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
11 (var_create): Use the new functions for default variable
13 (var_default_formats): New function.
14 (var_default_measure): New function.
15 (var_default_alignment): New function.
17 * format.h (macro FMT_MAX_NUMERIC_WIDTH): New macro.
19 2008-02-09 Ben Pfaff <blp@gnu.org>
21 Add a couple of extensions to GET DATA TYPE=TXT. Patch #6412.
22 Thanks to John Darrington for review.
24 * data-in.c (data_in): Add new argument to designate the last
25 column of the data field being parsed, for use in error messages.
28 * data-parser (struct data_parser): New member `quote_escape'.
29 (data_parser_create): Initialize quote_escape.
30 (data_parser_set_quotes): New function.
31 (cut_field): Support escaped quotes.
32 (parse_delimited_span): Ditto.
33 (parse_delimited_no_span): Ditto.
35 * get-data.c (parse_get_txt): Support ESCAPE extension subcommand
36 in enhanced mode. Only support multiple quote characters in
39 2008-02-06 John Darrington <john@darrington.wattle.id.au>
41 psql-reader.c psql-reader.h: Read more than one tuple at
42 once. Fix bug reading a query which returns no data. Fix bug
43 when transformation followed a reader.
44 Ask the server for the number of records in the query, for the
47 2008-02-05 John Darrington <john@darrington.wattle.id.au>
49 psql-reader.c: So yesterday they release postgresql 8.3.0
50 which has money represented with 64 bits. They must get
53 2008-02-02 John Darrington <john@darrington.wattle.id.au>
55 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
56 for reviewing this code.
58 2008-02-02 Ben Pfaff <blp@gnu.org>
62 * sys-file-reader.c (read_variable_record): Allow missing values
63 to be specified on long string variables, but warn about them
64 because PSPP does not yet support them.
65 (read_extension_record): Ignore extension records 20 and 21, which
66 PSPP does not yet support.
67 (read_header): Fix error message when floating-point format cannot
70 2008-02-01 Ben Pfaff <blp@gnu.org>
72 Patch #6386. Thanks to John Darrington for review and for the
73 updates to gnumeric-reader.c.
75 * dictionary.c (make_hinted_name): New function.
76 (make_numeric_name): New function.
77 (dict_make_unique_var_name): New function.
79 * gnumeric-reader.c (devise_name): Removed.
80 (munge_name): Removed.
81 (gnumeric_open_reader): Use new function
82 dict_make_unique_var_name.
84 * short-names.c (set_var_short_name_suffix): Use new function
87 2008-01-19 John Darrington <john@darrington.wattle.id.au>
89 * settings.c settings.h: Moved static variables into a
90 single struct. Renamed functions to have a settings_ prefix.
92 2008-01-14 John Darrington <john@darrington.wattle.id.au>
94 * data-out.c (output_AHEX): Corrected number of bytes in
95 call to output_hex. Closes bug #22011
97 2008-01-02 John Darrington <john@darrington.wattle.id.au>
99 * variable.c variable.h: Replaced var_get_value_name with
100 var_append_value_name which doesn't use any static data.
101 Thanks to Ben for review.
103 2007-12-07 Ben Pfaff <blp@gnu.org>
107 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
108 made from an empty casereader produce a casegrouper with no
109 groups, instead of one with one group that has no cases.
111 * casereader.c (casereader_is_empty): New function.
113 2007-12-06 Ben Pfaff <blp@gnu.org>
117 * sys-file-reader.c (read_display_parameters): Handle variable
118 display parameters record with only 2 data items per variable.
119 Reported by Guido Gay <gay@irer.it>.
121 2007-12-04 Ben Pfaff <blp@gnu.org>
123 * identifier.c (lex_id_match_n): New function.
124 (lex_id_match): Reimplement in terms of lex_id_match_n.
126 2007-11-24 Ben Pfaff <blp@gnu.org>
128 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
131 2007-11-24 Ben Pfaff <blp@gnu.org>
133 Fix warning reported when reading back system files that include
134 very long string variables. Thanks to Guido Gay <gay@irer.it> for
137 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
138 was written where `j' was meant.
140 2007-11-18 Ben Pfaff <blp@gnu.org>
142 Properly write variables that include a range of missing values to
143 system files. Thanks to Guido Gay <gay@irer.it> for reporting
146 * sys-file-writer.c (write_variable): Correctly calculate
147 n_missing_values field when writing variables that include a range
150 2007-11-10 Ben Pfaff <blp@gnu.org>
152 Cleanups and bug fixes devised while writing up documentation.
155 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
157 * dict-class.c: New file.
158 (dict_class_from_id): Move here.
159 (dict_class_to_name): Move here.
161 * dict-class.h: New file.
162 (enum dict_class): Move here. Change from consecutive integers to
163 consecutive bits, to make testing for any of multiple values
164 easier. Add new DC_ALL constant.
166 * dictionary.c (struct dictionary): Change `case_limit' from
167 size_t to casenumber.
168 (dict_get_vars): Make final argument an enum dict_class.
169 (dict_get_vars_mutable): Ditto.
170 (dict_get_case_limit): Change return value to casenumber.
171 (dict_set_case_limit): Change final argument to a casenumber.
172 (dict_unset_split_var): Add assertion. Rephrase slightly.
173 (dict_set_label): Use xstrndup to simplify.
175 * format.c (fmt_step_width): AHEX format also needs 2-byte
177 (fmt_set_style): Simplify assertions.
179 * missing-values.c (mv_add_num_range): Rename mv_add_range.
180 Simplify implementation.
181 (mv_has_value): Simplify implementation.
182 (mv_pop_value): Remove the first value, not the last, to avoid
183 having GET followed by SAVE reverse the order of missing values.
184 (mv_peek_value): Rename mv_get_value. Simplify assertion.
185 (mv_has_range): Simplify implementation.
186 (mv_peek_range): Rename mv_get_range.
187 (can_resize_string): Removed.
188 (mv_is_resizable): Use value_is_resizable.
189 (mv_resize): Use value_resize.
191 * short-names.h (SHORT_NAME_LEN): Move here.
193 * val-type.h: New file, for definitions related to type and width
194 of abstract values. Before, these definitions were mixed among
195 those related to "union value" and those related to variables.
196 (macro SYSMIS): Move here.
197 (macro LOWEST): Move here.
198 (macro HIGHEST): Move here.
199 (macro MAX_STRING); Move here.
200 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
201 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
202 All references updated.
203 (val_type_is_valid): New function. Replaces var_type_is_valid.
204 All references updated.
205 (val_type_from_width): New function. Replaces
206 var_type_from_width. All references updated.
208 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
210 (val_labs_can_set_width): Use value_is_resizable.
211 (val_labs_add): Simply return false if the value labels set is too
212 wide, instead of having undefined behavior.
213 (val_labs_replace): Ditto.
214 (val_labs_replace): Ditto.
215 (val_labs_first): Set iterator to null if iteration is complete.
216 (val_labs_first_sorted): Ditto.
217 (val_labs_done): Become a no-op if the iterator is null.
219 * value.c (value_is_resizable): New function.
220 (value_resize): New function.
222 * variable.c (var_get_dict_class): New function.
224 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
227 2007-11-08 Ben Pfaff <blp@gnu.org>
229 * data-in.c: Make formatted data parsing locale-independent.
230 (parse_number): Use c_strtod instead of strtod, to avoid
231 locale-specific behavior.
234 2007-11-06 Ben Pfaff <blp@gnu.org>
236 Patch #6256: add support for binary, 360 file formats. Reviewed
239 * data-in.c (struct data_in): Add `encoding' member.
240 (data_in): Add `encoding' parameter, and re-encode the data passed
241 in where appropriate. Update all callers to pass it in.
242 (parse_A): Implement EBCDIC recoding wart described in manual.
243 (parse_AHEX): Implement EBCDIC recoding.
245 * data-out.c (data_out_legacy): New function.
246 (data_out): Make into a wrapper around data_out_legacy.
248 * file-handle-def.c (struct file_handle): New member `encoding'.
249 (fh_create_file): Set encoding.
250 (fh_default_properties): Set default encoding.
251 (fh_get_legacy_encoding): New function.
253 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
254 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
255 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
256 (struct fh_properties): New member `encoding'.
258 2007-11-05 Ben Pfaff <blp@gnu.org>
260 Patch #6258. Reviewed by John Darrington.
262 * file-handle-def.c (fh_lock): Add comment that TYPE should be
263 marked with N_() in the caller. Added these markings to each
264 caller too. Should make i18n easier.
265 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
267 2007-11-03 Ben Pfaff <blp@gnu.org>
269 Allow output files to overwrite input files (bug #21280).
271 * file-handle-def.c: Separate locking of files for input and for
272 output, to allow both to take place at once. Also, distinguish a
273 file handle from the identity of the underlying file, because the
274 identity of a file changes over time and the file handle can't
275 represent two different identities.
276 (struct file_handle): Remove `next', `open_cnt', `deleted',
277 `type', `open_mode', `aux', `identity' members. Change `id' from
278 char array to char *. Add `ref_cnt' member.
279 (file_handle_from_ll) New function.
280 (file_handles) Removed.
281 (named_handles) New variable.
282 (fh_init) Initialize named_handles.
283 (fh_done) Remove name from all named_handles.
284 (free_handle) Updated for modified struct file_handle.
285 (unname_handle) New function.
286 (fh_ref) New function.
287 (fh_from_file_name) Removed.
288 (fh_unref) New function.
289 (fh_unname) New function.
290 (fh_from_id) Rewritten.
291 (create_handle) Updated for modified struct file_handle.
292 (fh_inline_file) Increment the handle's ref count.
293 (fh_create_file) Updated for modified struct file_handle.
294 (fh_create_scratch) Ditto.
299 (fh_is_open) Removed.
300 (fh_get_id) Updated for modified struct file_handle.
301 (fh_get_default_handle) Increment the handle's ref count.
302 (fh_set_default_handle) Handle ref counts.
303 (struct fh_lock) New structure.
304 (locks) New static var.
305 (fh_lock) New function.
306 (fh_unlock) New function.
307 (fh_lock_get_aux) New function.
308 (fh_lock_set_aux) New function.
309 (fh_is_locked) New function.
310 (make_key) New function.
311 (free_key) New function.
312 (compare_fh_locks) New function.
313 (hash_fh_lock) New function.
315 * file-handle-def.h (enum fh_access) New enum.
317 * file-name.c: Made file_identity the same in all supported
319 (struct file_identity): New `name' member.
320 (fn_get_identity): For a file that doesn't exist, get the
321 dev/inode of its directory plus its name. If even the directory
322 doesn't exist, just use its name. Merge the Windows
323 implementation into the Unix one.
324 (fn_compare_file_identities): Rewritten. Merge the Windows
325 implementation into the Unix one.
326 (fn_hash_identity): New function.
328 * make-file.c (struct replace_file): New structure.
329 (all_files): New static var.
330 (replace_file_start): New function.
331 (replace_file_commit): New function.
332 (replace_file_abort): New function.
333 (free_replace_file): New function.
334 (unlink_replace_files): New function.
336 * por-file-reader.c (struct pfm_reader): Add `lock' member.
337 (close_reader): Unlock file.
338 (pfm_open_reader): Lock file.
340 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
342 (pfm_open_writer): Lock file and prepare for its replacement.
343 (close_writer): Unlock file.
345 * scratch-handle.h (struct scratch_handle): Add unique_id so that
346 different generations of a scratch file can be distinguished.
348 * scratch-reader.c (scratch_reader_open): Verify that the file is
351 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
352 remove scratch_handle member.
353 (scratch_writer_open): Lock handle. Prepare to replace handle
354 data, instead of doing it immediately.
355 (scratch_writer_casewriter_destroy): Replace handle data and
358 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
359 (sfm_open_reader): Lock file.
360 (close_reader): Unlock file.
362 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
364 (sfm_open_writer): Lock file and prepare for its replacement.
365 (close_writer): Unlock file.
367 2007-11-02 Ben Pfaff <blp@gnu.org>
369 * data-out.c (output_number): Use isfinite (from C99) instead of
370 gsl_isfinite, since we now have universal support for it in
372 (output_infinite): Ditto.
373 [!HAVE_ROUND] (round): Remove definition, since we now have a
374 replacement in gnulib.
376 2007-10-27 John Darrington <john@darrington.wattle.id.au>
378 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
379 wasn't initialised correctly. Closes bugs #21330 and 21397
382 2007-10-08 Ben Pfaff <blp@gnu.org>
384 Fix bug #21280. Thanks to John Darrington for review.
386 * file-name.c (create_stream): New function.
388 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
389 file handle before creating the file, to ensure that we don't
390 truncate a file that we're reading. Make code easier to read by
393 * sys-file-write.c (sfm_open_writer): Ditto.
395 2007-10-01 Ben Pfaff <blp@gnu.org>
397 Fix bug #21192. Thanks to John Darrington for review.
399 * casereader.c (casereader_read): Decrement case_cnt before
400 calling the casereader's "read" member function, so that we
401 interact properly with lazy_casereader.
403 * datasheet.c: Add regression test for above bug fix.
404 (clone_datasheet): New function.
405 (lazy_callback): New function.
406 (check_datasheet_casereader): New function.
407 (check_datasheet): Check datasheet contents are reported correctly
408 through an ordinary casereader and a lazy casereader.
409 (clone_model): Use clone_datasheet.
411 2007-09-24 Ben Pfaff <blp@gnu.org>
413 Patch #6210. Reviewed by John Darrington.
414 * settings.c: Drop "long view", which is not needed any longer.
415 (static var long_view): Removed.
416 (force_long_view): Removed.
417 (get_viewwidth): Removed.
418 (init_viewport): Removed long_view reference.
420 2007-09-19 John Darrington <john@darrington.wattle.id.au>
422 * settings.c settings.h: Changed viewport's length and width to be
423 owned by the user interface which uses the data library. This allows
424 better abstraction, and makes dynamically adjustable dimensions easier.
426 2007-09-18 Ben Pfaff <blp@gnu.org>
428 * procedure.c (proc_extract_active_file_data): New function.
430 * lazy-casereader.h: New file.
432 * lazy-casereader.c: New file.
434 * casereader.c (casereader_dynamic_cast): New function.
436 2007-09-14 Ben Pfaff <blp@gnu.org>
438 * dictionary.c (dict_clone): Copy case indexes from cloned
439 dictionary. Fixes bug #21061. Reviewed by John Darrington.
441 2007-09-13 John Darrington <john@darrington.wattle.id.au>
443 * value.c value.h (create_value): New function.
445 2007-09-12 Ben Pfaff <blp@gnu.org>
447 Make it clear that translator casereader and casewriter translate
448 functions are supposed to destroy their input case.
450 * casereader-translator.c (struct casereader_translator): Change
451 input case parameter of translate member function from const
452 struct ccase * to struct ccase *.
453 (casereader_create_translator): Ditto, for translate parameter.
455 * casewriter-translator.c (struct casewriter_translator): Ditto.
456 (casewriter_create_translator): Ditto.
458 2007-08-27 John Darrington <john@darrington.wattle.id.au>
460 * sys-file-reader.c (read_display_parameters): Force display width
461 to 8, if the sys file says 0 (like SPSS does).
463 2007-08-12 Ben Pfaff <blp@gnu.org>
465 * dictionary.c (dict_dump): New function.
467 2007-08-12 Ben Pfaff <blp@gnu.org>
469 Drop dict_compactor in favor of using the new struct case_map.
471 * dictionary.c (struct copy_map): Removed.
472 (struct dict_compactor): Removed.
473 (dict_make_compactor): Removed.
474 (dict_compactor_compact): Removed.
475 (dict_compactor_destroy): Removed.
477 * procedure.c (struct data_set): Change `compactor' member to be a
479 (proc_open): Use case_map_to_compact_dict instead of
481 (proc_casereader_read): Use case_map_execute instead of
482 dict_compactor_compact.
483 (proc_commit): Use case_map_destroy instead of
484 dict_compactor_destroy.
486 * scratch-writer.c (struct scratch_writer): Change `compactor'
487 member to be a struct case_map *.
488 (scratch_writer_open): Use case_map_to_compact_dict instead of
490 (scratch_writer_casewriter_write): Use case_map_execute instead of
491 dict_compactor_compact.
493 2007-08-12 Ben Pfaff <blp@gnu.org>
495 * automake.mk: Add case-map.c, case-map.h.
497 * case-map.c: New file.
499 * case-map.h: New file.
501 2007-08-12 Ben Pfaff <blp@gnu.org>
503 * dictionary.c (dict_compact_values): Don't delete scratch
504 variables as well as compacting case indexes. Update all callers.
505 (dict_get_compacted_value_cnt): Rename dict_count_values and
506 change interface. Update all callers.
507 (dict_get_compacted_value_cnt): Remove.
508 (dict_compacting_would_shrink): Remove.
509 (dict_compacting_would_change): Remove.
510 (dict_make_compactor): Add new parameter. Update all callers.
512 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
513 instead of dict_count_values, changing an O(N) operation into
516 2007-08-12 Ben Pfaff <blp@gnu.org>
518 * casereader.c (casereader_read): Don't require cases read by a
519 casereader to be exactly the expected size: as long as they're big
522 2007-08-12 Ben Pfaff <blp@gnu.org>
524 Make casewriters keep track of the number of `union value's in
525 each case. This is useful for two reasons: casewriter_write can
526 then check that the case being written is large enough, and later
527 recipients of the casewriter can determine the size of the case.
529 * casewriter-translator.c (casewriter_create_translator): Add
532 * casewriter.c (struct casewriter): Add value_cnt member.
533 (casewriter_write): Check that the case passed in is big enough.
534 (casewriter_get_value_cnt): New function.
535 (casewriter_create): Add value_cnt parameter.
537 2007-08-09 Ben Pfaff <blp@gnu.org>
539 Fix bug reported by Jason Stover.
540 * settings.c: Move get_termcap_viewport and in particular the
541 #include for <curses.h> to the end of the file. curses.h
542 redefines bool on some systems (e.g. OpenBSD), which causes
543 disagreement between uses of bool before its inclusion and
545 Tested by Jason Stover.
547 2007-07-29 Ben Pfaff <blp@gnu.org>
549 Provisional fix for bug #18692 and bug #20161. Reviewed by John
552 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
553 argument (never "rb" or "wb") because SUSv3 says that only those
554 modes are defined, and glibc in fact rejects other modes.
556 Open portable files with fn_open so that they can be read from
557 pipes. Fix missing fh_close call to go along with fh_open.
558 Report an error if the file close reports an error.
559 * por-file-reader.c (close_reader): New function.
560 (por_file_casereader_destroy): Use close_reader.
561 (pfm_open_reader): Open file with fn_open.
563 2007-07-28 Ben Pfaff <blp@gnu.org>
565 Make PSPP able to read all the portable files I could find on the
566 web. Thanks to John Darrington for review. Bug #17620.
567 * por-file-reader.c (struct pfm_reader): New member `line_length'.
568 (error): Print file offset in hexadecimal.
569 (warning): New function.
570 (advance): Treat lines less than 80 bytes long as padded to 80
572 (pfm_open_reader): Call read_documents if we find an "E" record.
573 (convert_format): Convert invalid formats to the default format
574 instead of aborting reading the file.
575 (read_variables): Rename duplicate variable names instead of
576 aborting reading the file.
577 (read_value_label): Allow string variables of different widths to
578 be assigned value labels in the same record. Replace duplicate
579 value labels instead of aborting.
580 (read_documents): New function.
582 * por-file-writer.c (pfm_open_writer): Call write_documents if the
583 dictionary has documents.
584 (write_documents): New function.
586 2007-07-25 Ben Pfaff <blp@gnu.org>
588 Fix bugs related to bug #17213.
590 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
591 since the former is what config.h has. Include the needed ncurses
593 (static var echo) Rename to `do_echo' because the original name is
594 the same as an ncurses identifier.
595 (get_termcap_viewport) Use error instead of msg.
597 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
598 (fn_close): Don't close stdin, stdout, stderr.
600 2007-07-26 John Darrington <john@darrington.wattle.id.au>
602 * procedure.c procedure.h: Added callbacks which get invoked whenever
603 a dataset's transformation chain changes.
605 2007-07-24 Ben Pfaff <blp@gnu.org>
608 * sys-file-writer.c (write_variable_display_parameters): Use new
609 var_default_display_width function to choose display width of
610 segments after the first one in a given variable.
611 * variable.c (var_create): Use var_default_display_width to pick
612 new variable's display width.
613 (var_default_display_width): New function.
614 Reviewed by John Darrington.
616 2007-07-24 Ben Pfaff <blp@gnu.org>
619 * por-file-writer.c (write_variables): Write weight variable.
620 Reviewed by John Darrington.
622 2007-07-23 Ben Pfaff <blp@gnu.org>
624 Improvements to system file reader and writer.
626 First, move all detailed knowledge of very long strings into
627 sys-file-private.[ch], so that this nasty stuff can be isolated.
629 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
630 (EFFECTIVE_VLS_CHUNK): New macro.
631 (min_int): New function.
632 (max_int): New function.
633 (sfm_width_to_bytes): Rewrite.
634 (sfm_width_to_octs): New function.
635 (sfm_segment_alloc_width): New function.
636 (sfm_segment_alloc_bytes): New function.
637 (sfm_segment_used_bytes): New function.
638 (sfm_segment_offset): New function.
639 (sfm_segment_effective_offset): New function.
640 (sfm_dictionary_to_sfm_vars): New function.
642 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
643 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
644 (struct sfm_var): New structure.
646 Next, improvements to the system file reader.
648 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
649 type casenumber. Added `version_major', `version_minor',
652 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
653 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
654 Change `case_cnt' to type casenumber. Removed `has_vls'.
655 (struct sfm_var): Removed.
656 (sfm_open_reader): Don't warn on wrong case size if the file was
657 written by SPSS 13, which tends to get it wrong. Use
658 sfm_dictionary_to_sfm_vars.
659 (read_header): Always output system file info.
660 (read_variable_record): Simplify code for reading missing values.
661 (read_machine_int32_info): Save version numbers from system file
662 into info struct passed as new argument.
663 (read_long_string_map): Restructured to use new sys-file-private
665 (read_value_labels): Use size_overflow_p.
666 (sys_file_casereader_read): Get rid of distinction between fast
667 and slow paths. Use information provided by sys-file-primate's
668 struct sfm_var to simplify code.
669 (skip_whole_strings): New function.
670 (read_int32): Renamed read_int. Changed return value to int.
672 (read_flt64): Renamed read_float. Changed return value to
673 double. Updated all callers.
674 (int32_to_native): Removed. Changed callers to use
676 (flt64_to_double): Removed. Changed callers to use float_convert.
678 Finally, get rid of int32, flt64 terminology and types in system
679 file writer. The former wasn't very useful since a POSIX "int"
680 can hold the whole range of int32 and we generally didn't have a
681 need for it to be exactly-32-bits, just at-least-32-bits. The
682 latter was inconvenient because we had to assume that it could be
683 different from double and thereby convert special values SYSMIS,
684 HIGHEST, LOWEST to and from it in multiple places. Instead, now
685 we just use "int" and "double" in most places, and do conversions,
686 if necessary, very close to where we do I/O. This change meant
687 that the writer code couldn't represent records in the file as C
688 structs any longer, but that's no great loss. The code actually
689 seems to be more readable without them.
691 Simplify the compression buffering code: only buffer as much as
692 necessary, which is no more than eight 8-byte units at any given
695 * sys-file-writer.c (typedef flt64): Removed.
696 (macro second_lowest_flt64): Removed.
697 (struct sysfile_header): Removed.
698 (struct sysfile_variable): Removed.
699 (struct sfm_writer): Removed `needs_translation', `has_vls',
700 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
701 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
702 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
703 compression buffering by `opcodes', `opcode_cnt', `data',
704 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
705 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
706 data writer functions instead of structures.
707 (calc_oct_idx): New function.
708 (write_header): Use simple data writer functions instead of
710 (write_format_spec): Renamed write_format. New argument.
711 (write_variable_continuation_records): New function.
712 (write_variable): Use simple data writer functions instead of
713 structures. Use write_variable_continuation_records. Write
714 entire very long string instead of requiring caller to understand
716 (write_value_labels): Use simple data writer functions instead of
718 (write_documents): Ditto.
719 (write_variable_display_parameters): Use sys-file-private
720 functions to simplify. Use simple data writer functions instead
722 (write_vls_length_table): Use simple data writer functions instead
724 (write_longvar_table): Ditto.
725 (write_rec_7_34): Break into new functions
726 write_integer_info_record, write_float_info_record. Use simple
727 data writer functions instead of structures.
728 (buf_write): Removed.
729 (append_string_max): Removed.
730 (ensure_buf_space): Removed.
731 (sys_file_casewriter_write): Get rid of the distinction between
732 fast and slow paths, which didn't seem to be too useful. Use new
733 functions write_case_uncompressed, write_case_compressed.
734 (put_instruction): Removed.
735 (put_element): Removed.
736 (write_compressed_data): Removed.
737 (close_writer): Use flush_compressed. Only write case count to
738 system file if it will fit in the field.
739 (write_case_compressed): New function.
740 (write_case_uncompressed): New function.
741 (flush_compressed): New function.
742 (put_cmp_opcode): New function.
743 (put_cmp_number): New function.
744 (write_int): New function.
745 (convert_double_to_output_format): New function.
746 (write_float): New function.
747 (write_value): New function.
748 (write_string): New function.
749 (write_bytes): New function.
750 (write_zeros): New function.
751 (write_spaces): New function.
753 Reviewed by John Darrington.
755 2007-07-22 Ben Pfaff <blp@gnu.org>
757 Don't try to write very long strings to portable files. The
758 format does not support it.
760 * por-file-writer.c (MAX_POR_WIDTH): New macro.
761 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
762 (write_format): Add arg to take width to resize format to.
763 (write_value): Limit width of value written to MAX_POR_WIDTH.
764 (write_variables): Limit width of variable and its output formats
766 Reviewed by John Darrington.
768 2007-07-22 Ben Pfaff <blp@gnu.org>
770 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
771 local variable instead of literal 5.
772 Reviewed by John Darrington.
774 2007-07-22 Ben Pfaff <blp@gnu.org>
776 Fix problems with uniqueness of short names in system files with
777 very long string variables. Now a variable may have multiple
780 * automake.mk (src_data_libdata_a_SOURCES): Add new files
781 short-names.c, short-names.h.
783 * dictionary.c (dict_clone): Clone all the short names.
784 (compare_strings): Move into short-names.c.
785 (hash_strings): Ditto.
786 (set_var_short_name_suffix): Ditto.
787 (dict_assign_short_names): Ditto, rename short_names_assign,
788 change to assign all short names.
790 * por-file-writer.c (write_variables): Use short_names_assign
791 instead of dict_assign_short_names.
793 * short-names.c: New file.
795 * short-names.h: New file.
797 * sys-file-private.c (sfm_width_to_segments): New function.
799 * sys-file-reader.c (read_long_var_name_map): Save and restore all
800 the short names, not just the first one.
802 * sys-file-writer.c (cont_var_name): Removed.
803 (sfm_open_writer): Use short_names_assign instead of
804 dict_assign_short_names. Use unique short names assigned by
805 short_names_assign instead of those generated by cont_var_name.
807 * variable.c (struct variable): Remove `short_name' member,
808 replace by `short_names' and `short_name_cnt'.
809 (var_create) Initialize new members.
810 (var_get_short_name_cnt): New function.
811 (var_get_short_name): Now takes an index argument. Changed most
813 (var_set_short_name): Ditto.
814 (var_clear_short_name): Renamed var_clear_short_names, changed to
815 clear all short names.
817 Reviewed by John Darrington.
819 2007-07-22 Ben Pfaff <blp@gnu.org>
821 * variable.c (var_set_width): Use new var_set_width function.
823 * missing-values.c (mv_n_values): Drop assertion, which was not
826 * format.c (fmt_default_for_width): New function.
827 (fmt_resize): New function.
829 Reviewed by John Darrington.
831 2007-07-18 John Darrington <john@darrington.wattle.id.au>
833 * datasheet.c (datasheet_delete_columns): Added assertion to check
834 we're not deleting outside the range of the sheet.
837 * dictionary.c dictionary.h variable.c: Added the ability for string
838 variables to be resized.
840 * vardict.h: Added some prototypes (moved from dictionary.h) as
841 these should only be called by variable.c
844 2007-07-14 John Darrington <john@darrington.wattle.id.au>
846 * sfm-reader.c: Respect case_cnt field in file header.
848 2007-07-01 John Darrington <john@darrington.wattle.id.au>
850 * transformation.c transformation.h (trns_chain_execute): Changed the
851 signature (Patch #6057)
853 2007-06-10 Ben Pfaff <blp@gnu.org>
855 * casereader-filter.c (casereader_filter_destroy): Make sure to
856 write all the remaining excluded cases to the casewriter, if any.
858 * caseinit.c (init_list_destroy): Rewrite.
859 (init_list_clear): Ditto.
861 * casegrouper.c (casegrouper_get_next_group): Always set *reader
862 to null when returning false.
864 2007-06-06 Ben Pfaff <blp@gnu.org>
866 Actually implement the new procedure code and adapt all of its
867 clients to match. Also adapt all of the other case sources and
868 sinks in the tree and their clients to use the
869 casereader/casewriter infrastructure.
871 * automake.mk: Add and remove files.
873 * any-reader.c: Change into a casereader.
874 * por-file-reader.c: Ditto.
875 * scratch-reader.c: Ditto.
876 * sys-file-reader.c: Ditto.
878 * any-writer.c: Change into a casewriter.
879 * por-file-writer.c: Ditto.
880 * scratch-writer.c: Ditto.
881 * sys-file-writer.c: Ditto.
883 * procedure.c: Change to use casereader, casewriter, caseinit, and
884 other new infrastructure.
886 * scratch-handle.c: Adapt to new infrastructure.
888 * case-sink.c: Removed, now dead code.
889 * case-sink.h: Ditto.
890 * case-source.c: Ditto.
891 * case-source.h: Ditto.
892 * casefile-factory.c: Ditto.
893 * casefile-private.h: Ditto.
896 * casefilter.c: Ditto.
897 * casefilter.h: Ditto.
900 * fastfile-factory.c: Ditto.
901 * fastfile-factory.h: Ditto.
902 * storage-stream.c: Ditto.
903 * storage-stream.h: Ditto.
905 2007-06-06 Ben Pfaff <blp@gnu.org>
909 * automake.mk: Add new files.
911 * datasheet.c: New file.
913 * datasheet.h: New file.
915 2007-06-06 Ben Pfaff <blp@gnu.org>
917 Until now, the procedure code has provided a case to the
918 case_source, which has filled in the data values that come from
919 the active file. "Left" data values that don't come from the
920 active file naturally stay the same from case to case, because the
921 procedure code keeps using that same case.
923 One of the compromises that comes with the new procedure code is
924 that the active file allocates and provides its own case, which
925 the procedure code then has to resize to provide room for any
926 other variables that should go in the case and then fill in the
927 values of "left" variables. Then, when we're done with that case,
928 we have to save the values of "left" variables to copy into the
929 next case read from the active file.
931 The caseinit code helps with this.
933 * automake.mk: Add new files.
935 * caseinit.c: New file.
937 * caseinit.h: New file.
939 2007-06-06 Ben Pfaff <blp@gnu.org>
941 * value.h (value_cnt_from_width): New function.
943 * variable.c (var_get_value_cnt): Use new function.
945 2007-06-06 Ben Pfaff <blp@gnu.org>
947 Add casegrouper, to allow cases read from a given casereader to be
948 broken into groups, each of which has its own casereader.
949 Generally cases are grouped based on having equal values for some
952 * automake.mk: Add new files.
954 * casegrouper.c: New file.
956 * casegrouper.h: New file.
958 2007-06-06 Ben Pfaff <blp@gnu.org>
960 Add interface to lexicographical ordering of cases.
962 * automake.mk: Add new files.
964 * case-ordering.c: New file.
966 * case-ordering.h: New file.
968 2007-06-06 Ben Pfaff <blp@gnu.org>
970 Add casereaders and casewriters, the basis of the new data processing
971 implementation. A casereader is a uniform interface to reading cases
972 from a data source; a casewriter is a uniform interface to writing
973 cases to a data sink.
975 * automake.mk: Add new files.
977 * casereader-filter.c: New file.
979 * casereader-provider.h: New file.
981 * casereader-translator.c: New file.
983 * casereader.c: New file.
985 * casereader.h: New file.
987 * casewriter-provider.h: New file.
989 * casewriter-translator.c: New file.
991 * casewriter.c: New file.
993 * casewriter.h: New file.
995 2007-06-06 Ben Pfaff <blp@gnu.org>
997 "casewindow" data structure that extends the deque (from libpspp)
998 of cases with the ability to dump cases to disk if we get too many
1001 * automake.mk: Add new files.
1003 * casewindow.c: New file.
1005 * casewindow.h: New file.
1007 2007-06-06 Ben Pfaff <blp@gnu.org>
1009 sparse_cases data structure that augments a sparse_array of cases
1010 with the ability to dump cases to disk if we get too many cases in
1013 * automake.mk: Add new files.
1015 * sparse-cases.c: New file.
1017 * sparse-cases.h: New file.
1019 2007-06-06 Ben Pfaff <blp@gnu.org>
1021 Adds a low-level on-disk case array data structure.
1023 * automake.mk: Add new files.
1025 * case-tmpfile.c: New file.
1027 * case-tmpfile.h: New file.
1029 2007-06-06 Ben Pfaff <blp@gnu.org>
1031 In a couple of places we calculate the maximum number of cases to
1032 keep in memory based on the user-defined workspace. Enable
1033 centralizing the calculation through a new function.
1035 * settings.c (get_workspace_cases): New function.
1037 2007-06-06 Ben Pfaff <blp@gnu.org>
1039 The casenumber type is defined in transformations.h, but case.h is
1040 a more sensible place. Move it.
1042 * case.h (CASENUMBER_MAX): New macro.
1043 (typedef casenumber): Move here, from transformations.h.
1045 2007-06-03 Ben Pfaff <blp@gnu.org>
1047 Slightly generalize case_to_values and case_from_values functions.
1049 * case.c (case_to_values): Rename case_copy_out, change interface.
1050 (case_from_values): Rename case_copy_in, change interface.
1052 * fastfile.c (fastfilereader_get_next_case): Update caller.
1053 (write_case_to_disk): Ditto.
1055 2007-06-02 Ben Pfaff <blp@gnu.org>
1057 Clean up after a forgotten part of patch #5829.
1059 * casedeque.h: Remove unused file.
1061 * automake.mk: Remove casedeque.h from sources.
1063 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1065 * category.c: Removed redundant #include
1067 2007-05-06 Ben Pfaff <blp@gnu.org>
1069 Abstract the documents within a dictionary a little better.
1070 Thanks to John Darrington for suggestion, initial version, and
1071 review. Patch #5917.
1073 * dictionary.c (struct dictionary): Change `documents' member from
1074 char * to struct string.
1075 (dict_clear): Destroy struct string.
1076 (dict_get_documents): Convert struct string to char *.
1077 (dict_set_documents): Set struct string. Pad to 80-character
1079 (dict_clear_documents): New function.
1080 (dict_add_document_line): New function.
1081 (dict_get_document_line_cnt): New function.
1082 (dict_get_document_line): New function.
1084 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1086 * sys-file-reader.c (read_documents): Use new document functions.
1088 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1090 * sys-file-reader.c: When reading a system file which has no
1091 long name table, automatically create one where the long names
1092 are the lower case versions of the short names.
1094 2007-04-22 Ben Pfaff <blp@gnu.org>
1096 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1097 dict_clear will free most data related to the dictionary.
1098 dict_clear does a decent job, except that dict_set_split_vars on
1099 some systems won't actually free the dict's "split" member.
1100 Instead, it'll allocate a 1-byte region. Fix this.
1102 * value.c (value_copy): New function.
1103 (value_set_missing): Ditto.
1105 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1107 * Deleted existing category.h and moved cat-routines.h into
1108 category.h Encapsulated struct cat_vals better.
1110 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1112 * sys-file-reader.c: When reading a system file which has no
1113 long name table, automatically create one where the long names
1114 are the lower case versions of the short names.
1116 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1118 * sys-file-reader.c: Some versions of Other Software seem to
1119 produce system files with string variables' measure set to
1120 zero. We'll assume these are supposed to be nominal variables.
1122 2007-03-30 Ben Pfaff <blp@gnu.org>
1124 * procedure.c: Adapt to new deque data structure.
1126 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1127 Ben Pfaff <blp@gnu.org>
1129 * file-name.c: Mingw compatibility fixes.
1130 (fn_search_path): Use ISSLASH instead of comparing against '/'
1132 (fn_dir_name): Use dir_name from gnulib.
1133 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1134 (fn_get_identity): Use GetFullPathName instead of canonicalize
1135 from gnulib, because the latter does not fully support
1136 Windows-style path names. Use this implementation based on the
1137 detected presence of Windows instead of the absence of Unix, since
1138 the new implementation is Windows-specific.
1139 (fn_compare_file_identities): In Windows implementation, compare
1140 names case-insensitively.
1142 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1144 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1145 <stdlib.h> to have the same effect.
1147 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1149 * por-file-reader.c: Add missing _() around messages.
1151 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1153 * make-file.c: Include "mkstemp.h", without which linking on
1156 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1158 Reduce platform dependence.
1160 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1161 Everywhere we using this, we really should have just depended on
1162 the shell to expand tildes.
1163 (fn_search_path): Simplify, given that we don't do tilde expansion
1165 (fn_normalize): Removed. Caller changed to use the canonicalize
1167 (fn_get_cwd): Removed. Only user was fn_normalize.
1168 (fn_is_absolute): Really only test for absolute names.
1169 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1171 (fn_readlink): Removed, as it was only used fn_normalize.
1172 (fn_exists): Assume the stat function is available; gnulib does.
1173 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1175 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1177 * dictionary.c (dict_create_vector_assert): New function.
1179 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1181 * file-name.c (fn_normalize): Correct name of function
1182 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1183 for pointing this out.
1185 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1187 We are using a single member in struct file_handle, the "name"
1188 field, for more than one purpose. When it begins with '"', it's a
1189 file name; otherwise, it's a token that can be used to identify
1190 it. When that assertion fires, it's because we searched for the
1191 name case-sensitively as a file name (so that there was no match),
1192 and then we try to insert it case-insensitively as a token, which
1193 fails because duplicates aren't allowed.
1195 Solution: break the two purposes into two separate fields. This
1196 fixes the problem and likely makes the code easier to read too.
1198 Fixes bug #18922. Thanks to John Darrington for bug report and
1201 * file-handle-def.c (struct file_handle): New `id' member.
1202 (fh_from_name): Rename fh_from_id. Update all callers.
1203 (create_handle): New `id' parameter. Update all callers.
1204 (fh_create_file): Ditto.
1205 (fh_get_id): New function.
1207 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1209 * case.c (case_is_null): Change return type to bool.
1211 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1215 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1216 reference counting to share data. This makes it easy for
1217 valgrind, etc. to find accesses to cases that have been destroyed
1218 but have been kept around by another user's ref-count. This often
1219 happens when the data set is small enough to find in memory; if a
1220 bigger data set that would overflow to disk were used, then data
1221 corruption would occur.
1223 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1227 * case.c (case_unshare): Make it check internally whether the
1228 ref_cnt is greater than 1, so that the callers don't have to.
1229 Update callers not to check.
1231 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1233 Before, I was thinking that I might want to get rid of reference
1234 counting at some point. Now, I'm pretty sure that it's here to
1235 stay. Thus, because we have to store the value_cnt anyway for
1236 reference-counted cases, we might as well expose it to users.
1238 * case.c (case_get_value_cnt): New function.
1239 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1240 resize case if its size actually changed.
1242 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1243 instead of peeking inside struct case directly.
1244 (casefile_append): Ditto.
1246 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1248 Get rid of the inlines for the case functions, which made the
1249 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1250 the inlines didn't speed up "make check" at all, which is not a
1251 perfect benchmark but seems indicative.)
1253 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1254 definitions. Remove some assertions on nonnull pointers that were
1255 redundant with a pointer dereference soon after in the function.
1257 (struct case_data): Move definition here from case.h.
1261 (case_data_wr): Ditto.
1263 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1265 * automake.mk: Add casedeque.h to sources.
1267 * casedeque.h: New file.
1269 * procedure.c: (struct dataset) Change lag_count, lag_head,
1270 lag_queue member into single struct casedeque member. Update all
1271 users to use the casedeque instead.
1274 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1276 * procedure.c: Simplify lagged cases interface. Updated all
1277 clients--well, the only client--to use the simplified interface.
1278 (dataset_n_lag) Removed.
1279 (dataset_set_n_lag) Removed.
1280 (dataset_need_lag) New function.
1282 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1284 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1285 whenever appropriate, but only when the dataset/dictionary is in a
1288 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1290 * dictionary.c dictionary.h : Added callbacks for change of filter and
1291 split variables. Refactored some code to ensure that callbacks get
1292 invoked when appropriate.
1294 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1295 replace_dict callback occurs when permanent_dict replaces the current
1298 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1300 * dictionary.c dictionary.h : Added callback for when the weight
1301 variable of a dictionary changes.
1303 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1305 * dictionary.c dictionary.h : Added replace_source and replace_dict
1306 callbacks, and functions to deal with them.
1308 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1310 Simplify missing value handling.
1312 * missing-values.h (enum mv_class): New type.
1313 (enum mv_type): Moved definition into missing-values.c and renamed
1314 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1315 enums. Updated all uses.
1316 (struct missing_values): Changed type of `type' from `enum
1317 mv_type' to `int' because the definition is no longer exposed.
1319 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1320 parameter. Update all callers.
1321 (mv_is_num_missing): Ditto.
1322 (mv_is_str_missing): Ditto.
1323 (mv_is_value_user_missing): Removed. Changed callers to use
1324 mv_is_value_missing.
1325 (mv_is_num_user_missing): Removed. Changed callers to use
1327 (mv_is_str_user_missing): Removed. Changed callers to use
1329 (mv_is_value_system_missing): Removed. Changed callers to use
1330 mv_is_value_missing.
1331 (mv_set_type): Removed. Changed callers to use mv_clear.
1332 (mv_clear): New function.
1334 * variable.c (var_is_value_missing): Add new enum mv_class
1335 parameter. Update all callers.
1336 (var_is_num_missing): Ditto.
1337 (var_is_str_missing): Ditto.
1338 (var_is_value_user_missing): Removed. Changed callers to use
1339 var_is_value_missing.
1340 (var_is_num_user_missing): Removed. Changed callers to use
1342 (var_is_str_user_missing): Removed. Changed callers to use
1344 (var_is_value_system_missing): Removed. Changed callers to use
1345 var_is_value_missing.
1347 * casefilter.c (struct casefilter): Use enum mv_class in place of
1349 (casefilter_variable_missing): Adapt to new member.
1350 (casefilter_create): Change signature to take enum mv_class,
1353 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1355 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1357 * case-sink.c case-sink.h procedure.c procedure.h
1358 storage-stream.c: Now uses the factory.
1360 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1362 Make it possible to pull cases from the active file with a
1363 function call, instead of requiring indirection through a callback
1366 * case-source.h (struct case_source_class): Change ->read function
1367 to return a single case, instead of calling a callback function
1368 for each case. Change ->destroy function to return an error
1371 * case-source.c (free_case_source): Pass along the value returned
1372 by the case_source ->destroy function.
1374 * procedure.c (struct write_case_data): Removed.
1375 (struct dataset): Added some members to track procedure state.
1376 (procedure): Optimize the trivial case at this level.
1377 (internal_procedure): Re-implement in terms of proc_open,
1378 proc_read, proc_close.
1379 (proc_open) New function.
1380 (proc_read) New function.
1381 (proc_close) New function.
1382 (write_case) Moved into proc_read.
1383 (close_active_file) Moved closing of data source into proc_close.
1385 * storage-source.c: Rewrote to conform with modified
1386 case_source_class interface.
1388 * transformations.c (trns_chain_execute): Added argument to allow
1389 starting execution from an arbitrary transformation. Updated
1392 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1394 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1396 * sys-file-reader.c (read_display_parameters): Don't assume that
1397 MEASURE_* and ALIGN_* have the same values found in system files.
1399 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1401 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1402 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1403 the value of n_MEASURES, which was off by 1 (at least from my
1406 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1408 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1409 callbacks which are invoked when the dictionary or its
1410 variables are changed.
1412 * missing-values.c missing-values.h value-labels.c: Tidied up
1413 consistency checks, and made some of them return false
1414 instead of assert-failing.
1416 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1418 * calendar.c (calendar_days_in_month): New function.
1420 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1422 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1423 the value label's width.
1425 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1427 * sfm-private.h: Move contents into sys-file-writer.c, which is
1428 the only remaining user. Removed Borland C++-specific directives.
1430 * sys-file-reader.c: Clean up and rewrite entire file. The
1431 rewritten version is simpler and better abstracted, and should be
1432 easier to maintain and extend. It avoids using structures to read
1433 file data, which is prone to padding variations among compilers.
1434 It should also handle non-IEEE 754 system files, although I
1435 haven't been able to find any. It has been tested against many
1436 .sav files obtained from the Web and found to produce the same
1437 results as the earlier version of the code, or in some cases
1438 improved results. It is more tolerant of format variations found
1441 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1442 member, putting an enum integer_format in its place. New member
1443 `float_format'. Changed `compressed' member to type bool.
1445 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1447 * dictionary.c (dict_delete_consecutive_vars): New function.
1449 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1451 * file-name.c (fn_search_path): Remove prefix arg that was unused
1452 by any caller. Updated all callers.
1454 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1456 * format.c (fmt_dollar_template): Use user's decimal point
1457 character. Add assertion.
1459 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1461 * format.c (fmt_dollar_template): New function, based on
1462 dollar_format_template from var-type-dialog.c.
1464 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1466 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1469 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1471 Finish converting struct variable to an opaque type. In this
1472 phase, we add remaining setter and getter functions, convert the
1473 remaining PSPP code to use them, and do a bunch of cleanup. The
1474 resulting changes are pervasive but mostly trivial, and only the
1475 notable changes are logged.
1477 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1480 * case.c (case_data): Renamed case_data_idx.
1481 (case_num): Renamed case_num_idx.
1482 (case_str): Renamed case_str_idx.
1483 (case_data_rw): Renamed case_data_rw_idx.
1485 * case.h (case_data): New function with old name and an interface
1486 that takes a variable instead of an index, which is easier to
1487 use. Updated all callers to use the new interface, or to use the
1488 new *_idx function (see above).
1491 (case_data_rw): Ditto.
1493 * category.c (cat_stored_values_destroy): Changed interface to
1494 take a struct cat_vals * instead of a struct variable *.
1496 * dictionary.c (dict_clone): Use new vector_clone function.
1497 (dict_clear) Use new var_destroy function.
1498 (add_var) New function.
1499 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1500 (dict_create_var_assert) Rewrite in terms of add_var.
1501 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1502 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1503 (dict_lookup_var) Use new var_create, var_destroy functions.
1504 (dict_contains_var) Rewrite in terms of new vardict functionality.
1505 (set_var_dict_index) New function.
1506 (set_var_case_index) New function.
1507 (reindex_vars) New function.
1508 (dict_delete_var) Rewrite in terms of new vardict functionality.
1509 (dict_reorder_var) Ditto.
1510 (dict_reorder_vars) Ditto.
1511 (rename_var) New function.
1512 (dict_rename_var) Use rename_var.
1513 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1514 (dict_get_compacted_idx_to_fv) Rename
1515 dict_get_compacted_dict_index_to_case_index, update callers.
1516 (dict_create_vector) Use new vector_create function.
1517 (dict_clear_vectors) Use new vector_destroy function.
1518 (set_var_short_name_suffix) Move here from variable.c, renamed
1519 from var_set_short_name_suffix, make static, update caller.
1521 * sys-file-private.c: New file.
1522 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1523 width_to_bytes, update callers.
1525 * sys-file-private.h: New file. Later it will supplant
1526 sfm-private.h; for now it supplements it.
1527 (macro MIN_VERY_LONG_STRING) New macro.
1528 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1530 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1531 MAX_LONG_STRING was used before.
1533 * sys-file-writer.c: Ditto.
1535 * value-labels.c: Change the paradigm here to be that a null
1536 pointer is OK for a struct val_labs * in most cases; it just
1537 represents an empty set of value labels.
1538 (val_labs_copy) A copy of a null set is a null set.
1539 (val_labs_count) A null set has 0 labels.
1540 (val_labs_replace) Change return type to void. Rewrite for
1542 (val_labs_find) A null set does not contain the value.
1543 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1544 transposed argument order, updated all callers.
1546 * value.c: New file.
1547 (value_dup) Moved here from variable.c.
1548 (compare_values) Ditto.
1551 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1552 (macro MAX_LONG_STRING) Removed, because it was only interesting
1553 for system files, not for general code.
1554 (macro MAX_VERY_LONG_STRING) Ditto.
1555 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1556 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1558 * vardict.h: New file, for an interface between variables and
1561 * variable.c: A lot of functions were moved around, for better
1563 (struct variable) Move definition here, from variable.h.
1564 (var_type_adj) Removed--makes i18n hard.
1565 (var_type_noun) Ditto.
1566 (var_create) New function.
1567 (var_clone) New function.
1568 (var_destroy) New function.
1569 (var_set_name) Assert that variable is not in a dictionary.
1570 (compare_var_names) Rename compare_vars_by_name and fix a couple
1571 of callers who thought the args were strings.
1572 (hash_var_name) Rename hash_var_by_name.
1573 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1574 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1575 (var_is_very_long_string) Removed, because it was only interesting
1576 to system file code.
1577 (var_set_missing_values) Allow the argument to be the wrong width,
1578 as long as we can resize it. Simplify callers who were doing the
1579 resizing themselves.
1580 (var_get_value_labels) New function.
1581 (var_has_value_labels) New function.
1582 (var_set_value_labels) New function.
1583 (alloc_value_labels) New function.
1584 (var_add_value_label) New function.
1585 (var_replace_value_label) New function.
1586 (var_clear_value_labels) New function.
1587 (var_lookup_value_label) New function.
1588 (var_get_value_name) Moved here from variable.c, renamed from
1589 var_get_value_name, transposed argument order, updated all
1591 (var_to_string) Moved here, from variable-label.c.
1592 (var_set_leave) New function.
1593 (var_get_leave) New function.
1594 (var_must_leave) New function.
1595 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1596 set_var_short_name_suffix.
1597 (var_get_dict_index) New function.
1598 (var_get_case_index) New function.
1599 (var_get_obs_vals) New function.
1600 (var_set_obs_vals) New function.
1601 (var_has_obs_vals) New function.
1602 (var_get_vardict) New function.
1603 (var_set_vardict) New function.
1604 (var_has_vardict) New function.
1605 (var_clear_vardict) New function.
1606 (value_dup) Moved to value.c.
1607 (compare_values) Ditto.
1610 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1611 (enum ALPHA) Rename VAR_STRING, update all users.
1613 * vector.c: New file.
1614 (struct vector) Moved here, from variable.h.
1615 (check_widths) New function.
1616 (vector_create) New function.
1617 (vector_clone) New function.
1618 (vector_destroy) New function.
1619 (vector_get_name) New function.
1620 (vector_get_var) New function.
1621 (vector_get_var_cnt) New function.
1622 (compare_vector_ptrs_by_name) New function.
1624 * vector.h: New file.
1626 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1628 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1629 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1632 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1634 * value-labels.c (destroy_atoms): New function.
1635 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1637 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1639 Thanks to Jason Stover for pointing out this problem.
1641 * data-out.c (output_number): Use gsl_finite from GSL, which is
1642 portable, instead of isfinite, which is not.
1645 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1647 * variable.c variable.h (value_dup): New function.
1649 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1651 Start converting struct variable to an opaque type. In this
1652 phase, we add a bunch of setter and getter functions and convert
1653 most of the PSPP code to use them. The resulting changes are
1654 pervasive but mostly trivial, and only the notable changes are
1657 * format.c (fmt_equal): New function.
1659 * variable.c (var_type_is_valid): New function.
1660 (measure_is_valid) Moved here, from format.c.
1661 (alignment_is_valid) Moved here, from format.c.
1662 (var_get_name) New function.
1663 (var_set_name) New function.
1664 (width_to_type) New function.
1665 (var_get_type) New function.
1666 (var_get_width) New function.
1667 (var_set_width) New function.
1668 (var_is_numeric) New function.
1669 (var_is_alpha) New function.
1670 (var_is_short_string) New function.
1671 (var_is_long_string) New function.
1672 (var_is_very_long_string) New function.
1673 (var_get_missing_values) New function.
1674 (var_set_missing_values) New function.
1675 (var_clear_missing_values) New function.
1676 (var_has_missing_values) New function.
1677 (var_is_value_missing) New function.
1678 (var_is_num_missing) New function.
1679 (var_is_str_missing) New function.
1680 (var_is_value_user_missing) New function.
1681 (var_is_num_user_missing) New function.
1682 (var_is_str_user_missing) New function.
1683 (var_is_value_system_missing) New function.
1684 (var_get_print_format) New function.
1685 (var_set_print_format) New function.
1686 (var_get_write_format) New function.
1687 (var_set_write_format) New function.
1688 (var_set_both_formats) New function.
1689 (var_get_label) New function.
1690 (var_set_label) New function.
1691 (var_clear_label) New function.
1692 (var_has_label) New function.
1693 (var_get_measure) New function.
1694 (var_set_measure) New function.
1695 (var_get_display_width) New function.
1696 (var_set_display_width) New function.
1697 (var_get_alignment) New function.
1698 (var_set_alignment) New function.
1699 (var_get_value_cnt) New function.
1700 (var_get_leave) New function.
1701 (var_get_short_name) New function.
1703 * variable.h: (struct variable) Removed "type" and "nv" members;
1704 they are now computed from "width" where needed.
1706 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1708 * missing-values.c (mv_resize): Don't write beyond end of the
1709 allocated buffer when resizing a long string.
1711 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1713 Clean up identifier code: don't require identifier enumerations to
1714 be in a particular order; make better use of string library;
1715 expose less of the internals.
1717 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1718 change interface. Updated all callers.
1719 (lex_id_match) Change interface to use struct substring, update
1721 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1722 (global array keywords[]) Make static, change form. Update all
1723 users to use lex_id_name instead.
1724 (lex_is_keyword) New function.
1725 (lex_id_to_token) Change interface to use struct substring, update
1727 (lex_id_name) New function.
1729 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1730 lex_is_keyword instead.
1731 (T_LAST_KEYWORD) Removed.
1732 (T_N_KEYWORDS) Removed.
1734 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1736 * format.c: (fmt_date_template) Distinguish characters for which a
1737 space is output and any date delimiter is allowed on input, from
1738 those for which a space is output and only a space is allowed on
1739 input. The former is represented by X, the latter by a space.
1740 Also, drop distinction between h and H, changing the former to the
1743 * data-in.c: Completely rewrite internals to conform to SPSS input
1744 formats as closely as possible.
1745 (data_in) Changed external interface by replacing the structure
1746 that was used as a single argument by a set of arguments. Updated
1748 (data_in_finite_line) Removed. Converted all callers to use plain
1750 (data_in_get_integer_format) New function.
1751 (data_in_set_integer_format) New function.
1752 (data_in_get_float_format) New function.
1753 (data_in_set_float_format) New function.
1755 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1756 (struct data_in) Removed.
1758 * data-out.c: (output_date) Drop each component from the input as
1759 it is output, to allow us to drop the distinction between h (a
1760 count of hours) and H (the hour of day) template characters.
1761 Also, handle new X template character.
1762 (output_scientific) Follow more rational rule on when to drop
1763 fraction introduced between SPSS 13 and 15. Updated test case to
1766 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1768 Fix buffer overflow reported by John Darrington.
1770 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1771 realize that DIGITS is a count of nibbles, not of bytes.
1773 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1775 * calendar.c (calendar_offset_to_gregorian) Also return the
1776 year-of-day. Change callers to new interface.
1778 * data-out.c: Completely rewrite internals to conform to SPSS
1779 output formats as completely as possible.
1780 (data_out) Change interface to put input parameters before output
1781 parameters, for consistency with the style I now prefer. Update
1783 (data_out_get_integer_format) New public function.
1784 (data_out_set_integer_format) New public function.
1785 (data_out_get_float_format) New public function.
1786 (data_out_set_float_format) New public function.
1788 * data-out.h: New file. Move prototype for data_out here, from
1791 * format.c: (fmt_step_width) Use equality comparison instead of
1792 bitwise and, for clarity.
1793 (fmt_is_string) Ditto.
1794 (fmt_input_to_output) Fix categories that are translated to F
1797 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1799 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1800 casefile-private.h: Added casefilter to assist commands with missing
1803 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1805 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1807 * settings.c: (route_errors_to_terminal) New variable.
1808 (route_errors_to_listing) New variable.
1809 (get_error_routing_to_terminal) New function.
1810 (set_error_routing_to_terminal) New function.
1811 (get_error_routing_to_listing) New function.
1812 (set_error_routing_to_listing) New function.
1814 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1816 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1818 * format.c: Completely rewrite, to achieve better abstraction.
1819 Rewrite all references to formats in other files.
1821 * format.def: Rewrite and reorganize.
1823 * settings.c: Move everything related to custom currency formats
1824 into format.[ch], changing them in form, so as to group related
1825 code and definitions better. Changed all references to use the
1827 (static var decimal) Removed.
1828 (static var grouping) Removed.
1829 (static var cc) Removed.
1830 (get_decimal) Removed.
1831 (set_decimal) Removed.
1832 (get_grouping) Removed.
1833 (set_grouping) Removed.
1837 * settings.h: (macro CC_CNT) Removed.
1838 (macro CC_WIDTH) Removed.
1839 (struct custom_currency) Removed.
1841 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1843 * data-in.c (data_in): Use switch statement instead of table, to
1844 avoid dependence on the order of the FMT_* enums.
1846 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1848 * data-out.c: (num_to_string) Removed, because it was dead code.
1850 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1852 * data-in.c (parse_trailer): Fix error message.
1854 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1856 * format.c (fmt_is_binary): New function.
1858 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1860 * procedure.c procedure.h: Encapsulated the static data into a single
1863 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1865 * casefile.c (casereader_read_xfer): Always initialize the case,
1866 even on an error condition.
1868 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1870 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1871 entire function useless.
1873 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1875 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1876 casereader method casereader_clone.
1878 * procedure.c pransformations.h: Introduced new type casenum_t
1880 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1882 * variable.c: (width_to_bytes) Rephrase code for clarify.
1884 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1886 * format.c: (fmt_type_from_string) New function.
1887 (fmt_to_string) Include decimals in output if the format has
1888 decimals, even if the format type does not. This way, we can
1889 accurately reproduce incorrect formats in user output.
1890 (check_common_specifier) Make the check for a bad format type an
1891 assertion, so we get bug reports if they show up. Fix message.
1892 Check for decimal places with a format type that doesn't allow
1894 (check_input_specifier) Remove check for FMT_X, which has been
1896 (check_output_specifier) Ditto.
1898 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1900 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1901 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1902 (enum fmt_parse_flags) Removed.
1904 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1906 * casefile.c casefile.h: Converted to an abstract base class.
1907 * casefile-private.h fastfile.c fastfile.h: New files.
1908 * automake.mk procedure.c scratch-writer.c storage-stream.c
1910 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1912 * procedure.c (internal_procedure): Create sink_case with only as
1913 many values as the compacted dictionary.
1915 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1917 Remove "debugging" code that caused plenty of false positives and
1920 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1922 * case.c: Remove all references to `this' member.
1924 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1926 Fix link error noted by Jason Stover.
1928 * storage-stream.c: Include <assert.h>.
1930 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1932 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1933 ALL) and additional underlying system file issues.
1935 Thanks to John Darrington for review.
1937 First problem: var_hash points to variables not owned by the
1938 sys-file-reader, which the caller may free or modify. Use an
1939 array of sfm_vars instead, as done earlier (e.g. CVS version
1942 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1943 members and remove all code that references it. Add vars, var_cnt
1944 members. Remove fix_specials member, which was unused.
1945 (struct sfm_var) Remove name member, which was unused.
1946 (sfm_close_reader) Free vars member instead of var_hash.
1947 (compare_var_shortnames) Removed.
1948 (hash_var_shortname) Removed.
1949 (sfm_open_reader) Fill out vars array.
1950 (compare_var_index) Removed.
1951 (sfm_read_case) Use vars instead of var_hash.
1953 Second problem: we're confused about when we actually have very
1954 long strings, causing us to choose incorrectly between slow path
1955 and fast path in sfm_read_case.
1957 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1958 have very long strings, not when we have long variable names,
1959 which is an unrelated feature.
1961 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1963 * variable.h: Move var_set and variable parsing declarations to
1964 new header, src/language/lexer/variable-parser.h. Modified lots
1965 of files to include the new header.
1967 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1969 * value-labels.c (value_to_string): When there's no value label,
1970 format the variable according to its print format, instead of
1971 always effectively using A or F format.
1973 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1975 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1978 * format.c: Removed tortological assertion.
1980 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1982 Reform string library.
1984 * file-name.c (fn_interp_vars): Change interface to take a
1985 substring as input. Updated all users.
1987 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1989 * format.c (measure_is_valid): Really return false when m >=
1992 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1994 Implement random access to casefiles, for use in GUI.
1996 * casefile.c: (struct casereader) Add `random', `file_ofs',
1997 `buffer_ofs' members.
1998 (casefile_get_random_reader) New function.
1999 (read_open_file) Break part into new function
2000 seek_and_fill_buffer().
2001 (fill_buffer) Update buffer_ofs, file_ofs.
2002 (casereader_seek) New function.
2004 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2006 * settings.c: Added call to i18n{done, init}.
2008 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
2010 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
2013 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
2015 * casefile.c: Convert many uses of `int' to `bool'.
2017 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
2019 * transformations.c (trns_chain_destroy): Destroy chain's trns
2020 member, to fix memory leak.
2022 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
2024 * storage-stream.c (storage_source_decapsulate): Destroy case
2025 source to fix memory leak.
2027 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
2029 * scratch-reader.c (scratch_reader_read_case): Copy into existing
2030 case passed as argument instead of initializing the argument as a
2031 case. Fixes memory leak that showed up in
2032 tests/command/aggregate.sh with scratch files.
2034 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
2036 * procedure.c (proc_done): Destroy default_dict, to fix memory
2039 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2041 Simplify procedure_with_splits().
2043 * procedure.c (struct split_aux_data): Removed case_count member.
2044 (procedure_with_splits) Don't initialize case_count.
2045 (split_procedure_case_func) Check whether prev_case is null
2046 instead of case_count.
2047 (split_procedure_end_func) Ditto.
2049 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2051 * case.c (case_move): Do nothing if dst and src are the same
2053 (case_try_create) Merge two similar cases.
2054 (case_copy) Unshare only if data must be actually copied.
2056 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2058 * data-in.c data-out.c dictionary.c sys-file-reader.c
2059 sys-file-writer.c variable.c variable.h: Reworked very long string
2060 support for better encapsulation.
2062 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2064 * value-labels.c (val_labs_can_set_width): New function.
2065 (val_labs_set_width) Clear labels if increasing width to long
2067 (val_labs_destroy) Remove unneeded test for null.
2069 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2071 * value-labels.h: Remove unneeded dependency on variable.h.
2073 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2075 Get rid of `char *c' member in union value, for cleanliness.
2077 * value.h: (union value) Remove `c' member.
2079 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2081 Make dictionary compacting functions a little more general.
2083 * sys-file-writer.c (sfm_open_writer): Use
2084 dict_compacting_would_change().
2085 (does_dict_need_translation) Removed.
2087 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2089 Make dictionary compacting functions a little more general.
2091 * dictionary.c (dict_needs_compaction): Rename
2092 dict_compacting_would_shrink(). Update all callers.
2093 (dict_compacting_would_change) New function.
2095 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2097 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2098 inverted return value (!).
2100 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2102 Continue reforming procedure execution.
2104 * procedure.c: Search and replace "vfm" by "proc". Notably:
2105 (static var vfm_source) Rename proc_source. Update all
2107 (static var vfm_sink) Rename proc_sink. Update all references.
2109 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2111 Continue reforming procedure execution. In this phase, remove
2112 PROCESS IF, which was deprecated anyway and can be easily
2113 simulated with TEMPORARY followed by SELECT IF.
2115 * procedure.c: (open_active_file) Don't call
2116 add_process_if_trns().
2117 (discard_variables) Get rid of redundant call to
2118 proc_cancel_all_transformations().
2119 (add_process_if_trns) Removed.
2120 (process_if_trns_proc) Removed.
2121 (process_if_trns_free) Removed.
2123 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2125 Continue reforming procedure execution. In this phase, add
2126 `const' to the case passed to procedure()'s callback.
2128 Updated all users of procedure() as well.
2130 * procedure.c: (struct write_case_data) Add "const" to ccase
2131 parameter for case_func member.
2132 (procedure) Add "const" to ccase parameter for proc_func
2134 (multipass_case_func) Make ccase parameter const.
2135 (internal_procedure) Add "const" to ccase parameter for case_func
2137 (split_procedure_case_func) Make ccase parameter const.
2138 (multipass_split_case_func) Make ccase parameter const.
2140 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2142 Continue reforming procedure execution. In this phase, get rid of
2143 the output code for SPLIT FILE groups in procedure.c, which really
2144 shouldn't be doing any output. Move it into the individual
2145 procedures instead. This also adds some flexibility.
2147 Updated many users of procedure_with_splits() and
2148 multipass_procedure_with_splits() to call
2149 output_split_file_values() and to deal with increased use of
2152 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2153 parameter to begin_func member.
2154 (procedure_with_splits) Add "const struct ccase *" parameter to
2155 begin_func parameter. Make ccase parameter const in proc_func
2157 (split_procedure_case_func) Don't dump split file group. Pass
2159 (dump_splits) Moved to language/dictionary/split-file.c as
2160 output_split_file_values().
2161 (struct multipass_split_aux_data) Add "const struct ccase *"
2162 parameter to split_func member.
2163 (multipass_procedure_with_splits) Add "const struct ccase *"
2164 parameter to split_func parameter.
2165 (multipass_split_case_func) Save new SPLIT FILE case before
2167 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2169 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2171 Continue reforming procedure execution. Change
2172 internal_procedure() so that it calls open_active_file() and
2173 close_active_file(), which isolates most of the actual procedure
2176 * procedure.c: (struct write_case_data) Rename `proc_func' member
2177 to `case_func' and update all references.
2178 (procedure) Rewrite as one-line wrapper around
2179 internal_procedure().
2180 (struct multipass_aux_data) New.
2181 (multipass_callback) Renamed multipass_case_func(). Use struct
2182 multipass_aux_data as auxiliary data.
2183 (multipass_end_func) New function.
2184 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2185 that uses multipass_case_func, multipass_end_func.
2186 (internal_procedure) Add `end_func' argument. Move optimization
2187 of trivial case in here. Move call to open_active_file() and
2188 close_active_file() in here. Now assert that vfm_source is
2190 (procedure_with_splits_callback) Rename
2191 split_procedure_case_func().
2192 (split_procedure_end_func) New function.
2193 (multipass_split_callback) Rename multipass_split_case_func.
2194 (multipass_split_end_func) New function.
2195 (discard_variables) No need to test for nonnull vfm_source.
2197 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2199 Continue reforming procedure execution. Get rid of unused member.
2201 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2203 (write_case) Don't increment cases_analyzed.
2205 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2207 Continue reforming procedure execution. In this phase, move
2208 procedure.c and procedure.h from src to src/data. Update
2209 makefiles and #includes accordingly.
2211 * procedure.c: Moved here from src/.
2213 * procedure.h: Moved here from src/.
2215 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2217 Continue reforming procedure execution. In this phase, get rid of
2218 many global variables, consolidating procedure execution in
2219 procedure.c. Encapsulate transformations in new "struct
2220 trns_chain". Also, change implementation of N OF CASES, FILTER,
2221 and PROCESS IF from special cases to transformations.
2223 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2226 * dictionary.c: (global variable default_dict) Move to
2229 * variable.h: (TRNS_*) Move to transformations.h.
2230 (struct transformation) Move to transformations.c.
2232 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2234 * sys-file-reader.c: Fixed invalid read problems.
2236 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2238 * storage-stream.c: Add missing function comments.
2240 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2242 Continue reforming procedure execution. In this phase, add some
2243 new, needed functionality to storage-stream.
2245 * storage-stream.c: (storage_source_decapsulate) New function.
2247 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2249 * variable.c (width_to_bytes): Declarations must precede
2250 statements for C90 compliance.
2252 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2254 * data-out.c, data-in.c, variable.c, variable.h: New functions
2255 copy_mangle and copy_demangle for reading/writing cases; emulates the
2256 way SPSS deals with strings > 255 bytes.
2258 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2259 subtype 14 needed for strings longer than 255 bytes.
2261 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2262 instead of literal values. Also fixed some constness issues.
2264 * format.h: Constness
2266 * sfm-private.h: Renamed the case_size identifier, since I discovered
2267 that SPSS's respect for this variable is very nominal.
2269 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2271 Change case limit type from int to size_t.
2273 * dictionary.c: (struct dictionary) Change type of case_limit
2275 (dict_get_case_limit) Change return type.
2276 (dict_set_case_limit) Change parameter type.
2278 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2280 * variable.h: (struct variable) Rename `reinit' member as `leave'
2281 and invert sense. Fix up all references.
2283 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2285 Continue reforming procedure execution. In this phase, break
2286 procedure.c into multiple files.
2288 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2290 * case-sink.c: New file.
2292 * case-sink.h: New file.
2294 * case-source.c: New file.
2296 * case-source.h: New file.
2298 * storage-stream.c: New file.
2300 * storage-stream.h: New file.
2302 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2304 * variable.h: (struct variable) Remove `init' member and all
2305 references to it from other files. It was initialized in several
2306 places, but nothing really ever used it for anything worthwhile.
2307 Thanks to Jason Stover for pointing out how confusing this
2310 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2312 Continue reforming error message support. In this phase, get rid
2313 of message "titles" and put the message text in `struct error'.
2314 Now `struct error' encapsulates a message more properly.
2316 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2317 Format message ourselves.
2319 * data-in.c: (vdls_error) Ditto.
2321 * por-file-reader.c: (error) Ditto.
2323 * sys-file-reader.c: (corrupt_msg) Ditto.
2325 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2327 GNU standards require "file name" instead of "filename" in
2328 documentation. It's nice for our code to follow the convention
2331 * casefile.c: (struct casefile) Rename `filename' member to
2332 `file_name'. Updated all references.
2334 * file-name.c: [!unix] (struct file_identity) Rename
2335 normalized_filename member to normalized_file_name. Updated all
2338 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2340 We don't really support anything but Unix-like environments well,
2341 so we might as well de-obfuscate by writing directory and path
2342 separators explicitly.
2344 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2346 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2347 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2349 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2352 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2354 GNU standards require "file name" instead of "filename" in
2355 documentation. It's nice for our code to follow the convention
2358 * filename.c: Rename to file-name.c.
2360 * filename.h: Rename to file-name.h. Update all inclusions.
2361 Update header guards.
2363 * automake.mk: Update file names.
2365 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2367 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2369 (fn_basename) Removed (dead code).
2370 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2371 (fn_special_p) Renamed fn_is_special(), all references updated.
2372 (fn_exists_p) Renamed fn_exists(), all references updated.
2374 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2376 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2377 Also, now it only tilde-expands file names, not paths.
2378 (fn_search_path) Tilde-expand one directory at a time.
2380 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2382 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2383 get rid of non-Unixlike version of the code, which has probably
2385 (fn_prepend_dir) Removed (dead code).
2387 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2388 (macro PATH_SEPARATOR_STRING) New.
2389 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2391 Continue reforming error message support. In this phase, we get
2392 rid of VM() and the other msg() support for "verbosity", replacing
2393 it by a new function verbose_msg().
2395 * filename.c: (fn_search_path) Use verbose_msg() instead of
2398 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2400 * sfm-private.h: Get rid of #defines after #error, which makes no
2403 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2405 Get rid of our own int32 type in favor of the standard int32_t
2408 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2411 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2413 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2415 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2417 Remove ill-considered file routines that are no longer used.
2419 * filename.c: (fn_open_ext) Removed.
2420 (fn_close_ext) Removed.
2422 * filename.h: (struct file_ext) Removed.
2424 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2426 * variable.c (var_is_valid_name): Move declarations before code
2429 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2431 * filename.ch (fn_interp_vars): Fixed small buglet.
2433 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2435 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2436 so as to modify the string inline. Thus makeing it easier to
2437 destroy the results when no longer needed.
2439 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2441 * category.c (cat_stored_values_destroy): Fixed memory leak.
2443 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2445 Add some missing frees. Thanks to John Darrington for reporting
2448 * any-writer.c (any_writer_close): Free writer.
2450 * any-reader.c (any_reader_close): Free reader.
2452 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2454 * por-file-reader.c: (error) Mark as NO_RETURN.
2456 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2458 * settings.c: Changed default value of scompress to true.
2460 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2462 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2464 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2467 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2468 here from pref.h.orig.
2470 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2472 * sys-file-reader.c: Fixed bug reading compressed files.
2474 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2476 * Numerous renames. See src/ChangeLog for details.
2478 * Moved files from src directory