1 2008-07-16 Jason Stover <jhs@math.gcsu.edu>
3 * category.c (cat_get_category_count): New function.
5 2008-05-15 Ben Pfaff <blp@gnu.org>
9 * data-out.c (output_infinite): Use gsl_isnan instead of isnan,
10 and gsl_isinf instead of isinf, as a stopgap measure for
11 portability until appropriate gnulib modules are available.
13 * por-file-writer.c (format_trig_double): Similarly, use
14 gsl_finite instead of finite.
16 2008-03-18 John Darrington <john@darrington.wattle.id.au>
18 * data-in.c: If category is custom currency, then use
19 numeric format for input.
21 2008-03-06 Ben Pfaff <blp@gnu.org>
23 * dictionary.c (var_name_is_insertable): New function.
24 (make_hinted_name): Don't accept variable names that match PSPP
25 keywords. Thanks to Jason Stover for reporting the problem.
27 2008-03-06 Ben Pfaff <blp@gnu.org>
29 * format-guesser.c (syntax): Require month names to be spelled out
30 as English words, so that single characters that happen to be
31 Roman numerals don't get detected as months. Thanks to John
32 Darrington for reporting this bug.
34 2008-03-04 Ben Pfaff <blp@gnu.org>
36 Patch #6441. Reviewed by John Darrington.
38 * format.c (fmt_fix): New function.
39 (fmt_fix_input): New function.
40 (fmt_fix_output): New function.
42 * format.def: Correct minimum width for DATETIME format. It was
43 7, should have been 17.
45 * automake.mk: Add new files.
47 * format-guesser.c: New file.
49 * format-guesser.h: New file.
51 2008-02-18 Ben Pfaff <blp@gnu.org>
53 Patch #6426. Reviewed by John Darrington.
55 * format.c (min_width): Renamed fmt_min_width and made public.
56 Updated all references.
57 (max_width): Renamed fmt_max_width and made public. Updated all
59 (max_decimals): Renamed fmt_max_decimals and made public. Updated
61 (var_create): Use the new functions for default variable
63 (var_default_formats): New function.
64 (var_default_measure): New function.
65 (var_default_alignment): New function.
67 * format.h (macro FMT_MAX_NUMERIC_WIDTH): New macro.
69 2008-02-09 Ben Pfaff <blp@gnu.org>
71 Add a couple of extensions to GET DATA TYPE=TXT. Patch #6412.
72 Thanks to John Darrington for review.
74 * data-in.c (data_in): Add new argument to designate the last
75 column of the data field being parsed, for use in error messages.
78 * data-parser (struct data_parser): New member `quote_escape'.
79 (data_parser_create): Initialize quote_escape.
80 (data_parser_set_quotes): New function.
81 (cut_field): Support escaped quotes.
82 (parse_delimited_span): Ditto.
83 (parse_delimited_no_span): Ditto.
85 * get-data.c (parse_get_txt): Support ESCAPE extension subcommand
86 in enhanced mode. Only support multiple quote characters in
89 2008-02-06 John Darrington <john@darrington.wattle.id.au>
91 psql-reader.c psql-reader.h: Read more than one tuple at
92 once. Fix bug reading a query which returns no data. Fix bug
93 when transformation followed a reader.
94 Ask the server for the number of records in the query, for the
97 2008-02-05 John Darrington <john@darrington.wattle.id.au>
99 psql-reader.c: So yesterday they release postgresql 8.3.0
100 which has money represented with 64 bits. They must get
103 2008-02-02 John Darrington <john@darrington.wattle.id.au>
105 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
106 for reviewing this code.
108 2008-02-02 Ben Pfaff <blp@gnu.org>
112 * sys-file-reader.c (read_variable_record): Allow missing values
113 to be specified on long string variables, but warn about them
114 because PSPP does not yet support them.
115 (read_extension_record): Ignore extension records 20 and 21, which
116 PSPP does not yet support.
117 (read_header): Fix error message when floating-point format cannot
120 2008-02-01 Ben Pfaff <blp@gnu.org>
122 Patch #6386. Thanks to John Darrington for review and for the
123 updates to gnumeric-reader.c.
125 * dictionary.c (make_hinted_name): New function.
126 (make_numeric_name): New function.
127 (dict_make_unique_var_name): New function.
129 * gnumeric-reader.c (devise_name): Removed.
130 (munge_name): Removed.
131 (gnumeric_open_reader): Use new function
132 dict_make_unique_var_name.
134 * short-names.c (set_var_short_name_suffix): Use new function
137 2008-01-19 John Darrington <john@darrington.wattle.id.au>
139 * settings.c settings.h: Moved static variables into a
140 single struct. Renamed functions to have a settings_ prefix.
142 2008-01-14 John Darrington <john@darrington.wattle.id.au>
144 * data-out.c (output_AHEX): Corrected number of bytes in
145 call to output_hex. Closes bug #22011
147 2008-01-02 John Darrington <john@darrington.wattle.id.au>
149 * variable.c variable.h: Replaced var_get_value_name with
150 var_append_value_name which doesn't use any static data.
151 Thanks to Ben for review.
153 2007-12-07 Ben Pfaff <blp@gnu.org>
157 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
158 made from an empty casereader produce a casegrouper with no
159 groups, instead of one with one group that has no cases.
161 * casereader.c (casereader_is_empty): New function.
163 2007-12-06 Ben Pfaff <blp@gnu.org>
167 * sys-file-reader.c (read_display_parameters): Handle variable
168 display parameters record with only 2 data items per variable.
169 Reported by Guido Gay <gay@irer.it>.
171 2007-12-04 Ben Pfaff <blp@gnu.org>
173 * identifier.c (lex_id_match_n): New function.
174 (lex_id_match): Reimplement in terms of lex_id_match_n.
176 2007-11-24 Ben Pfaff <blp@gnu.org>
178 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
181 2007-11-24 Ben Pfaff <blp@gnu.org>
183 Fix warning reported when reading back system files that include
184 very long string variables. Thanks to Guido Gay <gay@irer.it> for
187 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
188 was written where `j' was meant.
190 2007-11-18 Ben Pfaff <blp@gnu.org>
192 Properly write variables that include a range of missing values to
193 system files. Thanks to Guido Gay <gay@irer.it> for reporting
196 * sys-file-writer.c (write_variable): Correctly calculate
197 n_missing_values field when writing variables that include a range
200 2007-11-10 Ben Pfaff <blp@gnu.org>
202 Cleanups and bug fixes devised while writing up documentation.
205 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
207 * dict-class.c: New file.
208 (dict_class_from_id): Move here.
209 (dict_class_to_name): Move here.
211 * dict-class.h: New file.
212 (enum dict_class): Move here. Change from consecutive integers to
213 consecutive bits, to make testing for any of multiple values
214 easier. Add new DC_ALL constant.
216 * dictionary.c (struct dictionary): Change `case_limit' from
217 size_t to casenumber.
218 (dict_get_vars): Make final argument an enum dict_class.
219 (dict_get_vars_mutable): Ditto.
220 (dict_get_case_limit): Change return value to casenumber.
221 (dict_set_case_limit): Change final argument to a casenumber.
222 (dict_unset_split_var): Add assertion. Rephrase slightly.
223 (dict_set_label): Use xstrndup to simplify.
225 * format.c (fmt_step_width): AHEX format also needs 2-byte
227 (fmt_set_style): Simplify assertions.
229 * missing-values.c (mv_add_num_range): Rename mv_add_range.
230 Simplify implementation.
231 (mv_has_value): Simplify implementation.
232 (mv_pop_value): Remove the first value, not the last, to avoid
233 having GET followed by SAVE reverse the order of missing values.
234 (mv_peek_value): Rename mv_get_value. Simplify assertion.
235 (mv_has_range): Simplify implementation.
236 (mv_peek_range): Rename mv_get_range.
237 (can_resize_string): Removed.
238 (mv_is_resizable): Use value_is_resizable.
239 (mv_resize): Use value_resize.
241 * short-names.h (SHORT_NAME_LEN): Move here.
243 * val-type.h: New file, for definitions related to type and width
244 of abstract values. Before, these definitions were mixed among
245 those related to "union value" and those related to variables.
246 (macro SYSMIS): Move here.
247 (macro LOWEST): Move here.
248 (macro HIGHEST): Move here.
249 (macro MAX_STRING); Move here.
250 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
251 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
252 All references updated.
253 (val_type_is_valid): New function. Replaces var_type_is_valid.
254 All references updated.
255 (val_type_from_width): New function. Replaces
256 var_type_from_width. All references updated.
258 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
260 (val_labs_can_set_width): Use value_is_resizable.
261 (val_labs_add): Simply return false if the value labels set is too
262 wide, instead of having undefined behavior.
263 (val_labs_replace): Ditto.
264 (val_labs_replace): Ditto.
265 (val_labs_first): Set iterator to null if iteration is complete.
266 (val_labs_first_sorted): Ditto.
267 (val_labs_done): Become a no-op if the iterator is null.
269 * value.c (value_is_resizable): New function.
270 (value_resize): New function.
272 * variable.c (var_get_dict_class): New function.
274 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
277 2007-11-08 Ben Pfaff <blp@gnu.org>
279 * data-in.c: Make formatted data parsing locale-independent.
280 (parse_number): Use c_strtod instead of strtod, to avoid
281 locale-specific behavior.
284 2007-11-06 Ben Pfaff <blp@gnu.org>
286 Patch #6256: add support for binary, 360 file formats. Reviewed
289 * data-in.c (struct data_in): Add `encoding' member.
290 (data_in): Add `encoding' parameter, and re-encode the data passed
291 in where appropriate. Update all callers to pass it in.
292 (parse_A): Implement EBCDIC recoding wart described in manual.
293 (parse_AHEX): Implement EBCDIC recoding.
295 * data-out.c (data_out_legacy): New function.
296 (data_out): Make into a wrapper around data_out_legacy.
298 * file-handle-def.c (struct file_handle): New member `encoding'.
299 (fh_create_file): Set encoding.
300 (fh_default_properties): Set default encoding.
301 (fh_get_legacy_encoding): New function.
303 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
304 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
305 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
306 (struct fh_properties): New member `encoding'.
308 2007-11-05 Ben Pfaff <blp@gnu.org>
310 Patch #6258. Reviewed by John Darrington.
312 * file-handle-def.c (fh_lock): Add comment that TYPE should be
313 marked with N_() in the caller. Added these markings to each
314 caller too. Should make i18n easier.
315 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
317 2007-11-03 Ben Pfaff <blp@gnu.org>
319 Allow output files to overwrite input files (bug #21280).
321 * file-handle-def.c: Separate locking of files for input and for
322 output, to allow both to take place at once. Also, distinguish a
323 file handle from the identity of the underlying file, because the
324 identity of a file changes over time and the file handle can't
325 represent two different identities.
326 (struct file_handle): Remove `next', `open_cnt', `deleted',
327 `type', `open_mode', `aux', `identity' members. Change `id' from
328 char array to char *. Add `ref_cnt' member.
329 (file_handle_from_ll) New function.
330 (file_handles) Removed.
331 (named_handles) New variable.
332 (fh_init) Initialize named_handles.
333 (fh_done) Remove name from all named_handles.
334 (free_handle) Updated for modified struct file_handle.
335 (unname_handle) New function.
336 (fh_ref) New function.
337 (fh_from_file_name) Removed.
338 (fh_unref) New function.
339 (fh_unname) New function.
340 (fh_from_id) Rewritten.
341 (create_handle) Updated for modified struct file_handle.
342 (fh_inline_file) Increment the handle's ref count.
343 (fh_create_file) Updated for modified struct file_handle.
344 (fh_create_scratch) Ditto.
349 (fh_is_open) Removed.
350 (fh_get_id) Updated for modified struct file_handle.
351 (fh_get_default_handle) Increment the handle's ref count.
352 (fh_set_default_handle) Handle ref counts.
353 (struct fh_lock) New structure.
354 (locks) New static var.
355 (fh_lock) New function.
356 (fh_unlock) New function.
357 (fh_lock_get_aux) New function.
358 (fh_lock_set_aux) New function.
359 (fh_is_locked) New function.
360 (make_key) New function.
361 (free_key) New function.
362 (compare_fh_locks) New function.
363 (hash_fh_lock) New function.
365 * file-handle-def.h (enum fh_access) New enum.
367 * file-name.c: Made file_identity the same in all supported
369 (struct file_identity): New `name' member.
370 (fn_get_identity): For a file that doesn't exist, get the
371 dev/inode of its directory plus its name. If even the directory
372 doesn't exist, just use its name. Merge the Windows
373 implementation into the Unix one.
374 (fn_compare_file_identities): Rewritten. Merge the Windows
375 implementation into the Unix one.
376 (fn_hash_identity): New function.
378 * make-file.c (struct replace_file): New structure.
379 (all_files): New static var.
380 (replace_file_start): New function.
381 (replace_file_commit): New function.
382 (replace_file_abort): New function.
383 (free_replace_file): New function.
384 (unlink_replace_files): New function.
386 * por-file-reader.c (struct pfm_reader): Add `lock' member.
387 (close_reader): Unlock file.
388 (pfm_open_reader): Lock file.
390 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
392 (pfm_open_writer): Lock file and prepare for its replacement.
393 (close_writer): Unlock file.
395 * scratch-handle.h (struct scratch_handle): Add unique_id so that
396 different generations of a scratch file can be distinguished.
398 * scratch-reader.c (scratch_reader_open): Verify that the file is
401 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
402 remove scratch_handle member.
403 (scratch_writer_open): Lock handle. Prepare to replace handle
404 data, instead of doing it immediately.
405 (scratch_writer_casewriter_destroy): Replace handle data and
408 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
409 (sfm_open_reader): Lock file.
410 (close_reader): Unlock file.
412 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
414 (sfm_open_writer): Lock file and prepare for its replacement.
415 (close_writer): Unlock file.
417 2007-11-02 Ben Pfaff <blp@gnu.org>
419 * data-out.c (output_number): Use isfinite (from C99) instead of
420 gsl_isfinite, since we now have universal support for it in
422 (output_infinite): Ditto.
423 [!HAVE_ROUND] (round): Remove definition, since we now have a
424 replacement in gnulib.
426 2007-10-27 John Darrington <john@darrington.wattle.id.au>
428 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
429 wasn't initialised correctly. Closes bugs #21330 and 21397
432 2007-10-08 Ben Pfaff <blp@gnu.org>
434 Fix bug #21280. Thanks to John Darrington for review.
436 * file-name.c (create_stream): New function.
438 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
439 file handle before creating the file, to ensure that we don't
440 truncate a file that we're reading. Make code easier to read by
443 * sys-file-write.c (sfm_open_writer): Ditto.
445 2007-10-01 Ben Pfaff <blp@gnu.org>
447 Fix bug #21192. Thanks to John Darrington for review.
449 * casereader.c (casereader_read): Decrement case_cnt before
450 calling the casereader's "read" member function, so that we
451 interact properly with lazy_casereader.
453 * datasheet.c: Add regression test for above bug fix.
454 (clone_datasheet): New function.
455 (lazy_callback): New function.
456 (check_datasheet_casereader): New function.
457 (check_datasheet): Check datasheet contents are reported correctly
458 through an ordinary casereader and a lazy casereader.
459 (clone_model): Use clone_datasheet.
461 2007-09-24 Ben Pfaff <blp@gnu.org>
463 Patch #6210. Reviewed by John Darrington.
464 * settings.c: Drop "long view", which is not needed any longer.
465 (static var long_view): Removed.
466 (force_long_view): Removed.
467 (get_viewwidth): Removed.
468 (init_viewport): Removed long_view reference.
470 2007-09-19 John Darrington <john@darrington.wattle.id.au>
472 * settings.c settings.h: Changed viewport's length and width to be
473 owned by the user interface which uses the data library. This allows
474 better abstraction, and makes dynamically adjustable dimensions easier.
476 2007-09-18 Ben Pfaff <blp@gnu.org>
478 * procedure.c (proc_extract_active_file_data): New function.
480 * lazy-casereader.h: New file.
482 * lazy-casereader.c: New file.
484 * casereader.c (casereader_dynamic_cast): New function.
486 2007-09-14 Ben Pfaff <blp@gnu.org>
488 * dictionary.c (dict_clone): Copy case indexes from cloned
489 dictionary. Fixes bug #21061. Reviewed by John Darrington.
491 2007-09-13 John Darrington <john@darrington.wattle.id.au>
493 * value.c value.h (create_value): New function.
495 2007-09-12 Ben Pfaff <blp@gnu.org>
497 Make it clear that translator casereader and casewriter translate
498 functions are supposed to destroy their input case.
500 * casereader-translator.c (struct casereader_translator): Change
501 input case parameter of translate member function from const
502 struct ccase * to struct ccase *.
503 (casereader_create_translator): Ditto, for translate parameter.
505 * casewriter-translator.c (struct casewriter_translator): Ditto.
506 (casewriter_create_translator): Ditto.
508 2007-08-27 John Darrington <john@darrington.wattle.id.au>
510 * sys-file-reader.c (read_display_parameters): Force display width
511 to 8, if the sys file says 0 (like SPSS does).
513 2007-08-12 Ben Pfaff <blp@gnu.org>
515 * dictionary.c (dict_dump): New function.
517 2007-08-12 Ben Pfaff <blp@gnu.org>
519 Drop dict_compactor in favor of using the new struct case_map.
521 * dictionary.c (struct copy_map): Removed.
522 (struct dict_compactor): Removed.
523 (dict_make_compactor): Removed.
524 (dict_compactor_compact): Removed.
525 (dict_compactor_destroy): Removed.
527 * procedure.c (struct data_set): Change `compactor' member to be a
529 (proc_open): Use case_map_to_compact_dict instead of
531 (proc_casereader_read): Use case_map_execute instead of
532 dict_compactor_compact.
533 (proc_commit): Use case_map_destroy instead of
534 dict_compactor_destroy.
536 * scratch-writer.c (struct scratch_writer): Change `compactor'
537 member to be a struct case_map *.
538 (scratch_writer_open): Use case_map_to_compact_dict instead of
540 (scratch_writer_casewriter_write): Use case_map_execute instead of
541 dict_compactor_compact.
543 2007-08-12 Ben Pfaff <blp@gnu.org>
545 * automake.mk: Add case-map.c, case-map.h.
547 * case-map.c: New file.
549 * case-map.h: New file.
551 2007-08-12 Ben Pfaff <blp@gnu.org>
553 * dictionary.c (dict_compact_values): Don't delete scratch
554 variables as well as compacting case indexes. Update all callers.
555 (dict_get_compacted_value_cnt): Rename dict_count_values and
556 change interface. Update all callers.
557 (dict_get_compacted_value_cnt): Remove.
558 (dict_compacting_would_shrink): Remove.
559 (dict_compacting_would_change): Remove.
560 (dict_make_compactor): Add new parameter. Update all callers.
562 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
563 instead of dict_count_values, changing an O(N) operation into
566 2007-08-12 Ben Pfaff <blp@gnu.org>
568 * casereader.c (casereader_read): Don't require cases read by a
569 casereader to be exactly the expected size: as long as they're big
572 2007-08-12 Ben Pfaff <blp@gnu.org>
574 Make casewriters keep track of the number of `union value's in
575 each case. This is useful for two reasons: casewriter_write can
576 then check that the case being written is large enough, and later
577 recipients of the casewriter can determine the size of the case.
579 * casewriter-translator.c (casewriter_create_translator): Add
582 * casewriter.c (struct casewriter): Add value_cnt member.
583 (casewriter_write): Check that the case passed in is big enough.
584 (casewriter_get_value_cnt): New function.
585 (casewriter_create): Add value_cnt parameter.
587 2007-08-09 Ben Pfaff <blp@gnu.org>
589 Fix bug reported by Jason Stover.
590 * settings.c: Move get_termcap_viewport and in particular the
591 #include for <curses.h> to the end of the file. curses.h
592 redefines bool on some systems (e.g. OpenBSD), which causes
593 disagreement between uses of bool before its inclusion and
595 Tested by Jason Stover.
597 2007-07-29 Ben Pfaff <blp@gnu.org>
599 Provisional fix for bug #18692 and bug #20161. Reviewed by John
602 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
603 argument (never "rb" or "wb") because SUSv3 says that only those
604 modes are defined, and glibc in fact rejects other modes.
606 Open portable files with fn_open so that they can be read from
607 pipes. Fix missing fh_close call to go along with fh_open.
608 Report an error if the file close reports an error.
609 * por-file-reader.c (close_reader): New function.
610 (por_file_casereader_destroy): Use close_reader.
611 (pfm_open_reader): Open file with fn_open.
613 2007-07-28 Ben Pfaff <blp@gnu.org>
615 Make PSPP able to read all the portable files I could find on the
616 web. Thanks to John Darrington for review. Bug #17620.
617 * por-file-reader.c (struct pfm_reader): New member `line_length'.
618 (error): Print file offset in hexadecimal.
619 (warning): New function.
620 (advance): Treat lines less than 80 bytes long as padded to 80
622 (pfm_open_reader): Call read_documents if we find an "E" record.
623 (convert_format): Convert invalid formats to the default format
624 instead of aborting reading the file.
625 (read_variables): Rename duplicate variable names instead of
626 aborting reading the file.
627 (read_value_label): Allow string variables of different widths to
628 be assigned value labels in the same record. Replace duplicate
629 value labels instead of aborting.
630 (read_documents): New function.
632 * por-file-writer.c (pfm_open_writer): Call write_documents if the
633 dictionary has documents.
634 (write_documents): New function.
636 2007-07-25 Ben Pfaff <blp@gnu.org>
638 Fix bugs related to bug #17213.
640 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
641 since the former is what config.h has. Include the needed ncurses
643 (static var echo) Rename to `do_echo' because the original name is
644 the same as an ncurses identifier.
645 (get_termcap_viewport) Use error instead of msg.
647 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
648 (fn_close): Don't close stdin, stdout, stderr.
650 2007-07-26 John Darrington <john@darrington.wattle.id.au>
652 * procedure.c procedure.h: Added callbacks which get invoked whenever
653 a dataset's transformation chain changes.
655 2007-07-24 Ben Pfaff <blp@gnu.org>
658 * sys-file-writer.c (write_variable_display_parameters): Use new
659 var_default_display_width function to choose display width of
660 segments after the first one in a given variable.
661 * variable.c (var_create): Use var_default_display_width to pick
662 new variable's display width.
663 (var_default_display_width): New function.
664 Reviewed by John Darrington.
666 2007-07-24 Ben Pfaff <blp@gnu.org>
669 * por-file-writer.c (write_variables): Write weight variable.
670 Reviewed by John Darrington.
672 2007-07-23 Ben Pfaff <blp@gnu.org>
674 Improvements to system file reader and writer.
676 First, move all detailed knowledge of very long strings into
677 sys-file-private.[ch], so that this nasty stuff can be isolated.
679 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
680 (EFFECTIVE_VLS_CHUNK): New macro.
681 (min_int): New function.
682 (max_int): New function.
683 (sfm_width_to_bytes): Rewrite.
684 (sfm_width_to_octs): New function.
685 (sfm_segment_alloc_width): New function.
686 (sfm_segment_alloc_bytes): New function.
687 (sfm_segment_used_bytes): New function.
688 (sfm_segment_offset): New function.
689 (sfm_segment_effective_offset): New function.
690 (sfm_dictionary_to_sfm_vars): New function.
692 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
693 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
694 (struct sfm_var): New structure.
696 Next, improvements to the system file reader.
698 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
699 type casenumber. Added `version_major', `version_minor',
702 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
703 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
704 Change `case_cnt' to type casenumber. Removed `has_vls'.
705 (struct sfm_var): Removed.
706 (sfm_open_reader): Don't warn on wrong case size if the file was
707 written by SPSS 13, which tends to get it wrong. Use
708 sfm_dictionary_to_sfm_vars.
709 (read_header): Always output system file info.
710 (read_variable_record): Simplify code for reading missing values.
711 (read_machine_int32_info): Save version numbers from system file
712 into info struct passed as new argument.
713 (read_long_string_map): Restructured to use new sys-file-private
715 (read_value_labels): Use size_overflow_p.
716 (sys_file_casereader_read): Get rid of distinction between fast
717 and slow paths. Use information provided by sys-file-primate's
718 struct sfm_var to simplify code.
719 (skip_whole_strings): New function.
720 (read_int32): Renamed read_int. Changed return value to int.
722 (read_flt64): Renamed read_float. Changed return value to
723 double. Updated all callers.
724 (int32_to_native): Removed. Changed callers to use
726 (flt64_to_double): Removed. Changed callers to use float_convert.
728 Finally, get rid of int32, flt64 terminology and types in system
729 file writer. The former wasn't very useful since a POSIX "int"
730 can hold the whole range of int32 and we generally didn't have a
731 need for it to be exactly-32-bits, just at-least-32-bits. The
732 latter was inconvenient because we had to assume that it could be
733 different from double and thereby convert special values SYSMIS,
734 HIGHEST, LOWEST to and from it in multiple places. Instead, now
735 we just use "int" and "double" in most places, and do conversions,
736 if necessary, very close to where we do I/O. This change meant
737 that the writer code couldn't represent records in the file as C
738 structs any longer, but that's no great loss. The code actually
739 seems to be more readable without them.
741 Simplify the compression buffering code: only buffer as much as
742 necessary, which is no more than eight 8-byte units at any given
745 * sys-file-writer.c (typedef flt64): Removed.
746 (macro second_lowest_flt64): Removed.
747 (struct sysfile_header): Removed.
748 (struct sysfile_variable): Removed.
749 (struct sfm_writer): Removed `needs_translation', `has_vls',
750 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
751 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
752 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
753 compression buffering by `opcodes', `opcode_cnt', `data',
754 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
755 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
756 data writer functions instead of structures.
757 (calc_oct_idx): New function.
758 (write_header): Use simple data writer functions instead of
760 (write_format_spec): Renamed write_format. New argument.
761 (write_variable_continuation_records): New function.
762 (write_variable): Use simple data writer functions instead of
763 structures. Use write_variable_continuation_records. Write
764 entire very long string instead of requiring caller to understand
766 (write_value_labels): Use simple data writer functions instead of
768 (write_documents): Ditto.
769 (write_variable_display_parameters): Use sys-file-private
770 functions to simplify. Use simple data writer functions instead
772 (write_vls_length_table): Use simple data writer functions instead
774 (write_longvar_table): Ditto.
775 (write_rec_7_34): Break into new functions
776 write_integer_info_record, write_float_info_record. Use simple
777 data writer functions instead of structures.
778 (buf_write): Removed.
779 (append_string_max): Removed.
780 (ensure_buf_space): Removed.
781 (sys_file_casewriter_write): Get rid of the distinction between
782 fast and slow paths, which didn't seem to be too useful. Use new
783 functions write_case_uncompressed, write_case_compressed.
784 (put_instruction): Removed.
785 (put_element): Removed.
786 (write_compressed_data): Removed.
787 (close_writer): Use flush_compressed. Only write case count to
788 system file if it will fit in the field.
789 (write_case_compressed): New function.
790 (write_case_uncompressed): New function.
791 (flush_compressed): New function.
792 (put_cmp_opcode): New function.
793 (put_cmp_number): New function.
794 (write_int): New function.
795 (convert_double_to_output_format): New function.
796 (write_float): New function.
797 (write_value): New function.
798 (write_string): New function.
799 (write_bytes): New function.
800 (write_zeros): New function.
801 (write_spaces): New function.
803 Reviewed by John Darrington.
805 2007-07-22 Ben Pfaff <blp@gnu.org>
807 Don't try to write very long strings to portable files. The
808 format does not support it.
810 * por-file-writer.c (MAX_POR_WIDTH): New macro.
811 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
812 (write_format): Add arg to take width to resize format to.
813 (write_value): Limit width of value written to MAX_POR_WIDTH.
814 (write_variables): Limit width of variable and its output formats
816 Reviewed by John Darrington.
818 2007-07-22 Ben Pfaff <blp@gnu.org>
820 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
821 local variable instead of literal 5.
822 Reviewed by John Darrington.
824 2007-07-22 Ben Pfaff <blp@gnu.org>
826 Fix problems with uniqueness of short names in system files with
827 very long string variables. Now a variable may have multiple
830 * automake.mk (src_data_libdata_a_SOURCES): Add new files
831 short-names.c, short-names.h.
833 * dictionary.c (dict_clone): Clone all the short names.
834 (compare_strings): Move into short-names.c.
835 (hash_strings): Ditto.
836 (set_var_short_name_suffix): Ditto.
837 (dict_assign_short_names): Ditto, rename short_names_assign,
838 change to assign all short names.
840 * por-file-writer.c (write_variables): Use short_names_assign
841 instead of dict_assign_short_names.
843 * short-names.c: New file.
845 * short-names.h: New file.
847 * sys-file-private.c (sfm_width_to_segments): New function.
849 * sys-file-reader.c (read_long_var_name_map): Save and restore all
850 the short names, not just the first one.
852 * sys-file-writer.c (cont_var_name): Removed.
853 (sfm_open_writer): Use short_names_assign instead of
854 dict_assign_short_names. Use unique short names assigned by
855 short_names_assign instead of those generated by cont_var_name.
857 * variable.c (struct variable): Remove `short_name' member,
858 replace by `short_names' and `short_name_cnt'.
859 (var_create) Initialize new members.
860 (var_get_short_name_cnt): New function.
861 (var_get_short_name): Now takes an index argument. Changed most
863 (var_set_short_name): Ditto.
864 (var_clear_short_name): Renamed var_clear_short_names, changed to
865 clear all short names.
867 Reviewed by John Darrington.
869 2007-07-22 Ben Pfaff <blp@gnu.org>
871 * variable.c (var_set_width): Use new var_set_width function.
873 * missing-values.c (mv_n_values): Drop assertion, which was not
876 * format.c (fmt_default_for_width): New function.
877 (fmt_resize): New function.
879 Reviewed by John Darrington.
881 2007-07-18 John Darrington <john@darrington.wattle.id.au>
883 * datasheet.c (datasheet_delete_columns): Added assertion to check
884 we're not deleting outside the range of the sheet.
887 * dictionary.c dictionary.h variable.c: Added the ability for string
888 variables to be resized.
890 * vardict.h: Added some prototypes (moved from dictionary.h) as
891 these should only be called by variable.c
894 2007-07-14 John Darrington <john@darrington.wattle.id.au>
896 * sfm-reader.c: Respect case_cnt field in file header.
898 2007-07-01 John Darrington <john@darrington.wattle.id.au>
900 * transformation.c transformation.h (trns_chain_execute): Changed the
901 signature (Patch #6057)
903 2007-06-10 Ben Pfaff <blp@gnu.org>
905 * casereader-filter.c (casereader_filter_destroy): Make sure to
906 write all the remaining excluded cases to the casewriter, if any.
908 * caseinit.c (init_list_destroy): Rewrite.
909 (init_list_clear): Ditto.
911 * casegrouper.c (casegrouper_get_next_group): Always set *reader
912 to null when returning false.
914 2007-06-06 Ben Pfaff <blp@gnu.org>
916 Actually implement the new procedure code and adapt all of its
917 clients to match. Also adapt all of the other case sources and
918 sinks in the tree and their clients to use the
919 casereader/casewriter infrastructure.
921 * automake.mk: Add and remove files.
923 * any-reader.c: Change into a casereader.
924 * por-file-reader.c: Ditto.
925 * scratch-reader.c: Ditto.
926 * sys-file-reader.c: Ditto.
928 * any-writer.c: Change into a casewriter.
929 * por-file-writer.c: Ditto.
930 * scratch-writer.c: Ditto.
931 * sys-file-writer.c: Ditto.
933 * procedure.c: Change to use casereader, casewriter, caseinit, and
934 other new infrastructure.
936 * scratch-handle.c: Adapt to new infrastructure.
938 * case-sink.c: Removed, now dead code.
939 * case-sink.h: Ditto.
940 * case-source.c: Ditto.
941 * case-source.h: Ditto.
942 * casefile-factory.c: Ditto.
943 * casefile-private.h: Ditto.
946 * casefilter.c: Ditto.
947 * casefilter.h: Ditto.
950 * fastfile-factory.c: Ditto.
951 * fastfile-factory.h: Ditto.
952 * storage-stream.c: Ditto.
953 * storage-stream.h: Ditto.
955 2007-06-06 Ben Pfaff <blp@gnu.org>
959 * automake.mk: Add new files.
961 * datasheet.c: New file.
963 * datasheet.h: New file.
965 2007-06-06 Ben Pfaff <blp@gnu.org>
967 Until now, the procedure code has provided a case to the
968 case_source, which has filled in the data values that come from
969 the active file. "Left" data values that don't come from the
970 active file naturally stay the same from case to case, because the
971 procedure code keeps using that same case.
973 One of the compromises that comes with the new procedure code is
974 that the active file allocates and provides its own case, which
975 the procedure code then has to resize to provide room for any
976 other variables that should go in the case and then fill in the
977 values of "left" variables. Then, when we're done with that case,
978 we have to save the values of "left" variables to copy into the
979 next case read from the active file.
981 The caseinit code helps with this.
983 * automake.mk: Add new files.
985 * caseinit.c: New file.
987 * caseinit.h: New file.
989 2007-06-06 Ben Pfaff <blp@gnu.org>
991 * value.h (value_cnt_from_width): New function.
993 * variable.c (var_get_value_cnt): Use new function.
995 2007-06-06 Ben Pfaff <blp@gnu.org>
997 Add casegrouper, to allow cases read from a given casereader to be
998 broken into groups, each of which has its own casereader.
999 Generally cases are grouped based on having equal values for some
1002 * automake.mk: Add new files.
1004 * casegrouper.c: New file.
1006 * casegrouper.h: New file.
1008 2007-06-06 Ben Pfaff <blp@gnu.org>
1010 Add interface to lexicographical ordering of cases.
1012 * automake.mk: Add new files.
1014 * case-ordering.c: New file.
1016 * case-ordering.h: New file.
1018 2007-06-06 Ben Pfaff <blp@gnu.org>
1020 Add casereaders and casewriters, the basis of the new data processing
1021 implementation. A casereader is a uniform interface to reading cases
1022 from a data source; a casewriter is a uniform interface to writing
1023 cases to a data sink.
1025 * automake.mk: Add new files.
1027 * casereader-filter.c: New file.
1029 * casereader-provider.h: New file.
1031 * casereader-translator.c: New file.
1033 * casereader.c: New file.
1035 * casereader.h: New file.
1037 * casewriter-provider.h: New file.
1039 * casewriter-translator.c: New file.
1041 * casewriter.c: New file.
1043 * casewriter.h: New file.
1045 2007-06-06 Ben Pfaff <blp@gnu.org>
1047 "casewindow" data structure that extends the deque (from libpspp)
1048 of cases with the ability to dump cases to disk if we get too many
1051 * automake.mk: Add new files.
1053 * casewindow.c: New file.
1055 * casewindow.h: New file.
1057 2007-06-06 Ben Pfaff <blp@gnu.org>
1059 sparse_cases data structure that augments a sparse_array of cases
1060 with the ability to dump cases to disk if we get too many cases in
1063 * automake.mk: Add new files.
1065 * sparse-cases.c: New file.
1067 * sparse-cases.h: New file.
1069 2007-06-06 Ben Pfaff <blp@gnu.org>
1071 Adds a low-level on-disk case array data structure.
1073 * automake.mk: Add new files.
1075 * case-tmpfile.c: New file.
1077 * case-tmpfile.h: New file.
1079 2007-06-06 Ben Pfaff <blp@gnu.org>
1081 In a couple of places we calculate the maximum number of cases to
1082 keep in memory based on the user-defined workspace. Enable
1083 centralizing the calculation through a new function.
1085 * settings.c (get_workspace_cases): New function.
1087 2007-06-06 Ben Pfaff <blp@gnu.org>
1089 The casenumber type is defined in transformations.h, but case.h is
1090 a more sensible place. Move it.
1092 * case.h (CASENUMBER_MAX): New macro.
1093 (typedef casenumber): Move here, from transformations.h.
1095 2007-06-03 Ben Pfaff <blp@gnu.org>
1097 Slightly generalize case_to_values and case_from_values functions.
1099 * case.c (case_to_values): Rename case_copy_out, change interface.
1100 (case_from_values): Rename case_copy_in, change interface.
1102 * fastfile.c (fastfilereader_get_next_case): Update caller.
1103 (write_case_to_disk): Ditto.
1105 2007-06-02 Ben Pfaff <blp@gnu.org>
1107 Clean up after a forgotten part of patch #5829.
1109 * casedeque.h: Remove unused file.
1111 * automake.mk: Remove casedeque.h from sources.
1113 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1115 * category.c: Removed redundant #include
1117 2007-05-06 Ben Pfaff <blp@gnu.org>
1119 Abstract the documents within a dictionary a little better.
1120 Thanks to John Darrington for suggestion, initial version, and
1121 review. Patch #5917.
1123 * dictionary.c (struct dictionary): Change `documents' member from
1124 char * to struct string.
1125 (dict_clear): Destroy struct string.
1126 (dict_get_documents): Convert struct string to char *.
1127 (dict_set_documents): Set struct string. Pad to 80-character
1129 (dict_clear_documents): New function.
1130 (dict_add_document_line): New function.
1131 (dict_get_document_line_cnt): New function.
1132 (dict_get_document_line): New function.
1134 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1136 * sys-file-reader.c (read_documents): Use new document functions.
1138 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1140 * sys-file-reader.c: When reading a system file which has no
1141 long name table, automatically create one where the long names
1142 are the lower case versions of the short names.
1144 2007-04-22 Ben Pfaff <blp@gnu.org>
1146 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1147 dict_clear will free most data related to the dictionary.
1148 dict_clear does a decent job, except that dict_set_split_vars on
1149 some systems won't actually free the dict's "split" member.
1150 Instead, it'll allocate a 1-byte region. Fix this.
1152 * value.c (value_copy): New function.
1153 (value_set_missing): Ditto.
1155 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1157 * Deleted existing category.h and moved cat-routines.h into
1158 category.h Encapsulated struct cat_vals better.
1160 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1162 * sys-file-reader.c: When reading a system file which has no
1163 long name table, automatically create one where the long names
1164 are the lower case versions of the short names.
1166 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1168 * sys-file-reader.c: Some versions of Other Software seem to
1169 produce system files with string variables' measure set to
1170 zero. We'll assume these are supposed to be nominal variables.
1172 2007-03-30 Ben Pfaff <blp@gnu.org>
1174 * procedure.c: Adapt to new deque data structure.
1176 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1177 Ben Pfaff <blp@gnu.org>
1179 * file-name.c: Mingw compatibility fixes.
1180 (fn_search_path): Use ISSLASH instead of comparing against '/'
1182 (fn_dir_name): Use dir_name from gnulib.
1183 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1184 (fn_get_identity): Use GetFullPathName instead of canonicalize
1185 from gnulib, because the latter does not fully support
1186 Windows-style path names. Use this implementation based on the
1187 detected presence of Windows instead of the absence of Unix, since
1188 the new implementation is Windows-specific.
1189 (fn_compare_file_identities): In Windows implementation, compare
1190 names case-insensitively.
1192 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1194 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1195 <stdlib.h> to have the same effect.
1197 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1199 * por-file-reader.c: Add missing _() around messages.
1201 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1203 * make-file.c: Include "mkstemp.h", without which linking on
1206 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1208 Reduce platform dependence.
1210 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1211 Everywhere we using this, we really should have just depended on
1212 the shell to expand tildes.
1213 (fn_search_path): Simplify, given that we don't do tilde expansion
1215 (fn_normalize): Removed. Caller changed to use the canonicalize
1217 (fn_get_cwd): Removed. Only user was fn_normalize.
1218 (fn_is_absolute): Really only test for absolute names.
1219 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1221 (fn_readlink): Removed, as it was only used fn_normalize.
1222 (fn_exists): Assume the stat function is available; gnulib does.
1223 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1225 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1227 * dictionary.c (dict_create_vector_assert): New function.
1229 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1231 * file-name.c (fn_normalize): Correct name of function
1232 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1233 for pointing this out.
1235 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1237 We are using a single member in struct file_handle, the "name"
1238 field, for more than one purpose. When it begins with '"', it's a
1239 file name; otherwise, it's a token that can be used to identify
1240 it. When that assertion fires, it's because we searched for the
1241 name case-sensitively as a file name (so that there was no match),
1242 and then we try to insert it case-insensitively as a token, which
1243 fails because duplicates aren't allowed.
1245 Solution: break the two purposes into two separate fields. This
1246 fixes the problem and likely makes the code easier to read too.
1248 Fixes bug #18922. Thanks to John Darrington for bug report and
1251 * file-handle-def.c (struct file_handle): New `id' member.
1252 (fh_from_name): Rename fh_from_id. Update all callers.
1253 (create_handle): New `id' parameter. Update all callers.
1254 (fh_create_file): Ditto.
1255 (fh_get_id): New function.
1257 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1259 * case.c (case_is_null): Change return type to bool.
1261 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1265 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1266 reference counting to share data. This makes it easy for
1267 valgrind, etc. to find accesses to cases that have been destroyed
1268 but have been kept around by another user's ref-count. This often
1269 happens when the data set is small enough to find in memory; if a
1270 bigger data set that would overflow to disk were used, then data
1271 corruption would occur.
1273 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1277 * case.c (case_unshare): Make it check internally whether the
1278 ref_cnt is greater than 1, so that the callers don't have to.
1279 Update callers not to check.
1281 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1283 Before, I was thinking that I might want to get rid of reference
1284 counting at some point. Now, I'm pretty sure that it's here to
1285 stay. Thus, because we have to store the value_cnt anyway for
1286 reference-counted cases, we might as well expose it to users.
1288 * case.c (case_get_value_cnt): New function.
1289 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1290 resize case if its size actually changed.
1292 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1293 instead of peeking inside struct case directly.
1294 (casefile_append): Ditto.
1296 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1298 Get rid of the inlines for the case functions, which made the
1299 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1300 the inlines didn't speed up "make check" at all, which is not a
1301 perfect benchmark but seems indicative.)
1303 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1304 definitions. Remove some assertions on nonnull pointers that were
1305 redundant with a pointer dereference soon after in the function.
1307 (struct case_data): Move definition here from case.h.
1311 (case_data_wr): Ditto.
1313 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1315 * automake.mk: Add casedeque.h to sources.
1317 * casedeque.h: New file.
1319 * procedure.c: (struct dataset) Change lag_count, lag_head,
1320 lag_queue member into single struct casedeque member. Update all
1321 users to use the casedeque instead.
1324 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1326 * procedure.c: Simplify lagged cases interface. Updated all
1327 clients--well, the only client--to use the simplified interface.
1328 (dataset_n_lag) Removed.
1329 (dataset_set_n_lag) Removed.
1330 (dataset_need_lag) New function.
1332 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1334 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1335 whenever appropriate, but only when the dataset/dictionary is in a
1338 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1340 * dictionary.c dictionary.h : Added callbacks for change of filter and
1341 split variables. Refactored some code to ensure that callbacks get
1342 invoked when appropriate.
1344 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1345 replace_dict callback occurs when permanent_dict replaces the current
1348 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1350 * dictionary.c dictionary.h : Added callback for when the weight
1351 variable of a dictionary changes.
1353 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1355 * dictionary.c dictionary.h : Added replace_source and replace_dict
1356 callbacks, and functions to deal with them.
1358 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1360 Simplify missing value handling.
1362 * missing-values.h (enum mv_class): New type.
1363 (enum mv_type): Moved definition into missing-values.c and renamed
1364 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1365 enums. Updated all uses.
1366 (struct missing_values): Changed type of `type' from `enum
1367 mv_type' to `int' because the definition is no longer exposed.
1369 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1370 parameter. Update all callers.
1371 (mv_is_num_missing): Ditto.
1372 (mv_is_str_missing): Ditto.
1373 (mv_is_value_user_missing): Removed. Changed callers to use
1374 mv_is_value_missing.
1375 (mv_is_num_user_missing): Removed. Changed callers to use
1377 (mv_is_str_user_missing): Removed. Changed callers to use
1379 (mv_is_value_system_missing): Removed. Changed callers to use
1380 mv_is_value_missing.
1381 (mv_set_type): Removed. Changed callers to use mv_clear.
1382 (mv_clear): New function.
1384 * variable.c (var_is_value_missing): Add new enum mv_class
1385 parameter. Update all callers.
1386 (var_is_num_missing): Ditto.
1387 (var_is_str_missing): Ditto.
1388 (var_is_value_user_missing): Removed. Changed callers to use
1389 var_is_value_missing.
1390 (var_is_num_user_missing): Removed. Changed callers to use
1392 (var_is_str_user_missing): Removed. Changed callers to use
1394 (var_is_value_system_missing): Removed. Changed callers to use
1395 var_is_value_missing.
1397 * casefilter.c (struct casefilter): Use enum mv_class in place of
1399 (casefilter_variable_missing): Adapt to new member.
1400 (casefilter_create): Change signature to take enum mv_class,
1403 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1405 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1407 * case-sink.c case-sink.h procedure.c procedure.h
1408 storage-stream.c: Now uses the factory.
1410 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1412 Make it possible to pull cases from the active file with a
1413 function call, instead of requiring indirection through a callback
1416 * case-source.h (struct case_source_class): Change ->read function
1417 to return a single case, instead of calling a callback function
1418 for each case. Change ->destroy function to return an error
1421 * case-source.c (free_case_source): Pass along the value returned
1422 by the case_source ->destroy function.
1424 * procedure.c (struct write_case_data): Removed.
1425 (struct dataset): Added some members to track procedure state.
1426 (procedure): Optimize the trivial case at this level.
1427 (internal_procedure): Re-implement in terms of proc_open,
1428 proc_read, proc_close.
1429 (proc_open) New function.
1430 (proc_read) New function.
1431 (proc_close) New function.
1432 (write_case) Moved into proc_read.
1433 (close_active_file) Moved closing of data source into proc_close.
1435 * storage-source.c: Rewrote to conform with modified
1436 case_source_class interface.
1438 * transformations.c (trns_chain_execute): Added argument to allow
1439 starting execution from an arbitrary transformation. Updated
1442 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1444 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1446 * sys-file-reader.c (read_display_parameters): Don't assume that
1447 MEASURE_* and ALIGN_* have the same values found in system files.
1449 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1451 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1452 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1453 the value of n_MEASURES, which was off by 1 (at least from my
1456 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1458 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1459 callbacks which are invoked when the dictionary or its
1460 variables are changed.
1462 * missing-values.c missing-values.h value-labels.c: Tidied up
1463 consistency checks, and made some of them return false
1464 instead of assert-failing.
1466 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1468 * calendar.c (calendar_days_in_month): New function.
1470 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1472 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1473 the value label's width.
1475 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1477 * sfm-private.h: Move contents into sys-file-writer.c, which is
1478 the only remaining user. Removed Borland C++-specific directives.
1480 * sys-file-reader.c: Clean up and rewrite entire file. The
1481 rewritten version is simpler and better abstracted, and should be
1482 easier to maintain and extend. It avoids using structures to read
1483 file data, which is prone to padding variations among compilers.
1484 It should also handle non-IEEE 754 system files, although I
1485 haven't been able to find any. It has been tested against many
1486 .sav files obtained from the Web and found to produce the same
1487 results as the earlier version of the code, or in some cases
1488 improved results. It is more tolerant of format variations found
1491 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1492 member, putting an enum integer_format in its place. New member
1493 `float_format'. Changed `compressed' member to type bool.
1495 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1497 * dictionary.c (dict_delete_consecutive_vars): New function.
1499 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1501 * file-name.c (fn_search_path): Remove prefix arg that was unused
1502 by any caller. Updated all callers.
1504 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1506 * format.c (fmt_dollar_template): Use user's decimal point
1507 character. Add assertion.
1509 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1511 * format.c (fmt_dollar_template): New function, based on
1512 dollar_format_template from var-type-dialog.c.
1514 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1516 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1519 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1521 Finish converting struct variable to an opaque type. In this
1522 phase, we add remaining setter and getter functions, convert the
1523 remaining PSPP code to use them, and do a bunch of cleanup. The
1524 resulting changes are pervasive but mostly trivial, and only the
1525 notable changes are logged.
1527 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1530 * case.c (case_data): Renamed case_data_idx.
1531 (case_num): Renamed case_num_idx.
1532 (case_str): Renamed case_str_idx.
1533 (case_data_rw): Renamed case_data_rw_idx.
1535 * case.h (case_data): New function with old name and an interface
1536 that takes a variable instead of an index, which is easier to
1537 use. Updated all callers to use the new interface, or to use the
1538 new *_idx function (see above).
1541 (case_data_rw): Ditto.
1543 * category.c (cat_stored_values_destroy): Changed interface to
1544 take a struct cat_vals * instead of a struct variable *.
1546 * dictionary.c (dict_clone): Use new vector_clone function.
1547 (dict_clear) Use new var_destroy function.
1548 (add_var) New function.
1549 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1550 (dict_create_var_assert) Rewrite in terms of add_var.
1551 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1552 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1553 (dict_lookup_var) Use new var_create, var_destroy functions.
1554 (dict_contains_var) Rewrite in terms of new vardict functionality.
1555 (set_var_dict_index) New function.
1556 (set_var_case_index) New function.
1557 (reindex_vars) New function.
1558 (dict_delete_var) Rewrite in terms of new vardict functionality.
1559 (dict_reorder_var) Ditto.
1560 (dict_reorder_vars) Ditto.
1561 (rename_var) New function.
1562 (dict_rename_var) Use rename_var.
1563 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1564 (dict_get_compacted_idx_to_fv) Rename
1565 dict_get_compacted_dict_index_to_case_index, update callers.
1566 (dict_create_vector) Use new vector_create function.
1567 (dict_clear_vectors) Use new vector_destroy function.
1568 (set_var_short_name_suffix) Move here from variable.c, renamed
1569 from var_set_short_name_suffix, make static, update caller.
1571 * sys-file-private.c: New file.
1572 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1573 width_to_bytes, update callers.
1575 * sys-file-private.h: New file. Later it will supplant
1576 sfm-private.h; for now it supplements it.
1577 (macro MIN_VERY_LONG_STRING) New macro.
1578 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1580 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1581 MAX_LONG_STRING was used before.
1583 * sys-file-writer.c: Ditto.
1585 * value-labels.c: Change the paradigm here to be that a null
1586 pointer is OK for a struct val_labs * in most cases; it just
1587 represents an empty set of value labels.
1588 (val_labs_copy) A copy of a null set is a null set.
1589 (val_labs_count) A null set has 0 labels.
1590 (val_labs_replace) Change return type to void. Rewrite for
1592 (val_labs_find) A null set does not contain the value.
1593 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1594 transposed argument order, updated all callers.
1596 * value.c: New file.
1597 (value_dup) Moved here from variable.c.
1598 (compare_values) Ditto.
1601 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1602 (macro MAX_LONG_STRING) Removed, because it was only interesting
1603 for system files, not for general code.
1604 (macro MAX_VERY_LONG_STRING) Ditto.
1605 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1606 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1608 * vardict.h: New file, for an interface between variables and
1611 * variable.c: A lot of functions were moved around, for better
1613 (struct variable) Move definition here, from variable.h.
1614 (var_type_adj) Removed--makes i18n hard.
1615 (var_type_noun) Ditto.
1616 (var_create) New function.
1617 (var_clone) New function.
1618 (var_destroy) New function.
1619 (var_set_name) Assert that variable is not in a dictionary.
1620 (compare_var_names) Rename compare_vars_by_name and fix a couple
1621 of callers who thought the args were strings.
1622 (hash_var_name) Rename hash_var_by_name.
1623 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1624 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1625 (var_is_very_long_string) Removed, because it was only interesting
1626 to system file code.
1627 (var_set_missing_values) Allow the argument to be the wrong width,
1628 as long as we can resize it. Simplify callers who were doing the
1629 resizing themselves.
1630 (var_get_value_labels) New function.
1631 (var_has_value_labels) New function.
1632 (var_set_value_labels) New function.
1633 (alloc_value_labels) New function.
1634 (var_add_value_label) New function.
1635 (var_replace_value_label) New function.
1636 (var_clear_value_labels) New function.
1637 (var_lookup_value_label) New function.
1638 (var_get_value_name) Moved here from variable.c, renamed from
1639 var_get_value_name, transposed argument order, updated all
1641 (var_to_string) Moved here, from variable-label.c.
1642 (var_set_leave) New function.
1643 (var_get_leave) New function.
1644 (var_must_leave) New function.
1645 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1646 set_var_short_name_suffix.
1647 (var_get_dict_index) New function.
1648 (var_get_case_index) New function.
1649 (var_get_obs_vals) New function.
1650 (var_set_obs_vals) New function.
1651 (var_has_obs_vals) New function.
1652 (var_get_vardict) New function.
1653 (var_set_vardict) New function.
1654 (var_has_vardict) New function.
1655 (var_clear_vardict) New function.
1656 (value_dup) Moved to value.c.
1657 (compare_values) Ditto.
1660 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1661 (enum ALPHA) Rename VAR_STRING, update all users.
1663 * vector.c: New file.
1664 (struct vector) Moved here, from variable.h.
1665 (check_widths) New function.
1666 (vector_create) New function.
1667 (vector_clone) New function.
1668 (vector_destroy) New function.
1669 (vector_get_name) New function.
1670 (vector_get_var) New function.
1671 (vector_get_var_cnt) New function.
1672 (compare_vector_ptrs_by_name) New function.
1674 * vector.h: New file.
1676 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1678 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1679 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1682 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1684 * value-labels.c (destroy_atoms): New function.
1685 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1687 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1689 Thanks to Jason Stover for pointing out this problem.
1691 * data-out.c (output_number): Use gsl_finite from GSL, which is
1692 portable, instead of isfinite, which is not.
1695 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1697 * variable.c variable.h (value_dup): New function.
1699 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1701 Start converting struct variable to an opaque type. In this
1702 phase, we add a bunch of setter and getter functions and convert
1703 most of the PSPP code to use them. The resulting changes are
1704 pervasive but mostly trivial, and only the notable changes are
1707 * format.c (fmt_equal): New function.
1709 * variable.c (var_type_is_valid): New function.
1710 (measure_is_valid) Moved here, from format.c.
1711 (alignment_is_valid) Moved here, from format.c.
1712 (var_get_name) New function.
1713 (var_set_name) New function.
1714 (width_to_type) New function.
1715 (var_get_type) New function.
1716 (var_get_width) New function.
1717 (var_set_width) New function.
1718 (var_is_numeric) New function.
1719 (var_is_alpha) New function.
1720 (var_is_short_string) New function.
1721 (var_is_long_string) New function.
1722 (var_is_very_long_string) New function.
1723 (var_get_missing_values) New function.
1724 (var_set_missing_values) New function.
1725 (var_clear_missing_values) New function.
1726 (var_has_missing_values) New function.
1727 (var_is_value_missing) New function.
1728 (var_is_num_missing) New function.
1729 (var_is_str_missing) New function.
1730 (var_is_value_user_missing) New function.
1731 (var_is_num_user_missing) New function.
1732 (var_is_str_user_missing) New function.
1733 (var_is_value_system_missing) New function.
1734 (var_get_print_format) New function.
1735 (var_set_print_format) New function.
1736 (var_get_write_format) New function.
1737 (var_set_write_format) New function.
1738 (var_set_both_formats) New function.
1739 (var_get_label) New function.
1740 (var_set_label) New function.
1741 (var_clear_label) New function.
1742 (var_has_label) New function.
1743 (var_get_measure) New function.
1744 (var_set_measure) New function.
1745 (var_get_display_width) New function.
1746 (var_set_display_width) New function.
1747 (var_get_alignment) New function.
1748 (var_set_alignment) New function.
1749 (var_get_value_cnt) New function.
1750 (var_get_leave) New function.
1751 (var_get_short_name) New function.
1753 * variable.h: (struct variable) Removed "type" and "nv" members;
1754 they are now computed from "width" where needed.
1756 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1758 * missing-values.c (mv_resize): Don't write beyond end of the
1759 allocated buffer when resizing a long string.
1761 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1763 Clean up identifier code: don't require identifier enumerations to
1764 be in a particular order; make better use of string library;
1765 expose less of the internals.
1767 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1768 change interface. Updated all callers.
1769 (lex_id_match) Change interface to use struct substring, update
1771 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1772 (global array keywords[]) Make static, change form. Update all
1773 users to use lex_id_name instead.
1774 (lex_is_keyword) New function.
1775 (lex_id_to_token) Change interface to use struct substring, update
1777 (lex_id_name) New function.
1779 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1780 lex_is_keyword instead.
1781 (T_LAST_KEYWORD) Removed.
1782 (T_N_KEYWORDS) Removed.
1784 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1786 * format.c: (fmt_date_template) Distinguish characters for which a
1787 space is output and any date delimiter is allowed on input, from
1788 those for which a space is output and only a space is allowed on
1789 input. The former is represented by X, the latter by a space.
1790 Also, drop distinction between h and H, changing the former to the
1793 * data-in.c: Completely rewrite internals to conform to SPSS input
1794 formats as closely as possible.
1795 (data_in) Changed external interface by replacing the structure
1796 that was used as a single argument by a set of arguments. Updated
1798 (data_in_finite_line) Removed. Converted all callers to use plain
1800 (data_in_get_integer_format) New function.
1801 (data_in_set_integer_format) New function.
1802 (data_in_get_float_format) New function.
1803 (data_in_set_float_format) New function.
1805 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1806 (struct data_in) Removed.
1808 * data-out.c: (output_date) Drop each component from the input as
1809 it is output, to allow us to drop the distinction between h (a
1810 count of hours) and H (the hour of day) template characters.
1811 Also, handle new X template character.
1812 (output_scientific) Follow more rational rule on when to drop
1813 fraction introduced between SPSS 13 and 15. Updated test case to
1816 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1818 Fix buffer overflow reported by John Darrington.
1820 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1821 realize that DIGITS is a count of nibbles, not of bytes.
1823 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1825 * calendar.c (calendar_offset_to_gregorian) Also return the
1826 year-of-day. Change callers to new interface.
1828 * data-out.c: Completely rewrite internals to conform to SPSS
1829 output formats as completely as possible.
1830 (data_out) Change interface to put input parameters before output
1831 parameters, for consistency with the style I now prefer. Update
1833 (data_out_get_integer_format) New public function.
1834 (data_out_set_integer_format) New public function.
1835 (data_out_get_float_format) New public function.
1836 (data_out_set_float_format) New public function.
1838 * data-out.h: New file. Move prototype for data_out here, from
1841 * format.c: (fmt_step_width) Use equality comparison instead of
1842 bitwise and, for clarity.
1843 (fmt_is_string) Ditto.
1844 (fmt_input_to_output) Fix categories that are translated to F
1847 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1849 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1850 casefile-private.h: Added casefilter to assist commands with missing
1853 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1855 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1857 * settings.c: (route_errors_to_terminal) New variable.
1858 (route_errors_to_listing) New variable.
1859 (get_error_routing_to_terminal) New function.
1860 (set_error_routing_to_terminal) New function.
1861 (get_error_routing_to_listing) New function.
1862 (set_error_routing_to_listing) New function.
1864 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1866 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1868 * format.c: Completely rewrite, to achieve better abstraction.
1869 Rewrite all references to formats in other files.
1871 * format.def: Rewrite and reorganize.
1873 * settings.c: Move everything related to custom currency formats
1874 into format.[ch], changing them in form, so as to group related
1875 code and definitions better. Changed all references to use the
1877 (static var decimal) Removed.
1878 (static var grouping) Removed.
1879 (static var cc) Removed.
1880 (get_decimal) Removed.
1881 (set_decimal) Removed.
1882 (get_grouping) Removed.
1883 (set_grouping) Removed.
1887 * settings.h: (macro CC_CNT) Removed.
1888 (macro CC_WIDTH) Removed.
1889 (struct custom_currency) Removed.
1891 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1893 * data-in.c (data_in): Use switch statement instead of table, to
1894 avoid dependence on the order of the FMT_* enums.
1896 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1898 * data-out.c: (num_to_string) Removed, because it was dead code.
1900 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1902 * data-in.c (parse_trailer): Fix error message.
1904 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1906 * format.c (fmt_is_binary): New function.
1908 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1910 * procedure.c procedure.h: Encapsulated the static data into a single
1913 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1915 * casefile.c (casereader_read_xfer): Always initialize the case,
1916 even on an error condition.
1918 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1920 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1921 entire function useless.
1923 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1925 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1926 casereader method casereader_clone.
1928 * procedure.c pransformations.h: Introduced new type casenum_t
1930 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1932 * variable.c: (width_to_bytes) Rephrase code for clarify.
1934 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1936 * format.c: (fmt_type_from_string) New function.
1937 (fmt_to_string) Include decimals in output if the format has
1938 decimals, even if the format type does not. This way, we can
1939 accurately reproduce incorrect formats in user output.
1940 (check_common_specifier) Make the check for a bad format type an
1941 assertion, so we get bug reports if they show up. Fix message.
1942 Check for decimal places with a format type that doesn't allow
1944 (check_input_specifier) Remove check for FMT_X, which has been
1946 (check_output_specifier) Ditto.
1948 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1950 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1951 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1952 (enum fmt_parse_flags) Removed.
1954 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1956 * casefile.c casefile.h: Converted to an abstract base class.
1957 * casefile-private.h fastfile.c fastfile.h: New files.
1958 * automake.mk procedure.c scratch-writer.c storage-stream.c
1960 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1962 * procedure.c (internal_procedure): Create sink_case with only as
1963 many values as the compacted dictionary.
1965 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1967 Remove "debugging" code that caused plenty of false positives and
1970 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1972 * case.c: Remove all references to `this' member.
1974 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1976 Fix link error noted by Jason Stover.
1978 * storage-stream.c: Include <assert.h>.
1980 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1982 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1983 ALL) and additional underlying system file issues.
1985 Thanks to John Darrington for review.
1987 First problem: var_hash points to variables not owned by the
1988 sys-file-reader, which the caller may free or modify. Use an
1989 array of sfm_vars instead, as done earlier (e.g. CVS version
1992 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1993 members and remove all code that references it. Add vars, var_cnt
1994 members. Remove fix_specials member, which was unused.
1995 (struct sfm_var) Remove name member, which was unused.
1996 (sfm_close_reader) Free vars member instead of var_hash.
1997 (compare_var_shortnames) Removed.
1998 (hash_var_shortname) Removed.
1999 (sfm_open_reader) Fill out vars array.
2000 (compare_var_index) Removed.
2001 (sfm_read_case) Use vars instead of var_hash.
2003 Second problem: we're confused about when we actually have very
2004 long strings, causing us to choose incorrectly between slow path
2005 and fast path in sfm_read_case.
2007 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
2008 have very long strings, not when we have long variable names,
2009 which is an unrelated feature.
2011 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
2013 * variable.h: Move var_set and variable parsing declarations to
2014 new header, src/language/lexer/variable-parser.h. Modified lots
2015 of files to include the new header.
2017 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
2019 * value-labels.c (value_to_string): When there's no value label,
2020 format the variable according to its print format, instead of
2021 always effectively using A or F format.
2023 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
2025 * casefile.c (casefile_get_random_reader): Nasty hack to get around
2028 * format.c: Removed tortological assertion.
2030 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
2032 Reform string library.
2034 * file-name.c (fn_interp_vars): Change interface to take a
2035 substring as input. Updated all users.
2037 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
2039 * format.c (measure_is_valid): Really return false when m >=
2042 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
2044 Implement random access to casefiles, for use in GUI.
2046 * casefile.c: (struct casereader) Add `random', `file_ofs',
2047 `buffer_ofs' members.
2048 (casefile_get_random_reader) New function.
2049 (read_open_file) Break part into new function
2050 seek_and_fill_buffer().
2051 (fill_buffer) Update buffer_ofs, file_ofs.
2052 (casereader_seek) New function.
2054 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2056 * settings.c: Added call to i18n{done, init}.
2058 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
2060 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
2063 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
2065 * casefile.c: Convert many uses of `int' to `bool'.
2067 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
2069 * transformations.c (trns_chain_destroy): Destroy chain's trns
2070 member, to fix memory leak.
2072 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
2074 * storage-stream.c (storage_source_decapsulate): Destroy case
2075 source to fix memory leak.
2077 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
2079 * scratch-reader.c (scratch_reader_read_case): Copy into existing
2080 case passed as argument instead of initializing the argument as a
2081 case. Fixes memory leak that showed up in
2082 tests/command/aggregate.sh with scratch files.
2084 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
2086 * procedure.c (proc_done): Destroy default_dict, to fix memory
2089 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2091 Simplify procedure_with_splits().
2093 * procedure.c (struct split_aux_data): Removed case_count member.
2094 (procedure_with_splits) Don't initialize case_count.
2095 (split_procedure_case_func) Check whether prev_case is null
2096 instead of case_count.
2097 (split_procedure_end_func) Ditto.
2099 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2101 * case.c (case_move): Do nothing if dst and src are the same
2103 (case_try_create) Merge two similar cases.
2104 (case_copy) Unshare only if data must be actually copied.
2106 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2108 * data-in.c data-out.c dictionary.c sys-file-reader.c
2109 sys-file-writer.c variable.c variable.h: Reworked very long string
2110 support for better encapsulation.
2112 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2114 * value-labels.c (val_labs_can_set_width): New function.
2115 (val_labs_set_width) Clear labels if increasing width to long
2117 (val_labs_destroy) Remove unneeded test for null.
2119 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2121 * value-labels.h: Remove unneeded dependency on variable.h.
2123 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2125 Get rid of `char *c' member in union value, for cleanliness.
2127 * value.h: (union value) Remove `c' member.
2129 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2131 Make dictionary compacting functions a little more general.
2133 * sys-file-writer.c (sfm_open_writer): Use
2134 dict_compacting_would_change().
2135 (does_dict_need_translation) Removed.
2137 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2139 Make dictionary compacting functions a little more general.
2141 * dictionary.c (dict_needs_compaction): Rename
2142 dict_compacting_would_shrink(). Update all callers.
2143 (dict_compacting_would_change) New function.
2145 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2147 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2148 inverted return value (!).
2150 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2152 Continue reforming procedure execution.
2154 * procedure.c: Search and replace "vfm" by "proc". Notably:
2155 (static var vfm_source) Rename proc_source. Update all
2157 (static var vfm_sink) Rename proc_sink. Update all references.
2159 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2161 Continue reforming procedure execution. In this phase, remove
2162 PROCESS IF, which was deprecated anyway and can be easily
2163 simulated with TEMPORARY followed by SELECT IF.
2165 * procedure.c: (open_active_file) Don't call
2166 add_process_if_trns().
2167 (discard_variables) Get rid of redundant call to
2168 proc_cancel_all_transformations().
2169 (add_process_if_trns) Removed.
2170 (process_if_trns_proc) Removed.
2171 (process_if_trns_free) Removed.
2173 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2175 Continue reforming procedure execution. In this phase, add
2176 `const' to the case passed to procedure()'s callback.
2178 Updated all users of procedure() as well.
2180 * procedure.c: (struct write_case_data) Add "const" to ccase
2181 parameter for case_func member.
2182 (procedure) Add "const" to ccase parameter for proc_func
2184 (multipass_case_func) Make ccase parameter const.
2185 (internal_procedure) Add "const" to ccase parameter for case_func
2187 (split_procedure_case_func) Make ccase parameter const.
2188 (multipass_split_case_func) Make ccase parameter const.
2190 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2192 Continue reforming procedure execution. In this phase, get rid of
2193 the output code for SPLIT FILE groups in procedure.c, which really
2194 shouldn't be doing any output. Move it into the individual
2195 procedures instead. This also adds some flexibility.
2197 Updated many users of procedure_with_splits() and
2198 multipass_procedure_with_splits() to call
2199 output_split_file_values() and to deal with increased use of
2202 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2203 parameter to begin_func member.
2204 (procedure_with_splits) Add "const struct ccase *" parameter to
2205 begin_func parameter. Make ccase parameter const in proc_func
2207 (split_procedure_case_func) Don't dump split file group. Pass
2209 (dump_splits) Moved to language/dictionary/split-file.c as
2210 output_split_file_values().
2211 (struct multipass_split_aux_data) Add "const struct ccase *"
2212 parameter to split_func member.
2213 (multipass_procedure_with_splits) Add "const struct ccase *"
2214 parameter to split_func parameter.
2215 (multipass_split_case_func) Save new SPLIT FILE case before
2217 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2219 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2221 Continue reforming procedure execution. Change
2222 internal_procedure() so that it calls open_active_file() and
2223 close_active_file(), which isolates most of the actual procedure
2226 * procedure.c: (struct write_case_data) Rename `proc_func' member
2227 to `case_func' and update all references.
2228 (procedure) Rewrite as one-line wrapper around
2229 internal_procedure().
2230 (struct multipass_aux_data) New.
2231 (multipass_callback) Renamed multipass_case_func(). Use struct
2232 multipass_aux_data as auxiliary data.
2233 (multipass_end_func) New function.
2234 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2235 that uses multipass_case_func, multipass_end_func.
2236 (internal_procedure) Add `end_func' argument. Move optimization
2237 of trivial case in here. Move call to open_active_file() and
2238 close_active_file() in here. Now assert that vfm_source is
2240 (procedure_with_splits_callback) Rename
2241 split_procedure_case_func().
2242 (split_procedure_end_func) New function.
2243 (multipass_split_callback) Rename multipass_split_case_func.
2244 (multipass_split_end_func) New function.
2245 (discard_variables) No need to test for nonnull vfm_source.
2247 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2249 Continue reforming procedure execution. Get rid of unused member.
2251 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2253 (write_case) Don't increment cases_analyzed.
2255 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2257 Continue reforming procedure execution. In this phase, move
2258 procedure.c and procedure.h from src to src/data. Update
2259 makefiles and #includes accordingly.
2261 * procedure.c: Moved here from src/.
2263 * procedure.h: Moved here from src/.
2265 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2267 Continue reforming procedure execution. In this phase, get rid of
2268 many global variables, consolidating procedure execution in
2269 procedure.c. Encapsulate transformations in new "struct
2270 trns_chain". Also, change implementation of N OF CASES, FILTER,
2271 and PROCESS IF from special cases to transformations.
2273 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2276 * dictionary.c: (global variable default_dict) Move to
2279 * variable.h: (TRNS_*) Move to transformations.h.
2280 (struct transformation) Move to transformations.c.
2282 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2284 * sys-file-reader.c: Fixed invalid read problems.
2286 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2288 * storage-stream.c: Add missing function comments.
2290 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2292 Continue reforming procedure execution. In this phase, add some
2293 new, needed functionality to storage-stream.
2295 * storage-stream.c: (storage_source_decapsulate) New function.
2297 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2299 * variable.c (width_to_bytes): Declarations must precede
2300 statements for C90 compliance.
2302 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2304 * data-out.c, data-in.c, variable.c, variable.h: New functions
2305 copy_mangle and copy_demangle for reading/writing cases; emulates the
2306 way SPSS deals with strings > 255 bytes.
2308 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2309 subtype 14 needed for strings longer than 255 bytes.
2311 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2312 instead of literal values. Also fixed some constness issues.
2314 * format.h: Constness
2316 * sfm-private.h: Renamed the case_size identifier, since I discovered
2317 that SPSS's respect for this variable is very nominal.
2319 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2321 Change case limit type from int to size_t.
2323 * dictionary.c: (struct dictionary) Change type of case_limit
2325 (dict_get_case_limit) Change return type.
2326 (dict_set_case_limit) Change parameter type.
2328 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2330 * variable.h: (struct variable) Rename `reinit' member as `leave'
2331 and invert sense. Fix up all references.
2333 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2335 Continue reforming procedure execution. In this phase, break
2336 procedure.c into multiple files.
2338 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2340 * case-sink.c: New file.
2342 * case-sink.h: New file.
2344 * case-source.c: New file.
2346 * case-source.h: New file.
2348 * storage-stream.c: New file.
2350 * storage-stream.h: New file.
2352 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2354 * variable.h: (struct variable) Remove `init' member and all
2355 references to it from other files. It was initialized in several
2356 places, but nothing really ever used it for anything worthwhile.
2357 Thanks to Jason Stover for pointing out how confusing this
2360 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2362 Continue reforming error message support. In this phase, get rid
2363 of message "titles" and put the message text in `struct error'.
2364 Now `struct error' encapsulates a message more properly.
2366 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2367 Format message ourselves.
2369 * data-in.c: (vdls_error) Ditto.
2371 * por-file-reader.c: (error) Ditto.
2373 * sys-file-reader.c: (corrupt_msg) Ditto.
2375 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2377 GNU standards require "file name" instead of "filename" in
2378 documentation. It's nice for our code to follow the convention
2381 * casefile.c: (struct casefile) Rename `filename' member to
2382 `file_name'. Updated all references.
2384 * file-name.c: [!unix] (struct file_identity) Rename
2385 normalized_filename member to normalized_file_name. Updated all
2388 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2390 We don't really support anything but Unix-like environments well,
2391 so we might as well de-obfuscate by writing directory and path
2392 separators explicitly.
2394 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2396 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2397 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2399 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2402 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2404 GNU standards require "file name" instead of "filename" in
2405 documentation. It's nice for our code to follow the convention
2408 * filename.c: Rename to file-name.c.
2410 * filename.h: Rename to file-name.h. Update all inclusions.
2411 Update header guards.
2413 * automake.mk: Update file names.
2415 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2417 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2419 (fn_basename) Removed (dead code).
2420 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2421 (fn_special_p) Renamed fn_is_special(), all references updated.
2422 (fn_exists_p) Renamed fn_exists(), all references updated.
2424 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2426 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2427 Also, now it only tilde-expands file names, not paths.
2428 (fn_search_path) Tilde-expand one directory at a time.
2430 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2432 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2433 get rid of non-Unixlike version of the code, which has probably
2435 (fn_prepend_dir) Removed (dead code).
2437 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2438 (macro PATH_SEPARATOR_STRING) New.
2439 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2441 Continue reforming error message support. In this phase, we get
2442 rid of VM() and the other msg() support for "verbosity", replacing
2443 it by a new function verbose_msg().
2445 * filename.c: (fn_search_path) Use verbose_msg() instead of
2448 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2450 * sfm-private.h: Get rid of #defines after #error, which makes no
2453 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2455 Get rid of our own int32 type in favor of the standard int32_t
2458 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2461 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2463 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2465 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2467 Remove ill-considered file routines that are no longer used.
2469 * filename.c: (fn_open_ext) Removed.
2470 (fn_close_ext) Removed.
2472 * filename.h: (struct file_ext) Removed.
2474 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2476 * variable.c (var_is_valid_name): Move declarations before code
2479 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2481 * filename.ch (fn_interp_vars): Fixed small buglet.
2483 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2485 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2486 so as to modify the string inline. Thus makeing it easier to
2487 destroy the results when no longer needed.
2489 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2491 * category.c (cat_stored_values_destroy): Fixed memory leak.
2493 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2495 Add some missing frees. Thanks to John Darrington for reporting
2498 * any-writer.c (any_writer_close): Free writer.
2500 * any-reader.c (any_reader_close): Free reader.
2502 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2504 * por-file-reader.c: (error) Mark as NO_RETURN.
2506 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2508 * settings.c: Changed default value of scompress to true.
2510 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2512 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2514 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2517 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2518 here from pref.h.orig.
2520 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2522 * sys-file-reader.c: Fixed bug reading compressed files.
2524 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2526 * Numerous renames. See src/ChangeLog for details.
2528 * Moved files from src directory