1 2008-07-26 John Darrington <john@darrington.wattle.id.au>
3 * case-ordering.c case-ordering.h: Removed the value_cnt
4 associated with the ordering structure.
6 2008-07-16 Jason Stover <jhs@math.gcsu.edu>
8 * category.c (cat_get_category_count): New function.
10 2008-05-15 Ben Pfaff <blp@gnu.org>
14 * data-out.c (output_infinite): Use gsl_isnan instead of isnan,
15 and gsl_isinf instead of isinf, as a stopgap measure for
16 portability until appropriate gnulib modules are available.
18 * por-file-writer.c (format_trig_double): Similarly, use
19 gsl_finite instead of finite.
21 2008-03-18 John Darrington <john@darrington.wattle.id.au>
23 * data-in.c: If category is custom currency, then use
24 numeric format for input.
26 2008-03-06 Ben Pfaff <blp@gnu.org>
28 * dictionary.c (var_name_is_insertable): New function.
29 (make_hinted_name): Don't accept variable names that match PSPP
30 keywords. Thanks to Jason Stover for reporting the problem.
32 2008-03-06 Ben Pfaff <blp@gnu.org>
34 * format-guesser.c (syntax): Require month names to be spelled out
35 as English words, so that single characters that happen to be
36 Roman numerals don't get detected as months. Thanks to John
37 Darrington for reporting this bug.
39 2008-03-04 Ben Pfaff <blp@gnu.org>
41 Patch #6441. Reviewed by John Darrington.
43 * format.c (fmt_fix): New function.
44 (fmt_fix_input): New function.
45 (fmt_fix_output): New function.
47 * format.def: Correct minimum width for DATETIME format. It was
48 7, should have been 17.
50 * automake.mk: Add new files.
52 * format-guesser.c: New file.
54 * format-guesser.h: New file.
56 2008-02-18 Ben Pfaff <blp@gnu.org>
58 Patch #6426. Reviewed by John Darrington.
60 * format.c (min_width): Renamed fmt_min_width and made public.
61 Updated all references.
62 (max_width): Renamed fmt_max_width and made public. Updated all
64 (max_decimals): Renamed fmt_max_decimals and made public. Updated
66 (var_create): Use the new functions for default variable
68 (var_default_formats): New function.
69 (var_default_measure): New function.
70 (var_default_alignment): New function.
72 * format.h (macro FMT_MAX_NUMERIC_WIDTH): New macro.
74 2008-02-09 Ben Pfaff <blp@gnu.org>
76 Add a couple of extensions to GET DATA TYPE=TXT. Patch #6412.
77 Thanks to John Darrington for review.
79 * data-in.c (data_in): Add new argument to designate the last
80 column of the data field being parsed, for use in error messages.
83 * data-parser (struct data_parser): New member `quote_escape'.
84 (data_parser_create): Initialize quote_escape.
85 (data_parser_set_quotes): New function.
86 (cut_field): Support escaped quotes.
87 (parse_delimited_span): Ditto.
88 (parse_delimited_no_span): Ditto.
90 * get-data.c (parse_get_txt): Support ESCAPE extension subcommand
91 in enhanced mode. Only support multiple quote characters in
94 2008-02-06 John Darrington <john@darrington.wattle.id.au>
96 psql-reader.c psql-reader.h: Read more than one tuple at
97 once. Fix bug reading a query which returns no data. Fix bug
98 when transformation followed a reader.
99 Ask the server for the number of records in the query, for the
102 2008-02-05 John Darrington <john@darrington.wattle.id.au>
104 psql-reader.c: So yesterday they release postgresql 8.3.0
105 which has money represented with 64 bits. They must get
108 2008-02-02 John Darrington <john@darrington.wattle.id.au>
110 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
111 for reviewing this code.
113 2008-02-02 Ben Pfaff <blp@gnu.org>
117 * sys-file-reader.c (read_variable_record): Allow missing values
118 to be specified on long string variables, but warn about them
119 because PSPP does not yet support them.
120 (read_extension_record): Ignore extension records 20 and 21, which
121 PSPP does not yet support.
122 (read_header): Fix error message when floating-point format cannot
125 2008-02-01 Ben Pfaff <blp@gnu.org>
127 Patch #6386. Thanks to John Darrington for review and for the
128 updates to gnumeric-reader.c.
130 * dictionary.c (make_hinted_name): New function.
131 (make_numeric_name): New function.
132 (dict_make_unique_var_name): New function.
134 * gnumeric-reader.c (devise_name): Removed.
135 (munge_name): Removed.
136 (gnumeric_open_reader): Use new function
137 dict_make_unique_var_name.
139 * short-names.c (set_var_short_name_suffix): Use new function
142 2008-01-19 John Darrington <john@darrington.wattle.id.au>
144 * settings.c settings.h: Moved static variables into a
145 single struct. Renamed functions to have a settings_ prefix.
147 2008-01-14 John Darrington <john@darrington.wattle.id.au>
149 * data-out.c (output_AHEX): Corrected number of bytes in
150 call to output_hex. Closes bug #22011
152 2008-01-02 John Darrington <john@darrington.wattle.id.au>
154 * variable.c variable.h: Replaced var_get_value_name with
155 var_append_value_name which doesn't use any static data.
156 Thanks to Ben for review.
158 2007-12-07 Ben Pfaff <blp@gnu.org>
162 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
163 made from an empty casereader produce a casegrouper with no
164 groups, instead of one with one group that has no cases.
166 * casereader.c (casereader_is_empty): New function.
168 2007-12-06 Ben Pfaff <blp@gnu.org>
172 * sys-file-reader.c (read_display_parameters): Handle variable
173 display parameters record with only 2 data items per variable.
174 Reported by Guido Gay <gay@irer.it>.
176 2007-12-04 Ben Pfaff <blp@gnu.org>
178 * identifier.c (lex_id_match_n): New function.
179 (lex_id_match): Reimplement in terms of lex_id_match_n.
181 2007-11-24 Ben Pfaff <blp@gnu.org>
183 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
186 2007-11-24 Ben Pfaff <blp@gnu.org>
188 Fix warning reported when reading back system files that include
189 very long string variables. Thanks to Guido Gay <gay@irer.it> for
192 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
193 was written where `j' was meant.
195 2007-11-18 Ben Pfaff <blp@gnu.org>
197 Properly write variables that include a range of missing values to
198 system files. Thanks to Guido Gay <gay@irer.it> for reporting
201 * sys-file-writer.c (write_variable): Correctly calculate
202 n_missing_values field when writing variables that include a range
205 2007-11-10 Ben Pfaff <blp@gnu.org>
207 Cleanups and bug fixes devised while writing up documentation.
210 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
212 * dict-class.c: New file.
213 (dict_class_from_id): Move here.
214 (dict_class_to_name): Move here.
216 * dict-class.h: New file.
217 (enum dict_class): Move here. Change from consecutive integers to
218 consecutive bits, to make testing for any of multiple values
219 easier. Add new DC_ALL constant.
221 * dictionary.c (struct dictionary): Change `case_limit' from
222 size_t to casenumber.
223 (dict_get_vars): Make final argument an enum dict_class.
224 (dict_get_vars_mutable): Ditto.
225 (dict_get_case_limit): Change return value to casenumber.
226 (dict_set_case_limit): Change final argument to a casenumber.
227 (dict_unset_split_var): Add assertion. Rephrase slightly.
228 (dict_set_label): Use xstrndup to simplify.
230 * format.c (fmt_step_width): AHEX format also needs 2-byte
232 (fmt_set_style): Simplify assertions.
234 * missing-values.c (mv_add_num_range): Rename mv_add_range.
235 Simplify implementation.
236 (mv_has_value): Simplify implementation.
237 (mv_pop_value): Remove the first value, not the last, to avoid
238 having GET followed by SAVE reverse the order of missing values.
239 (mv_peek_value): Rename mv_get_value. Simplify assertion.
240 (mv_has_range): Simplify implementation.
241 (mv_peek_range): Rename mv_get_range.
242 (can_resize_string): Removed.
243 (mv_is_resizable): Use value_is_resizable.
244 (mv_resize): Use value_resize.
246 * short-names.h (SHORT_NAME_LEN): Move here.
248 * val-type.h: New file, for definitions related to type and width
249 of abstract values. Before, these definitions were mixed among
250 those related to "union value" and those related to variables.
251 (macro SYSMIS): Move here.
252 (macro LOWEST): Move here.
253 (macro HIGHEST): Move here.
254 (macro MAX_STRING); Move here.
255 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
256 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
257 All references updated.
258 (val_type_is_valid): New function. Replaces var_type_is_valid.
259 All references updated.
260 (val_type_from_width): New function. Replaces
261 var_type_from_width. All references updated.
263 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
265 (val_labs_can_set_width): Use value_is_resizable.
266 (val_labs_add): Simply return false if the value labels set is too
267 wide, instead of having undefined behavior.
268 (val_labs_replace): Ditto.
269 (val_labs_replace): Ditto.
270 (val_labs_first): Set iterator to null if iteration is complete.
271 (val_labs_first_sorted): Ditto.
272 (val_labs_done): Become a no-op if the iterator is null.
274 * value.c (value_is_resizable): New function.
275 (value_resize): New function.
277 * variable.c (var_get_dict_class): New function.
279 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
282 2007-11-08 Ben Pfaff <blp@gnu.org>
284 * data-in.c: Make formatted data parsing locale-independent.
285 (parse_number): Use c_strtod instead of strtod, to avoid
286 locale-specific behavior.
289 2007-11-06 Ben Pfaff <blp@gnu.org>
291 Patch #6256: add support for binary, 360 file formats. Reviewed
294 * data-in.c (struct data_in): Add `encoding' member.
295 (data_in): Add `encoding' parameter, and re-encode the data passed
296 in where appropriate. Update all callers to pass it in.
297 (parse_A): Implement EBCDIC recoding wart described in manual.
298 (parse_AHEX): Implement EBCDIC recoding.
300 * data-out.c (data_out_legacy): New function.
301 (data_out): Make into a wrapper around data_out_legacy.
303 * file-handle-def.c (struct file_handle): New member `encoding'.
304 (fh_create_file): Set encoding.
305 (fh_default_properties): Set default encoding.
306 (fh_get_legacy_encoding): New function.
308 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
309 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
310 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
311 (struct fh_properties): New member `encoding'.
313 2007-11-05 Ben Pfaff <blp@gnu.org>
315 Patch #6258. Reviewed by John Darrington.
317 * file-handle-def.c (fh_lock): Add comment that TYPE should be
318 marked with N_() in the caller. Added these markings to each
319 caller too. Should make i18n easier.
320 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
322 2007-11-03 Ben Pfaff <blp@gnu.org>
324 Allow output files to overwrite input files (bug #21280).
326 * file-handle-def.c: Separate locking of files for input and for
327 output, to allow both to take place at once. Also, distinguish a
328 file handle from the identity of the underlying file, because the
329 identity of a file changes over time and the file handle can't
330 represent two different identities.
331 (struct file_handle): Remove `next', `open_cnt', `deleted',
332 `type', `open_mode', `aux', `identity' members. Change `id' from
333 char array to char *. Add `ref_cnt' member.
334 (file_handle_from_ll) New function.
335 (file_handles) Removed.
336 (named_handles) New variable.
337 (fh_init) Initialize named_handles.
338 (fh_done) Remove name from all named_handles.
339 (free_handle) Updated for modified struct file_handle.
340 (unname_handle) New function.
341 (fh_ref) New function.
342 (fh_from_file_name) Removed.
343 (fh_unref) New function.
344 (fh_unname) New function.
345 (fh_from_id) Rewritten.
346 (create_handle) Updated for modified struct file_handle.
347 (fh_inline_file) Increment the handle's ref count.
348 (fh_create_file) Updated for modified struct file_handle.
349 (fh_create_scratch) Ditto.
354 (fh_is_open) Removed.
355 (fh_get_id) Updated for modified struct file_handle.
356 (fh_get_default_handle) Increment the handle's ref count.
357 (fh_set_default_handle) Handle ref counts.
358 (struct fh_lock) New structure.
359 (locks) New static var.
360 (fh_lock) New function.
361 (fh_unlock) New function.
362 (fh_lock_get_aux) New function.
363 (fh_lock_set_aux) New function.
364 (fh_is_locked) New function.
365 (make_key) New function.
366 (free_key) New function.
367 (compare_fh_locks) New function.
368 (hash_fh_lock) New function.
370 * file-handle-def.h (enum fh_access) New enum.
372 * file-name.c: Made file_identity the same in all supported
374 (struct file_identity): New `name' member.
375 (fn_get_identity): For a file that doesn't exist, get the
376 dev/inode of its directory plus its name. If even the directory
377 doesn't exist, just use its name. Merge the Windows
378 implementation into the Unix one.
379 (fn_compare_file_identities): Rewritten. Merge the Windows
380 implementation into the Unix one.
381 (fn_hash_identity): New function.
383 * make-file.c (struct replace_file): New structure.
384 (all_files): New static var.
385 (replace_file_start): New function.
386 (replace_file_commit): New function.
387 (replace_file_abort): New function.
388 (free_replace_file): New function.
389 (unlink_replace_files): New function.
391 * por-file-reader.c (struct pfm_reader): Add `lock' member.
392 (close_reader): Unlock file.
393 (pfm_open_reader): Lock file.
395 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
397 (pfm_open_writer): Lock file and prepare for its replacement.
398 (close_writer): Unlock file.
400 * scratch-handle.h (struct scratch_handle): Add unique_id so that
401 different generations of a scratch file can be distinguished.
403 * scratch-reader.c (scratch_reader_open): Verify that the file is
406 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
407 remove scratch_handle member.
408 (scratch_writer_open): Lock handle. Prepare to replace handle
409 data, instead of doing it immediately.
410 (scratch_writer_casewriter_destroy): Replace handle data and
413 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
414 (sfm_open_reader): Lock file.
415 (close_reader): Unlock file.
417 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
419 (sfm_open_writer): Lock file and prepare for its replacement.
420 (close_writer): Unlock file.
422 2007-11-02 Ben Pfaff <blp@gnu.org>
424 * data-out.c (output_number): Use isfinite (from C99) instead of
425 gsl_isfinite, since we now have universal support for it in
427 (output_infinite): Ditto.
428 [!HAVE_ROUND] (round): Remove definition, since we now have a
429 replacement in gnulib.
431 2007-10-27 John Darrington <john@darrington.wattle.id.au>
433 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
434 wasn't initialised correctly. Closes bugs #21330 and 21397
437 2007-10-08 Ben Pfaff <blp@gnu.org>
439 Fix bug #21280. Thanks to John Darrington for review.
441 * file-name.c (create_stream): New function.
443 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
444 file handle before creating the file, to ensure that we don't
445 truncate a file that we're reading. Make code easier to read by
448 * sys-file-write.c (sfm_open_writer): Ditto.
450 2007-10-01 Ben Pfaff <blp@gnu.org>
452 Fix bug #21192. Thanks to John Darrington for review.
454 * casereader.c (casereader_read): Decrement case_cnt before
455 calling the casereader's "read" member function, so that we
456 interact properly with lazy_casereader.
458 * datasheet.c: Add regression test for above bug fix.
459 (clone_datasheet): New function.
460 (lazy_callback): New function.
461 (check_datasheet_casereader): New function.
462 (check_datasheet): Check datasheet contents are reported correctly
463 through an ordinary casereader and a lazy casereader.
464 (clone_model): Use clone_datasheet.
466 2007-09-24 Ben Pfaff <blp@gnu.org>
468 Patch #6210. Reviewed by John Darrington.
469 * settings.c: Drop "long view", which is not needed any longer.
470 (static var long_view): Removed.
471 (force_long_view): Removed.
472 (get_viewwidth): Removed.
473 (init_viewport): Removed long_view reference.
475 2007-09-19 John Darrington <john@darrington.wattle.id.au>
477 * settings.c settings.h: Changed viewport's length and width to be
478 owned by the user interface which uses the data library. This allows
479 better abstraction, and makes dynamically adjustable dimensions easier.
481 2007-09-18 Ben Pfaff <blp@gnu.org>
483 * procedure.c (proc_extract_active_file_data): New function.
485 * lazy-casereader.h: New file.
487 * lazy-casereader.c: New file.
489 * casereader.c (casereader_dynamic_cast): New function.
491 2007-09-14 Ben Pfaff <blp@gnu.org>
493 * dictionary.c (dict_clone): Copy case indexes from cloned
494 dictionary. Fixes bug #21061. Reviewed by John Darrington.
496 2007-09-13 John Darrington <john@darrington.wattle.id.au>
498 * value.c value.h (create_value): New function.
500 2007-09-12 Ben Pfaff <blp@gnu.org>
502 Make it clear that translator casereader and casewriter translate
503 functions are supposed to destroy their input case.
505 * casereader-translator.c (struct casereader_translator): Change
506 input case parameter of translate member function from const
507 struct ccase * to struct ccase *.
508 (casereader_create_translator): Ditto, for translate parameter.
510 * casewriter-translator.c (struct casewriter_translator): Ditto.
511 (casewriter_create_translator): Ditto.
513 2007-08-27 John Darrington <john@darrington.wattle.id.au>
515 * sys-file-reader.c (read_display_parameters): Force display width
516 to 8, if the sys file says 0 (like SPSS does).
518 2007-08-12 Ben Pfaff <blp@gnu.org>
520 * dictionary.c (dict_dump): New function.
522 2007-08-12 Ben Pfaff <blp@gnu.org>
524 Drop dict_compactor in favor of using the new struct case_map.
526 * dictionary.c (struct copy_map): Removed.
527 (struct dict_compactor): Removed.
528 (dict_make_compactor): Removed.
529 (dict_compactor_compact): Removed.
530 (dict_compactor_destroy): Removed.
532 * procedure.c (struct data_set): Change `compactor' member to be a
534 (proc_open): Use case_map_to_compact_dict instead of
536 (proc_casereader_read): Use case_map_execute instead of
537 dict_compactor_compact.
538 (proc_commit): Use case_map_destroy instead of
539 dict_compactor_destroy.
541 * scratch-writer.c (struct scratch_writer): Change `compactor'
542 member to be a struct case_map *.
543 (scratch_writer_open): Use case_map_to_compact_dict instead of
545 (scratch_writer_casewriter_write): Use case_map_execute instead of
546 dict_compactor_compact.
548 2007-08-12 Ben Pfaff <blp@gnu.org>
550 * automake.mk: Add case-map.c, case-map.h.
552 * case-map.c: New file.
554 * case-map.h: New file.
556 2007-08-12 Ben Pfaff <blp@gnu.org>
558 * dictionary.c (dict_compact_values): Don't delete scratch
559 variables as well as compacting case indexes. Update all callers.
560 (dict_get_compacted_value_cnt): Rename dict_count_values and
561 change interface. Update all callers.
562 (dict_get_compacted_value_cnt): Remove.
563 (dict_compacting_would_shrink): Remove.
564 (dict_compacting_would_change): Remove.
565 (dict_make_compactor): Add new parameter. Update all callers.
567 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
568 instead of dict_count_values, changing an O(N) operation into
571 2007-08-12 Ben Pfaff <blp@gnu.org>
573 * casereader.c (casereader_read): Don't require cases read by a
574 casereader to be exactly the expected size: as long as they're big
577 2007-08-12 Ben Pfaff <blp@gnu.org>
579 Make casewriters keep track of the number of `union value's in
580 each case. This is useful for two reasons: casewriter_write can
581 then check that the case being written is large enough, and later
582 recipients of the casewriter can determine the size of the case.
584 * casewriter-translator.c (casewriter_create_translator): Add
587 * casewriter.c (struct casewriter): Add value_cnt member.
588 (casewriter_write): Check that the case passed in is big enough.
589 (casewriter_get_value_cnt): New function.
590 (casewriter_create): Add value_cnt parameter.
592 2007-08-09 Ben Pfaff <blp@gnu.org>
594 Fix bug reported by Jason Stover.
595 * settings.c: Move get_termcap_viewport and in particular the
596 #include for <curses.h> to the end of the file. curses.h
597 redefines bool on some systems (e.g. OpenBSD), which causes
598 disagreement between uses of bool before its inclusion and
600 Tested by Jason Stover.
602 2007-07-29 Ben Pfaff <blp@gnu.org>
604 Provisional fix for bug #18692 and bug #20161. Reviewed by John
607 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
608 argument (never "rb" or "wb") because SUSv3 says that only those
609 modes are defined, and glibc in fact rejects other modes.
611 Open portable files with fn_open so that they can be read from
612 pipes. Fix missing fh_close call to go along with fh_open.
613 Report an error if the file close reports an error.
614 * por-file-reader.c (close_reader): New function.
615 (por_file_casereader_destroy): Use close_reader.
616 (pfm_open_reader): Open file with fn_open.
618 2007-07-28 Ben Pfaff <blp@gnu.org>
620 Make PSPP able to read all the portable files I could find on the
621 web. Thanks to John Darrington for review. Bug #17620.
622 * por-file-reader.c (struct pfm_reader): New member `line_length'.
623 (error): Print file offset in hexadecimal.
624 (warning): New function.
625 (advance): Treat lines less than 80 bytes long as padded to 80
627 (pfm_open_reader): Call read_documents if we find an "E" record.
628 (convert_format): Convert invalid formats to the default format
629 instead of aborting reading the file.
630 (read_variables): Rename duplicate variable names instead of
631 aborting reading the file.
632 (read_value_label): Allow string variables of different widths to
633 be assigned value labels in the same record. Replace duplicate
634 value labels instead of aborting.
635 (read_documents): New function.
637 * por-file-writer.c (pfm_open_writer): Call write_documents if the
638 dictionary has documents.
639 (write_documents): New function.
641 2007-07-25 Ben Pfaff <blp@gnu.org>
643 Fix bugs related to bug #17213.
645 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
646 since the former is what config.h has. Include the needed ncurses
648 (static var echo) Rename to `do_echo' because the original name is
649 the same as an ncurses identifier.
650 (get_termcap_viewport) Use error instead of msg.
652 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
653 (fn_close): Don't close stdin, stdout, stderr.
655 2007-07-26 John Darrington <john@darrington.wattle.id.au>
657 * procedure.c procedure.h: Added callbacks which get invoked whenever
658 a dataset's transformation chain changes.
660 2007-07-24 Ben Pfaff <blp@gnu.org>
663 * sys-file-writer.c (write_variable_display_parameters): Use new
664 var_default_display_width function to choose display width of
665 segments after the first one in a given variable.
666 * variable.c (var_create): Use var_default_display_width to pick
667 new variable's display width.
668 (var_default_display_width): New function.
669 Reviewed by John Darrington.
671 2007-07-24 Ben Pfaff <blp@gnu.org>
674 * por-file-writer.c (write_variables): Write weight variable.
675 Reviewed by John Darrington.
677 2007-07-23 Ben Pfaff <blp@gnu.org>
679 Improvements to system file reader and writer.
681 First, move all detailed knowledge of very long strings into
682 sys-file-private.[ch], so that this nasty stuff can be isolated.
684 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
685 (EFFECTIVE_VLS_CHUNK): New macro.
686 (min_int): New function.
687 (max_int): New function.
688 (sfm_width_to_bytes): Rewrite.
689 (sfm_width_to_octs): New function.
690 (sfm_segment_alloc_width): New function.
691 (sfm_segment_alloc_bytes): New function.
692 (sfm_segment_used_bytes): New function.
693 (sfm_segment_offset): New function.
694 (sfm_segment_effective_offset): New function.
695 (sfm_dictionary_to_sfm_vars): New function.
697 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
698 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
699 (struct sfm_var): New structure.
701 Next, improvements to the system file reader.
703 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
704 type casenumber. Added `version_major', `version_minor',
707 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
708 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
709 Change `case_cnt' to type casenumber. Removed `has_vls'.
710 (struct sfm_var): Removed.
711 (sfm_open_reader): Don't warn on wrong case size if the file was
712 written by SPSS 13, which tends to get it wrong. Use
713 sfm_dictionary_to_sfm_vars.
714 (read_header): Always output system file info.
715 (read_variable_record): Simplify code for reading missing values.
716 (read_machine_int32_info): Save version numbers from system file
717 into info struct passed as new argument.
718 (read_long_string_map): Restructured to use new sys-file-private
720 (read_value_labels): Use size_overflow_p.
721 (sys_file_casereader_read): Get rid of distinction between fast
722 and slow paths. Use information provided by sys-file-primate's
723 struct sfm_var to simplify code.
724 (skip_whole_strings): New function.
725 (read_int32): Renamed read_int. Changed return value to int.
727 (read_flt64): Renamed read_float. Changed return value to
728 double. Updated all callers.
729 (int32_to_native): Removed. Changed callers to use
731 (flt64_to_double): Removed. Changed callers to use float_convert.
733 Finally, get rid of int32, flt64 terminology and types in system
734 file writer. The former wasn't very useful since a POSIX "int"
735 can hold the whole range of int32 and we generally didn't have a
736 need for it to be exactly-32-bits, just at-least-32-bits. The
737 latter was inconvenient because we had to assume that it could be
738 different from double and thereby convert special values SYSMIS,
739 HIGHEST, LOWEST to and from it in multiple places. Instead, now
740 we just use "int" and "double" in most places, and do conversions,
741 if necessary, very close to where we do I/O. This change meant
742 that the writer code couldn't represent records in the file as C
743 structs any longer, but that's no great loss. The code actually
744 seems to be more readable without them.
746 Simplify the compression buffering code: only buffer as much as
747 necessary, which is no more than eight 8-byte units at any given
750 * sys-file-writer.c (typedef flt64): Removed.
751 (macro second_lowest_flt64): Removed.
752 (struct sysfile_header): Removed.
753 (struct sysfile_variable): Removed.
754 (struct sfm_writer): Removed `needs_translation', `has_vls',
755 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
756 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
757 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
758 compression buffering by `opcodes', `opcode_cnt', `data',
759 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
760 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
761 data writer functions instead of structures.
762 (calc_oct_idx): New function.
763 (write_header): Use simple data writer functions instead of
765 (write_format_spec): Renamed write_format. New argument.
766 (write_variable_continuation_records): New function.
767 (write_variable): Use simple data writer functions instead of
768 structures. Use write_variable_continuation_records. Write
769 entire very long string instead of requiring caller to understand
771 (write_value_labels): Use simple data writer functions instead of
773 (write_documents): Ditto.
774 (write_variable_display_parameters): Use sys-file-private
775 functions to simplify. Use simple data writer functions instead
777 (write_vls_length_table): Use simple data writer functions instead
779 (write_longvar_table): Ditto.
780 (write_rec_7_34): Break into new functions
781 write_integer_info_record, write_float_info_record. Use simple
782 data writer functions instead of structures.
783 (buf_write): Removed.
784 (append_string_max): Removed.
785 (ensure_buf_space): Removed.
786 (sys_file_casewriter_write): Get rid of the distinction between
787 fast and slow paths, which didn't seem to be too useful. Use new
788 functions write_case_uncompressed, write_case_compressed.
789 (put_instruction): Removed.
790 (put_element): Removed.
791 (write_compressed_data): Removed.
792 (close_writer): Use flush_compressed. Only write case count to
793 system file if it will fit in the field.
794 (write_case_compressed): New function.
795 (write_case_uncompressed): New function.
796 (flush_compressed): New function.
797 (put_cmp_opcode): New function.
798 (put_cmp_number): New function.
799 (write_int): New function.
800 (convert_double_to_output_format): New function.
801 (write_float): New function.
802 (write_value): New function.
803 (write_string): New function.
804 (write_bytes): New function.
805 (write_zeros): New function.
806 (write_spaces): New function.
808 Reviewed by John Darrington.
810 2007-07-22 Ben Pfaff <blp@gnu.org>
812 Don't try to write very long strings to portable files. The
813 format does not support it.
815 * por-file-writer.c (MAX_POR_WIDTH): New macro.
816 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
817 (write_format): Add arg to take width to resize format to.
818 (write_value): Limit width of value written to MAX_POR_WIDTH.
819 (write_variables): Limit width of variable and its output formats
821 Reviewed by John Darrington.
823 2007-07-22 Ben Pfaff <blp@gnu.org>
825 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
826 local variable instead of literal 5.
827 Reviewed by John Darrington.
829 2007-07-22 Ben Pfaff <blp@gnu.org>
831 Fix problems with uniqueness of short names in system files with
832 very long string variables. Now a variable may have multiple
835 * automake.mk (src_data_libdata_a_SOURCES): Add new files
836 short-names.c, short-names.h.
838 * dictionary.c (dict_clone): Clone all the short names.
839 (compare_strings): Move into short-names.c.
840 (hash_strings): Ditto.
841 (set_var_short_name_suffix): Ditto.
842 (dict_assign_short_names): Ditto, rename short_names_assign,
843 change to assign all short names.
845 * por-file-writer.c (write_variables): Use short_names_assign
846 instead of dict_assign_short_names.
848 * short-names.c: New file.
850 * short-names.h: New file.
852 * sys-file-private.c (sfm_width_to_segments): New function.
854 * sys-file-reader.c (read_long_var_name_map): Save and restore all
855 the short names, not just the first one.
857 * sys-file-writer.c (cont_var_name): Removed.
858 (sfm_open_writer): Use short_names_assign instead of
859 dict_assign_short_names. Use unique short names assigned by
860 short_names_assign instead of those generated by cont_var_name.
862 * variable.c (struct variable): Remove `short_name' member,
863 replace by `short_names' and `short_name_cnt'.
864 (var_create) Initialize new members.
865 (var_get_short_name_cnt): New function.
866 (var_get_short_name): Now takes an index argument. Changed most
868 (var_set_short_name): Ditto.
869 (var_clear_short_name): Renamed var_clear_short_names, changed to
870 clear all short names.
872 Reviewed by John Darrington.
874 2007-07-22 Ben Pfaff <blp@gnu.org>
876 * variable.c (var_set_width): Use new var_set_width function.
878 * missing-values.c (mv_n_values): Drop assertion, which was not
881 * format.c (fmt_default_for_width): New function.
882 (fmt_resize): New function.
884 Reviewed by John Darrington.
886 2007-07-18 John Darrington <john@darrington.wattle.id.au>
888 * datasheet.c (datasheet_delete_columns): Added assertion to check
889 we're not deleting outside the range of the sheet.
892 * dictionary.c dictionary.h variable.c: Added the ability for string
893 variables to be resized.
895 * vardict.h: Added some prototypes (moved from dictionary.h) as
896 these should only be called by variable.c
899 2007-07-14 John Darrington <john@darrington.wattle.id.au>
901 * sfm-reader.c: Respect case_cnt field in file header.
903 2007-07-01 John Darrington <john@darrington.wattle.id.au>
905 * transformation.c transformation.h (trns_chain_execute): Changed the
906 signature (Patch #6057)
908 2007-06-10 Ben Pfaff <blp@gnu.org>
910 * casereader-filter.c (casereader_filter_destroy): Make sure to
911 write all the remaining excluded cases to the casewriter, if any.
913 * caseinit.c (init_list_destroy): Rewrite.
914 (init_list_clear): Ditto.
916 * casegrouper.c (casegrouper_get_next_group): Always set *reader
917 to null when returning false.
919 2007-06-06 Ben Pfaff <blp@gnu.org>
921 Actually implement the new procedure code and adapt all of its
922 clients to match. Also adapt all of the other case sources and
923 sinks in the tree and their clients to use the
924 casereader/casewriter infrastructure.
926 * automake.mk: Add and remove files.
928 * any-reader.c: Change into a casereader.
929 * por-file-reader.c: Ditto.
930 * scratch-reader.c: Ditto.
931 * sys-file-reader.c: Ditto.
933 * any-writer.c: Change into a casewriter.
934 * por-file-writer.c: Ditto.
935 * scratch-writer.c: Ditto.
936 * sys-file-writer.c: Ditto.
938 * procedure.c: Change to use casereader, casewriter, caseinit, and
939 other new infrastructure.
941 * scratch-handle.c: Adapt to new infrastructure.
943 * case-sink.c: Removed, now dead code.
944 * case-sink.h: Ditto.
945 * case-source.c: Ditto.
946 * case-source.h: Ditto.
947 * casefile-factory.c: Ditto.
948 * casefile-private.h: Ditto.
951 * casefilter.c: Ditto.
952 * casefilter.h: Ditto.
955 * fastfile-factory.c: Ditto.
956 * fastfile-factory.h: Ditto.
957 * storage-stream.c: Ditto.
958 * storage-stream.h: Ditto.
960 2007-06-06 Ben Pfaff <blp@gnu.org>
964 * automake.mk: Add new files.
966 * datasheet.c: New file.
968 * datasheet.h: New file.
970 2007-06-06 Ben Pfaff <blp@gnu.org>
972 Until now, the procedure code has provided a case to the
973 case_source, which has filled in the data values that come from
974 the active file. "Left" data values that don't come from the
975 active file naturally stay the same from case to case, because the
976 procedure code keeps using that same case.
978 One of the compromises that comes with the new procedure code is
979 that the active file allocates and provides its own case, which
980 the procedure code then has to resize to provide room for any
981 other variables that should go in the case and then fill in the
982 values of "left" variables. Then, when we're done with that case,
983 we have to save the values of "left" variables to copy into the
984 next case read from the active file.
986 The caseinit code helps with this.
988 * automake.mk: Add new files.
990 * caseinit.c: New file.
992 * caseinit.h: New file.
994 2007-06-06 Ben Pfaff <blp@gnu.org>
996 * value.h (value_cnt_from_width): New function.
998 * variable.c (var_get_value_cnt): Use new function.
1000 2007-06-06 Ben Pfaff <blp@gnu.org>
1002 Add casegrouper, to allow cases read from a given casereader to be
1003 broken into groups, each of which has its own casereader.
1004 Generally cases are grouped based on having equal values for some
1007 * automake.mk: Add new files.
1009 * casegrouper.c: New file.
1011 * casegrouper.h: New file.
1013 2007-06-06 Ben Pfaff <blp@gnu.org>
1015 Add interface to lexicographical ordering of cases.
1017 * automake.mk: Add new files.
1019 * case-ordering.c: New file.
1021 * case-ordering.h: New file.
1023 2007-06-06 Ben Pfaff <blp@gnu.org>
1025 Add casereaders and casewriters, the basis of the new data processing
1026 implementation. A casereader is a uniform interface to reading cases
1027 from a data source; a casewriter is a uniform interface to writing
1028 cases to a data sink.
1030 * automake.mk: Add new files.
1032 * casereader-filter.c: New file.
1034 * casereader-provider.h: New file.
1036 * casereader-translator.c: New file.
1038 * casereader.c: New file.
1040 * casereader.h: New file.
1042 * casewriter-provider.h: New file.
1044 * casewriter-translator.c: New file.
1046 * casewriter.c: New file.
1048 * casewriter.h: New file.
1050 2007-06-06 Ben Pfaff <blp@gnu.org>
1052 "casewindow" data structure that extends the deque (from libpspp)
1053 of cases with the ability to dump cases to disk if we get too many
1056 * automake.mk: Add new files.
1058 * casewindow.c: New file.
1060 * casewindow.h: New file.
1062 2007-06-06 Ben Pfaff <blp@gnu.org>
1064 sparse_cases data structure that augments a sparse_array of cases
1065 with the ability to dump cases to disk if we get too many cases in
1068 * automake.mk: Add new files.
1070 * sparse-cases.c: New file.
1072 * sparse-cases.h: New file.
1074 2007-06-06 Ben Pfaff <blp@gnu.org>
1076 Adds a low-level on-disk case array data structure.
1078 * automake.mk: Add new files.
1080 * case-tmpfile.c: New file.
1082 * case-tmpfile.h: New file.
1084 2007-06-06 Ben Pfaff <blp@gnu.org>
1086 In a couple of places we calculate the maximum number of cases to
1087 keep in memory based on the user-defined workspace. Enable
1088 centralizing the calculation through a new function.
1090 * settings.c (get_workspace_cases): New function.
1092 2007-06-06 Ben Pfaff <blp@gnu.org>
1094 The casenumber type is defined in transformations.h, but case.h is
1095 a more sensible place. Move it.
1097 * case.h (CASENUMBER_MAX): New macro.
1098 (typedef casenumber): Move here, from transformations.h.
1100 2007-06-03 Ben Pfaff <blp@gnu.org>
1102 Slightly generalize case_to_values and case_from_values functions.
1104 * case.c (case_to_values): Rename case_copy_out, change interface.
1105 (case_from_values): Rename case_copy_in, change interface.
1107 * fastfile.c (fastfilereader_get_next_case): Update caller.
1108 (write_case_to_disk): Ditto.
1110 2007-06-02 Ben Pfaff <blp@gnu.org>
1112 Clean up after a forgotten part of patch #5829.
1114 * casedeque.h: Remove unused file.
1116 * automake.mk: Remove casedeque.h from sources.
1118 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1120 * category.c: Removed redundant #include
1122 2007-05-06 Ben Pfaff <blp@gnu.org>
1124 Abstract the documents within a dictionary a little better.
1125 Thanks to John Darrington for suggestion, initial version, and
1126 review. Patch #5917.
1128 * dictionary.c (struct dictionary): Change `documents' member from
1129 char * to struct string.
1130 (dict_clear): Destroy struct string.
1131 (dict_get_documents): Convert struct string to char *.
1132 (dict_set_documents): Set struct string. Pad to 80-character
1134 (dict_clear_documents): New function.
1135 (dict_add_document_line): New function.
1136 (dict_get_document_line_cnt): New function.
1137 (dict_get_document_line): New function.
1139 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1141 * sys-file-reader.c (read_documents): Use new document functions.
1143 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1145 * sys-file-reader.c: When reading a system file which has no
1146 long name table, automatically create one where the long names
1147 are the lower case versions of the short names.
1149 2007-04-22 Ben Pfaff <blp@gnu.org>
1151 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1152 dict_clear will free most data related to the dictionary.
1153 dict_clear does a decent job, except that dict_set_split_vars on
1154 some systems won't actually free the dict's "split" member.
1155 Instead, it'll allocate a 1-byte region. Fix this.
1157 * value.c (value_copy): New function.
1158 (value_set_missing): Ditto.
1160 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1162 * Deleted existing category.h and moved cat-routines.h into
1163 category.h Encapsulated struct cat_vals better.
1165 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1167 * sys-file-reader.c: When reading a system file which has no
1168 long name table, automatically create one where the long names
1169 are the lower case versions of the short names.
1171 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1173 * sys-file-reader.c: Some versions of Other Software seem to
1174 produce system files with string variables' measure set to
1175 zero. We'll assume these are supposed to be nominal variables.
1177 2007-03-30 Ben Pfaff <blp@gnu.org>
1179 * procedure.c: Adapt to new deque data structure.
1181 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1182 Ben Pfaff <blp@gnu.org>
1184 * file-name.c: Mingw compatibility fixes.
1185 (fn_search_path): Use ISSLASH instead of comparing against '/'
1187 (fn_dir_name): Use dir_name from gnulib.
1188 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1189 (fn_get_identity): Use GetFullPathName instead of canonicalize
1190 from gnulib, because the latter does not fully support
1191 Windows-style path names. Use this implementation based on the
1192 detected presence of Windows instead of the absence of Unix, since
1193 the new implementation is Windows-specific.
1194 (fn_compare_file_identities): In Windows implementation, compare
1195 names case-insensitively.
1197 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1199 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1200 <stdlib.h> to have the same effect.
1202 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1204 * por-file-reader.c: Add missing _() around messages.
1206 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1208 * make-file.c: Include "mkstemp.h", without which linking on
1211 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1213 Reduce platform dependence.
1215 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1216 Everywhere we using this, we really should have just depended on
1217 the shell to expand tildes.
1218 (fn_search_path): Simplify, given that we don't do tilde expansion
1220 (fn_normalize): Removed. Caller changed to use the canonicalize
1222 (fn_get_cwd): Removed. Only user was fn_normalize.
1223 (fn_is_absolute): Really only test for absolute names.
1224 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1226 (fn_readlink): Removed, as it was only used fn_normalize.
1227 (fn_exists): Assume the stat function is available; gnulib does.
1228 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1230 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1232 * dictionary.c (dict_create_vector_assert): New function.
1234 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1236 * file-name.c (fn_normalize): Correct name of function
1237 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1238 for pointing this out.
1240 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1242 We are using a single member in struct file_handle, the "name"
1243 field, for more than one purpose. When it begins with '"', it's a
1244 file name; otherwise, it's a token that can be used to identify
1245 it. When that assertion fires, it's because we searched for the
1246 name case-sensitively as a file name (so that there was no match),
1247 and then we try to insert it case-insensitively as a token, which
1248 fails because duplicates aren't allowed.
1250 Solution: break the two purposes into two separate fields. This
1251 fixes the problem and likely makes the code easier to read too.
1253 Fixes bug #18922. Thanks to John Darrington for bug report and
1256 * file-handle-def.c (struct file_handle): New `id' member.
1257 (fh_from_name): Rename fh_from_id. Update all callers.
1258 (create_handle): New `id' parameter. Update all callers.
1259 (fh_create_file): Ditto.
1260 (fh_get_id): New function.
1262 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1264 * case.c (case_is_null): Change return type to bool.
1266 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1270 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1271 reference counting to share data. This makes it easy for
1272 valgrind, etc. to find accesses to cases that have been destroyed
1273 but have been kept around by another user's ref-count. This often
1274 happens when the data set is small enough to find in memory; if a
1275 bigger data set that would overflow to disk were used, then data
1276 corruption would occur.
1278 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1282 * case.c (case_unshare): Make it check internally whether the
1283 ref_cnt is greater than 1, so that the callers don't have to.
1284 Update callers not to check.
1286 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1288 Before, I was thinking that I might want to get rid of reference
1289 counting at some point. Now, I'm pretty sure that it's here to
1290 stay. Thus, because we have to store the value_cnt anyway for
1291 reference-counted cases, we might as well expose it to users.
1293 * case.c (case_get_value_cnt): New function.
1294 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1295 resize case if its size actually changed.
1297 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1298 instead of peeking inside struct case directly.
1299 (casefile_append): Ditto.
1301 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1303 Get rid of the inlines for the case functions, which made the
1304 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1305 the inlines didn't speed up "make check" at all, which is not a
1306 perfect benchmark but seems indicative.)
1308 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1309 definitions. Remove some assertions on nonnull pointers that were
1310 redundant with a pointer dereference soon after in the function.
1312 (struct case_data): Move definition here from case.h.
1316 (case_data_wr): Ditto.
1318 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1320 * automake.mk: Add casedeque.h to sources.
1322 * casedeque.h: New file.
1324 * procedure.c: (struct dataset) Change lag_count, lag_head,
1325 lag_queue member into single struct casedeque member. Update all
1326 users to use the casedeque instead.
1329 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1331 * procedure.c: Simplify lagged cases interface. Updated all
1332 clients--well, the only client--to use the simplified interface.
1333 (dataset_n_lag) Removed.
1334 (dataset_set_n_lag) Removed.
1335 (dataset_need_lag) New function.
1337 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1339 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1340 whenever appropriate, but only when the dataset/dictionary is in a
1343 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1345 * dictionary.c dictionary.h : Added callbacks for change of filter and
1346 split variables. Refactored some code to ensure that callbacks get
1347 invoked when appropriate.
1349 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1350 replace_dict callback occurs when permanent_dict replaces the current
1353 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1355 * dictionary.c dictionary.h : Added callback for when the weight
1356 variable of a dictionary changes.
1358 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1360 * dictionary.c dictionary.h : Added replace_source and replace_dict
1361 callbacks, and functions to deal with them.
1363 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1365 Simplify missing value handling.
1367 * missing-values.h (enum mv_class): New type.
1368 (enum mv_type): Moved definition into missing-values.c and renamed
1369 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1370 enums. Updated all uses.
1371 (struct missing_values): Changed type of `type' from `enum
1372 mv_type' to `int' because the definition is no longer exposed.
1374 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1375 parameter. Update all callers.
1376 (mv_is_num_missing): Ditto.
1377 (mv_is_str_missing): Ditto.
1378 (mv_is_value_user_missing): Removed. Changed callers to use
1379 mv_is_value_missing.
1380 (mv_is_num_user_missing): Removed. Changed callers to use
1382 (mv_is_str_user_missing): Removed. Changed callers to use
1384 (mv_is_value_system_missing): Removed. Changed callers to use
1385 mv_is_value_missing.
1386 (mv_set_type): Removed. Changed callers to use mv_clear.
1387 (mv_clear): New function.
1389 * variable.c (var_is_value_missing): Add new enum mv_class
1390 parameter. Update all callers.
1391 (var_is_num_missing): Ditto.
1392 (var_is_str_missing): Ditto.
1393 (var_is_value_user_missing): Removed. Changed callers to use
1394 var_is_value_missing.
1395 (var_is_num_user_missing): Removed. Changed callers to use
1397 (var_is_str_user_missing): Removed. Changed callers to use
1399 (var_is_value_system_missing): Removed. Changed callers to use
1400 var_is_value_missing.
1402 * casefilter.c (struct casefilter): Use enum mv_class in place of
1404 (casefilter_variable_missing): Adapt to new member.
1405 (casefilter_create): Change signature to take enum mv_class,
1408 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1410 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1412 * case-sink.c case-sink.h procedure.c procedure.h
1413 storage-stream.c: Now uses the factory.
1415 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1417 Make it possible to pull cases from the active file with a
1418 function call, instead of requiring indirection through a callback
1421 * case-source.h (struct case_source_class): Change ->read function
1422 to return a single case, instead of calling a callback function
1423 for each case. Change ->destroy function to return an error
1426 * case-source.c (free_case_source): Pass along the value returned
1427 by the case_source ->destroy function.
1429 * procedure.c (struct write_case_data): Removed.
1430 (struct dataset): Added some members to track procedure state.
1431 (procedure): Optimize the trivial case at this level.
1432 (internal_procedure): Re-implement in terms of proc_open,
1433 proc_read, proc_close.
1434 (proc_open) New function.
1435 (proc_read) New function.
1436 (proc_close) New function.
1437 (write_case) Moved into proc_read.
1438 (close_active_file) Moved closing of data source into proc_close.
1440 * storage-source.c: Rewrote to conform with modified
1441 case_source_class interface.
1443 * transformations.c (trns_chain_execute): Added argument to allow
1444 starting execution from an arbitrary transformation. Updated
1447 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1449 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1451 * sys-file-reader.c (read_display_parameters): Don't assume that
1452 MEASURE_* and ALIGN_* have the same values found in system files.
1454 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1456 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1457 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1458 the value of n_MEASURES, which was off by 1 (at least from my
1461 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1463 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1464 callbacks which are invoked when the dictionary or its
1465 variables are changed.
1467 * missing-values.c missing-values.h value-labels.c: Tidied up
1468 consistency checks, and made some of them return false
1469 instead of assert-failing.
1471 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1473 * calendar.c (calendar_days_in_month): New function.
1475 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1477 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1478 the value label's width.
1480 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1482 * sfm-private.h: Move contents into sys-file-writer.c, which is
1483 the only remaining user. Removed Borland C++-specific directives.
1485 * sys-file-reader.c: Clean up and rewrite entire file. The
1486 rewritten version is simpler and better abstracted, and should be
1487 easier to maintain and extend. It avoids using structures to read
1488 file data, which is prone to padding variations among compilers.
1489 It should also handle non-IEEE 754 system files, although I
1490 haven't been able to find any. It has been tested against many
1491 .sav files obtained from the Web and found to produce the same
1492 results as the earlier version of the code, or in some cases
1493 improved results. It is more tolerant of format variations found
1496 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1497 member, putting an enum integer_format in its place. New member
1498 `float_format'. Changed `compressed' member to type bool.
1500 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1502 * dictionary.c (dict_delete_consecutive_vars): New function.
1504 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1506 * file-name.c (fn_search_path): Remove prefix arg that was unused
1507 by any caller. Updated all callers.
1509 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1511 * format.c (fmt_dollar_template): Use user's decimal point
1512 character. Add assertion.
1514 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1516 * format.c (fmt_dollar_template): New function, based on
1517 dollar_format_template from var-type-dialog.c.
1519 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1521 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1524 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1526 Finish converting struct variable to an opaque type. In this
1527 phase, we add remaining setter and getter functions, convert the
1528 remaining PSPP code to use them, and do a bunch of cleanup. The
1529 resulting changes are pervasive but mostly trivial, and only the
1530 notable changes are logged.
1532 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1535 * case.c (case_data): Renamed case_data_idx.
1536 (case_num): Renamed case_num_idx.
1537 (case_str): Renamed case_str_idx.
1538 (case_data_rw): Renamed case_data_rw_idx.
1540 * case.h (case_data): New function with old name and an interface
1541 that takes a variable instead of an index, which is easier to
1542 use. Updated all callers to use the new interface, or to use the
1543 new *_idx function (see above).
1546 (case_data_rw): Ditto.
1548 * category.c (cat_stored_values_destroy): Changed interface to
1549 take a struct cat_vals * instead of a struct variable *.
1551 * dictionary.c (dict_clone): Use new vector_clone function.
1552 (dict_clear) Use new var_destroy function.
1553 (add_var) New function.
1554 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1555 (dict_create_var_assert) Rewrite in terms of add_var.
1556 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1557 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1558 (dict_lookup_var) Use new var_create, var_destroy functions.
1559 (dict_contains_var) Rewrite in terms of new vardict functionality.
1560 (set_var_dict_index) New function.
1561 (set_var_case_index) New function.
1562 (reindex_vars) New function.
1563 (dict_delete_var) Rewrite in terms of new vardict functionality.
1564 (dict_reorder_var) Ditto.
1565 (dict_reorder_vars) Ditto.
1566 (rename_var) New function.
1567 (dict_rename_var) Use rename_var.
1568 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1569 (dict_get_compacted_idx_to_fv) Rename
1570 dict_get_compacted_dict_index_to_case_index, update callers.
1571 (dict_create_vector) Use new vector_create function.
1572 (dict_clear_vectors) Use new vector_destroy function.
1573 (set_var_short_name_suffix) Move here from variable.c, renamed
1574 from var_set_short_name_suffix, make static, update caller.
1576 * sys-file-private.c: New file.
1577 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1578 width_to_bytes, update callers.
1580 * sys-file-private.h: New file. Later it will supplant
1581 sfm-private.h; for now it supplements it.
1582 (macro MIN_VERY_LONG_STRING) New macro.
1583 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1585 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1586 MAX_LONG_STRING was used before.
1588 * sys-file-writer.c: Ditto.
1590 * value-labels.c: Change the paradigm here to be that a null
1591 pointer is OK for a struct val_labs * in most cases; it just
1592 represents an empty set of value labels.
1593 (val_labs_copy) A copy of a null set is a null set.
1594 (val_labs_count) A null set has 0 labels.
1595 (val_labs_replace) Change return type to void. Rewrite for
1597 (val_labs_find) A null set does not contain the value.
1598 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1599 transposed argument order, updated all callers.
1601 * value.c: New file.
1602 (value_dup) Moved here from variable.c.
1603 (compare_values) Ditto.
1606 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1607 (macro MAX_LONG_STRING) Removed, because it was only interesting
1608 for system files, not for general code.
1609 (macro MAX_VERY_LONG_STRING) Ditto.
1610 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1611 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1613 * vardict.h: New file, for an interface between variables and
1616 * variable.c: A lot of functions were moved around, for better
1618 (struct variable) Move definition here, from variable.h.
1619 (var_type_adj) Removed--makes i18n hard.
1620 (var_type_noun) Ditto.
1621 (var_create) New function.
1622 (var_clone) New function.
1623 (var_destroy) New function.
1624 (var_set_name) Assert that variable is not in a dictionary.
1625 (compare_var_names) Rename compare_vars_by_name and fix a couple
1626 of callers who thought the args were strings.
1627 (hash_var_name) Rename hash_var_by_name.
1628 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1629 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1630 (var_is_very_long_string) Removed, because it was only interesting
1631 to system file code.
1632 (var_set_missing_values) Allow the argument to be the wrong width,
1633 as long as we can resize it. Simplify callers who were doing the
1634 resizing themselves.
1635 (var_get_value_labels) New function.
1636 (var_has_value_labels) New function.
1637 (var_set_value_labels) New function.
1638 (alloc_value_labels) New function.
1639 (var_add_value_label) New function.
1640 (var_replace_value_label) New function.
1641 (var_clear_value_labels) New function.
1642 (var_lookup_value_label) New function.
1643 (var_get_value_name) Moved here from variable.c, renamed from
1644 var_get_value_name, transposed argument order, updated all
1646 (var_to_string) Moved here, from variable-label.c.
1647 (var_set_leave) New function.
1648 (var_get_leave) New function.
1649 (var_must_leave) New function.
1650 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1651 set_var_short_name_suffix.
1652 (var_get_dict_index) New function.
1653 (var_get_case_index) New function.
1654 (var_get_obs_vals) New function.
1655 (var_set_obs_vals) New function.
1656 (var_has_obs_vals) New function.
1657 (var_get_vardict) New function.
1658 (var_set_vardict) New function.
1659 (var_has_vardict) New function.
1660 (var_clear_vardict) New function.
1661 (value_dup) Moved to value.c.
1662 (compare_values) Ditto.
1665 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1666 (enum ALPHA) Rename VAR_STRING, update all users.
1668 * vector.c: New file.
1669 (struct vector) Moved here, from variable.h.
1670 (check_widths) New function.
1671 (vector_create) New function.
1672 (vector_clone) New function.
1673 (vector_destroy) New function.
1674 (vector_get_name) New function.
1675 (vector_get_var) New function.
1676 (vector_get_var_cnt) New function.
1677 (compare_vector_ptrs_by_name) New function.
1679 * vector.h: New file.
1681 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1683 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1684 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1687 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1689 * value-labels.c (destroy_atoms): New function.
1690 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1692 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1694 Thanks to Jason Stover for pointing out this problem.
1696 * data-out.c (output_number): Use gsl_finite from GSL, which is
1697 portable, instead of isfinite, which is not.
1700 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1702 * variable.c variable.h (value_dup): New function.
1704 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1706 Start converting struct variable to an opaque type. In this
1707 phase, we add a bunch of setter and getter functions and convert
1708 most of the PSPP code to use them. The resulting changes are
1709 pervasive but mostly trivial, and only the notable changes are
1712 * format.c (fmt_equal): New function.
1714 * variable.c (var_type_is_valid): New function.
1715 (measure_is_valid) Moved here, from format.c.
1716 (alignment_is_valid) Moved here, from format.c.
1717 (var_get_name) New function.
1718 (var_set_name) New function.
1719 (width_to_type) New function.
1720 (var_get_type) New function.
1721 (var_get_width) New function.
1722 (var_set_width) New function.
1723 (var_is_numeric) New function.
1724 (var_is_alpha) New function.
1725 (var_is_short_string) New function.
1726 (var_is_long_string) New function.
1727 (var_is_very_long_string) New function.
1728 (var_get_missing_values) New function.
1729 (var_set_missing_values) New function.
1730 (var_clear_missing_values) New function.
1731 (var_has_missing_values) New function.
1732 (var_is_value_missing) New function.
1733 (var_is_num_missing) New function.
1734 (var_is_str_missing) New function.
1735 (var_is_value_user_missing) New function.
1736 (var_is_num_user_missing) New function.
1737 (var_is_str_user_missing) New function.
1738 (var_is_value_system_missing) New function.
1739 (var_get_print_format) New function.
1740 (var_set_print_format) New function.
1741 (var_get_write_format) New function.
1742 (var_set_write_format) New function.
1743 (var_set_both_formats) New function.
1744 (var_get_label) New function.
1745 (var_set_label) New function.
1746 (var_clear_label) New function.
1747 (var_has_label) New function.
1748 (var_get_measure) New function.
1749 (var_set_measure) New function.
1750 (var_get_display_width) New function.
1751 (var_set_display_width) New function.
1752 (var_get_alignment) New function.
1753 (var_set_alignment) New function.
1754 (var_get_value_cnt) New function.
1755 (var_get_leave) New function.
1756 (var_get_short_name) New function.
1758 * variable.h: (struct variable) Removed "type" and "nv" members;
1759 they are now computed from "width" where needed.
1761 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1763 * missing-values.c (mv_resize): Don't write beyond end of the
1764 allocated buffer when resizing a long string.
1766 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1768 Clean up identifier code: don't require identifier enumerations to
1769 be in a particular order; make better use of string library;
1770 expose less of the internals.
1772 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1773 change interface. Updated all callers.
1774 (lex_id_match) Change interface to use struct substring, update
1776 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1777 (global array keywords[]) Make static, change form. Update all
1778 users to use lex_id_name instead.
1779 (lex_is_keyword) New function.
1780 (lex_id_to_token) Change interface to use struct substring, update
1782 (lex_id_name) New function.
1784 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1785 lex_is_keyword instead.
1786 (T_LAST_KEYWORD) Removed.
1787 (T_N_KEYWORDS) Removed.
1789 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1791 * format.c: (fmt_date_template) Distinguish characters for which a
1792 space is output and any date delimiter is allowed on input, from
1793 those for which a space is output and only a space is allowed on
1794 input. The former is represented by X, the latter by a space.
1795 Also, drop distinction between h and H, changing the former to the
1798 * data-in.c: Completely rewrite internals to conform to SPSS input
1799 formats as closely as possible.
1800 (data_in) Changed external interface by replacing the structure
1801 that was used as a single argument by a set of arguments. Updated
1803 (data_in_finite_line) Removed. Converted all callers to use plain
1805 (data_in_get_integer_format) New function.
1806 (data_in_set_integer_format) New function.
1807 (data_in_get_float_format) New function.
1808 (data_in_set_float_format) New function.
1810 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1811 (struct data_in) Removed.
1813 * data-out.c: (output_date) Drop each component from the input as
1814 it is output, to allow us to drop the distinction between h (a
1815 count of hours) and H (the hour of day) template characters.
1816 Also, handle new X template character.
1817 (output_scientific) Follow more rational rule on when to drop
1818 fraction introduced between SPSS 13 and 15. Updated test case to
1821 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1823 Fix buffer overflow reported by John Darrington.
1825 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1826 realize that DIGITS is a count of nibbles, not of bytes.
1828 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1830 * calendar.c (calendar_offset_to_gregorian) Also return the
1831 year-of-day. Change callers to new interface.
1833 * data-out.c: Completely rewrite internals to conform to SPSS
1834 output formats as completely as possible.
1835 (data_out) Change interface to put input parameters before output
1836 parameters, for consistency with the style I now prefer. Update
1838 (data_out_get_integer_format) New public function.
1839 (data_out_set_integer_format) New public function.
1840 (data_out_get_float_format) New public function.
1841 (data_out_set_float_format) New public function.
1843 * data-out.h: New file. Move prototype for data_out here, from
1846 * format.c: (fmt_step_width) Use equality comparison instead of
1847 bitwise and, for clarity.
1848 (fmt_is_string) Ditto.
1849 (fmt_input_to_output) Fix categories that are translated to F
1852 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1854 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1855 casefile-private.h: Added casefilter to assist commands with missing
1858 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1860 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1862 * settings.c: (route_errors_to_terminal) New variable.
1863 (route_errors_to_listing) New variable.
1864 (get_error_routing_to_terminal) New function.
1865 (set_error_routing_to_terminal) New function.
1866 (get_error_routing_to_listing) New function.
1867 (set_error_routing_to_listing) New function.
1869 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1871 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1873 * format.c: Completely rewrite, to achieve better abstraction.
1874 Rewrite all references to formats in other files.
1876 * format.def: Rewrite and reorganize.
1878 * settings.c: Move everything related to custom currency formats
1879 into format.[ch], changing them in form, so as to group related
1880 code and definitions better. Changed all references to use the
1882 (static var decimal) Removed.
1883 (static var grouping) Removed.
1884 (static var cc) Removed.
1885 (get_decimal) Removed.
1886 (set_decimal) Removed.
1887 (get_grouping) Removed.
1888 (set_grouping) Removed.
1892 * settings.h: (macro CC_CNT) Removed.
1893 (macro CC_WIDTH) Removed.
1894 (struct custom_currency) Removed.
1896 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1898 * data-in.c (data_in): Use switch statement instead of table, to
1899 avoid dependence on the order of the FMT_* enums.
1901 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1903 * data-out.c: (num_to_string) Removed, because it was dead code.
1905 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1907 * data-in.c (parse_trailer): Fix error message.
1909 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1911 * format.c (fmt_is_binary): New function.
1913 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1915 * procedure.c procedure.h: Encapsulated the static data into a single
1918 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1920 * casefile.c (casereader_read_xfer): Always initialize the case,
1921 even on an error condition.
1923 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1925 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1926 entire function useless.
1928 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1930 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1931 casereader method casereader_clone.
1933 * procedure.c pransformations.h: Introduced new type casenum_t
1935 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1937 * variable.c: (width_to_bytes) Rephrase code for clarify.
1939 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1941 * format.c: (fmt_type_from_string) New function.
1942 (fmt_to_string) Include decimals in output if the format has
1943 decimals, even if the format type does not. This way, we can
1944 accurately reproduce incorrect formats in user output.
1945 (check_common_specifier) Make the check for a bad format type an
1946 assertion, so we get bug reports if they show up. Fix message.
1947 Check for decimal places with a format type that doesn't allow
1949 (check_input_specifier) Remove check for FMT_X, which has been
1951 (check_output_specifier) Ditto.
1953 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1955 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1956 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1957 (enum fmt_parse_flags) Removed.
1959 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1961 * casefile.c casefile.h: Converted to an abstract base class.
1962 * casefile-private.h fastfile.c fastfile.h: New files.
1963 * automake.mk procedure.c scratch-writer.c storage-stream.c
1965 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1967 * procedure.c (internal_procedure): Create sink_case with only as
1968 many values as the compacted dictionary.
1970 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1972 Remove "debugging" code that caused plenty of false positives and
1975 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1977 * case.c: Remove all references to `this' member.
1979 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1981 Fix link error noted by Jason Stover.
1983 * storage-stream.c: Include <assert.h>.
1985 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1987 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1988 ALL) and additional underlying system file issues.
1990 Thanks to John Darrington for review.
1992 First problem: var_hash points to variables not owned by the
1993 sys-file-reader, which the caller may free or modify. Use an
1994 array of sfm_vars instead, as done earlier (e.g. CVS version
1997 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1998 members and remove all code that references it. Add vars, var_cnt
1999 members. Remove fix_specials member, which was unused.
2000 (struct sfm_var) Remove name member, which was unused.
2001 (sfm_close_reader) Free vars member instead of var_hash.
2002 (compare_var_shortnames) Removed.
2003 (hash_var_shortname) Removed.
2004 (sfm_open_reader) Fill out vars array.
2005 (compare_var_index) Removed.
2006 (sfm_read_case) Use vars instead of var_hash.
2008 Second problem: we're confused about when we actually have very
2009 long strings, causing us to choose incorrectly between slow path
2010 and fast path in sfm_read_case.
2012 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
2013 have very long strings, not when we have long variable names,
2014 which is an unrelated feature.
2016 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
2018 * variable.h: Move var_set and variable parsing declarations to
2019 new header, src/language/lexer/variable-parser.h. Modified lots
2020 of files to include the new header.
2022 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
2024 * value-labels.c (value_to_string): When there's no value label,
2025 format the variable according to its print format, instead of
2026 always effectively using A or F format.
2028 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
2030 * casefile.c (casefile_get_random_reader): Nasty hack to get around
2033 * format.c: Removed tortological assertion.
2035 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
2037 Reform string library.
2039 * file-name.c (fn_interp_vars): Change interface to take a
2040 substring as input. Updated all users.
2042 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
2044 * format.c (measure_is_valid): Really return false when m >=
2047 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
2049 Implement random access to casefiles, for use in GUI.
2051 * casefile.c: (struct casereader) Add `random', `file_ofs',
2052 `buffer_ofs' members.
2053 (casefile_get_random_reader) New function.
2054 (read_open_file) Break part into new function
2055 seek_and_fill_buffer().
2056 (fill_buffer) Update buffer_ofs, file_ofs.
2057 (casereader_seek) New function.
2059 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2061 * settings.c: Added call to i18n{done, init}.
2063 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
2065 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
2068 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
2070 * casefile.c: Convert many uses of `int' to `bool'.
2072 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
2074 * transformations.c (trns_chain_destroy): Destroy chain's trns
2075 member, to fix memory leak.
2077 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
2079 * storage-stream.c (storage_source_decapsulate): Destroy case
2080 source to fix memory leak.
2082 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
2084 * scratch-reader.c (scratch_reader_read_case): Copy into existing
2085 case passed as argument instead of initializing the argument as a
2086 case. Fixes memory leak that showed up in
2087 tests/command/aggregate.sh with scratch files.
2089 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
2091 * procedure.c (proc_done): Destroy default_dict, to fix memory
2094 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2096 Simplify procedure_with_splits().
2098 * procedure.c (struct split_aux_data): Removed case_count member.
2099 (procedure_with_splits) Don't initialize case_count.
2100 (split_procedure_case_func) Check whether prev_case is null
2101 instead of case_count.
2102 (split_procedure_end_func) Ditto.
2104 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2106 * case.c (case_move): Do nothing if dst and src are the same
2108 (case_try_create) Merge two similar cases.
2109 (case_copy) Unshare only if data must be actually copied.
2111 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2113 * data-in.c data-out.c dictionary.c sys-file-reader.c
2114 sys-file-writer.c variable.c variable.h: Reworked very long string
2115 support for better encapsulation.
2117 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2119 * value-labels.c (val_labs_can_set_width): New function.
2120 (val_labs_set_width) Clear labels if increasing width to long
2122 (val_labs_destroy) Remove unneeded test for null.
2124 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2126 * value-labels.h: Remove unneeded dependency on variable.h.
2128 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2130 Get rid of `char *c' member in union value, for cleanliness.
2132 * value.h: (union value) Remove `c' member.
2134 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2136 Make dictionary compacting functions a little more general.
2138 * sys-file-writer.c (sfm_open_writer): Use
2139 dict_compacting_would_change().
2140 (does_dict_need_translation) Removed.
2142 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2144 Make dictionary compacting functions a little more general.
2146 * dictionary.c (dict_needs_compaction): Rename
2147 dict_compacting_would_shrink(). Update all callers.
2148 (dict_compacting_would_change) New function.
2150 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2152 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2153 inverted return value (!).
2155 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2157 Continue reforming procedure execution.
2159 * procedure.c: Search and replace "vfm" by "proc". Notably:
2160 (static var vfm_source) Rename proc_source. Update all
2162 (static var vfm_sink) Rename proc_sink. Update all references.
2164 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2166 Continue reforming procedure execution. In this phase, remove
2167 PROCESS IF, which was deprecated anyway and can be easily
2168 simulated with TEMPORARY followed by SELECT IF.
2170 * procedure.c: (open_active_file) Don't call
2171 add_process_if_trns().
2172 (discard_variables) Get rid of redundant call to
2173 proc_cancel_all_transformations().
2174 (add_process_if_trns) Removed.
2175 (process_if_trns_proc) Removed.
2176 (process_if_trns_free) Removed.
2178 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2180 Continue reforming procedure execution. In this phase, add
2181 `const' to the case passed to procedure()'s callback.
2183 Updated all users of procedure() as well.
2185 * procedure.c: (struct write_case_data) Add "const" to ccase
2186 parameter for case_func member.
2187 (procedure) Add "const" to ccase parameter for proc_func
2189 (multipass_case_func) Make ccase parameter const.
2190 (internal_procedure) Add "const" to ccase parameter for case_func
2192 (split_procedure_case_func) Make ccase parameter const.
2193 (multipass_split_case_func) Make ccase parameter const.
2195 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2197 Continue reforming procedure execution. In this phase, get rid of
2198 the output code for SPLIT FILE groups in procedure.c, which really
2199 shouldn't be doing any output. Move it into the individual
2200 procedures instead. This also adds some flexibility.
2202 Updated many users of procedure_with_splits() and
2203 multipass_procedure_with_splits() to call
2204 output_split_file_values() and to deal with increased use of
2207 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2208 parameter to begin_func member.
2209 (procedure_with_splits) Add "const struct ccase *" parameter to
2210 begin_func parameter. Make ccase parameter const in proc_func
2212 (split_procedure_case_func) Don't dump split file group. Pass
2214 (dump_splits) Moved to language/dictionary/split-file.c as
2215 output_split_file_values().
2216 (struct multipass_split_aux_data) Add "const struct ccase *"
2217 parameter to split_func member.
2218 (multipass_procedure_with_splits) Add "const struct ccase *"
2219 parameter to split_func parameter.
2220 (multipass_split_case_func) Save new SPLIT FILE case before
2222 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2224 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2226 Continue reforming procedure execution. Change
2227 internal_procedure() so that it calls open_active_file() and
2228 close_active_file(), which isolates most of the actual procedure
2231 * procedure.c: (struct write_case_data) Rename `proc_func' member
2232 to `case_func' and update all references.
2233 (procedure) Rewrite as one-line wrapper around
2234 internal_procedure().
2235 (struct multipass_aux_data) New.
2236 (multipass_callback) Renamed multipass_case_func(). Use struct
2237 multipass_aux_data as auxiliary data.
2238 (multipass_end_func) New function.
2239 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2240 that uses multipass_case_func, multipass_end_func.
2241 (internal_procedure) Add `end_func' argument. Move optimization
2242 of trivial case in here. Move call to open_active_file() and
2243 close_active_file() in here. Now assert that vfm_source is
2245 (procedure_with_splits_callback) Rename
2246 split_procedure_case_func().
2247 (split_procedure_end_func) New function.
2248 (multipass_split_callback) Rename multipass_split_case_func.
2249 (multipass_split_end_func) New function.
2250 (discard_variables) No need to test for nonnull vfm_source.
2252 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2254 Continue reforming procedure execution. Get rid of unused member.
2256 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2258 (write_case) Don't increment cases_analyzed.
2260 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2262 Continue reforming procedure execution. In this phase, move
2263 procedure.c and procedure.h from src to src/data. Update
2264 makefiles and #includes accordingly.
2266 * procedure.c: Moved here from src/.
2268 * procedure.h: Moved here from src/.
2270 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2272 Continue reforming procedure execution. In this phase, get rid of
2273 many global variables, consolidating procedure execution in
2274 procedure.c. Encapsulate transformations in new "struct
2275 trns_chain". Also, change implementation of N OF CASES, FILTER,
2276 and PROCESS IF from special cases to transformations.
2278 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2281 * dictionary.c: (global variable default_dict) Move to
2284 * variable.h: (TRNS_*) Move to transformations.h.
2285 (struct transformation) Move to transformations.c.
2287 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2289 * sys-file-reader.c: Fixed invalid read problems.
2291 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2293 * storage-stream.c: Add missing function comments.
2295 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2297 Continue reforming procedure execution. In this phase, add some
2298 new, needed functionality to storage-stream.
2300 * storage-stream.c: (storage_source_decapsulate) New function.
2302 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2304 * variable.c (width_to_bytes): Declarations must precede
2305 statements for C90 compliance.
2307 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2309 * data-out.c, data-in.c, variable.c, variable.h: New functions
2310 copy_mangle and copy_demangle for reading/writing cases; emulates the
2311 way SPSS deals with strings > 255 bytes.
2313 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2314 subtype 14 needed for strings longer than 255 bytes.
2316 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2317 instead of literal values. Also fixed some constness issues.
2319 * format.h: Constness
2321 * sfm-private.h: Renamed the case_size identifier, since I discovered
2322 that SPSS's respect for this variable is very nominal.
2324 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2326 Change case limit type from int to size_t.
2328 * dictionary.c: (struct dictionary) Change type of case_limit
2330 (dict_get_case_limit) Change return type.
2331 (dict_set_case_limit) Change parameter type.
2333 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2335 * variable.h: (struct variable) Rename `reinit' member as `leave'
2336 and invert sense. Fix up all references.
2338 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2340 Continue reforming procedure execution. In this phase, break
2341 procedure.c into multiple files.
2343 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2345 * case-sink.c: New file.
2347 * case-sink.h: New file.
2349 * case-source.c: New file.
2351 * case-source.h: New file.
2353 * storage-stream.c: New file.
2355 * storage-stream.h: New file.
2357 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2359 * variable.h: (struct variable) Remove `init' member and all
2360 references to it from other files. It was initialized in several
2361 places, but nothing really ever used it for anything worthwhile.
2362 Thanks to Jason Stover for pointing out how confusing this
2365 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2367 Continue reforming error message support. In this phase, get rid
2368 of message "titles" and put the message text in `struct error'.
2369 Now `struct error' encapsulates a message more properly.
2371 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2372 Format message ourselves.
2374 * data-in.c: (vdls_error) Ditto.
2376 * por-file-reader.c: (error) Ditto.
2378 * sys-file-reader.c: (corrupt_msg) Ditto.
2380 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2382 GNU standards require "file name" instead of "filename" in
2383 documentation. It's nice for our code to follow the convention
2386 * casefile.c: (struct casefile) Rename `filename' member to
2387 `file_name'. Updated all references.
2389 * file-name.c: [!unix] (struct file_identity) Rename
2390 normalized_filename member to normalized_file_name. Updated all
2393 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2395 We don't really support anything but Unix-like environments well,
2396 so we might as well de-obfuscate by writing directory and path
2397 separators explicitly.
2399 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2401 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2402 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2404 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2407 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2409 GNU standards require "file name" instead of "filename" in
2410 documentation. It's nice for our code to follow the convention
2413 * filename.c: Rename to file-name.c.
2415 * filename.h: Rename to file-name.h. Update all inclusions.
2416 Update header guards.
2418 * automake.mk: Update file names.
2420 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2422 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2424 (fn_basename) Removed (dead code).
2425 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2426 (fn_special_p) Renamed fn_is_special(), all references updated.
2427 (fn_exists_p) Renamed fn_exists(), all references updated.
2429 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2431 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2432 Also, now it only tilde-expands file names, not paths.
2433 (fn_search_path) Tilde-expand one directory at a time.
2435 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2437 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2438 get rid of non-Unixlike version of the code, which has probably
2440 (fn_prepend_dir) Removed (dead code).
2442 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2443 (macro PATH_SEPARATOR_STRING) New.
2444 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2446 Continue reforming error message support. In this phase, we get
2447 rid of VM() and the other msg() support for "verbosity", replacing
2448 it by a new function verbose_msg().
2450 * filename.c: (fn_search_path) Use verbose_msg() instead of
2453 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2455 * sfm-private.h: Get rid of #defines after #error, which makes no
2458 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2460 Get rid of our own int32 type in favor of the standard int32_t
2463 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2466 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2468 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2470 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2472 Remove ill-considered file routines that are no longer used.
2474 * filename.c: (fn_open_ext) Removed.
2475 (fn_close_ext) Removed.
2477 * filename.h: (struct file_ext) Removed.
2479 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2481 * variable.c (var_is_valid_name): Move declarations before code
2484 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2486 * filename.ch (fn_interp_vars): Fixed small buglet.
2488 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2490 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2491 so as to modify the string inline. Thus makeing it easier to
2492 destroy the results when no longer needed.
2494 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2496 * category.c (cat_stored_values_destroy): Fixed memory leak.
2498 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2500 Add some missing frees. Thanks to John Darrington for reporting
2503 * any-writer.c (any_writer_close): Free writer.
2505 * any-reader.c (any_reader_close): Free reader.
2507 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2509 * por-file-reader.c: (error) Mark as NO_RETURN.
2511 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2513 * settings.c: Changed default value of scompress to true.
2515 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2517 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2519 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2522 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2523 here from pref.h.orig.
2525 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2527 * sys-file-reader.c: Fixed bug reading compressed files.
2529 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2531 * Numerous renames. See src/ChangeLog for details.
2533 * Moved files from src directory