto.
937 * storage-stream.c: Ditto.
938 * storage-stream.h: Ditto.
940 2007-06-06 Ben Pfaff <blp@gnu.org>
944 * automake.mk: Add new files.
946 * datasheet.c: New file.
948 * datasheet.h: New file.
950 2007-06-06 Ben Pfaff <blp@gnu.org>
952 Until now, the procedure code has provided a case to the
953 case_source, which has filled in the data values that come from
954 the active file. "Left" data values that don't come from the
955 active file naturally stay the same from case to case, because the
956 procedure code keeps using that same case.
958 One of the compromises that comes with the new procedure code is
959 that the active file allocates and provides its own case, which
960 the procedure code then has to resize to provide room for any
961 other variables that should go in the case and then fill in the
962 values of "left" variables. Then, when we're done with that case,
963 we have to save the values of "left" variables to copy into the
964 next case read from the active file.
966 The caseinit code helps with this.
968 * automake.mk: Add new files.
970 * caseinit.c: New file.
972 * caseinit.h: New file.
974 2007-06-06 Ben Pfaff <blp@gnu.org>
976 * value.h (value_cnt_from_width): New function.
978 * variable.c (var_get_value_cnt): Use new function.
980 2007-06-06 Ben Pfaff <blp@gnu.org>
982 Add casegrouper, to allow cases read from a given casereader to be
983 broken into groups, each of which has its own casereader.
984 Generally cases are grouped based on having equal values for some
987 * automake.mk: Add new files.
989 * casegrouper.c: New file.
991 * casegrouper.h: New file.
993 2007-06-06 Ben Pfaff <blp@gnu.org>
995 Add interface to lexicographical ordering of cases.
997 * automake.mk: Add new files.
999 * case-ordering.c: New file.
1001 * case-ordering.h: New file.
1003 2007-06-06 Ben Pfaff <blp@gnu.org>
1005 Add casereaders and casewriters, the basis of the new data processing
1006 implementation. A casereader is a uniform interface to reading cases
1007 from a data source; a casewriter is a uniform interface to writing
1008 cases to a data sink.
1010 * automake.mk: Add new files.
1012 * casereader-filter.c: New file.
1014 * casereader-provider.h: New file.
1016 * casereader-translator.c: New file.
1018 * casereader.c: New file.
1020 * casereader.h: New file.
1022 * casewriter-provider.h: New file.
1024 * casewriter-translator.c: New file.
1026 * casewriter.c: New file.
1028 * casewriter.h: New file.
1030 2007-06-06 Ben Pfaff <blp@gnu.org>
1032 "casewindow" data structure that extends the deque (from libpspp)
1033 of cases with the ability to dump cases to disk if we get too many
1036 * automake.mk: Add new files.
1038 * casewindow.c: New file.
1040 * casewindow.h: New file.
1042 2007-06-06 Ben Pfaff <blp@gnu.org>
1044 sparse_cases data structure that augments a sparse_array of cases
1045 with the ability to dump cases to disk if we get too many cases in
1048 * automake.mk: Add new files.
1050 * sparse-cases.c: New file.
1052 * sparse-cases.h: New file.
1054 2007-06-06 Ben Pfaff <blp@gnu.org>
1056 Adds a low-level on-disk case array data structure.
1058 * automake.mk: Add new files.
1060 * case-tmpfile.c: New file.
1062 * case-tmpfile.h: New file.
1064 2007-06-06 Ben Pfaff <blp@gnu.org>
1066 In a couple of places we calculate the maximum number of cases to
1067 keep in memory based on the user-defined workspace. Enable
1068 centralizing the calculation through a new function.
1070 * settings.c (get_workspace_cases): New function.
1072 2007-06-06 Ben Pfaff <blp@gnu.org>
1074 The casenumber type is defined in transformations.h, but case.h is
1075 a more sensible place. Move it.
1077 * case.h (CASENUMBER_MAX): New macro.
1078 (typedef casenumber): Move here, from transformations.h.
1080 2007-06-03 Ben Pfaff <blp@gnu.org>
1082 Slightly generalize case_to_values and case_from_values functions.
1084 * case.c (case_to_values): Rename case_copy_out, change interface.
1085 (case_from_values): Rename case_copy_in, change interface.
1087 * fastfile.c (fastfilereader_get_next_case): Update caller.
1088 (write_case_to_disk): Ditto.
1090 2007-06-02 Ben Pfaff <blp@gnu.org>
1092 Clean up after a forgotten part of patch #5829.
1094 * casedeque.h: Remove unused file.
1096 * automake.mk: Remove casedeque.h from sources.
1098 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1100 * category.c: Removed redundant #include
1102 2007-05-06 Ben Pfaff <blp@gnu.org>
1104 Abstract the documents within a dictionary a little better.
1105 Thanks to John Darrington for suggestion, initial version, and
1106 review. Patch #5917.
1108 * dictionary.c (struct dictionary): Change `documents' member from
1109 char * to struct string.
1110 (dict_clear): Destroy struct string.
1111 (dict_get_documents): Convert struct string to char *.
1112 (dict_set_documents): Set struct string. Pad to 80-character
1114 (dict_clear_documents): New function.
1115 (dict_add_document_line): New function.
1116 (dict_get_document_line_cnt): New function.
1117 (dict_get_document_line): New function.
1119 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1121 * sys-file-reader.c (read_documents): Use new document functions.
1123 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1125 * sys-file-reader.c: When reading a system file which has no
1126 long name table, automatically create one where the long names
1127 are the lower case versions of the short names.
1129 2007-04-22 Ben Pfaff <blp@gnu.org>
1131 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1132 dict_clear will free most data related to the dictionary.
1133 dict_clear does a decent job, except that dict_set_split_vars on
1134 some systems won't actually free the dict's "split" member.
1135 Instead, it'll allocate a 1-byte region. Fix this.
1137 * value.c (value_copy): New function.
1138 (value_set_missing): Ditto.
1140 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1142 * Deleted existing category.h and moved cat-routines.h into
1143 category.h Encapsulated struct cat_vals better.
1145 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1147 * sys-file-reader.c: When reading a system file which has no
1148 long name table, automatically create one where the long names
1149 are the lower case versions of the short names.
1151 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1153 * sys-file-reader.c: Some versions of Other Software seem to
1154 produce system files with string variables' measure set to
1155 zero. We'll assume these are supposed to be nominal variables.
1157 2007-03-30 Ben Pfaff <blp@gnu.org>
1159 * procedure.c: Adapt to new deque data structure.
1161 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1162 Ben Pfaff <blp@gnu.org>
1164 * file-name.c: Mingw compatibility fixes.
1165 (fn_search_path): Use ISSLASH instead of comparing against '/'
1167 (fn_dir_name): Use dir_name from gnulib.
1168 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1169 (fn_get_identity): Use GetFullPathName instead of canonicalize
1170 from gnulib, because the latter does not fully support
1171 Windows-style path names. Use this implementation based on the
1172 detected presence of Windows instead of the absence of Unix, since
1173 the new implementation is Windows-specific.
1174 (fn_compare_file_identities): In Windows implementation, compare
1175 names case-insensitively.
1177 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1179 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1180 <stdlib.h> to have the same effect.
1182 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1184 * por-file-reader.c: Add missing _() around messages.
1186 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1188 * make-file.c: Include "mkstemp.h", without which linking on
1191 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1193 Reduce platform dependence.
1195 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1196 Everywhere we using this, we really should have just depended on
1197 the shell to expand tildes.
1198 (fn_search_path): Simplify, given that we don't do tilde expansion
1200 (fn_normalize): Removed. Caller changed to use the canonicalize
1202 (fn_get_cwd): Removed. Only user was fn_normalize.
1203 (fn_is_absolute): Really only test for absolute names.
1204 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1206 (fn_readlink): Removed, as it was only used fn_normalize.
1207 (fn_exists): Assume the stat function is available; gnulib does.
1208 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1210 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1212 * dictionary.c (dict_create_vector_assert): New function.
1214 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1216 * file-name.c (fn_normalize): Correct name of function
1217 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1218 for pointing this out.
1220 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1222 We are using a single member in struct file_handle, the "name"
1223 field, for more than one purpose. When it begins with '"', it's a
1224 file name; otherwise, it's a token that can be used to identify
1225 it. When that assertion fires, it's because we searched for the
1226 name case-sensitively as a file name (so that there was no match),
1227 and then we try to insert it case-insensitively as a token, which
1228 fails because duplicates aren't allowed.
1230 Solution: break the two purposes into two separate fields. This
1231 fixes the problem and likely makes the code easier to read too.
1233 Fixes bug #18922. Thanks to John Darrington for bug report and
1236 * file-handle-def.c (struct file_handle): New `id' member.
1237 (fh_from_name): Rename fh_from_id. Update all callers.
1238 (create_handle): New `id' parameter. Update all callers.
1239 (fh_create_file): Ditto.
1240 (fh_get_id): New function.
1242 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1244 * case.c (case_is_null): Change return type to bool.
1246 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1250 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1251 reference counting to share data. This makes it easy for
1252 valgrind, etc. to find accesses to cases that have been destroyed
1253 but have been kept around by another user's ref-count. This often
1254 happens when the data set is small enough to find in memory; if a
1255 bigger data set that would overflow to disk were used, then data
1256 corruption would occur.
1258 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1262 * case.c (case_unshare): Make it check internally whether the
1263 ref_cnt is greater than 1, so that the callers don't have to.
1264 Update callers not to check.
1266 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1268 Before, I was thinking that I might want to get rid of reference
1269 counting at some point. Now, I'm pretty sure that it's here to
1270 stay. Thus, because we have to store the value_cnt anyway for
1271 reference-counted cases, we might as well expose it to users.
1273 * case.c (case_get_value_cnt): New function.
1274 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1275 resize case if its size actually changed.
1277 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1278 instead of peeking inside struct case directly.
1279 (casefile_append): Ditto.
1281 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1283 Get rid of the inlines for the case functions, which made the
1284 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1285 the inlines didn't speed up "make check" at all, which is not a
1286 perfect benchmark but seems indicative.)
1288 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1289 definitions. Remove some assertions on nonnull pointers that were
1290 redundant with a pointer dereference soon after in the function.
1292 (struct case_data): Move definition here from case.h.
1296 (case_data_wr): Ditto.
1298 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1300 * automake.mk: Add casedeque.h to sources.
1302 * casedeque.h: New file.
1304 * procedure.c: (struct dataset) Change lag_count, lag_head,
1305 lag_queue member into single struct casedeque member. Update all
1306 users to use the casedeque instead.
1309 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1311 * procedure.c: Simplify lagged cases interface. Updated all
1312 clients--well, the only client--to use the simplified interface.
1313 (dataset_n_lag) Removed.
1314 (dataset_set_n_lag) Removed.
1315 (dataset_need_lag) New function.
1317 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1319 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1320 whenever appropriate, but only when the dataset/dictionary is in a
1323 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1325 * dictionary.c dictionary.h : Added callbacks for change of filter and
1326 split variables. Refactored some code to ensure that callbacks get
1327 invoked when appropriate.
1329 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1330 replace_dict callback occurs when permanent_dict replaces the current
1333 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1335 * dictionary.c dictionary.h : Added callback for when the weight
1336 variable of a dictionary changes.
1338 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1340 * dictionary.c dictionary.h : Added replace_source and replace_dict
1341 callbacks, and functions to deal with them.
1343 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1345 Simplify missing value handling.
1347 * missing-values.h (enum mv_class): New type.
1348 (enum mv_type): Moved definition into missing-values.c and renamed
1349 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1350 enums. Updated all uses.
1351 (struct missing_values): Changed type of `type' from `enum
1352 mv_type' to `int' because the definition is no longer exposed.
1354 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1355 parameter. Update all callers.
1356 (mv_is_num_missing): Ditto.
1357 (mv_is_str_missing): Ditto.
1358 (mv_is_value_user_missing): Removed. Changed callers to use
1359 mv_is_value_missing.
1360 (mv_is_num_user_missing): Removed. Changed callers to use
1362 (mv_is_str_user_missing): Removed. Changed callers to use
1364 (mv_is_value_system_missing): Removed. Changed callers to use
1365 mv_is_value_missing.
1366 (mv_set_type): Removed. Changed callers to use mv_clear.
1367 (mv_clear): New function.
1369 * variable.c (var_is_value_missing): Add new enum mv_class
1370 parameter. Update all callers.
1371 (var_is_num_missing): Ditto.
1372 (var_is_str_missing): Ditto.
1373 (var_is_value_user_missing): Removed. Changed callers to use
1374 var_is_value_missing.
1375 (var_is_num_user_missing): Removed. Changed callers to use
1377 (var_is_str_user_missing): Removed. Changed callers to use
1379 (var_is_value_system_missing): Removed. Changed callers to use
1380 var_is_value_missing.
1382 * casefilter.c (struct casefilter): Use enum mv_class in place of
1384 (casefilter_variable_missing): Adapt to new member.
1385 (casefilter_create): Change signature to take enum mv_class,
1388 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1390 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1392 * case-sink.c case-sink.h procedure.c procedure.h
1393 storage-stream.c: Now uses the factory.
1395 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1397 Make it possible to pull cases from the active file with a
1398 function call, instead of requiring indirection through a callback
1401 * case-source.h (struct case_source_class): Change ->read function
1402 to return a single case, instead of calling a callback function
1403 for each case. Change ->destroy function to return an error
1406 * case-source.c (free_case_source): Pass along the value returned
1407 by the case_source ->destroy function.
1409 * procedure.c (struct write_case_data): Removed.
1410 (struct dataset): Added some members to track procedure state.
1411 (procedure): Optimize the trivial case at this level.
1412 (internal_procedure): Re-implement in terms of proc_open,
1413 proc_read, proc_close.
1414 (proc_open) New function.
1415 (proc_read) New function.
1416 (proc_close) New function.
1417 (write_case) Moved into proc_read.
1418 (close_active_file) Moved closing of data source into proc_close.
1420 * storage-source.c: Rewrote to conform with modified
1421 case_source_class interface.
1423 * transformations.c (trns_chain_execute): Added argument to allow
1424 starting execution from an arbitrary transformation. Updated
1427 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1429 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1431 * sys-file-reader.c (read_display_parameters): Don't assume that
1432 MEASURE_* and ALIGN_* have the same values found in system files.
1434 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1436 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1437 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1438 the value of n_MEASURES, which was off by 1 (at least from my
1441 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1443 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1444 callbacks which are invoked when the dictionary or its
1445 variables are changed.
1447 * missing-values.c missing-values.h value-labels.c: Tidied up
1448 consistency checks, and made some of them return false
1449 instead of assert-failing.
1451 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1453 * calendar.c (calendar_days_in_month): New function.
1455 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1457 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1458 the value label's width.
1460 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1462 * sfm-private.h: Move contents into sys-file-writer.c, which is
1463 the only remaining user. Removed Borland C++-specific directives.
1465 * sys-file-reader.c: Clean up and rewrite entire file. The
1466 rewritten version is simpler and better abstracted, and should be
1467 easier to maintain and extend. It avoids using structures to read
1468 file data, which is prone to padding variations among compilers.
1469 It should also handle non-IEEE 754 system files, although I
1470 haven't been able to find any. It has been tested against many
1471 .sav files obtained from the Web and found to produce the same
1472 results as the earlier version of the code, or in some cases
1473 improved results. It is more tolerant of format variations found
1476 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1477 member, putting an enum integer_format in its place. New member
1478 `float_format'. Changed `compressed' member to type bool.
1480 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1482 * dictionary.c (dict_delete_consecutive_vars): New function.
1484 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1486 * file-name.c (fn_search_path): Remove prefix arg that was unused
1487 by any caller. Updated all callers.
1489 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1491 * format.c (fmt_dollar_template): Use user's decimal point
1492 character. Add assertion.
1494 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1496 * format.c (fmt_dollar_template): New function, based on
1497 dollar_format_template from var-type-dialog.c.
1499 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1501 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1504 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1506 Finish converting struct variable to an opaque type. In this
1507 phase, we add remaining setter and getter functions, convert the
1508 remaining PSPP code to use them, and do a bunch of cleanup. The
1509 resulting changes are pervasive but mostly trivial, and only the
1510 notable changes are logged.
1512 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1515 * case.c (case_data): Renamed case_data_idx.
1516 (case_num): Renamed case_num_idx.
1517 (case_str): Renamed case_str_idx.
1518 (case_data_rw): Renamed case_data_rw_idx.
1520 * case.h (case_data): New function with old name and an interface
1521 that takes a variable instead of an index, which is easier to
1522 use. Updated all callers to use the new interface, or to use the
1523 new *_idx function (see above).
1526 (case_data_rw): Ditto.
1528 * category.c (cat_stored_values_destroy): Changed interface to
1529 take a struct cat_vals * instead of a struct variable *.
1531 * dictionary.c (dict_clone): Use new vector_clone function.
1532 (dict_clear) Use new var_destroy function.
1533 (add_var) New function.
1534 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1535 (dict_create_var_assert) Rewrite in terms of add_var.
1536 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1537 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1538 (dict_lookup_var) Use new var_create, var_destroy functions.
1539 (dict_contains_var) Rewrite in terms of new vardict functionality.
1540 (set_var_dict_index) New function.
1541 (set_var_case_index) New function.
1542 (reindex_vars) New function.
1543 (dict_delete_var) Rewrite in terms of new vardict functionality.
1544 (dict_reorder_var) Ditto.
1545 (dict_reorder_vars) Ditto.
1546 (rename_var) New function.
1547 (dict_rename_var) Use rename_var.
1548 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1549 (dict_get_compacted_idx_to_fv) Rename
1550 dict_get_compacted_dict_index_to_case_index, update callers.
1551 (dict_create_vector) Use new vector_create function.
1552 (dict_clear_vectors) Use new vector_destroy function.
1553 (set_var_short_name_suffix) Move here from variable.c, renamed
1554 from var_set_short_name_suffix, make static, update caller.
1556 * sys-file-private.c: New file.
1557 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1558 width_to_bytes, update callers.
1560 * sys-file-private.h: New file. Later it will supplant
1561 sfm-private.h; for now it supplements it.
1562 (macro MIN_VERY_LONG_STRING) New macro.
1563 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1565 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1566 MAX_LONG_STRING was used before.
1568 * sys-file-writer.c: Ditto.
1570 * value-labels.c: Change the paradigm here to be that a null
1571 pointer is OK for a struct val_labs * in most cases; it just
1572 represents an empty set of value labels.
1573 (val_labs_copy) A copy of a null set is a null set.
1574 (val_labs_count) A null set has 0 labels.
1575 (val_labs_replace) Change return type to void. Rewrite for
1577 (val_labs_find) A null set does not contain the value.
1578 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1579 transposed argument order, updated all callers.
1581 * value.c: New file.
1582 (value_dup) Moved here from variable.c.
1583 (compare_values) Ditto.
1586 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1587 (macro MAX_LONG_STRING) Removed, because it was only interesting
1588 for system files, not for general code.
1589 (macro MAX_VERY_LONG_STRING) Ditto.
1590 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1591 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1593 * vardict.h: New file, for an interface between variables and
1596 * variable.c: A lot of functions were moved around, for better
1598 (struct variable) Move definition here, from variable.h.
1599 (var_type_adj) Removed--makes i18n hard.
1600 (var_type_noun) Ditto.
1601 (var_create) New function.
1602 (var_clone) New function.
1603 (var_destroy) New function.
1604 (var_set_name) Assert that variable is not in a dictionary.
1605 (compare_var_names) Rename compare_vars_by_name and fix a couple
1606 of callers who thought the args were strings.
1607 (hash_var_name) Rename hash_var_by_name.
1608 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1609 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1610 (var_is_very_long_string) Removed, because it was only interesting
1611 to system file code.
1612 (var_set_missing_values) Allow the argument to be the wrong width,
1613 as long as we can resize it. Simplify callers who were doing the
1614 resizing themselves.
1615 (var_get_value_labels) New function.
1616 (var_has_value_labels) New function.
1617 (var_set_value_labels) New function.
1618 (alloc_value_labels) New function.
1619 (var_add_value_label) New function.
1620 (var_replace_value_label) New function.
1621 (var_clear_value_labels) New function.
1622 (var_lookup_value_label) New function.
1623 (var_get_value_name) Moved here from variable.c, renamed from
1624 var_get_value_name, transposed argument order, updated all
1626 (var_to_string) Moved here, from variable-label.c.
1627 (var_set_leave) New function.
1628 (var_get_leave) New function.
1629 (var_must_leave) New function.
1630 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1631 set_var_short_name_suffix.
1632 (var_get_dict_index) New function.
1633 (var_get_case_index) New function.
1634 (var_get_obs_vals) New function.
1635 (var_set_obs_vals) New function.
1636 (var_has_obs_vals) New function.
1637 (var_get_vardict) New function.
1638 (var_set_vardict) New function.
1639 (var_has_vardict) New function.
1640 (var_clear_vardict) New function.
1641 (value_dup) Moved to value.c.
1642 (compare_values) Ditto.
1645 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1646 (enum ALPHA) Rename VAR_STRING, update all users.
1648 * vector.c: New file.
1649 (struct vector) Moved here, from variable.h.
1650 (check_widths) New function.
1651 (vector_create) New function.
1652 (vector_clone) New function.
1653 (vector_destroy) New function.
1654 (vector_get_name) New function.
1655 (vector_get_var) New function.
1656 (vector_get_var_cnt) New function.
1657 (compare_vector_ptrs_by_name) New function.
1659 * vector.h: New file.
1661 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1663 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1664 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1667 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1669 * value-labels.c (destroy_atoms): New function.
1670 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1672 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1674 Thanks to Jason Stover for pointing out this problem.
1676 * data-out.c (output_number): Use gsl_finite from GSL, which is
1677 portable, instead of isfinite, which is not.
1680 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1682 * variable.c variable.h (value_dup): New function.
1684 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1686 Start converting struct variable to an opaque type. In this
1687 phase, we add a bunch of setter and getter functions and convert
1688 most of the PSPP code to use them. The resulting changes are
1689 pervasive but mostly trivial, and only the notable changes are
1692 * format.c (fmt_equal): New function.
1694 * variable.c (var_type_is_valid): New function.
1695 (measure_is_valid) Moved here, from format.c.
1696 (alignment_is_valid) Moved here, from format.c.
1697 (var_get_name) New function.
1698 (var_set_name) New function.
1699 (width_to_type) New function.
1700 (var_get_type) New function.
1701 (var_get_width) New function.
1702 (var_set_width) New function.
1703 (var_is_numeric) New function.
1704 (var_is_alpha) New function.
1705 (var_is_short_string) New function.
1706 (var_is_long_string) New function.
1707 (var_is_very_long_string) New function.
1708 (var_get_missing_values) New function.
1709 (var_set_missing_values) New function.
1710 (var_clear_missing_values) New function.
1711 (var_has_missing_values) New function.
1712 (var_is_value_missing) New function.
1713 (var_is_num_missing) New function.
1714 (var_is_str_missing) New function.
1715 (var_is_value_user_missing) New function.
1716 (var_is_num_user_missing) New function.
1717 (var_is_str_user_missing) New function.
1718 (var_is_value_system_missing) New function.
1719 (var_get_print_format) New function.
1720 (var_set_print_format) New function.
1721 (var_get_write_format) New function.
1722 (var_set_write_format) New function.
1723 (var_set_both_formats) New function.
1724 (var_get_label) New function.
1725 (var_set_label) New function.
1726 (var_clear_label) New function.
1727 (var_has_label) New function.
1728 (var_get_measure) New function.
1729 (var_set_measure) New function.
1730 (var_get_display_width) New function.
1731 (var_set_display_width) New function.
1732 (var_get_alignment) New function.
1733 (var_set_alignment) New function.
1734 (var_get_value_cnt) New function.
1735 (var_get_leave) New function.
1736 (var_get_short_name) New function.
1738 * variable.h: (struct variable) Removed "type" and "nv" members;
1739 they are now computed from "width" where needed.
1741 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1743 * missing-values.c (mv_resize): Don't write beyond end of the
1744 allocated buffer when resizing a long string.
1746 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1748 Clean up identifier code: don't require identifier enumerations to
1749 be in a particular order; make better use of string library;
1750 expose less of the internals.
1752 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1753 change interface. Updated all callers.
1754 (lex_id_match) Change interface to use struct substring, update
1756 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1757 (global array keywords[]) Make static, change form. Update all
1758 users to use lex_id_name instead.
1759 (lex_is_keyword) New function.
1760 (lex_id_to_token) Change interface to use struct substring, update
1762 (lex_id_name) New function.
1764 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1765 lex_is_keyword instead.
1766 (T_LAST_KEYWORD) Removed.
1767 (T_N_KEYWORDS) Removed.
1769 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1771 * format.c: (fmt_date_template) Distinguish characters for which a
1772 space is output and any date delimiter is allowed on input, from
1773 those for which a space is output and only a space is allowed on
1774 input. The former is represented by X, the latter by a space.
1775 Also, drop distinction between h and H, changing the former to the
1778 * data-in.c: Completely rewrite internals to conform to SPSS input
1779 formats as closely as possible.
1780 (data_in) Changed external interface by replacing the structure
1781 that was used as a single argument by a set of arguments. Updated
1783 (data_in_finite_line) Removed. Converted all callers to use plain
1785 (data_in_get_integer_format) New function.
1786 (data_in_set_integer_format) New function.
1787 (data_in_get_float_format) New function.
1788 (data_in_set_float_format) New function.
1790 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1791 (struct data_in) Removed.
1793 * data-out.c: (output_date) Drop each component from the input as
1794 it is output, to allow us to drop the distinction between h (a
1795 count of hours) and H (the hour of day) template characters.
1796 Also, handle new X template character.
1797 (output_scientific) Follow more rational rule on when to drop
1798 fraction introduced between SPSS 13 and 15. Updated test case to
1801 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1803 Fix buffer overflow reported by John Darrington.
1805 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1806 realize that DIGITS is a count of nibbles, not of bytes.
1808 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1810 * calendar.c (calendar_offset_to_gregorian) Also return the
1811 year-of-day. Change callers to new interface.
1813 * data-out.c: Completely rewrite internals to conform to SPSS
1814 output formats as completely as possible.
1815 (data_out) Change interface to put input parameters before output
1816 parameters, for consistency with the style I now prefer. Update
1818 (data_out_get_integer_format) New public function.
1819 (data_out_set_integer_format) New public function.
1820 (data_out_get_float_format) New public function.
1821 (data_out_set_float_format) New public function.
1823 * data-out.h: New file. Move prototype for data_out here, from
1826 * format.c: (fmt_step_width) Use equality comparison instead of
1827 bitwise and, for clarity.
1828 (fmt_is_string) Ditto.
1829 (fmt_input_to_output) Fix categories that are translated to F
1832 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1834 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1835 casefile-private.h: Added casefilter to assist commands with missing
1838 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1840 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1842 * settings.c: (route_errors_to_terminal) New variable.
1843 (route_errors_to_listing) New variable.
1844 (get_error_routing_to_terminal) New function.
1845 (set_error_routing_to_terminal) New function.
1846 (get_error_routing_to_listing) New function.
1847 (set_error_routing_to_listing) New function.
1849 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1851 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1853 * format.c: Completely rewrite, to achieve better abstraction.
1854 Rewrite all references to formats in other files.
1856 * format.def: Rewrite and reorganize.
1858 * settings.c: Move everything related to custom currency formats
1859 into format.[ch], changing them in form, so as to group related
1860 code and definitions better. Changed all references to use the
1862 (static var decimal) Removed.
1863 (static var grouping) Removed.
1864 (static var cc) Removed.
1865 (get_decimal) Removed.
1866 (set_decimal) Removed.
1867 (get_grouping) Removed.
1868 (set_grouping) Removed.
1872 * settings.h: (macro CC_CNT) Removed.
1873 (macro CC_WIDTH) Removed.
1874 (struct custom_currency) Removed.
1876 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1878 * data-in.c (data_in): Use switch statement instead of table, to
1879 avoid dependence on the order of the FMT_* enums.
1881 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1883 * data-out.c: (num_to_string) Removed, because it was dead code.
1885 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1887 * data-in.c (parse_trailer): Fix error message.
1889 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1891 * format.c (fmt_is_binary): New function.
1893 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1895 * procedure.c procedure.h: Encapsulated the static data into a single
1898 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1900 * casefile.c (casereader_read_xfer): Always initialize the case,
1901 even on an error condition.
1903 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1905 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1906 entire function useless.
1908 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1910 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1911 casereader method casereader_clone.
1913 * procedure.c pransformations.h: Introduced new type casenum_t
1915 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1917 * variable.c: (width_to_bytes) Rephrase code for clarify.
1919 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1921 * format.c: (fmt_type_from_string) New function.
1922 (fmt_to_string) Include decimals in output if the format has
1923 decimals, even if the format type does not. This way, we can
1924 accurately reproduce incorrect formats in user output.
1925 (check_common_specifier) Make the check for a bad format type an
1926 assertion, so we get bug reports if they show up. Fix message.
1927 Check for decimal places with a format type that doesn't allow
1929 (check_input_specifier) Remove check for FMT_X, which has been
1931 (check_output_specifier) Ditto.
1933 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1935 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1936 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1937 (enum fmt_parse_flags) Removed.
1939 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1941 * casefile.c casefile.h: Converted to an abstract base class.
1942 * casefile-private.h fastfile.c fastfile.h: New files.
1943 * automake.mk procedure.c scratch-writer.c storage-stream.c
1945 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1947 * procedure.c (internal_procedure): Create sink_case with only as
1948 many values as the compacted dictionary.
1950 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1952 Remove "debugging" code that caused plenty of false positives and
1955 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1957 * case.c: Remove all references to `this' member.
1959 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1961 Fix link error noted by Jason Stover.
1963 * storage-stream.c: Include <assert.h>.
1965 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1967 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1968 ALL) and additional underlying system file issues.
1970 Thanks to John Darrington for review.
1972 First problem: var_hash points to variables not owned by the
1973 sys-file-reader, which the caller may free or modify. Use an
1974 array of sfm_vars instead, as done earlier (e.g. CVS version
1977 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1978 members and remove all code that references it. Add vars, var_cnt
1979 members. Remove fix_specials member, which was unused.
1980 (struct sfm_var) Remove name member, which was unused.
1981 (sfm_close_reader) Free vars member instead of var_hash.
1982 (compare_var_shortnames) Removed.
1983 (hash_var_shortname) Removed.
1984 (sfm_open_reader) Fill out vars array.
1985 (compare_var_index) Removed.
1986 (sfm_read_case) Use vars instead of var_hash.
1988 Second problem: we're confused about when we actually have very
1989 long strings, causing us to choose incorrectly between slow path
1990 and fast path in sfm_read_case.
1992 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1993 have very long strings, not when we have long variable names,
1994 which is an unrelated feature.
1996 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1998 * variable.h: Move var_set and variable parsing declarations to
1999 new header, src/language/lexer/variable-parser.h. Modified lots
2000 of files to include the new header.
2002 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
2004 * value-labels.c (value_to_string): When there's no value label,
2005 format the variable according to its print format, instead of
2006 always effectively using A or F format.
2008 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
2010 * casefile.c (casefile_get_random_reader): Nasty hack to get around
2013 * format.c: Removed tortological assertion.
2015 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
2017 Reform string library.
2019 * file-name.c (fn_interp_vars): Change interface to take a
2020 substring as input. Updated all users.
2022 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
2024 * format.c (measure_is_valid): Really return false when m >=
2027 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
2029 Implement random access to casefiles, for use in GUI.
2031 * casefile.c: (struct casereader) Add `random', `file_ofs',
2032 `buffer_ofs' members.
2033 (casefile_get_random_reader) New function.
2034 (read_open_file) Break part into new function
2035 seek_and_fill_buffer().
2036 (fill_buffer) Update buffer_ofs, file_ofs.
2037 (casereader_seek) New function.
2039 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2041 * settings.c: Added call to i18n{done, init}.
2043 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
2045 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
2048 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
2050 * casefile.c: Convert many uses of `int' to `bool'.
2052 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
2054 * transformations.c (trns_chain_destroy): Destroy chain's trns
2055 member, to fix memory leak.
2057 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
2059 * storage-stream.c (storage_source_decapsulate): Destroy case
2060 source to fix memory leak.
2062 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
2064 * scratch-reader.c (scratch_reader_read_case): Copy into existing
2065 case passed as argument instead of initializing the argument as a
2066 case. Fixes memory leak that showed up in
2067 tests/command/aggregate.sh with scratch files.
2069 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
2071 * procedure.c (proc_done): Destroy default_dict, to fix memory
2074 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2076 Simplify procedure_with_splits().
2078 * procedure.c (struct split_aux_data): Removed case_count member.
2079 (procedure_with_splits) Don't initialize case_count.
2080 (split_procedure_case_func) Check whether prev_case is null
2081 instead of case_count.
2082 (split_procedure_end_func) Ditto.
2084 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2086 * case.c (case_move): Do nothing if dst and src are the same
2088 (case_try_create) Merge two similar cases.
2089 (case_copy) Unshare only if data must be actually copied.
2091 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2093 * data-in.c data-out.c dictionary.c sys-file-reader.c
2094 sys-file-writer.c variable.c variable.h: Reworked very long string
2095 support for better encapsulation.
2097 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2099 * value-labels.c (val_labs_can_set_width): New function.
2100 (val_labs_set_width) Clear labels if increasing width to long
2102 (val_labs_destroy) Remove unneeded test for null.
2104 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2106 * value-labels.h: Remove unneeded dependency on variable.h.
2108 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2110 Get rid of `char *c' member in union value, for cleanliness.
2112 * value.h: (union value) Remove `c' member.
2114 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2116 Make dictionary compacting functions a little more general.
2118 * sys-file-writer.c (sfm_open_writer): Use
2119 dict_compacting_would_change().
2120 (does_dict_need_translation) Removed.
2122 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2124 Make dictionary compacting functions a little more general.
2126 * dictionary.c (dict_needs_compaction): Rename
2127 dict_compacting_would_shrink(). Update all callers.
2128 (dict_compacting_would_change) New function.
2130 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2132 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2133 inverted return value (!).
2135 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2137 Continue reforming procedure execution.
2139 * procedure.c: Search and replace "vfm" by "proc". Notably:
2140 (static var vfm_source) Rename proc_source. Update all
2142 (static var vfm_sink) Rename proc_sink. Update all references.
2144 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2146 Continue reforming procedure execution. In this phase, remove
2147 PROCESS IF, which was deprecated anyway and can be easily
2148 simulated with TEMPORARY followed by SELECT IF.
2150 * procedure.c: (open_active_file) Don't call
2151 add_process_if_trns().
2152 (discard_variables) Get rid of redundant call to
2153 proc_cancel_all_transformations().
2154 (add_process_if_trns) Removed.
2155 (process_if_trns_proc) Removed.
2156 (process_if_trns_free) Removed.
2158 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2160 Continue reforming procedure execution. In this phase, add
2161 `const' to the case passed to procedure()'s callback.
2163 Updated all users of procedure() as well.
2165 * procedure.c: (struct write_case_data) Add "const" to ccase
2166 parameter for case_func member.
2167 (procedure) Add "const" to ccase parameter for proc_func
2169 (multipass_case_func) Make ccase parameter const.
2170 (internal_procedure) Add "const" to ccase parameter for case_func
2172 (split_procedure_case_func) Make ccase parameter const.
2173 (multipass_split_case_func) Make ccase parameter const.
2175 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2177 Continue reforming procedure execution. In this phase, get rid of
2178 the output code for SPLIT FILE groups in procedure.c, which really
2179 shouldn't be doing any output. Move it into the individual
2180 procedures instead. This also adds some flexibility.
2182 Updated many users of procedure_with_splits() and
2183 multipass_procedure_with_splits() to call
2184 output_split_file_values() and to deal with increased use of
2187 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2188 parameter to begin_func member.
2189 (procedure_with_splits) Add "const struct ccase *" parameter to
2190 begin_func parameter. Make ccase parameter const in proc_func
2192 (split_procedure_case_func) Don't dump split file group. Pass
2194 (dump_splits) Moved to language/dictionary/split-file.c as
2195 output_split_file_values().
2196 (struct multipass_split_aux_data) Add "const struct ccase *"
2197 parameter to split_func member.
2198 (multipass_procedure_with_splits) Add "const struct ccase *"
2199 parameter to split_func parameter.
2200 (multipass_split_case_func) Save new SPLIT FILE case before
2202 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2204 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2206 Continue reforming procedure execution. Change
2207 internal_procedure() so that it calls open_active_file() and
2208 close_active_file(), which isolates most of the actual procedure
2211 * procedure.c: (struct write_case_data) Rename `proc_func' member
2212 to `case_func' and update all references.
2213 (procedure) Rewrite as one-line wrapper around
2214 internal_procedure().
2215 (struct multipass_aux_data) New.
2216 (multipass_callback) Renamed multipass_case_func(). Use struct
2217 multipass_aux_data as auxiliary data.
2218 (multipass_end_func) New function.
2219 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2220 that uses multipass_case_func, multipass_end_func.
2221 (internal_procedure) Add `end_func' argument. Move optimization
2222 of trivial case in here. Move call to open_active_file() and
2223 close_active_file() in here. Now assert that vfm_source is
2225 (procedure_with_splits_callback) Rename
2226 split_procedure_case_func().
2227 (split_procedure_end_func) New function.
2228 (multipass_split_callback) Rename multipass_split_case_func.
2229 (multipass_split_end_func) New function.
2230 (discard_variables) No need to test for nonnull vfm_source.
2232 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2234 Continue reforming procedure execution. Get rid of unused member.
2236 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2238 (write_case) Don't increment cases_analyzed.
2240 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2242 Continue reforming procedure execution. In this phase, move
2243 procedure.c and procedure.h from src to src/data. Update
2244 makefiles and #includes accordingly.
2246 * procedure.c: Moved here from src/.
2248 * procedure.h: Moved here from src/.
2250 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2252 Continue reforming procedure execution. In this phase, get rid of
2253 many global variables, consolidating procedure execution in
2254 procedure.c. Encapsulate transformations in new "struct
2255 trns_chain". Also, change implementation of N OF CASES, FILTER,
2256 and PROCESS IF from special cases to transformations.
2258 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2261 * dictionary.c: (global variable default_dict) Move to
2264 * variable.h: (TRNS_*) Move to transformations.h.
2265 (struct transformation) Move to transformations.c.
2267 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2269 * sys-file-reader.c: Fixed invalid read problems.
2271 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2273 * storage-stream.c: Add missing function comments.
2275 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2277 Continue reforming procedure execution. In this phase, add some
2278 new, needed functionality to storage-stream.
2280 * storage-stream.c: (storage_source_decapsulate) New function.
2282 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2284 * variable.c (width_to_bytes): Declarations must precede
2285 statements for C90 compliance.
2287 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2289 * data-out.c, data-in.c, variable.c, variable.h: New functions
2290 copy_mangle and copy_demangle for reading/writing cases; emulates the
2291 way SPSS deals with strings > 255 bytes.
2293 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2294 subtype 14 needed for strings longer than 255 bytes.
2296 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2297 instead of literal values. Also fixed some constness issues.
2299 * format.h: Constness
2301 * sfm-private.h: Renamed the case_size identifier, since I discovered
2302 that SPSS's respect for this variable is very nominal.
2304 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2306 Change case limit type from int to size_t.
2308 * dictionary.c: (struct dictionary) Change type of case_limit
2310 (dict_get_case_limit) Change return type.
2311 (dict_set_case_limit) Change parameter type.
2313 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2315 * variable.h: (struct variable) Rename `reinit' member as `leave'
2316 and invert sense. Fix up all references.
2318 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2320 Continue reforming procedure execution. In this phase, break
2321 procedure.c into multiple files.
2323 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2325 * case-sink.c: New file.
2327 * case-sink.h: New file.
2329 * case-source.c: New file.
2331 * case-source.h: New file.
2333 * storage-stream.c: New file.
2335 * storage-stream.h: New file.
2337 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2339 * variable.h: (struct variable) Remove `init' member and all
2340 references to it from other files. It was initialized in several
2341 places, but nothing really ever used it for anything worthwhile.
2342 Thanks to Jason Stover for pointing out how confusing this
2345 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2347 Continue reforming error message support. In this phase, get rid
2348 of message "titles" and put the message text in `struct error'.
2349 Now `struct error' encapsulates a message more properly.
2351 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2352 Format message ourselves.
2354 * data-in.c: (vdls_error) Ditto.
2356 * por-file-reader.c: (error) Ditto.
2358 * sys-file-reader.c: (corrupt_msg) Ditto.
2360 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2362 GNU standards require "file name" instead of "filename" in
2363 documentation. It's nice for our code to follow the convention
2366 * casefile.c: (struct casefile) Rename `filename' member to
2367 `file_name'. Updated all references.
2369 * file-name.c: [!unix] (struct file_identity) Rename
2370 normalized_filename member to normalized_file_name. Updated all
2373 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2375 We don't really support anything but Unix-like environments well,
2376 so we might as well de-obfuscate by writing directory and path
2377 separators explicitly.
2379 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2381 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2382 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2384 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2387 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2389 GNU standards require "file name" instead of "filename" in
2390 documentation. It's nice for our code to follow the convention
2393 * filename.c: Rename to file-name.c.
2395 * filename.h: Rename to file-name.h. Update all inclusions.
2396 Update header guards.
2398 * automake.mk: Update file names.
2400 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2402 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2404 (fn_basename) Removed (dead code).
2405 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2406 (fn_special_p) Renamed fn_is_special(), all references updated.
2407 (fn_exists_p) Renamed fn_exists(), all references updated.
2409 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2411 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2412 Also, now it only tilde-expands file names, not paths.
2413 (fn_search_path) Tilde-expand one directory at a time.
2415 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2417 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2418 get rid of non-Unixlike version of the code, which has probably
2420 (fn_prepend_dir) Removed (dead code).
2422 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2423 (macro PATH_SEPARATOR_STRING) New.
2424 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2426 Continue reforming error message support. In this phase, we get
2427 rid of VM() and the other msg() support for "verbosity", replacing
2428 it by a new function verbose_msg().
2430 * filename.c: (fn_search_path) Use verbose_msg() instead of
2433 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2435 * sfm-private.h: Get rid of #defines after #error, which makes no
2438 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2440 Get rid of our own int32 type in favor of the standard int32_t
2443 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2446 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2448 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2450 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2452 Remove ill-considered file routines that are no longer used.
2454 * filename.c: (fn_open_ext) Removed.
2455 (fn_close_ext) Removed.
2457 * filename.h: (struct file_ext) Removed.
2459 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2461 * variable.c (var_is_valid_name): Move declarations before code
2464 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2466 * filename.ch (fn_interp_vars): Fixed small buglet.
2468 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2470 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2471 so as to modify the string inline. Thus makeing it easier to
2472 destroy the results when no longer needed.
2474 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2476 * category.c (cat_stored_values_destroy): Fixed memory leak.
2478 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2480 Add some missing frees. Thanks to John Darrington for reporting
2483 * any-writer.c (any_writer_close): Free writer.
2485 * any-reader.c (any_reader_close): Free reader.
2487 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2489 * por-file-reader.c: (error) Mark as NO_RETURN.
2491 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2493 * settings.c: Changed default value of scompress to true.
2495 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2497 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2499 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2502 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2503 here from pref.h.orig.
2505 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2507 * sys-file-reader.c: Fixed bug reading compressed files.
2509 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2511 * Numerous renames. See src/ChangeLog for details.
2513 * Moved files from src directory