1 2008-02-18 Ben Pfaff <blp@gnu.org>
3 Patch #6426. Reviewed by John Darrington.
5 * format.c (min_width): Renamed fmt_min_width and made public.
6 Updated all references.
7 (max_width): Renamed fmt_max_width and made public. Updated all
9 (max_decimals): Renamed fmt_max_decimals and made public. Updated
12 * format.h (macro FMT_MAX_NUMERIC_WIDTH): New macro.
14 2008-02-09 Ben Pfaff <blp@gnu.org>
16 Add a couple of extensions to GET DATA TYPE=TXT. Patch #6412.
17 Thanks to John Darrington for review.
19 * data-in.c (data_in): Add new argument to designate the last
20 column of the data field being parsed, for use in error messages.
23 * data-parser (struct data_parser): New member `quote_escape'.
24 (data_parser_create): Initialize quote_escape.
25 (data_parser_set_quotes): New function.
26 (cut_field): Support escaped quotes.
27 (parse_delimited_span): Ditto.
28 (parse_delimited_no_span): Ditto.
30 * get-data.c (parse_get_txt): Support ESCAPE extension subcommand
31 in enhanced mode. Only support multiple quote characters in
34 2008-02-06 John Darrington <john@darrington.wattle.id.au>
36 psql-reader.c psql-reader.h: Read more than one tuple at
37 once. Fix bug reading a query which returns no data. Fix bug
38 when transformation followed a reader.
39 Ask the server for the number of records in the query, for the
42 2008-02-05 John Darrington <john@darrington.wattle.id.au>
44 psql-reader.c: So yesterday they release postgresql 8.3.0
45 which has money represented with 64 bits. They must get
48 2008-02-02 John Darrington <john@darrington.wattle.id.au>
50 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
51 for reviewing this code.
53 2008-02-02 Ben Pfaff <blp@gnu.org>
57 * sys-file-reader.c (read_variable_record): Allow missing values
58 to be specified on long string variables, but warn about them
59 because PSPP does not yet support them.
60 (read_extension_record): Ignore extension records 20 and 21, which
61 PSPP does not yet support.
62 (read_header): Fix error message when floating-point format cannot
65 2008-02-01 Ben Pfaff <blp@gnu.org>
67 Patch #6386. Thanks to John Darrington for review and for the
68 updates to gnumeric-reader.c.
70 * dictionary.c (make_hinted_name): New function.
71 (make_numeric_name): New function.
72 (dict_make_unique_var_name): New function.
74 * gnumeric-reader.c (devise_name): Removed.
75 (munge_name): Removed.
76 (gnumeric_open_reader): Use new function
77 dict_make_unique_var_name.
79 * short-names.c (set_var_short_name_suffix): Use new function
82 2008-01-19 John Darrington <john@darrington.wattle.id.au>
84 * settings.c settings.h: Moved static variables into a
85 single struct. Renamed functions to have a settings_ prefix.
87 2008-01-14 John Darrington <john@darrington.wattle.id.au>
89 * data-out.c (output_AHEX): Corrected number of bytes in
90 call to output_hex. Closes bug #22011
92 2008-01-02 John Darrington <john@darrington.wattle.id.au>
94 * variable.c variable.h: Replaced var_get_value_name with
95 var_append_value_name which doesn't use any static data.
96 Thanks to Ben for review.
98 2007-12-07 Ben Pfaff <blp@gnu.org>
102 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
103 made from an empty casereader produce a casegrouper with no
104 groups, instead of one with one group that has no cases.
106 * casereader.c (casereader_is_empty): New function.
108 2007-12-06 Ben Pfaff <blp@gnu.org>
112 * sys-file-reader.c (read_display_parameters): Handle variable
113 display parameters record with only 2 data items per variable.
114 Reported by Guido Gay <gay@irer.it>.
116 2007-12-04 Ben Pfaff <blp@gnu.org>
118 * identifier.c (lex_id_match_n): New function.
119 (lex_id_match): Reimplement in terms of lex_id_match_n.
121 2007-11-24 Ben Pfaff <blp@gnu.org>
123 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
126 2007-11-24 Ben Pfaff <blp@gnu.org>
128 Fix warning reported when reading back system files that include
129 very long string variables. Thanks to Guido Gay <gay@irer.it> for
132 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
133 was written where `j' was meant.
135 2007-11-18 Ben Pfaff <blp@gnu.org>
137 Properly write variables that include a range of missing values to
138 system files. Thanks to Guido Gay <gay@irer.it> for reporting
141 * sys-file-writer.c (write_variable): Correctly calculate
142 n_missing_values field when writing variables that include a range
145 2007-11-10 Ben Pfaff <blp@gnu.org>
147 Cleanups and bug fixes devised while writing up documentation.
150 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
152 * dict-class.c: New file.
153 (dict_class_from_id): Move here.
154 (dict_class_to_name): Move here.
156 * dict-class.h: New file.
157 (enum dict_class): Move here. Change from consecutive integers to
158 consecutive bits, to make testing for any of multiple values
159 easier. Add new DC_ALL constant.
161 * dictionary.c (struct dictionary): Change `case_limit' from
162 size_t to casenumber.
163 (dict_get_vars): Make final argument an enum dict_class.
164 (dict_get_vars_mutable): Ditto.
165 (dict_get_case_limit): Change return value to casenumber.
166 (dict_set_case_limit): Change final argument to a casenumber.
167 (dict_unset_split_var): Add assertion. Rephrase slightly.
168 (dict_set_label): Use xstrndup to simplify.
170 * format.c (fmt_step_width): AHEX format also needs 2-byte
172 (fmt_set_style): Simplify assertions.
174 * missing-values.c (mv_add_num_range): Rename mv_add_range.
175 Simplify implementation.
176 (mv_has_value): Simplify implementation.
177 (mv_pop_value): Remove the first value, not the last, to avoid
178 having GET followed by SAVE reverse the order of missing values.
179 (mv_peek_value): Rename mv_get_value. Simplify assertion.
180 (mv_has_range): Simplify implementation.
181 (mv_peek_range): Rename mv_get_range.
182 (can_resize_string): Removed.
183 (mv_is_resizable): Use value_is_resizable.
184 (mv_resize): Use value_resize.
186 * short-names.h (SHORT_NAME_LEN): Move here.
188 * val-type.h: New file, for definitions related to type and width
189 of abstract values. Before, these definitions were mixed among
190 those related to "union value" and those related to variables.
191 (macro SYSMIS): Move here.
192 (macro LOWEST): Move here.
193 (macro HIGHEST): Move here.
194 (macro MAX_STRING); Move here.
195 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
196 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
197 All references updated.
198 (val_type_is_valid): New function. Replaces var_type_is_valid.
199 All references updated.
200 (val_type_from_width): New function. Replaces
201 var_type_from_width. All references updated.
203 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
205 (val_labs_can_set_width): Use value_is_resizable.
206 (val_labs_add): Simply return false if the value labels set is too
207 wide, instead of having undefined behavior.
208 (val_labs_replace): Ditto.
209 (val_labs_replace): Ditto.
210 (val_labs_first): Set iterator to null if iteration is complete.
211 (val_labs_first_sorted): Ditto.
212 (val_labs_done): Become a no-op if the iterator is null.
214 * value.c (value_is_resizable): New function.
215 (value_resize): New function.
217 * variable.c (var_get_dict_class): New function.
219 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
222 2007-11-08 Ben Pfaff <blp@gnu.org>
224 * data-in.c: Make formatted data parsing locale-independent.
225 (parse_number): Use c_strtod instead of strtod, to avoid
226 locale-specific behavior.
229 2007-11-06 Ben Pfaff <blp@gnu.org>
231 Patch #6256: add support for binary, 360 file formats. Reviewed
234 * data-in.c (struct data_in): Add `encoding' member.
235 (data_in): Add `encoding' parameter, and re-encode the data passed
236 in where appropriate. Update all callers to pass it in.
237 (parse_A): Implement EBCDIC recoding wart described in manual.
238 (parse_AHEX): Implement EBCDIC recoding.
240 * data-out.c (data_out_legacy): New function.
241 (data_out): Make into a wrapper around data_out_legacy.
243 * file-handle-def.c (struct file_handle): New member `encoding'.
244 (fh_create_file): Set encoding.
245 (fh_default_properties): Set default encoding.
246 (fh_get_legacy_encoding): New function.
248 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
249 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
250 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
251 (struct fh_properties): New member `encoding'.
253 2007-11-05 Ben Pfaff <blp@gnu.org>
255 Patch #6258. Reviewed by John Darrington.
257 * file-handle-def.c (fh_lock): Add comment that TYPE should be
258 marked with N_() in the caller. Added these markings to each
259 caller too. Should make i18n easier.
260 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
262 2007-11-03 Ben Pfaff <blp@gnu.org>
264 Allow output files to overwrite input files (bug #21280).
266 * file-handle-def.c: Separate locking of files for input and for
267 output, to allow both to take place at once. Also, distinguish a
268 file handle from the identity of the underlying file, because the
269 identity of a file changes over time and the file handle can't
270 represent two different identities.
271 (struct file_handle): Remove `next', `open_cnt', `deleted',
272 `type', `open_mode', `aux', `identity' members. Change `id' from
273 char array to char *. Add `ref_cnt' member.
274 (file_handle_from_ll) New function.
275 (file_handles) Removed.
276 (named_handles) New variable.
277 (fh_init) Initialize named_handles.
278 (fh_done) Remove name from all named_handles.
279 (free_handle) Updated for modified struct file_handle.
280 (unname_handle) New function.
281 (fh_ref) New function.
282 (fh_from_file_name) Removed.
283 (fh_unref) New function.
284 (fh_unname) New function.
285 (fh_from_id) Rewritten.
286 (create_handle) Updated for modified struct file_handle.
287 (fh_inline_file) Increment the handle's ref count.
288 (fh_create_file) Updated for modified struct file_handle.
289 (fh_create_scratch) Ditto.
294 (fh_is_open) Removed.
295 (fh_get_id) Updated for modified struct file_handle.
296 (fh_get_default_handle) Increment the handle's ref count.
297 (fh_set_default_handle) Handle ref counts.
298 (struct fh_lock) New structure.
299 (locks) New static var.
300 (fh_lock) New function.
301 (fh_unlock) New function.
302 (fh_lock_get_aux) New function.
303 (fh_lock_set_aux) New function.
304 (fh_is_locked) New function.
305 (make_key) New function.
306 (free_key) New function.
307 (compare_fh_locks) New function.
308 (hash_fh_lock) New function.
310 * file-handle-def.h (enum fh_access) New enum.
312 * file-name.c: Made file_identity the same in all supported
314 (struct file_identity): New `name' member.
315 (fn_get_identity): For a file that doesn't exist, get the
316 dev/inode of its directory plus its name. If even the directory
317 doesn't exist, just use its name. Merge the Windows
318 implementation into the Unix one.
319 (fn_compare_file_identities): Rewritten. Merge the Windows
320 implementation into the Unix one.
321 (fn_hash_identity): New function.
323 * make-file.c (struct replace_file): New structure.
324 (all_files): New static var.
325 (replace_file_start): New function.
326 (replace_file_commit): New function.
327 (replace_file_abort): New function.
328 (free_replace_file): New function.
329 (unlink_replace_files): New function.
331 * por-file-reader.c (struct pfm_reader): Add `lock' member.
332 (close_reader): Unlock file.
333 (pfm_open_reader): Lock file.
335 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
337 (pfm_open_writer): Lock file and prepare for its replacement.
338 (close_writer): Unlock file.
340 * scratch-handle.h (struct scratch_handle): Add unique_id so that
341 different generations of a scratch file can be distinguished.
343 * scratch-reader.c (scratch_reader_open): Verify that the file is
346 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
347 remove scratch_handle member.
348 (scratch_writer_open): Lock handle. Prepare to replace handle
349 data, instead of doing it immediately.
350 (scratch_writer_casewriter_destroy): Replace handle data and
353 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
354 (sfm_open_reader): Lock file.
355 (close_reader): Unlock file.
357 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
359 (sfm_open_writer): Lock file and prepare for its replacement.
360 (close_writer): Unlock file.
362 2007-11-02 Ben Pfaff <blp@gnu.org>
364 * data-out.c (output_number): Use isfinite (from C99) instead of
365 gsl_isfinite, since we now have universal support for it in
367 (output_infinite): Ditto.
368 [!HAVE_ROUND] (round): Remove definition, since we now have a
369 replacement in gnulib.
371 2007-10-27 John Darrington <john@darrington.wattle.id.au>
373 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
374 wasn't initialised correctly. Closes bugs #21330 and 21397
377 2007-10-08 Ben Pfaff <blp@gnu.org>
379 Fix bug #21280. Thanks to John Darrington for review.
381 * file-name.c (create_stream): New function.
383 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
384 file handle before creating the file, to ensure that we don't
385 truncate a file that we're reading. Make code easier to read by
388 * sys-file-write.c (sfm_open_writer): Ditto.
390 2007-10-01 Ben Pfaff <blp@gnu.org>
392 Fix bug #21192. Thanks to John Darrington for review.
394 * casereader.c (casereader_read): Decrement case_cnt before
395 calling the casereader's "read" member function, so that we
396 interact properly with lazy_casereader.
398 * datasheet.c: Add regression test for above bug fix.
399 (clone_datasheet): New function.
400 (lazy_callback): New function.
401 (check_datasheet_casereader): New function.
402 (check_datasheet): Check datasheet contents are reported correctly
403 through an ordinary casereader and a lazy casereader.
404 (clone_model): Use clone_datasheet.
406 2007-09-24 Ben Pfaff <blp@gnu.org>
408 Patch #6210. Reviewed by John Darrington.
409 * settings.c: Drop "long view", which is not needed any longer.
410 (static var long_view): Removed.
411 (force_long_view): Removed.
412 (get_viewwidth): Removed.
413 (init_viewport): Removed long_view reference.
415 2007-09-19 John Darrington <john@darrington.wattle.id.au>
417 * settings.c settings.h: Changed viewport's length and width to be
418 owned by the user interface which uses the data library. This allows
419 better abstraction, and makes dynamically adjustable dimensions easier.
421 2007-09-18 Ben Pfaff <blp@gnu.org>
423 * procedure.c (proc_extract_active_file_data): New function.
425 * lazy-casereader.h: New file.
427 * lazy-casereader.c: New file.
429 * casereader.c (casereader_dynamic_cast): New function.
431 2007-09-14 Ben Pfaff <blp@gnu.org>
433 * dictionary.c (dict_clone): Copy case indexes from cloned
434 dictionary. Fixes bug #21061. Reviewed by John Darrington.
436 2007-09-13 John Darrington <john@darrington.wattle.id.au>
438 * value.c value.h (create_value): New function.
440 2007-09-12 Ben Pfaff <blp@gnu.org>
442 Make it clear that translator casereader and casewriter translate
443 functions are supposed to destroy their input case.
445 * casereader-translator.c (struct casereader_translator): Change
446 input case parameter of translate member function from const
447 struct ccase * to struct ccase *.
448 (casereader_create_translator): Ditto, for translate parameter.
450 * casewriter-translator.c (struct casewriter_translator): Ditto.
451 (casewriter_create_translator): Ditto.
453 2007-08-27 John Darrington <john@darrington.wattle.id.au>
455 * sys-file-reader.c (read_display_parameters): Force display width
456 to 8, if the sys file says 0 (like SPSS does).
458 2007-08-12 Ben Pfaff <blp@gnu.org>
460 * dictionary.c (dict_dump): New function.
462 2007-08-12 Ben Pfaff <blp@gnu.org>
464 Drop dict_compactor in favor of using the new struct case_map.
466 * dictionary.c (struct copy_map): Removed.
467 (struct dict_compactor): Removed.
468 (dict_make_compactor): Removed.
469 (dict_compactor_compact): Removed.
470 (dict_compactor_destroy): Removed.
472 * procedure.c (struct data_set): Change `compactor' member to be a
474 (proc_open): Use case_map_to_compact_dict instead of
476 (proc_casereader_read): Use case_map_execute instead of
477 dict_compactor_compact.
478 (proc_commit): Use case_map_destroy instead of
479 dict_compactor_destroy.
481 * scratch-writer.c (struct scratch_writer): Change `compactor'
482 member to be a struct case_map *.
483 (scratch_writer_open): Use case_map_to_compact_dict instead of
485 (scratch_writer_casewriter_write): Use case_map_execute instead of
486 dict_compactor_compact.
488 2007-08-12 Ben Pfaff <blp@gnu.org>
490 * automake.mk: Add case-map.c, case-map.h.
492 * case-map.c: New file.
494 * case-map.h: New file.
496 2007-08-12 Ben Pfaff <blp@gnu.org>
498 * dictionary.c (dict_compact_values): Don't delete scratch
499 variables as well as compacting case indexes. Update all callers.
500 (dict_get_compacted_value_cnt): Rename dict_count_values and
501 change interface. Update all callers.
502 (dict_get_compacted_value_cnt): Remove.
503 (dict_compacting_would_shrink): Remove.
504 (dict_compacting_would_change): Remove.
505 (dict_make_compactor): Add new parameter. Update all callers.
507 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
508 instead of dict_count_values, changing an O(N) operation into
511 2007-08-12 Ben Pfaff <blp@gnu.org>
513 * casereader.c (casereader_read): Don't require cases read by a
514 casereader to be exactly the expected size: as long as they're big
517 2007-08-12 Ben Pfaff <blp@gnu.org>
519 Make casewriters keep track of the number of `union value's in
520 each case. This is useful for two reasons: casewriter_write can
521 then check that the case being written is large enough, and later
522 recipients of the casewriter can determine the size of the case.
524 * casewriter-translator.c (casewriter_create_translator): Add
527 * casewriter.c (struct casewriter): Add value_cnt member.
528 (casewriter_write): Check that the case passed in is big enough.
529 (casewriter_get_value_cnt): New function.
530 (casewriter_create): Add value_cnt parameter.
532 2007-08-09 Ben Pfaff <blp@gnu.org>
534 Fix bug reported by Jason Stover.
535 * settings.c: Move get_termcap_viewport and in particular the
536 #include for <curses.h> to the end of the file. curses.h
537 redefines bool on some systems (e.g. OpenBSD), which causes
538 disagreement between uses of bool before its inclusion and
540 Tested by Jason Stover.
542 2007-07-29 Ben Pfaff <blp@gnu.org>
544 Provisional fix for bug #18692 and bug #20161. Reviewed by John
547 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
548 argument (never "rb" or "wb") because SUSv3 says that only those
549 modes are defined, and glibc in fact rejects other modes.
551 Open portable files with fn_open so that they can be read from
552 pipes. Fix missing fh_close call to go along with fh_open.
553 Report an error if the file close reports an error.
554 * por-file-reader.c (close_reader): New function.
555 (por_file_casereader_destroy): Use close_reader.
556 (pfm_open_reader): Open file with fn_open.
558 2007-07-28 Ben Pfaff <blp@gnu.org>
560 Make PSPP able to read all the portable files I could find on the
561 web. Thanks to John Darrington for review. Bug #17620.
562 * por-file-reader.c (struct pfm_reader): New member `line_length'.
563 (error): Print file offset in hexadecimal.
564 (warning): New function.
565 (advance): Treat lines less than 80 bytes long as padded to 80
567 (pfm_open_reader): Call read_documents if we find an "E" record.
568 (convert_format): Convert invalid formats to the default format
569 instead of aborting reading the file.
570 (read_variables): Rename duplicate variable names instead of
571 aborting reading the file.
572 (read_value_label): Allow string variables of different widths to
573 be assigned value labels in the same record. Replace duplicate
574 value labels instead of aborting.
575 (read_documents): New function.
577 * por-file-writer.c (pfm_open_writer): Call write_documents if the
578 dictionary has documents.
579 (write_documents): New function.
581 2007-07-25 Ben Pfaff <blp@gnu.org>
583 Fix bugs related to bug #17213.
585 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
586 since the former is what config.h has. Include the needed ncurses
588 (static var echo) Rename to `do_echo' because the original name is
589 the same as an ncurses identifier.
590 (get_termcap_viewport) Use error instead of msg.
592 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
593 (fn_close): Don't close stdin, stdout, stderr.
595 2007-07-26 John Darrington <john@darrington.wattle.id.au>
597 * procedure.c procedure.h: Added callbacks which get invoked whenever
598 a dataset's transformation chain changes.
600 2007-07-24 Ben Pfaff <blp@gnu.org>
603 * sys-file-writer.c (write_variable_display_parameters): Use new
604 var_default_display_width function to choose display width of
605 segments after the first one in a given variable.
606 * variable.c (var_create): Use var_default_display_width to pick
607 new variable's display width.
608 (var_default_display_width): New function.
609 Reviewed by John Darrington.
611 2007-07-24 Ben Pfaff <blp@gnu.org>
614 * por-file-writer.c (write_variables): Write weight variable.
615 Reviewed by John Darrington.
617 2007-07-23 Ben Pfaff <blp@gnu.org>
619 Improvements to system file reader and writer.
621 First, move all detailed knowledge of very long strings into
622 sys-file-private.[ch], so that this nasty stuff can be isolated.
624 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
625 (EFFECTIVE_VLS_CHUNK): New macro.
626 (min_int): New function.
627 (max_int): New function.
628 (sfm_width_to_bytes): Rewrite.
629 (sfm_width_to_octs): New function.
630 (sfm_segment_alloc_width): New function.
631 (sfm_segment_alloc_bytes): New function.
632 (sfm_segment_used_bytes): New function.
633 (sfm_segment_offset): New function.
634 (sfm_segment_effective_offset): New function.
635 (sfm_dictionary_to_sfm_vars): New function.
637 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
638 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
639 (struct sfm_var): New structure.
641 Next, improvements to the system file reader.
643 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
644 type casenumber. Added `version_major', `version_minor',
647 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
648 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
649 Change `case_cnt' to type casenumber. Removed `has_vls'.
650 (struct sfm_var): Removed.
651 (sfm_open_reader): Don't warn on wrong case size if the file was
652 written by SPSS 13, which tends to get it wrong. Use
653 sfm_dictionary_to_sfm_vars.
654 (read_header): Always output system file info.
655 (read_variable_record): Simplify code for reading missing values.
656 (read_machine_int32_info): Save version numbers from system file
657 into info struct passed as new argument.
658 (read_long_string_map): Restructured to use new sys-file-private
660 (read_value_labels): Use size_overflow_p.
661 (sys_file_casereader_read): Get rid of distinction between fast
662 and slow paths. Use information provided by sys-file-primate's
663 struct sfm_var to simplify code.
664 (skip_whole_strings): New function.
665 (read_int32): Renamed read_int. Changed return value to int.
667 (read_flt64): Renamed read_float. Changed return value to
668 double. Updated all callers.
669 (int32_to_native): Removed. Changed callers to use
671 (flt64_to_double): Removed. Changed callers to use float_convert.
673 Finally, get rid of int32, flt64 terminology and types in system
674 file writer. The former wasn't very useful since a POSIX "int"
675 can hold the whole range of int32 and we generally didn't have a
676 need for it to be exactly-32-bits, just at-least-32-bits. The
677 latter was inconvenient because we had to assume that it could be
678 different from double and thereby convert special values SYSMIS,
679 HIGHEST, LOWEST to and from it in multiple places. Instead, now
680 we just use "int" and "double" in most places, and do conversions,
681 if necessary, very close to where we do I/O. This change meant
682 that the writer code couldn't represent records in the file as C
683 structs any longer, but that's no great loss. The code actually
684 seems to be more readable without them.
686 Simplify the compression buffering code: only buffer as much as
687 necessary, which is no more than eight 8-byte units at any given
690 * sys-file-writer.c (typedef flt64): Removed.
691 (macro second_lowest_flt64): Removed.
692 (struct sysfile_header): Removed.
693 (struct sysfile_variable): Removed.
694 (struct sfm_writer): Removed `needs_translation', `has_vls',
695 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
696 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
697 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
698 compression buffering by `opcodes', `opcode_cnt', `data',
699 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
700 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
701 data writer functions instead of structures.
702 (calc_oct_idx): New function.
703 (write_header): Use simple data writer functions instead of
705 (write_format_spec): Renamed write_format. New argument.
706 (write_variable_continuation_records): New function.
707 (write_variable): Use simple data writer functions instead of
708 structures. Use write_variable_continuation_records. Write
709 entire very long string instead of requiring caller to understand
711 (write_value_labels): Use simple data writer functions instead of
713 (write_documents): Ditto.
714 (write_variable_display_parameters): Use sys-file-private
715 functions to simplify. Use simple data writer functions instead
717 (write_vls_length_table): Use simple data writer functions instead
719 (write_longvar_table): Ditto.
720 (write_rec_7_34): Break into new functions
721 write_integer_info_record, write_float_info_record. Use simple
722 data writer functions instead of structures.
723 (buf_write): Removed.
724 (append_string_max): Removed.
725 (ensure_buf_space): Removed.
726 (sys_file_casewriter_write): Get rid of the distinction between
727 fast and slow paths, which didn't seem to be too useful. Use new
728 functions write_case_uncompressed, write_case_compressed.
729 (put_instruction): Removed.
730 (put_element): Removed.
731 (write_compressed_data): Removed.
732 (close_writer): Use flush_compressed. Only write case count to
733 system file if it will fit in the field.
734 (write_case_compressed): New function.
735 (write_case_uncompressed): New function.
736 (flush_compressed): New function.
737 (put_cmp_opcode): New function.
738 (put_cmp_number): New function.
739 (write_int): New function.
740 (convert_double_to_output_format): New function.
741 (write_float): New function.
742 (write_value): New function.
743 (write_string): New function.
744 (write_bytes): New function.
745 (write_zeros): New function.
746 (write_spaces): New function.
748 Reviewed by John Darrington.
750 2007-07-22 Ben Pfaff <blp@gnu.org>
752 Don't try to write very long strings to portable files. The
753 format does not support it.
755 * por-file-writer.c (MAX_POR_WIDTH): New macro.
756 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
757 (write_format): Add arg to take width to resize format to.
758 (write_value): Limit width of value written to MAX_POR_WIDTH.
759 (write_variables): Limit width of variable and its output formats
761 Reviewed by John Darrington.
763 2007-07-22 Ben Pfaff <blp@gnu.org>
765 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
766 local variable instead of literal 5.
767 Reviewed by John Darrington.
769 2007-07-22 Ben Pfaff <blp@gnu.org>
771 Fix problems with uniqueness of short names in system files with
772 very long string variables. Now a variable may have multiple
775 * automake.mk (src_data_libdata_a_SOURCES): Add new files
776 short-names.c, short-names.h.
778 * dictionary.c (dict_clone): Clone all the short names.
779 (compare_strings): Move into short-names.c.
780 (hash_strings): Ditto.
781 (set_var_short_name_suffix): Ditto.
782 (dict_assign_short_names): Ditto, rename short_names_assign,
783 change to assign all short names.
785 * por-file-writer.c (write_variables): Use short_names_assign
786 instead of dict_assign_short_names.
788 * short-names.c: New file.
790 * short-names.h: New file.
792 * sys-file-private.c (sfm_width_to_segments): New function.
794 * sys-file-reader.c (read_long_var_name_map): Save and restore all
795 the short names, not just the first one.
797 * sys-file-writer.c (cont_var_name): Removed.
798 (sfm_open_writer): Use short_names_assign instead of
799 dict_assign_short_names. Use unique short names assigned by
800 short_names_assign instead of those generated by cont_var_name.
802 * variable.c (struct variable): Remove `short_name' member,
803 replace by `short_names' and `short_name_cnt'.
804 (var_create) Initialize new members.
805 (var_get_short_name_cnt): New function.
806 (var_get_short_name): Now takes an index argument. Changed most
808 (var_set_short_name): Ditto.
809 (var_clear_short_name): Renamed var_clear_short_names, changed to
810 clear all short names.
812 Reviewed by John Darrington.
814 2007-07-22 Ben Pfaff <blp@gnu.org>
816 * variable.c (var_set_width): Use new var_set_width function.
818 * missing-values.c (mv_n_values): Drop assertion, which was not
821 * format.c (fmt_default_for_width): New function.
822 (fmt_resize): New function.
824 Reviewed by John Darrington.
826 2007-07-18 John Darrington <john@darrington.wattle.id.au>
828 * datasheet.c (datasheet_delete_columns): Added assertion to check
829 we're not deleting outside the range of the sheet.
832 * dictionary.c dictionary.h variable.c: Added the ability for string
833 variables to be resized.
835 * vardict.h: Added some prototypes (moved from dictionary.h) as
836 these should only be called by variable.c
839 2007-07-14 John Darrington <john@darrington.wattle.id.au>
841 * sfm-reader.c: Respect case_cnt field in file header.
843 2007-07-01 John Darrington <john@darrington.wattle.id.au>
845 * transformation.c transformation.h (trns_chain_execute): Changed the
846 signature (Patch #6057)
848 2007-06-10 Ben Pfaff <blp@gnu.org>
850 * casereader-filter.c (casereader_filter_destroy): Make sure to
851 write all the remaining excluded cases to the casewriter, if any.
853 * caseinit.c (init_list_destroy): Rewrite.
854 (init_list_clear): Ditto.
856 * casegrouper.c (casegrouper_get_next_group): Always set *reader
857 to null when returning false.
859 2007-06-06 Ben Pfaff <blp@gnu.org>
861 Actually implement the new procedure code and adapt all of its
862 clients to match. Also adapt all of the other case sources and
863 sinks in the tree and their clients to use the
864 casereader/casewriter infrastructure.
866 * automake.mk: Add and remove files.
868 * any-reader.c: Change into a casereader.
869 * por-file-reader.c: Ditto.
870 * scratch-reader.c: Ditto.
871 * sys-file-reader.c: Ditto.
873 * any-writer.c: Change into a casewriter.
874 * por-file-writer.c: Ditto.
875 * scratch-writer.c: Ditto.
876 * sys-file-writer.c: Ditto.
878 * procedure.c: Change to use casereader, casewriter, caseinit, and
879 other new infrastructure.
881 * scratch-handle.c: Adapt to new infrastructure.
883 * case-sink.c: Removed, now dead code.
884 * case-sink.h: Ditto.
885 * case-source.c: Ditto.
886 * case-source.h: Ditto.
887 * casefile-factory.c: Ditto.
888 * casefile-private.h: Ditto.
891 * casefilter.c: Ditto.
892 * casefilter.h: Ditto.
895 * fastfile-factory.c: Ditto.
896 * fastfile-factory.h: Ditto.
897 * storage-stream.c: Ditto.
898 * storage-stream.h: Ditto.
900 2007-06-06 Ben Pfaff <blp@gnu.org>
904 * automake.mk: Add new files.
906 * datasheet.c: New file.
908 * datasheet.h: New file.
910 2007-06-06 Ben Pfaff <blp@gnu.org>
912 Until now, the procedure code has provided a case to the
913 case_source, which has filled in the data values that come from
914 the active file. "Left" data values that don't come from the
915 active file naturally stay the same from case to case, because the
916 procedure code keeps using that same case.
918 One of the compromises that comes with the new procedure code is
919 that the active file allocates and provides its own case, which
920 the procedure code then has to resize to provide room for any
921 other variables that should go in the case and then fill in the
922 values of "left" variables. Then, when we're done with that case,
923 we have to save the values of "left" variables to copy into the
924 next case read from the active file.
926 The caseinit code helps with this.
928 * automake.mk: Add new files.
930 * caseinit.c: New file.
932 * caseinit.h: New file.
934 2007-06-06 Ben Pfaff <blp@gnu.org>
936 * value.h (value_cnt_from_width): New function.
938 * variable.c (var_get_value_cnt): Use new function.
940 2007-06-06 Ben Pfaff <blp@gnu.org>
942 Add casegrouper, to allow cases read from a given casereader to be
943 broken into groups, each of which has its own casereader.
944 Generally cases are grouped based on having equal values for some
947 * automake.mk: Add new files.
949 * casegrouper.c: New file.
951 * casegrouper.h: New file.
953 2007-06-06 Ben Pfaff <blp@gnu.org>
955 Add interface to lexicographical ordering of cases.
957 * automake.mk: Add new files.
959 * case-ordering.c: New file.
961 * case-ordering.h: New file.
963 2007-06-06 Ben Pfaff <blp@gnu.org>
965 Add casereaders and casewriters, the basis of the new data processing
966 implementation. A casereader is a uniform interface to reading cases
967 from a data source; a casewriter is a uniform interface to writing
968 cases to a data sink.
970 * automake.mk: Add new files.
972 * casereader-filter.c: New file.
974 * casereader-provider.h: New file.
976 * casereader-translator.c: New file.
978 * casereader.c: New file.
980 * casereader.h: New file.
982 * casewriter-provider.h: New file.
984 * casewriter-translator.c: New file.
986 * casewriter.c: New file.
988 * casewriter.h: New file.
990 2007-06-06 Ben Pfaff <blp@gnu.org>
992 "casewindow" data structure that extends the deque (from libpspp)
993 of cases with the ability to dump cases to disk if we get too many
996 * automake.mk: Add new files.
998 * casewindow.c: New file.
1000 * casewindow.h: New file.
1002 2007-06-06 Ben Pfaff <blp@gnu.org>
1004 sparse_cases data structure that augments a sparse_array of cases
1005 with the ability to dump cases to disk if we get too many cases in
1008 * automake.mk: Add new files.
1010 * sparse-cases.c: New file.
1012 * sparse-cases.h: New file.
1014 2007-06-06 Ben Pfaff <blp@gnu.org>
1016 Adds a low-level on-disk case array data structure.
1018 * automake.mk: Add new files.
1020 * case-tmpfile.c: New file.
1022 * case-tmpfile.h: New file.
1024 2007-06-06 Ben Pfaff <blp@gnu.org>
1026 In a couple of places we calculate the maximum number of cases to
1027 keep in memory based on the user-defined workspace. Enable
1028 centralizing the calculation through a new function.
1030 * settings.c (get_workspace_cases): New function.
1032 2007-06-06 Ben Pfaff <blp@gnu.org>
1034 The casenumber type is defined in transformations.h, but case.h is
1035 a more sensible place. Move it.
1037 * case.h (CASENUMBER_MAX): New macro.
1038 (typedef casenumber): Move here, from transformations.h.
1040 2007-06-03 Ben Pfaff <blp@gnu.org>
1042 Slightly generalize case_to_values and case_from_values functions.
1044 * case.c (case_to_values): Rename case_copy_out, change interface.
1045 (case_from_values): Rename case_copy_in, change interface.
1047 * fastfile.c (fastfilereader_get_next_case): Update caller.
1048 (write_case_to_disk): Ditto.
1050 2007-06-02 Ben Pfaff <blp@gnu.org>
1052 Clean up after a forgotten part of patch #5829.
1054 * casedeque.h: Remove unused file.
1056 * automake.mk: Remove casedeque.h from sources.
1058 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1060 * category.c: Removed redundant #include
1062 2007-05-06 Ben Pfaff <blp@gnu.org>
1064 Abstract the documents within a dictionary a little better.
1065 Thanks to John Darrington for suggestion, initial version, and
1066 review. Patch #5917.
1068 * dictionary.c (struct dictionary): Change `documents' member from
1069 char * to struct string.
1070 (dict_clear): Destroy struct string.
1071 (dict_get_documents): Convert struct string to char *.
1072 (dict_set_documents): Set struct string. Pad to 80-character
1074 (dict_clear_documents): New function.
1075 (dict_add_document_line): New function.
1076 (dict_get_document_line_cnt): New function.
1077 (dict_get_document_line): New function.
1079 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1081 * sys-file-reader.c (read_documents): Use new document functions.
1083 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1085 * sys-file-reader.c: When reading a system file which has no
1086 long name table, automatically create one where the long names
1087 are the lower case versions of the short names.
1089 2007-04-22 Ben Pfaff <blp@gnu.org>
1091 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1092 dict_clear will free most data related to the dictionary.
1093 dict_clear does a decent job, except that dict_set_split_vars on
1094 some systems won't actually free the dict's "split" member.
1095 Instead, it'll allocate a 1-byte region. Fix this.
1097 * value.c (value_copy): New function.
1098 (value_set_missing): Ditto.
1100 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1102 * Deleted existing category.h and moved cat-routines.h into
1103 category.h Encapsulated struct cat_vals better.
1105 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1107 * sys-file-reader.c: When reading a system file which has no
1108 long name table, automatically create one where the long names
1109 are the lower case versions of the short names.
1111 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1113 * sys-file-reader.c: Some versions of Other Software seem to
1114 produce system files with string variables' measure set to
1115 zero. We'll assume these are supposed to be nominal variables.
1117 2007-03-30 Ben Pfaff <blp@gnu.org>
1119 * procedure.c: Adapt to new deque data structure.
1121 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1122 Ben Pfaff <blp@gnu.org>
1124 * file-name.c: Mingw compatibility fixes.
1125 (fn_search_path): Use ISSLASH instead of comparing against '/'
1127 (fn_dir_name): Use dir_name from gnulib.
1128 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1129 (fn_get_identity): Use GetFullPathName instead of canonicalize
1130 from gnulib, because the latter does not fully support
1131 Windows-style path names. Use this implementation based on the
1132 detected presence of Windows instead of the absence of Unix, since
1133 the new implementation is Windows-specific.
1134 (fn_compare_file_identities): In Windows implementation, compare
1135 names case-insensitively.
1137 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1139 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1140 <stdlib.h> to have the same effect.
1142 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1144 * por-file-reader.c: Add missing _() around messages.
1146 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1148 * make-file.c: Include "mkstemp.h", without which linking on
1151 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1153 Reduce platform dependence.
1155 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1156 Everywhere we using this, we really should have just depended on
1157 the shell to expand tildes.
1158 (fn_search_path): Simplify, given that we don't do tilde expansion
1160 (fn_normalize): Removed. Caller changed to use the canonicalize
1162 (fn_get_cwd): Removed. Only user was fn_normalize.
1163 (fn_is_absolute): Really only test for absolute names.
1164 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1166 (fn_readlink): Removed, as it was only used fn_normalize.
1167 (fn_exists): Assume the stat function is available; gnulib does.
1168 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1170 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1172 * dictionary.c (dict_create_vector_assert): New function.
1174 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1176 * file-name.c (fn_normalize): Correct name of function
1177 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1178 for pointing this out.
1180 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1182 We are using a single member in struct file_handle, the "name"
1183 field, for more than one purpose. When it begins with '"', it's a
1184 file name; otherwise, it's a token that can be used to identify
1185 it. When that assertion fires, it's because we searched for the
1186 name case-sensitively as a file name (so that there was no match),
1187 and then we try to insert it case-insensitively as a token, which
1188 fails because duplicates aren't allowed.
1190 Solution: break the two purposes into two separate fields. This
1191 fixes the problem and likely makes the code easier to read too.
1193 Fixes bug #18922. Thanks to John Darrington for bug report and
1196 * file-handle-def.c (struct file_handle): New `id' member.
1197 (fh_from_name): Rename fh_from_id. Update all callers.
1198 (create_handle): New `id' parameter. Update all callers.
1199 (fh_create_file): Ditto.
1200 (fh_get_id): New function.
1202 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1204 * case.c (case_is_null): Change return type to bool.
1206 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1210 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1211 reference counting to share data. This makes it easy for
1212 valgrind, etc. to find accesses to cases that have been destroyed
1213 but have been kept around by another user's ref-count. This often
1214 happens when the data set is small enough to find in memory; if a
1215 bigger data set that would overflow to disk were used, then data
1216 corruption would occur.
1218 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1222 * case.c (case_unshare): Make it check internally whether the
1223 ref_cnt is greater than 1, so that the callers don't have to.
1224 Update callers not to check.
1226 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1228 Before, I was thinking that I might want to get rid of reference
1229 counting at some point. Now, I'm pretty sure that it's here to
1230 stay. Thus, because we have to store the value_cnt anyway for
1231 reference-counted cases, we might as well expose it to users.
1233 * case.c (case_get_value_cnt): New function.
1234 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1235 resize case if its size actually changed.
1237 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1238 instead of peeking inside struct case directly.
1239 (casefile_append): Ditto.
1241 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1243 Get rid of the inlines for the case functions, which made the
1244 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1245 the inlines didn't speed up "make check" at all, which is not a
1246 perfect benchmark but seems indicative.)
1248 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1249 definitions. Remove some assertions on nonnull pointers that were
1250 redundant with a pointer dereference soon after in the function.
1252 (struct case_data): Move definition here from case.h.
1256 (case_data_wr): Ditto.
1258 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1260 * automake.mk: Add casedeque.h to sources.
1262 * casedeque.h: New file.
1264 * procedure.c: (struct dataset) Change lag_count, lag_head,
1265 lag_queue member into single struct casedeque member. Update all
1266 users to use the casedeque instead.
1269 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1271 * procedure.c: Simplify lagged cases interface. Updated all
1272 clients--well, the only client--to use the simplified interface.
1273 (dataset_n_lag) Removed.
1274 (dataset_set_n_lag) Removed.
1275 (dataset_need_lag) New function.
1277 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1279 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1280 whenever appropriate, but only when the dataset/dictionary is in a
1283 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1285 * dictionary.c dictionary.h : Added callbacks for change of filter and
1286 split variables. Refactored some code to ensure that callbacks get
1287 invoked when appropriate.
1289 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1290 replace_dict callback occurs when permanent_dict replaces the current
1293 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1295 * dictionary.c dictionary.h : Added callback for when the weight
1296 variable of a dictionary changes.
1298 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1300 * dictionary.c dictionary.h : Added replace_source and replace_dict
1301 callbacks, and functions to deal with them.
1303 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1305 Simplify missing value handling.
1307 * missing-values.h (enum mv_class): New type.
1308 (enum mv_type): Moved definition into missing-values.c and renamed
1309 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1310 enums. Updated all uses.
1311 (struct missing_values): Changed type of `type' from `enum
1312 mv_type' to `int' because the definition is no longer exposed.
1314 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1315 parameter. Update all callers.
1316 (mv_is_num_missing): Ditto.
1317 (mv_is_str_missing): Ditto.
1318 (mv_is_value_user_missing): Removed. Changed callers to use
1319 mv_is_value_missing.
1320 (mv_is_num_user_missing): Removed. Changed callers to use
1322 (mv_is_str_user_missing): Removed. Changed callers to use
1324 (mv_is_value_system_missing): Removed. Changed callers to use
1325 mv_is_value_missing.
1326 (mv_set_type): Removed. Changed callers to use mv_clear.
1327 (mv_clear): New function.
1329 * variable.c (var_is_value_missing): Add new enum mv_class
1330 parameter. Update all callers.
1331 (var_is_num_missing): Ditto.
1332 (var_is_str_missing): Ditto.
1333 (var_is_value_user_missing): Removed. Changed callers to use
1334 var_is_value_missing.
1335 (var_is_num_user_missing): Removed. Changed callers to use
1337 (var_is_str_user_missing): Removed. Changed callers to use
1339 (var_is_value_system_missing): Removed. Changed callers to use
1340 var_is_value_missing.
1342 * casefilter.c (struct casefilter): Use enum mv_class in place of
1344 (casefilter_variable_missing): Adapt to new member.
1345 (casefilter_create): Change signature to take enum mv_class,
1348 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1350 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1352 * case-sink.c case-sink.h procedure.c procedure.h
1353 storage-stream.c: Now uses the factory.
1355 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1357 Make it possible to pull cases from the active file with a
1358 function call, instead of requiring indirection through a callback
1361 * case-source.h (struct case_source_class): Change ->read function
1362 to return a single case, instead of calling a callback function
1363 for each case. Change ->destroy function to return an error
1366 * case-source.c (free_case_source): Pass along the value returned
1367 by the case_source ->destroy function.
1369 * procedure.c (struct write_case_data): Removed.
1370 (struct dataset): Added some members to track procedure state.
1371 (procedure): Optimize the trivial case at this level.
1372 (internal_procedure): Re-implement in terms of proc_open,
1373 proc_read, proc_close.
1374 (proc_open) New function.
1375 (proc_read) New function.
1376 (proc_close) New function.
1377 (write_case) Moved into proc_read.
1378 (close_active_file) Moved closing of data source into proc_close.
1380 * storage-source.c: Rewrote to conform with modified
1381 case_source_class interface.
1383 * transformations.c (trns_chain_execute): Added argument to allow
1384 starting execution from an arbitrary transformation. Updated
1387 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1389 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1391 * sys-file-reader.c (read_display_parameters): Don't assume that
1392 MEASURE_* and ALIGN_* have the same values found in system files.
1394 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1396 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1397 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1398 the value of n_MEASURES, which was off by 1 (at least from my
1401 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1403 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1404 callbacks which are invoked when the dictionary or its
1405 variables are changed.
1407 * missing-values.c missing-values.h value-labels.c: Tidied up
1408 consistency checks, and made some of them return false
1409 instead of assert-failing.
1411 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1413 * calendar.c (calendar_days_in_month): New function.
1415 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1417 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1418 the value label's width.
1420 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1422 * sfm-private.h: Move contents into sys-file-writer.c, which is
1423 the only remaining user. Removed Borland C++-specific directives.
1425 * sys-file-reader.c: Clean up and rewrite entire file. The
1426 rewritten version is simpler and better abstracted, and should be
1427 easier to maintain and extend. It avoids using structures to read
1428 file data, which is prone to padding variations among compilers.
1429 It should also handle non-IEEE 754 system files, although I
1430 haven't been able to find any. It has been tested against many
1431 .sav files obtained from the Web and found to produce the same
1432 results as the earlier version of the code, or in some cases
1433 improved results. It is more tolerant of format variations found
1436 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1437 member, putting an enum integer_format in its place. New member
1438 `float_format'. Changed `compressed' member to type bool.
1440 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1442 * dictionary.c (dict_delete_consecutive_vars): New function.
1444 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1446 * file-name.c (fn_search_path): Remove prefix arg that was unused
1447 by any caller. Updated all callers.
1449 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1451 * format.c (fmt_dollar_template): Use user's decimal point
1452 character. Add assertion.
1454 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1456 * format.c (fmt_dollar_template): New function, based on
1457 dollar_format_template from var-type-dialog.c.
1459 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1461 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1464 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1466 Finish converting struct variable to an opaque type. In this
1467 phase, we add remaining setter and getter functions, convert the
1468 remaining PSPP code to use them, and do a bunch of cleanup. The
1469 resulting changes are pervasive but mostly trivial, and only the
1470 notable changes are logged.
1472 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1475 * case.c (case_data): Renamed case_data_idx.
1476 (case_num): Renamed case_num_idx.
1477 (case_str): Renamed case_str_idx.
1478 (case_data_rw): Renamed case_data_rw_idx.
1480 * case.h (case_data): New function with old name and an interface
1481 that takes a variable instead of an index, which is easier to
1482 use. Updated all callers to use the new interface, or to use the
1483 new *_idx function (see above).
1486 (case_data_rw): Ditto.
1488 * category.c (cat_stored_values_destroy): Changed interface to
1489 take a struct cat_vals * instead of a struct variable *.
1491 * dictionary.c (dict_clone): Use new vector_clone function.
1492 (dict_clear) Use new var_destroy function.
1493 (add_var) New function.
1494 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1495 (dict_create_var_assert) Rewrite in terms of add_var.
1496 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1497 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1498 (dict_lookup_var) Use new var_create, var_destroy functions.
1499 (dict_contains_var) Rewrite in terms of new vardict functionality.
1500 (set_var_dict_index) New function.
1501 (set_var_case_index) New function.
1502 (reindex_vars) New function.
1503 (dict_delete_var) Rewrite in terms of new vardict functionality.
1504 (dict_reorder_var) Ditto.
1505 (dict_reorder_vars) Ditto.
1506 (rename_var) New function.
1507 (dict_rename_var) Use rename_var.
1508 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1509 (dict_get_compacted_idx_to_fv) Rename
1510 dict_get_compacted_dict_index_to_case_index, update callers.
1511 (dict_create_vector) Use new vector_create function.
1512 (dict_clear_vectors) Use new vector_destroy function.
1513 (set_var_short_name_suffix) Move here from variable.c, renamed
1514 from var_set_short_name_suffix, make static, update caller.
1516 * sys-file-private.c: New file.
1517 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1518 width_to_bytes, update callers.
1520 * sys-file-private.h: New file. Later it will supplant
1521 sfm-private.h; for now it supplements it.
1522 (macro MIN_VERY_LONG_STRING) New macro.
1523 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1525 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1526 MAX_LONG_STRING was used before.
1528 * sys-file-writer.c: Ditto.
1530 * value-labels.c: Change the paradigm here to be that a null
1531 pointer is OK for a struct val_labs * in most cases; it just
1532 represents an empty set of value labels.
1533 (val_labs_copy) A copy of a null set is a null set.
1534 (val_labs_count) A null set has 0 labels.
1535 (val_labs_replace) Change return type to void. Rewrite for
1537 (val_labs_find) A null set does not contain the value.
1538 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1539 transposed argument order, updated all callers.
1541 * value.c: New file.
1542 (value_dup) Moved here from variable.c.
1543 (compare_values) Ditto.
1546 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1547 (macro MAX_LONG_STRING) Removed, because it was only interesting
1548 for system files, not for general code.
1549 (macro MAX_VERY_LONG_STRING) Ditto.
1550 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1551 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1553 * vardict.h: New file, for an interface between variables and
1556 * variable.c: A lot of functions were moved around, for better
1558 (struct variable) Move definition here, from variable.h.
1559 (var_type_adj) Removed--makes i18n hard.
1560 (var_type_noun) Ditto.
1561 (var_create) New function.
1562 (var_clone) New function.
1563 (var_destroy) New function.
1564 (var_set_name) Assert that variable is not in a dictionary.
1565 (compare_var_names) Rename compare_vars_by_name and fix a couple
1566 of callers who thought the args were strings.
1567 (hash_var_name) Rename hash_var_by_name.
1568 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1569 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1570 (var_is_very_long_string) Removed, because it was only interesting
1571 to system file code.
1572 (var_set_missing_values) Allow the argument to be the wrong width,
1573 as long as we can resize it. Simplify callers who were doing the
1574 resizing themselves.
1575 (var_get_value_labels) New function.
1576 (var_has_value_labels) New function.
1577 (var_set_value_labels) New function.
1578 (alloc_value_labels) New function.
1579 (var_add_value_label) New function.
1580 (var_replace_value_label) New function.
1581 (var_clear_value_labels) New function.
1582 (var_lookup_value_label) New function.
1583 (var_get_value_name) Moved here from variable.c, renamed from
1584 var_get_value_name, transposed argument order, updated all
1586 (var_to_string) Moved here, from variable-label.c.
1587 (var_set_leave) New function.
1588 (var_get_leave) New function.
1589 (var_must_leave) New function.
1590 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1591 set_var_short_name_suffix.
1592 (var_get_dict_index) New function.
1593 (var_get_case_index) New function.
1594 (var_get_obs_vals) New function.
1595 (var_set_obs_vals) New function.
1596 (var_has_obs_vals) New function.
1597 (var_get_vardict) New function.
1598 (var_set_vardict) New function.
1599 (var_has_vardict) New function.
1600 (var_clear_vardict) New function.
1601 (value_dup) Moved to value.c.
1602 (compare_values) Ditto.
1605 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1606 (enum ALPHA) Rename VAR_STRING, update all users.
1608 * vector.c: New file.
1609 (struct vector) Moved here, from variable.h.
1610 (check_widths) New function.
1611 (vector_create) New function.
1612 (vector_clone) New function.
1613 (vector_destroy) New function.
1614 (vector_get_name) New function.
1615 (vector_get_var) New function.
1616 (vector_get_var_cnt) New function.
1617 (compare_vector_ptrs_by_name) New function.
1619 * vector.h: New file.
1621 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1623 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1624 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1627 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1629 * value-labels.c (destroy_atoms): New function.
1630 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1632 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1634 Thanks to Jason Stover for pointing out this problem.
1636 * data-out.c (output_number): Use gsl_finite from GSL, which is
1637 portable, instead of isfinite, which is not.
1640 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1642 * variable.c variable.h (value_dup): New function.
1644 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1646 Start converting struct variable to an opaque type. In this
1647 phase, we add a bunch of setter and getter functions and convert
1648 most of the PSPP code to use them. The resulting changes are
1649 pervasive but mostly trivial, and only the notable changes are
1652 * format.c (fmt_equal): New function.
1654 * variable.c (var_type_is_valid): New function.
1655 (measure_is_valid) Moved here, from format.c.
1656 (alignment_is_valid) Moved here, from format.c.
1657 (var_get_name) New function.
1658 (var_set_name) New function.
1659 (width_to_type) New function.
1660 (var_get_type) New function.
1661 (var_get_width) New function.
1662 (var_set_width) New function.
1663 (var_is_numeric) New function.
1664 (var_is_alpha) New function.
1665 (var_is_short_string) New function.
1666 (var_is_long_string) New function.
1667 (var_is_very_long_string) New function.
1668 (var_get_missing_values) New function.
1669 (var_set_missing_values) New function.
1670 (var_clear_missing_values) New function.
1671 (var_has_missing_values) New function.
1672 (var_is_value_missing) New function.
1673 (var_is_num_missing) New function.
1674 (var_is_str_missing) New function.
1675 (var_is_value_user_missing) New function.
1676 (var_is_num_user_missing) New function.
1677 (var_is_str_user_missing) New function.
1678 (var_is_value_system_missing) New function.
1679 (var_get_print_format) New function.
1680 (var_set_print_format) New function.
1681 (var_get_write_format) New function.
1682 (var_set_write_format) New function.
1683 (var_set_both_formats) New function.
1684 (var_get_label) New function.
1685 (var_set_label) New function.
1686 (var_clear_label) New function.
1687 (var_has_label) New function.
1688 (var_get_measure) New function.
1689 (var_set_measure) New function.
1690 (var_get_display_width) New function.
1691 (var_set_display_width) New function.
1692 (var_get_alignment) New function.
1693 (var_set_alignment) New function.
1694 (var_get_value_cnt) New function.
1695 (var_get_leave) New function.
1696 (var_get_short_name) New function.
1698 * variable.h: (struct variable) Removed "type" and "nv" members;
1699 they are now computed from "width" where needed.
1701 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1703 * missing-values.c (mv_resize): Don't write beyond end of the
1704 allocated buffer when resizing a long string.
1706 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1708 Clean up identifier code: don't require identifier enumerations to
1709 be in a particular order; make better use of string library;
1710 expose less of the internals.
1712 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1713 change interface. Updated all callers.
1714 (lex_id_match) Change interface to use struct substring, update
1716 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1717 (global array keywords[]) Make static, change form. Update all
1718 users to use lex_id_name instead.
1719 (lex_is_keyword) New function.
1720 (lex_id_to_token) Change interface to use struct substring, update
1722 (lex_id_name) New function.
1724 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1725 lex_is_keyword instead.
1726 (T_LAST_KEYWORD) Removed.
1727 (T_N_KEYWORDS) Removed.
1729 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1731 * format.c: (fmt_date_template) Distinguish characters for which a
1732 space is output and any date delimiter is allowed on input, from
1733 those for which a space is output and only a space is allowed on
1734 input. The former is represented by X, the latter by a space.
1735 Also, drop distinction between h and H, changing the former to the
1738 * data-in.c: Completely rewrite internals to conform to SPSS input
1739 formats as closely as possible.
1740 (data_in) Changed external interface by replacing the structure
1741 that was used as a single argument by a set of arguments. Updated
1743 (data_in_finite_line) Removed. Converted all callers to use plain
1745 (data_in_get_integer_format) New function.
1746 (data_in_set_integer_format) New function.
1747 (data_in_get_float_format) New function.
1748 (data_in_set_float_format) New function.
1750 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1751 (struct data_in) Removed.
1753 * data-out.c: (output_date) Drop each component from the input as
1754 it is output, to allow us to drop the distinction between h (a
1755 count of hours) and H (the hour of day) template characters.
1756 Also, handle new X template character.
1757 (output_scientific) Follow more rational rule on when to drop
1758 fraction introduced between SPSS 13 and 15. Updated test case to
1761 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1763 Fix buffer overflow reported by John Darrington.
1765 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1766 realize that DIGITS is a count of nibbles, not of bytes.
1768 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1770 * calendar.c (calendar_offset_to_gregorian) Also return the
1771 year-of-day. Change callers to new interface.
1773 * data-out.c: Completely rewrite internals to conform to SPSS
1774 output formats as completely as possible.
1775 (data_out) Change interface to put input parameters before output
1776 parameters, for consistency with the style I now prefer. Update
1778 (data_out_get_integer_format) New public function.
1779 (data_out_set_integer_format) New public function.
1780 (data_out_get_float_format) New public function.
1781 (data_out_set_float_format) New public function.
1783 * data-out.h: New file. Move prototype for data_out here, from
1786 * format.c: (fmt_step_width) Use equality comparison instead of
1787 bitwise and, for clarity.
1788 (fmt_is_string) Ditto.
1789 (fmt_input_to_output) Fix categories that are translated to F
1792 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1794 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1795 casefile-private.h: Added casefilter to assist commands with missing
1798 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1800 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1802 * settings.c: (route_errors_to_terminal) New variable.
1803 (route_errors_to_listing) New variable.
1804 (get_error_routing_to_terminal) New function.
1805 (set_error_routing_to_terminal) New function.
1806 (get_error_routing_to_listing) New function.
1807 (set_error_routing_to_listing) New function.
1809 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1811 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1813 * format.c: Completely rewrite, to achieve better abstraction.
1814 Rewrite all references to formats in other files.
1816 * format.def: Rewrite and reorganize.
1818 * settings.c: Move everything related to custom currency formats
1819 into format.[ch], changing them in form, so as to group related
1820 code and definitions better. Changed all references to use the
1822 (static var decimal) Removed.
1823 (static var grouping) Removed.
1824 (static var cc) Removed.
1825 (get_decimal) Removed.
1826 (set_decimal) Removed.
1827 (get_grouping) Removed.
1828 (set_grouping) Removed.
1832 * settings.h: (macro CC_CNT) Removed.
1833 (macro CC_WIDTH) Removed.
1834 (struct custom_currency) Removed.
1836 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1838 * data-in.c (data_in): Use switch statement instead of table, to
1839 avoid dependence on the order of the FMT_* enums.
1841 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1843 * data-out.c: (num_to_string) Removed, because it was dead code.
1845 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1847 * data-in.c (parse_trailer): Fix error message.
1849 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1851 * format.c (fmt_is_binary): New function.
1853 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1855 * procedure.c procedure.h: Encapsulated the static data into a single
1858 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1860 * casefile.c (casereader_read_xfer): Always initialize the case,
1861 even on an error condition.
1863 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1865 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1866 entire function useless.
1868 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1870 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1871 casereader method casereader_clone.
1873 * procedure.c pransformations.h: Introduced new type casenum_t
1875 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1877 * variable.c: (width_to_bytes) Rephrase code for clarify.
1879 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1881 * format.c: (fmt_type_from_string) New function.
1882 (fmt_to_string) Include decimals in output if the format has
1883 decimals, even if the format type does not. This way, we can
1884 accurately reproduce incorrect formats in user output.
1885 (check_common_specifier) Make the check for a bad format type an
1886 assertion, so we get bug reports if they show up. Fix message.
1887 Check for decimal places with a format type that doesn't allow
1889 (check_input_specifier) Remove check for FMT_X, which has been
1891 (check_output_specifier) Ditto.
1893 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1895 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1896 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1897 (enum fmt_parse_flags) Removed.
1899 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1901 * casefile.c casefile.h: Converted to an abstract base class.
1902 * casefile-private.h fastfile.c fastfile.h: New files.
1903 * automake.mk procedure.c scratch-writer.c storage-stream.c
1905 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1907 * procedure.c (internal_procedure): Create sink_case with only as
1908 many values as the compacted dictionary.
1910 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1912 Remove "debugging" code that caused plenty of false positives and
1915 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1917 * case.c: Remove all references to `this' member.
1919 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1921 Fix link error noted by Jason Stover.
1923 * storage-stream.c: Include <assert.h>.
1925 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1927 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1928 ALL) and additional underlying system file issues.
1930 Thanks to John Darrington for review.
1932 First problem: var_hash points to variables not owned by the
1933 sys-file-reader, which the caller may free or modify. Use an
1934 array of sfm_vars instead, as done earlier (e.g. CVS version
1937 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1938 members and remove all code that references it. Add vars, var_cnt
1939 members. Remove fix_specials member, which was unused.
1940 (struct sfm_var) Remove name member, which was unused.
1941 (sfm_close_reader) Free vars member instead of var_hash.
1942 (compare_var_shortnames) Removed.
1943 (hash_var_shortname) Removed.
1944 (sfm_open_reader) Fill out vars array.
1945 (compare_var_index) Removed.
1946 (sfm_read_case) Use vars instead of var_hash.
1948 Second problem: we're confused about when we actually have very
1949 long strings, causing us to choose incorrectly between slow path
1950 and fast path in sfm_read_case.
1952 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1953 have very long strings, not when we have long variable names,
1954 which is an unrelated feature.
1956 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1958 * variable.h: Move var_set and variable parsing declarations to
1959 new header, src/language/lexer/variable-parser.h. Modified lots
1960 of files to include the new header.
1962 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1964 * value-labels.c (value_to_string): When there's no value label,
1965 format the variable according to its print format, instead of
1966 always effectively using A or F format.
1968 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1970 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1973 * format.c: Removed tortological assertion.
1975 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1977 Reform string library.
1979 * file-name.c (fn_interp_vars): Change interface to take a
1980 substring as input. Updated all users.
1982 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1984 * format.c (measure_is_valid): Really return false when m >=
1987 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1989 Implement random access to casefiles, for use in GUI.
1991 * casefile.c: (struct casereader) Add `random', `file_ofs',
1992 `buffer_ofs' members.
1993 (casefile_get_random_reader) New function.
1994 (read_open_file) Break part into new function
1995 seek_and_fill_buffer().
1996 (fill_buffer) Update buffer_ofs, file_ofs.
1997 (casereader_seek) New function.
1999 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2001 * settings.c: Added call to i18n{done, init}.
2003 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
2005 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
2008 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
2010 * casefile.c: Convert many uses of `int' to `bool'.
2012 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
2014 * transformations.c (trns_chain_destroy): Destroy chain's trns
2015 member, to fix memory leak.
2017 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
2019 * storage-stream.c (storage_source_decapsulate): Destroy case
2020 source to fix memory leak.
2022 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
2024 * scratch-reader.c (scratch_reader_read_case): Copy into existing
2025 case passed as argument instead of initializing the argument as a
2026 case. Fixes memory leak that showed up in
2027 tests/command/aggregate.sh with scratch files.
2029 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
2031 * procedure.c (proc_done): Destroy default_dict, to fix memory
2034 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2036 Simplify procedure_with_splits().
2038 * procedure.c (struct split_aux_data): Removed case_count member.
2039 (procedure_with_splits) Don't initialize case_count.
2040 (split_procedure_case_func) Check whether prev_case is null
2041 instead of case_count.
2042 (split_procedure_end_func) Ditto.
2044 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2046 * case.c (case_move): Do nothing if dst and src are the same
2048 (case_try_create) Merge two similar cases.
2049 (case_copy) Unshare only if data must be actually copied.
2051 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2053 * data-in.c data-out.c dictionary.c sys-file-reader.c
2054 sys-file-writer.c variable.c variable.h: Reworked very long string
2055 support for better encapsulation.
2057 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2059 * value-labels.c (val_labs_can_set_width): New function.
2060 (val_labs_set_width) Clear labels if increasing width to long
2062 (val_labs_destroy) Remove unneeded test for null.
2064 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2066 * value-labels.h: Remove unneeded dependency on variable.h.
2068 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2070 Get rid of `char *c' member in union value, for cleanliness.
2072 * value.h: (union value) Remove `c' member.
2074 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2076 Make dictionary compacting functions a little more general.
2078 * sys-file-writer.c (sfm_open_writer): Use
2079 dict_compacting_would_change().
2080 (does_dict_need_translation) Removed.
2082 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2084 Make dictionary compacting functions a little more general.
2086 * dictionary.c (dict_needs_compaction): Rename
2087 dict_compacting_would_shrink(). Update all callers.
2088 (dict_compacting_would_change) New function.
2090 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2092 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2093 inverted return value (!).
2095 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2097 Continue reforming procedure execution.
2099 * procedure.c: Search and replace "vfm" by "proc". Notably:
2100 (static var vfm_source) Rename proc_source. Update all
2102 (static var vfm_sink) Rename proc_sink. Update all references.
2104 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2106 Continue reforming procedure execution. In this phase, remove
2107 PROCESS IF, which was deprecated anyway and can be easily
2108 simulated with TEMPORARY followed by SELECT IF.
2110 * procedure.c: (open_active_file) Don't call
2111 add_process_if_trns().
2112 (discard_variables) Get rid of redundant call to
2113 proc_cancel_all_transformations().
2114 (add_process_if_trns) Removed.
2115 (process_if_trns_proc) Removed.
2116 (process_if_trns_free) Removed.
2118 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2120 Continue reforming procedure execution. In this phase, add
2121 `const' to the case passed to procedure()'s callback.
2123 Updated all users of procedure() as well.
2125 * procedure.c: (struct write_case_data) Add "const" to ccase
2126 parameter for case_func member.
2127 (procedure) Add "const" to ccase parameter for proc_func
2129 (multipass_case_func) Make ccase parameter const.
2130 (internal_procedure) Add "const" to ccase parameter for case_func
2132 (split_procedure_case_func) Make ccase parameter const.
2133 (multipass_split_case_func) Make ccase parameter const.
2135 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2137 Continue reforming procedure execution. In this phase, get rid of
2138 the output code for SPLIT FILE groups in procedure.c, which really
2139 shouldn't be doing any output. Move it into the individual
2140 procedures instead. This also adds some flexibility.
2142 Updated many users of procedure_with_splits() and
2143 multipass_procedure_with_splits() to call
2144 output_split_file_values() and to deal with increased use of
2147 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2148 parameter to begin_func member.
2149 (procedure_with_splits) Add "const struct ccase *" parameter to
2150 begin_func parameter. Make ccase parameter const in proc_func
2152 (split_procedure_case_func) Don't dump split file group. Pass
2154 (dump_splits) Moved to language/dictionary/split-file.c as
2155 output_split_file_values().
2156 (struct multipass_split_aux_data) Add "const struct ccase *"
2157 parameter to split_func member.
2158 (multipass_procedure_with_splits) Add "const struct ccase *"
2159 parameter to split_func parameter.
2160 (multipass_split_case_func) Save new SPLIT FILE case before
2162 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2164 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2166 Continue reforming procedure execution. Change
2167 internal_procedure() so that it calls open_active_file() and
2168 close_active_file(), which isolates most of the actual procedure
2171 * procedure.c: (struct write_case_data) Rename `proc_func' member
2172 to `case_func' and update all references.
2173 (procedure) Rewrite as one-line wrapper around
2174 internal_procedure().
2175 (struct multipass_aux_data) New.
2176 (multipass_callback) Renamed multipass_case_func(). Use struct
2177 multipass_aux_data as auxiliary data.
2178 (multipass_end_func) New function.
2179 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2180 that uses multipass_case_func, multipass_end_func.
2181 (internal_procedure) Add `end_func' argument. Move optimization
2182 of trivial case in here. Move call to open_active_file() and
2183 close_active_file() in here. Now assert that vfm_source is
2185 (procedure_with_splits_callback) Rename
2186 split_procedure_case_func().
2187 (split_procedure_end_func) New function.
2188 (multipass_split_callback) Rename multipass_split_case_func.
2189 (multipass_split_end_func) New function.
2190 (discard_variables) No need to test for nonnull vfm_source.
2192 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2194 Continue reforming procedure execution. Get rid of unused member.
2196 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2198 (write_case) Don't increment cases_analyzed.
2200 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2202 Continue reforming procedure execution. In this phase, move
2203 procedure.c and procedure.h from src to src/data. Update
2204 makefiles and #includes accordingly.
2206 * procedure.c: Moved here from src/.
2208 * procedure.h: Moved here from src/.
2210 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2212 Continue reforming procedure execution. In this phase, get rid of
2213 many global variables, consolidating procedure execution in
2214 procedure.c. Encapsulate transformations in new "struct
2215 trns_chain". Also, change implementation of N OF CASES, FILTER,
2216 and PROCESS IF from special cases to transformations.
2218 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2221 * dictionary.c: (global variable default_dict) Move to
2224 * variable.h: (TRNS_*) Move to transformations.h.
2225 (struct transformation) Move to transformations.c.
2227 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2229 * sys-file-reader.c: Fixed invalid read problems.
2231 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2233 * storage-stream.c: Add missing function comments.
2235 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2237 Continue reforming procedure execution. In this phase, add some
2238 new, needed functionality to storage-stream.
2240 * storage-stream.c: (storage_source_decapsulate) New function.
2242 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2244 * variable.c (width_to_bytes): Declarations must precede
2245 statements for C90 compliance.
2247 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2249 * data-out.c, data-in.c, variable.c, variable.h: New functions
2250 copy_mangle and copy_demangle for reading/writing cases; emulates the
2251 way SPSS deals with strings > 255 bytes.
2253 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2254 subtype 14 needed for strings longer than 255 bytes.
2256 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2257 instead of literal values. Also fixed some constness issues.
2259 * format.h: Constness
2261 * sfm-private.h: Renamed the case_size identifier, since I discovered
2262 that SPSS's respect for this variable is very nominal.
2264 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2266 Change case limit type from int to size_t.
2268 * dictionary.c: (struct dictionary) Change type of case_limit
2270 (dict_get_case_limit) Change return type.
2271 (dict_set_case_limit) Change parameter type.
2273 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2275 * variable.h: (struct variable) Rename `reinit' member as `leave'
2276 and invert sense. Fix up all references.
2278 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2280 Continue reforming procedure execution. In this phase, break
2281 procedure.c into multiple files.
2283 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2285 * case-sink.c: New file.
2287 * case-sink.h: New file.
2289 * case-source.c: New file.
2291 * case-source.h: New file.
2293 * storage-stream.c: New file.
2295 * storage-stream.h: New file.
2297 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2299 * variable.h: (struct variable) Remove `init' member and all
2300 references to it from other files. It was initialized in several
2301 places, but nothing really ever used it for anything worthwhile.
2302 Thanks to Jason Stover for pointing out how confusing this
2305 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2307 Continue reforming error message support. In this phase, get rid
2308 of message "titles" and put the message text in `struct error'.
2309 Now `struct error' encapsulates a message more properly.
2311 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2312 Format message ourselves.
2314 * data-in.c: (vdls_error) Ditto.
2316 * por-file-reader.c: (error) Ditto.
2318 * sys-file-reader.c: (corrupt_msg) Ditto.
2320 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2322 GNU standards require "file name" instead of "filename" in
2323 documentation. It's nice for our code to follow the convention
2326 * casefile.c: (struct casefile) Rename `filename' member to
2327 `file_name'. Updated all references.
2329 * file-name.c: [!unix] (struct file_identity) Rename
2330 normalized_filename member to normalized_file_name. Updated all
2333 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2335 We don't really support anything but Unix-like environments well,
2336 so we might as well de-obfuscate by writing directory and path
2337 separators explicitly.
2339 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2341 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2342 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2344 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2347 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2349 GNU standards require "file name" instead of "filename" in
2350 documentation. It's nice for our code to follow the convention
2353 * filename.c: Rename to file-name.c.
2355 * filename.h: Rename to file-name.h. Update all inclusions.
2356 Update header guards.
2358 * automake.mk: Update file names.
2360 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2362 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2364 (fn_basename) Removed (dead code).
2365 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2366 (fn_special_p) Renamed fn_is_special(), all references updated.
2367 (fn_exists_p) Renamed fn_exists(), all references updated.
2369 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2371 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2372 Also, now it only tilde-expands file names, not paths.
2373 (fn_search_path) Tilde-expand one directory at a time.
2375 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2377 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2378 get rid of non-Unixlike version of the code, which has probably
2380 (fn_prepend_dir) Removed (dead code).
2382 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2383 (macro PATH_SEPARATOR_STRING) New.
2384 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2386 Continue reforming error message support. In this phase, we get
2387 rid of VM() and the other msg() support for "verbosity", replacing
2388 it by a new function verbose_msg().
2390 * filename.c: (fn_search_path) Use verbose_msg() instead of
2393 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2395 * sfm-private.h: Get rid of #defines after #error, which makes no
2398 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2400 Get rid of our own int32 type in favor of the standard int32_t
2403 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2406 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2408 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2410 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2412 Remove ill-considered file routines that are no longer used.
2414 * filename.c: (fn_open_ext) Removed.
2415 (fn_close_ext) Removed.
2417 * filename.h: (struct file_ext) Removed.
2419 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2421 * variable.c (var_is_valid_name): Move declarations before code
2424 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2426 * filename.ch (fn_interp_vars): Fixed small buglet.
2428 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2430 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2431 so as to modify the string inline. Thus makeing it easier to
2432 destroy the results when no longer needed.
2434 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2436 * category.c (cat_stored_values_destroy): Fixed memory leak.
2438 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2440 Add some missing frees. Thanks to John Darrington for reporting
2443 * any-writer.c (any_writer_close): Free writer.
2445 * any-reader.c (any_reader_close): Free reader.
2447 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2449 * por-file-reader.c: (error) Mark as NO_RETURN.
2451 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2453 * settings.c: Changed default value of scompress to true.
2455 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2457 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2459 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2462 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2463 here from pref.h.orig.
2465 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2467 * sys-file-reader.c: Fixed bug reading compressed files.
2469 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2471 * Numerous renames. See src/ChangeLog for details.
2473 * Moved files from src directory