1 2008-02-05 John Darrington <john@darrington.wattle.id.au>
3 psql-reader.c: So yesterday they release postgresql 8.3.0
4 which has money represented with 64 bits. They must get
7 2008-02-02 John Darrington <john@darrington.wattle.id.au>
9 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
10 for reviewing this code.
12 2008-02-02 Ben Pfaff <blp@gnu.org>
16 * sys-file-reader.c (read_variable_record): Allow missing values
17 to be specified on long string variables, but warn about them
18 because PSPP does not yet support them.
19 (read_extension_record): Ignore extension records 20 and 21, which
20 PSPP does not yet support.
21 (read_header): Fix error message when floating-point format cannot
24 2008-02-01 Ben Pfaff <blp@gnu.org>
26 Patch #6386. Thanks to John Darrington for review and for the
27 updates to gnumeric-reader.c.
29 * dictionary.c (make_hinted_name): New function.
30 (make_numeric_name): New function.
31 (dict_make_unique_var_name): New function.
33 * gnumeric-reader.c (devise_name): Removed.
34 (munge_name): Removed.
35 (gnumeric_open_reader): Use new function
36 dict_make_unique_var_name.
38 * short-names.c (set_var_short_name_suffix): Use new function
41 2008-01-19 John Darrington <john@darrington.wattle.id.au>
43 * settings.c settings.h: Moved static variables into a
44 single struct. Renamed functions to have a settings_ prefix.
46 2008-01-14 John Darrington <john@darrington.wattle.id.au>
48 * data-out.c (output_AHEX): Corrected number of bytes in
49 call to output_hex. Closes bug #22011
51 2008-01-02 John Darrington <john@darrington.wattle.id.au>
53 * variable.c variable.h: Replaced var_get_value_name with
54 var_append_value_name which doesn't use any static data.
55 Thanks to Ben for review.
57 2007-12-07 Ben Pfaff <blp@gnu.org>
61 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
62 made from an empty casereader produce a casegrouper with no
63 groups, instead of one with one group that has no cases.
65 * casereader.c (casereader_is_empty): New function.
67 2007-12-06 Ben Pfaff <blp@gnu.org>
71 * sys-file-reader.c (read_display_parameters): Handle variable
72 display parameters record with only 2 data items per variable.
73 Reported by Guido Gay <gay@irer.it>.
75 2007-12-04 Ben Pfaff <blp@gnu.org>
77 * identifier.c (lex_id_match_n): New function.
78 (lex_id_match): Reimplement in terms of lex_id_match_n.
80 2007-11-24 Ben Pfaff <blp@gnu.org>
82 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
85 2007-11-24 Ben Pfaff <blp@gnu.org>
87 Fix warning reported when reading back system files that include
88 very long string variables. Thanks to Guido Gay <gay@irer.it> for
91 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
92 was written where `j' was meant.
94 2007-11-18 Ben Pfaff <blp@gnu.org>
96 Properly write variables that include a range of missing values to
97 system files. Thanks to Guido Gay <gay@irer.it> for reporting
100 * sys-file-writer.c (write_variable): Correctly calculate
101 n_missing_values field when writing variables that include a range
104 2007-11-10 Ben Pfaff <blp@gnu.org>
106 Cleanups and bug fixes devised while writing up documentation.
109 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
111 * dict-class.c: New file.
112 (dict_class_from_id): Move here.
113 (dict_class_to_name): Move here.
115 * dict-class.h: New file.
116 (enum dict_class): Move here. Change from consecutive integers to
117 consecutive bits, to make testing for any of multiple values
118 easier. Add new DC_ALL constant.
120 * dictionary.c (struct dictionary): Change `case_limit' from
121 size_t to casenumber.
122 (dict_get_vars): Make final argument an enum dict_class.
123 (dict_get_vars_mutable): Ditto.
124 (dict_get_case_limit): Change return value to casenumber.
125 (dict_set_case_limit): Change final argument to a casenumber.
126 (dict_unset_split_var): Add assertion. Rephrase slightly.
127 (dict_set_label): Use xstrndup to simplify.
129 * format.c (fmt_step_width): AHEX format also needs 2-byte
131 (fmt_set_style): Simplify assertions.
133 * missing-values.c (mv_add_num_range): Rename mv_add_range.
134 Simplify implementation.
135 (mv_has_value): Simplify implementation.
136 (mv_pop_value): Remove the first value, not the last, to avoid
137 having GET followed by SAVE reverse the order of missing values.
138 (mv_peek_value): Rename mv_get_value. Simplify assertion.
139 (mv_has_range): Simplify implementation.
140 (mv_peek_range): Rename mv_get_range.
141 (can_resize_string): Removed.
142 (mv_is_resizable): Use value_is_resizable.
143 (mv_resize): Use value_resize.
145 * short-names.h (SHORT_NAME_LEN): Move here.
147 * val-type.h: New file, for definitions related to type and width
148 of abstract values. Before, these definitions were mixed among
149 those related to "union value" and those related to variables.
150 (macro SYSMIS): Move here.
151 (macro LOWEST): Move here.
152 (macro HIGHEST): Move here.
153 (macro MAX_STRING); Move here.
154 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
155 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
156 All references updated.
157 (val_type_is_valid): New function. Replaces var_type_is_valid.
158 All references updated.
159 (val_type_from_width): New function. Replaces
160 var_type_from_width. All references updated.
162 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
164 (val_labs_can_set_width): Use value_is_resizable.
165 (val_labs_add): Simply return false if the value labels set is too
166 wide, instead of having undefined behavior.
167 (val_labs_replace): Ditto.
168 (val_labs_replace): Ditto.
169 (val_labs_first): Set iterator to null if iteration is complete.
170 (val_labs_first_sorted): Ditto.
171 (val_labs_done): Become a no-op if the iterator is null.
173 * value.c (value_is_resizable): New function.
174 (value_resize): New function.
176 * variable.c (var_get_dict_class): New function.
178 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
181 2007-11-08 Ben Pfaff <blp@gnu.org>
183 * data-in.c: Make formatted data parsing locale-independent.
184 (parse_number): Use c_strtod instead of strtod, to avoid
185 locale-specific behavior.
188 2007-11-06 Ben Pfaff <blp@gnu.org>
190 Patch #6256: add support for binary, 360 file formats. Reviewed
193 * data-in.c (struct data_in): Add `encoding' member.
194 (data_in): Add `encoding' parameter, and re-encode the data passed
195 in where appropriate. Update all callers to pass it in.
196 (parse_A): Implement EBCDIC recoding wart described in manual.
197 (parse_AHEX): Implement EBCDIC recoding.
199 * data-out.c (data_out_legacy): New function.
200 (data_out): Make into a wrapper around data_out_legacy.
202 * file-handle-def.c (struct file_handle): New member `encoding'.
203 (fh_create_file): Set encoding.
204 (fh_default_properties): Set default encoding.
205 (fh_get_legacy_encoding): New function.
207 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
208 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
209 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
210 (struct fh_properties): New member `encoding'.
212 2007-11-05 Ben Pfaff <blp@gnu.org>
214 Patch #6258. Reviewed by John Darrington.
216 * file-handle-def.c (fh_lock): Add comment that TYPE should be
217 marked with N_() in the caller. Added these markings to each
218 caller too. Should make i18n easier.
219 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
221 2007-11-03 Ben Pfaff <blp@gnu.org>
223 Allow output files to overwrite input files (bug #21280).
225 * file-handle-def.c: Separate locking of files for input and for
226 output, to allow both to take place at once. Also, distinguish a
227 file handle from the identity of the underlying file, because the
228 identity of a file changes over time and the file handle can't
229 represent two different identities.
230 (struct file_handle): Remove `next', `open_cnt', `deleted',
231 `type', `open_mode', `aux', `identity' members. Change `id' from
232 char array to char *. Add `ref_cnt' member.
233 (file_handle_from_ll) New function.
234 (file_handles) Removed.
235 (named_handles) New variable.
236 (fh_init) Initialize named_handles.
237 (fh_done) Remove name from all named_handles.
238 (free_handle) Updated for modified struct file_handle.
239 (unname_handle) New function.
240 (fh_ref) New function.
241 (fh_from_file_name) Removed.
242 (fh_unref) New function.
243 (fh_unname) New function.
244 (fh_from_id) Rewritten.
245 (create_handle) Updated for modified struct file_handle.
246 (fh_inline_file) Increment the handle's ref count.
247 (fh_create_file) Updated for modified struct file_handle.
248 (fh_create_scratch) Ditto.
253 (fh_is_open) Removed.
254 (fh_get_id) Updated for modified struct file_handle.
255 (fh_get_default_handle) Increment the handle's ref count.
256 (fh_set_default_handle) Handle ref counts.
257 (struct fh_lock) New structure.
258 (locks) New static var.
259 (fh_lock) New function.
260 (fh_unlock) New function.
261 (fh_lock_get_aux) New function.
262 (fh_lock_set_aux) New function.
263 (fh_is_locked) New function.
264 (make_key) New function.
265 (free_key) New function.
266 (compare_fh_locks) New function.
267 (hash_fh_lock) New function.
269 * file-handle-def.h (enum fh_access) New enum.
271 * file-name.c: Made file_identity the same in all supported
273 (struct file_identity): New `name' member.
274 (fn_get_identity): For a file that doesn't exist, get the
275 dev/inode of its directory plus its name. If even the directory
276 doesn't exist, just use its name. Merge the Windows
277 implementation into the Unix one.
278 (fn_compare_file_identities): Rewritten. Merge the Windows
279 implementation into the Unix one.
280 (fn_hash_identity): New function.
282 * make-file.c (struct replace_file): New structure.
283 (all_files): New static var.
284 (replace_file_start): New function.
285 (replace_file_commit): New function.
286 (replace_file_abort): New function.
287 (free_replace_file): New function.
288 (unlink_replace_files): New function.
290 * por-file-reader.c (struct pfm_reader): Add `lock' member.
291 (close_reader): Unlock file.
292 (pfm_open_reader): Lock file.
294 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
296 (pfm_open_writer): Lock file and prepare for its replacement.
297 (close_writer): Unlock file.
299 * scratch-handle.h (struct scratch_handle): Add unique_id so that
300 different generations of a scratch file can be distinguished.
302 * scratch-reader.c (scratch_reader_open): Verify that the file is
305 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
306 remove scratch_handle member.
307 (scratch_writer_open): Lock handle. Prepare to replace handle
308 data, instead of doing it immediately.
309 (scratch_writer_casewriter_destroy): Replace handle data and
312 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
313 (sfm_open_reader): Lock file.
314 (close_reader): Unlock file.
316 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
318 (sfm_open_writer): Lock file and prepare for its replacement.
319 (close_writer): Unlock file.
321 2007-11-02 Ben Pfaff <blp@gnu.org>
323 * data-out.c (output_number): Use isfinite (from C99) instead of
324 gsl_isfinite, since we now have universal support for it in
326 (output_infinite): Ditto.
327 [!HAVE_ROUND] (round): Remove definition, since we now have a
328 replacement in gnulib.
330 2007-10-27 John Darrington <john@darrington.wattle.id.au>
332 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
333 wasn't initialised correctly. Closes bugs #21330 and 21397
336 2007-10-08 Ben Pfaff <blp@gnu.org>
338 Fix bug #21280. Thanks to John Darrington for review.
340 * file-name.c (create_stream): New function.
342 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
343 file handle before creating the file, to ensure that we don't
344 truncate a file that we're reading. Make code easier to read by
347 * sys-file-write.c (sfm_open_writer): Ditto.
349 2007-10-01 Ben Pfaff <blp@gnu.org>
351 Fix bug #21192. Thanks to John Darrington for review.
353 * casereader.c (casereader_read): Decrement case_cnt before
354 calling the casereader's "read" member function, so that we
355 interact properly with lazy_casereader.
357 * datasheet.c: Add regression test for above bug fix.
358 (clone_datasheet): New function.
359 (lazy_callback): New function.
360 (check_datasheet_casereader): New function.
361 (check_datasheet): Check datasheet contents are reported correctly
362 through an ordinary casereader and a lazy casereader.
363 (clone_model): Use clone_datasheet.
365 2007-09-24 Ben Pfaff <blp@gnu.org>
367 Patch #6210. Reviewed by John Darrington.
368 * settings.c: Drop "long view", which is not needed any longer.
369 (static var long_view): Removed.
370 (force_long_view): Removed.
371 (get_viewwidth): Removed.
372 (init_viewport): Removed long_view reference.
374 2007-09-19 John Darrington <john@darrington.wattle.id.au>
376 * settings.c settings.h: Changed viewport's length and width to be
377 owned by the user interface which uses the data library. This allows
378 better abstraction, and makes dynamically adjustable dimensions easier.
380 2007-09-18 Ben Pfaff <blp@gnu.org>
382 * procedure.c (proc_extract_active_file_data): New function.
384 * lazy-casereader.h: New file.
386 * lazy-casereader.c: New file.
388 * casereader.c (casereader_dynamic_cast): New function.
390 2007-09-14 Ben Pfaff <blp@gnu.org>
392 * dictionary.c (dict_clone): Copy case indexes from cloned
393 dictionary. Fixes bug #21061. Reviewed by John Darrington.
395 2007-09-13 John Darrington <john@darrington.wattle.id.au>
397 * value.c value.h (create_value): New function.
399 2007-09-12 Ben Pfaff <blp@gnu.org>
401 Make it clear that translator casereader and casewriter translate
402 functions are supposed to destroy their input case.
404 * casereader-translator.c (struct casereader_translator): Change
405 input case parameter of translate member function from const
406 struct ccase * to struct ccase *.
407 (casereader_create_translator): Ditto, for translate parameter.
409 * casewriter-translator.c (struct casewriter_translator): Ditto.
410 (casewriter_create_translator): Ditto.
412 2007-08-27 John Darrington <john@darrington.wattle.id.au>
414 * sys-file-reader.c (read_display_parameters): Force display width
415 to 8, if the sys file says 0 (like SPSS does).
417 2007-08-12 Ben Pfaff <blp@gnu.org>
419 * dictionary.c (dict_dump): New function.
421 2007-08-12 Ben Pfaff <blp@gnu.org>
423 Drop dict_compactor in favor of using the new struct case_map.
425 * dictionary.c (struct copy_map): Removed.
426 (struct dict_compactor): Removed.
427 (dict_make_compactor): Removed.
428 (dict_compactor_compact): Removed.
429 (dict_compactor_destroy): Removed.
431 * procedure.c (struct data_set): Change `compactor' member to be a
433 (proc_open): Use case_map_to_compact_dict instead of
435 (proc_casereader_read): Use case_map_execute instead of
436 dict_compactor_compact.
437 (proc_commit): Use case_map_destroy instead of
438 dict_compactor_destroy.
440 * scratch-writer.c (struct scratch_writer): Change `compactor'
441 member to be a struct case_map *.
442 (scratch_writer_open): Use case_map_to_compact_dict instead of
444 (scratch_writer_casewriter_write): Use case_map_execute instead of
445 dict_compactor_compact.
447 2007-08-12 Ben Pfaff <blp@gnu.org>
449 * automake.mk: Add case-map.c, case-map.h.
451 * case-map.c: New file.
453 * case-map.h: New file.
455 2007-08-12 Ben Pfaff <blp@gnu.org>
457 * dictionary.c (dict_compact_values): Don't delete scratch
458 variables as well as compacting case indexes. Update all callers.
459 (dict_get_compacted_value_cnt): Rename dict_count_values and
460 change interface. Update all callers.
461 (dict_get_compacted_value_cnt): Remove.
462 (dict_compacting_would_shrink): Remove.
463 (dict_compacting_would_change): Remove.
464 (dict_make_compactor): Add new parameter. Update all callers.
466 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
467 instead of dict_count_values, changing an O(N) operation into
470 2007-08-12 Ben Pfaff <blp@gnu.org>
472 * casereader.c (casereader_read): Don't require cases read by a
473 casereader to be exactly the expected size: as long as they're big
476 2007-08-12 Ben Pfaff <blp@gnu.org>
478 Make casewriters keep track of the number of `union value's in
479 each case. This is useful for two reasons: casewriter_write can
480 then check that the case being written is large enough, and later
481 recipients of the casewriter can determine the size of the case.
483 * casewriter-translator.c (casewriter_create_translator): Add
486 * casewriter.c (struct casewriter): Add value_cnt member.
487 (casewriter_write): Check that the case passed in is big enough.
488 (casewriter_get_value_cnt): New function.
489 (casewriter_create): Add value_cnt parameter.
491 2007-08-09 Ben Pfaff <blp@gnu.org>
493 Fix bug reported by Jason Stover.
494 * settings.c: Move get_termcap_viewport and in particular the
495 #include for <curses.h> to the end of the file. curses.h
496 redefines bool on some systems (e.g. OpenBSD), which causes
497 disagreement between uses of bool before its inclusion and
499 Tested by Jason Stover.
501 2007-07-29 Ben Pfaff <blp@gnu.org>
503 Provisional fix for bug #18692 and bug #20161. Reviewed by John
506 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
507 argument (never "rb" or "wb") because SUSv3 says that only those
508 modes are defined, and glibc in fact rejects other modes.
510 Open portable files with fn_open so that they can be read from
511 pipes. Fix missing fh_close call to go along with fh_open.
512 Report an error if the file close reports an error.
513 * por-file-reader.c (close_reader): New function.
514 (por_file_casereader_destroy): Use close_reader.
515 (pfm_open_reader): Open file with fn_open.
517 2007-07-28 Ben Pfaff <blp@gnu.org>
519 Make PSPP able to read all the portable files I could find on the
520 web. Thanks to John Darrington for review. Bug #17620.
521 * por-file-reader.c (struct pfm_reader): New member `line_length'.
522 (error): Print file offset in hexadecimal.
523 (warning): New function.
524 (advance): Treat lines less than 80 bytes long as padded to 80
526 (pfm_open_reader): Call read_documents if we find an "E" record.
527 (convert_format): Convert invalid formats to the default format
528 instead of aborting reading the file.
529 (read_variables): Rename duplicate variable names instead of
530 aborting reading the file.
531 (read_value_label): Allow string variables of different widths to
532 be assigned value labels in the same record. Replace duplicate
533 value labels instead of aborting.
534 (read_documents): New function.
536 * por-file-writer.c (pfm_open_writer): Call write_documents if the
537 dictionary has documents.
538 (write_documents): New function.
540 2007-07-25 Ben Pfaff <blp@gnu.org>
542 Fix bugs related to bug #17213.
544 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
545 since the former is what config.h has. Include the needed ncurses
547 (static var echo) Rename to `do_echo' because the original name is
548 the same as an ncurses identifier.
549 (get_termcap_viewport) Use error instead of msg.
551 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
552 (fn_close): Don't close stdin, stdout, stderr.
554 2007-07-26 John Darrington <john@darrington.wattle.id.au>
556 * procedure.c procedure.h: Added callbacks which get invoked whenever
557 a dataset's transformation chain changes.
559 2007-07-24 Ben Pfaff <blp@gnu.org>
562 * sys-file-writer.c (write_variable_display_parameters): Use new
563 var_default_display_width function to choose display width of
564 segments after the first one in a given variable.
565 * variable.c (var_create): Use var_default_display_width to pick
566 new variable's display width.
567 (var_default_display_width): New function.
568 Reviewed by John Darrington.
570 2007-07-24 Ben Pfaff <blp@gnu.org>
573 * por-file-writer.c (write_variables): Write weight variable.
574 Reviewed by John Darrington.
576 2007-07-23 Ben Pfaff <blp@gnu.org>
578 Improvements to system file reader and writer.
580 First, move all detailed knowledge of very long strings into
581 sys-file-private.[ch], so that this nasty stuff can be isolated.
583 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
584 (EFFECTIVE_VLS_CHUNK): New macro.
585 (min_int): New function.
586 (max_int): New function.
587 (sfm_width_to_bytes): Rewrite.
588 (sfm_width_to_octs): New function.
589 (sfm_segment_alloc_width): New function.
590 (sfm_segment_alloc_bytes): New function.
591 (sfm_segment_used_bytes): New function.
592 (sfm_segment_offset): New function.
593 (sfm_segment_effective_offset): New function.
594 (sfm_dictionary_to_sfm_vars): New function.
596 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
597 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
598 (struct sfm_var): New structure.
600 Next, improvements to the system file reader.
602 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
603 type casenumber. Added `version_major', `version_minor',
606 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
607 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
608 Change `case_cnt' to type casenumber. Removed `has_vls'.
609 (struct sfm_var): Removed.
610 (sfm_open_reader): Don't warn on wrong case size if the file was
611 written by SPSS 13, which tends to get it wrong. Use
612 sfm_dictionary_to_sfm_vars.
613 (read_header): Always output system file info.
614 (read_variable_record): Simplify code for reading missing values.
615 (read_machine_int32_info): Save version numbers from system file
616 into info struct passed as new argument.
617 (read_long_string_map): Restructured to use new sys-file-private
619 (read_value_labels): Use size_overflow_p.
620 (sys_file_casereader_read): Get rid of distinction between fast
621 and slow paths. Use information provided by sys-file-primate's
622 struct sfm_var to simplify code.
623 (skip_whole_strings): New function.
624 (read_int32): Renamed read_int. Changed return value to int.
626 (read_flt64): Renamed read_float. Changed return value to
627 double. Updated all callers.
628 (int32_to_native): Removed. Changed callers to use
630 (flt64_to_double): Removed. Changed callers to use float_convert.
632 Finally, get rid of int32, flt64 terminology and types in system
633 file writer. The former wasn't very useful since a POSIX "int"
634 can hold the whole range of int32 and we generally didn't have a
635 need for it to be exactly-32-bits, just at-least-32-bits. The
636 latter was inconvenient because we had to assume that it could be
637 different from double and thereby convert special values SYSMIS,
638 HIGHEST, LOWEST to and from it in multiple places. Instead, now
639 we just use "int" and "double" in most places, and do conversions,
640 if necessary, very close to where we do I/O. This change meant
641 that the writer code couldn't represent records in the file as C
642 structs any longer, but that's no great loss. The code actually
643 seems to be more readable without them.
645 Simplify the compression buffering code: only buffer as much as
646 necessary, which is no more than eight 8-byte units at any given
649 * sys-file-writer.c (typedef flt64): Removed.
650 (macro second_lowest_flt64): Removed.
651 (struct sysfile_header): Removed.
652 (struct sysfile_variable): Removed.
653 (struct sfm_writer): Removed `needs_translation', `has_vls',
654 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
655 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
656 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
657 compression buffering by `opcodes', `opcode_cnt', `data',
658 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
659 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
660 data writer functions instead of structures.
661 (calc_oct_idx): New function.
662 (write_header): Use simple data writer functions instead of
664 (write_format_spec): Renamed write_format. New argument.
665 (write_variable_continuation_records): New function.
666 (write_variable): Use simple data writer functions instead of
667 structures. Use write_variable_continuation_records. Write
668 entire very long string instead of requiring caller to understand
670 (write_value_labels): Use simple data writer functions instead of
672 (write_documents): Ditto.
673 (write_variable_display_parameters): Use sys-file-private
674 functions to simplify. Use simple data writer functions instead
676 (write_vls_length_table): Use simple data writer functions instead
678 (write_longvar_table): Ditto.
679 (write_rec_7_34): Break into new functions
680 write_integer_info_record, write_float_info_record. Use simple
681 data writer functions instead of structures.
682 (buf_write): Removed.
683 (append_string_max): Removed.
684 (ensure_buf_space): Removed.
685 (sys_file_casewriter_write): Get rid of the distinction between
686 fast and slow paths, which didn't seem to be too useful. Use new
687 functions write_case_uncompressed, write_case_compressed.
688 (put_instruction): Removed.
689 (put_element): Removed.
690 (write_compressed_data): Removed.
691 (close_writer): Use flush_compressed. Only write case count to
692 system file if it will fit in the field.
693 (write_case_compressed): New function.
694 (write_case_uncompressed): New function.
695 (flush_compressed): New function.
696 (put_cmp_opcode): New function.
697 (put_cmp_number): New function.
698 (write_int): New function.
699 (convert_double_to_output_format): New function.
700 (write_float): New function.
701 (write_value): New function.
702 (write_string): New function.
703 (write_bytes): New function.
704 (write_zeros): New function.
705 (write_spaces): New function.
707 Reviewed by John Darrington.
709 2007-07-22 Ben Pfaff <blp@gnu.org>
711 Don't try to write very long strings to portable files. The
712 format does not support it.
714 * por-file-writer.c (MAX_POR_WIDTH): New macro.
715 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
716 (write_format): Add arg to take width to resize format to.
717 (write_value): Limit width of value written to MAX_POR_WIDTH.
718 (write_variables): Limit width of variable and its output formats
720 Reviewed by John Darrington.
722 2007-07-22 Ben Pfaff <blp@gnu.org>
724 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
725 local variable instead of literal 5.
726 Reviewed by John Darrington.
728 2007-07-22 Ben Pfaff <blp@gnu.org>
730 Fix problems with uniqueness of short names in system files with
731 very long string variables. Now a variable may have multiple
734 * automake.mk (src_data_libdata_a_SOURCES): Add new files
735 short-names.c, short-names.h.
737 * dictionary.c (dict_clone): Clone all the short names.
738 (compare_strings): Move into short-names.c.
739 (hash_strings): Ditto.
740 (set_var_short_name_suffix): Ditto.
741 (dict_assign_short_names): Ditto, rename short_names_assign,
742 change to assign all short names.
744 * por-file-writer.c (write_variables): Use short_names_assign
745 instead of dict_assign_short_names.
747 * short-names.c: New file.
749 * short-names.h: New file.
751 * sys-file-private.c (sfm_width_to_segments): New function.
753 * sys-file-reader.c (read_long_var_name_map): Save and restore all
754 the short names, not just the first one.
756 * sys-file-writer.c (cont_var_name): Removed.
757 (sfm_open_writer): Use short_names_assign instead of
758 dict_assign_short_names. Use unique short names assigned by
759 short_names_assign instead of those generated by cont_var_name.
761 * variable.c (struct variable): Remove `short_name' member,
762 replace by `short_names' and `short_name_cnt'.
763 (var_create) Initialize new members.
764 (var_get_short_name_cnt): New function.
765 (var_get_short_name): Now takes an index argument. Changed most
767 (var_set_short_name): Ditto.
768 (var_clear_short_name): Renamed var_clear_short_names, changed to
769 clear all short names.
771 Reviewed by John Darrington.
773 2007-07-22 Ben Pfaff <blp@gnu.org>
775 * variable.c (var_set_width): Use new var_set_width function.
777 * missing-values.c (mv_n_values): Drop assertion, which was not
780 * format.c (fmt_default_for_width): New function.
781 (fmt_resize): New function.
783 Reviewed by John Darrington.
785 2007-07-18 John Darrington <john@darrington.wattle.id.au>
787 * datasheet.c (datasheet_delete_columns): Added assertion to check
788 we're not deleting outside the range of the sheet.
791 * dictionary.c dictionary.h variable.c: Added the ability for string
792 variables to be resized.
794 * vardict.h: Added some prototypes (moved from dictionary.h) as
795 these should only be called by variable.c
798 2007-07-14 John Darrington <john@darrington.wattle.id.au>
800 * sfm-reader.c: Respect case_cnt field in file header.
802 2007-07-01 John Darrington <john@darrington.wattle.id.au>
804 * transformation.c transformation.h (trns_chain_execute): Changed the
805 signature (Patch #6057)
807 2007-06-10 Ben Pfaff <blp@gnu.org>
809 * casereader-filter.c (casereader_filter_destroy): Make sure to
810 write all the remaining excluded cases to the casewriter, if any.
812 * caseinit.c (init_list_destroy): Rewrite.
813 (init_list_clear): Ditto.
815 * casegrouper.c (casegrouper_get_next_group): Always set *reader
816 to null when returning false.
818 2007-06-06 Ben Pfaff <blp@gnu.org>
820 Actually implement the new procedure code and adapt all of its
821 clients to match. Also adapt all of the other case sources and
822 sinks in the tree and their clients to use the
823 casereader/casewriter infrastructure.
825 * automake.mk: Add and remove files.
827 * any-reader.c: Change into a casereader.
828 * por-file-reader.c: Ditto.
829 * scratch-reader.c: Ditto.
830 * sys-file-reader.c: Ditto.
832 * any-writer.c: Change into a casewriter.
833 * por-file-writer.c: Ditto.
834 * scratch-writer.c: Ditto.
835 * sys-file-writer.c: Ditto.
837 * procedure.c: Change to use casereader, casewriter, caseinit, and
838 other new infrastructure.
840 * scratch-handle.c: Adapt to new infrastructure.
842 * case-sink.c: Removed, now dead code.
843 * case-sink.h: Ditto.
844 * case-source.c: Ditto.
845 * case-source.h: Ditto.
846 * casefile-factory.c: Ditto.
847 * casefile-private.h: Ditto.
850 * casefilter.c: Ditto.
851 * casefilter.h: Ditto.
854 * fastfile-factory.c: Ditto.
855 * fastfile-factory.h: Ditto.
856 * storage-stream.c: Ditto.
857 * storage-stream.h: Ditto.
859 2007-06-06 Ben Pfaff <blp@gnu.org>
863 * automake.mk: Add new files.
865 * datasheet.c: New file.
867 * datasheet.h: New file.
869 2007-06-06 Ben Pfaff <blp@gnu.org>
871 Until now, the procedure code has provided a case to the
872 case_source, which has filled in the data values that come from
873 the active file. "Left" data values that don't come from the
874 active file naturally stay the same from case to case, because the
875 procedure code keeps using that same case.
877 One of the compromises that comes with the new procedure code is
878 that the active file allocates and provides its own case, which
879 the procedure code then has to resize to provide room for any
880 other variables that should go in the case and then fill in the
881 values of "left" variables. Then, when we're done with that case,
882 we have to save the values of "left" variables to copy into the
883 next case read from the active file.
885 The caseinit code helps with this.
887 * automake.mk: Add new files.
889 * caseinit.c: New file.
891 * caseinit.h: New file.
893 2007-06-06 Ben Pfaff <blp@gnu.org>
895 * value.h (value_cnt_from_width): New function.
897 * variable.c (var_get_value_cnt): Use new function.
899 2007-06-06 Ben Pfaff <blp@gnu.org>
901 Add casegrouper, to allow cases read from a given casereader to be
902 broken into groups, each of which has its own casereader.
903 Generally cases are grouped based on having equal values for some
906 * automake.mk: Add new files.
908 * casegrouper.c: New file.
910 * casegrouper.h: New file.
912 2007-06-06 Ben Pfaff <blp@gnu.org>
914 Add interface to lexicographical ordering of cases.
916 * automake.mk: Add new files.
918 * case-ordering.c: New file.
920 * case-ordering.h: New file.
922 2007-06-06 Ben Pfaff <blp@gnu.org>
924 Add casereaders and casewriters, the basis of the new data processing
925 implementation. A casereader is a uniform interface to reading cases
926 from a data source; a casewriter is a uniform interface to writing
927 cases to a data sink.
929 * automake.mk: Add new files.
931 * casereader-filter.c: New file.
933 * casereader-provider.h: New file.
935 * casereader-translator.c: New file.
937 * casereader.c: New file.
939 * casereader.h: New file.
941 * casewriter-provider.h: New file.
943 * casewriter-translator.c: New file.
945 * casewriter.c: New file.
947 * casewriter.h: New file.
949 2007-06-06 Ben Pfaff <blp@gnu.org>
951 "casewindow" data structure that extends the deque (from libpspp)
952 of cases with the ability to dump cases to disk if we get too many
955 * automake.mk: Add new files.
957 * casewindow.c: New file.
959 * casewindow.h: New file.
961 2007-06-06 Ben Pfaff <blp@gnu.org>
963 sparse_cases data structure that augments a sparse_array of cases
964 with the ability to dump cases to disk if we get too many cases in
967 * automake.mk: Add new files.
969 * sparse-cases.c: New file.
971 * sparse-cases.h: New file.
973 2007-06-06 Ben Pfaff <blp@gnu.org>
975 Adds a low-level on-disk case array data structure.
977 * automake.mk: Add new files.
979 * case-tmpfile.c: New file.
981 * case-tmpfile.h: New file.
983 2007-06-06 Ben Pfaff <blp@gnu.org>
985 In a couple of places we calculate the maximum number of cases to
986 keep in memory based on the user-defined workspace. Enable
987 centralizing the calculation through a new function.
989 * settings.c (get_workspace_cases): New function.
991 2007-06-06 Ben Pfaff <blp@gnu.org>
993 The casenumber type is defined in transformations.h, but case.h is
994 a more sensible place. Move it.
996 * case.h (CASENUMBER_MAX): New macro.
997 (typedef casenumber): Move here, from transformations.h.
999 2007-06-03 Ben Pfaff <blp@gnu.org>
1001 Slightly generalize case_to_values and case_from_values functions.
1003 * case.c (case_to_values): Rename case_copy_out, change interface.
1004 (case_from_values): Rename case_copy_in, change interface.
1006 * fastfile.c (fastfilereader_get_next_case): Update caller.
1007 (write_case_to_disk): Ditto.
1009 2007-06-02 Ben Pfaff <blp@gnu.org>
1011 Clean up after a forgotten part of patch #5829.
1013 * casedeque.h: Remove unused file.
1015 * automake.mk: Remove casedeque.h from sources.
1017 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1019 * category.c: Removed redundant #include
1021 2007-05-06 Ben Pfaff <blp@gnu.org>
1023 Abstract the documents within a dictionary a little better.
1024 Thanks to John Darrington for suggestion, initial version, and
1025 review. Patch #5917.
1027 * dictionary.c (struct dictionary): Change `documents' member from
1028 char * to struct string.
1029 (dict_clear): Destroy struct string.
1030 (dict_get_documents): Convert struct string to char *.
1031 (dict_set_documents): Set struct string. Pad to 80-character
1033 (dict_clear_documents): New function.
1034 (dict_add_document_line): New function.
1035 (dict_get_document_line_cnt): New function.
1036 (dict_get_document_line): New function.
1038 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1040 * sys-file-reader.c (read_documents): Use new document functions.
1042 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1044 * sys-file-reader.c: When reading a system file which has no
1045 long name table, automatically create one where the long names
1046 are the lower case versions of the short names.
1048 2007-04-22 Ben Pfaff <blp@gnu.org>
1050 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1051 dict_clear will free most data related to the dictionary.
1052 dict_clear does a decent job, except that dict_set_split_vars on
1053 some systems won't actually free the dict's "split" member.
1054 Instead, it'll allocate a 1-byte region. Fix this.
1056 * value.c (value_copy): New function.
1057 (value_set_missing): Ditto.
1059 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1061 * Deleted existing category.h and moved cat-routines.h into
1062 category.h Encapsulated struct cat_vals better.
1064 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1066 * sys-file-reader.c: When reading a system file which has no
1067 long name table, automatically create one where the long names
1068 are the lower case versions of the short names.
1070 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1072 * sys-file-reader.c: Some versions of Other Software seem to
1073 produce system files with string variables' measure set to
1074 zero. We'll assume these are supposed to be nominal variables.
1076 2007-03-30 Ben Pfaff <blp@gnu.org>
1078 * procedure.c: Adapt to new deque data structure.
1080 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1081 Ben Pfaff <blp@gnu.org>
1083 * file-name.c: Mingw compatibility fixes.
1084 (fn_search_path): Use ISSLASH instead of comparing against '/'
1086 (fn_dir_name): Use dir_name from gnulib.
1087 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1088 (fn_get_identity): Use GetFullPathName instead of canonicalize
1089 from gnulib, because the latter does not fully support
1090 Windows-style path names. Use this implementation based on the
1091 detected presence of Windows instead of the absence of Unix, since
1092 the new implementation is Windows-specific.
1093 (fn_compare_file_identities): In Windows implementation, compare
1094 names case-insensitively.
1096 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1098 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1099 <stdlib.h> to have the same effect.
1101 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1103 * por-file-reader.c: Add missing _() around messages.
1105 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1107 * make-file.c: Include "mkstemp.h", without which linking on
1110 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1112 Reduce platform dependence.
1114 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1115 Everywhere we using this, we really should have just depended on
1116 the shell to expand tildes.
1117 (fn_search_path): Simplify, given that we don't do tilde expansion
1119 (fn_normalize): Removed. Caller changed to use the canonicalize
1121 (fn_get_cwd): Removed. Only user was fn_normalize.
1122 (fn_is_absolute): Really only test for absolute names.
1123 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1125 (fn_readlink): Removed, as it was only used fn_normalize.
1126 (fn_exists): Assume the stat function is available; gnulib does.
1127 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1129 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1131 * dictionary.c (dict_create_vector_assert): New function.
1133 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1135 * file-name.c (fn_normalize): Correct name of function
1136 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1137 for pointing this out.
1139 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1141 We are using a single member in struct file_handle, the "name"
1142 field, for more than one purpose. When it begins with '"', it's a
1143 file name; otherwise, it's a token that can be used to identify
1144 it. When that assertion fires, it's because we searched for the
1145 name case-sensitively as a file name (so that there was no match),
1146 and then we try to insert it case-insensitively as a token, which
1147 fails because duplicates aren't allowed.
1149 Solution: break the two purposes into two separate fields. This
1150 fixes the problem and likely makes the code easier to read too.
1152 Fixes bug #18922. Thanks to John Darrington for bug report and
1155 * file-handle-def.c (struct file_handle): New `id' member.
1156 (fh_from_name): Rename fh_from_id. Update all callers.
1157 (create_handle): New `id' parameter. Update all callers.
1158 (fh_create_file): Ditto.
1159 (fh_get_id): New function.
1161 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1163 * case.c (case_is_null): Change return type to bool.
1165 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1169 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1170 reference counting to share data. This makes it easy for
1171 valgrind, etc. to find accesses to cases that have been destroyed
1172 but have been kept around by another user's ref-count. This often
1173 happens when the data set is small enough to find in memory; if a
1174 bigger data set that would overflow to disk were used, then data
1175 corruption would occur.
1177 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1181 * case.c (case_unshare): Make it check internally whether the
1182 ref_cnt is greater than 1, so that the callers don't have to.
1183 Update callers not to check.
1185 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1187 Before, I was thinking that I might want to get rid of reference
1188 counting at some point. Now, I'm pretty sure that it's here to
1189 stay. Thus, because we have to store the value_cnt anyway for
1190 reference-counted cases, we might as well expose it to users.
1192 * case.c (case_get_value_cnt): New function.
1193 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1194 resize case if its size actually changed.
1196 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1197 instead of peeking inside struct case directly.
1198 (casefile_append): Ditto.
1200 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1202 Get rid of the inlines for the case functions, which made the
1203 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1204 the inlines didn't speed up "make check" at all, which is not a
1205 perfect benchmark but seems indicative.)
1207 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1208 definitions. Remove some assertions on nonnull pointers that were
1209 redundant with a pointer dereference soon after in the function.
1211 (struct case_data): Move definition here from case.h.
1215 (case_data_wr): Ditto.
1217 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1219 * automake.mk: Add casedeque.h to sources.
1221 * casedeque.h: New file.
1223 * procedure.c: (struct dataset) Change lag_count, lag_head,
1224 lag_queue member into single struct casedeque member. Update all
1225 users to use the casedeque instead.
1228 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1230 * procedure.c: Simplify lagged cases interface. Updated all
1231 clients--well, the only client--to use the simplified interface.
1232 (dataset_n_lag) Removed.
1233 (dataset_set_n_lag) Removed.
1234 (dataset_need_lag) New function.
1236 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1238 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1239 whenever appropriate, but only when the dataset/dictionary is in a
1242 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1244 * dictionary.c dictionary.h : Added callbacks for change of filter and
1245 split variables. Refactored some code to ensure that callbacks get
1246 invoked when appropriate.
1248 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1249 replace_dict callback occurs when permanent_dict replaces the current
1252 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1254 * dictionary.c dictionary.h : Added callback for when the weight
1255 variable of a dictionary changes.
1257 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1259 * dictionary.c dictionary.h : Added replace_source and replace_dict
1260 callbacks, and functions to deal with them.
1262 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1264 Simplify missing value handling.
1266 * missing-values.h (enum mv_class): New type.
1267 (enum mv_type): Moved definition into missing-values.c and renamed
1268 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1269 enums. Updated all uses.
1270 (struct missing_values): Changed type of `type' from `enum
1271 mv_type' to `int' because the definition is no longer exposed.
1273 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1274 parameter. Update all callers.
1275 (mv_is_num_missing): Ditto.
1276 (mv_is_str_missing): Ditto.
1277 (mv_is_value_user_missing): Removed. Changed callers to use
1278 mv_is_value_missing.
1279 (mv_is_num_user_missing): Removed. Changed callers to use
1281 (mv_is_str_user_missing): Removed. Changed callers to use
1283 (mv_is_value_system_missing): Removed. Changed callers to use
1284 mv_is_value_missing.
1285 (mv_set_type): Removed. Changed callers to use mv_clear.
1286 (mv_clear): New function.
1288 * variable.c (var_is_value_missing): Add new enum mv_class
1289 parameter. Update all callers.
1290 (var_is_num_missing): Ditto.
1291 (var_is_str_missing): Ditto.
1292 (var_is_value_user_missing): Removed. Changed callers to use
1293 var_is_value_missing.
1294 (var_is_num_user_missing): Removed. Changed callers to use
1296 (var_is_str_user_missing): Removed. Changed callers to use
1298 (var_is_value_system_missing): Removed. Changed callers to use
1299 var_is_value_missing.
1301 * casefilter.c (struct casefilter): Use enum mv_class in place of
1303 (casefilter_variable_missing): Adapt to new member.
1304 (casefilter_create): Change signature to take enum mv_class,
1307 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1309 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1311 * case-sink.c case-sink.h procedure.c procedure.h
1312 storage-stream.c: Now uses the factory.
1314 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1316 Make it possible to pull cases from the active file with a
1317 function call, instead of requiring indirection through a callback
1320 * case-source.h (struct case_source_class): Change ->read function
1321 to return a single case, instead of calling a callback function
1322 for each case. Change ->destroy function to return an error
1325 * case-source.c (free_case_source): Pass along the value returned
1326 by the case_source ->destroy function.
1328 * procedure.c (struct write_case_data): Removed.
1329 (struct dataset): Added some members to track procedure state.
1330 (procedure): Optimize the trivial case at this level.
1331 (internal_procedure): Re-implement in terms of proc_open,
1332 proc_read, proc_close.
1333 (proc_open) New function.
1334 (proc_read) New function.
1335 (proc_close) New function.
1336 (write_case) Moved into proc_read.
1337 (close_active_file) Moved closing of data source into proc_close.
1339 * storage-source.c: Rewrote to conform with modified
1340 case_source_class interface.
1342 * transformations.c (trns_chain_execute): Added argument to allow
1343 starting execution from an arbitrary transformation. Updated
1346 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1348 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1350 * sys-file-reader.c (read_display_parameters): Don't assume that
1351 MEASURE_* and ALIGN_* have the same values found in system files.
1353 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1355 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1356 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1357 the value of n_MEASURES, which was off by 1 (at least from my
1360 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1362 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1363 callbacks which are invoked when the dictionary or its
1364 variables are changed.
1366 * missing-values.c missing-values.h value-labels.c: Tidied up
1367 consistency checks, and made some of them return false
1368 instead of assert-failing.
1370 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1372 * calendar.c (calendar_days_in_month): New function.
1374 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1376 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1377 the value label's width.
1379 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1381 * sfm-private.h: Move contents into sys-file-writer.c, which is
1382 the only remaining user. Removed Borland C++-specific directives.
1384 * sys-file-reader.c: Clean up and rewrite entire file. The
1385 rewritten version is simpler and better abstracted, and should be
1386 easier to maintain and extend. It avoids using structures to read
1387 file data, which is prone to padding variations among compilers.
1388 It should also handle non-IEEE 754 system files, although I
1389 haven't been able to find any. It has been tested against many
1390 .sav files obtained from the Web and found to produce the same
1391 results as the earlier version of the code, or in some cases
1392 improved results. It is more tolerant of format variations found
1395 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1396 member, putting an enum integer_format in its place. New member
1397 `float_format'. Changed `compressed' member to type bool.
1399 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1401 * dictionary.c (dict_delete_consecutive_vars): New function.
1403 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1405 * file-name.c (fn_search_path): Remove prefix arg that was unused
1406 by any caller. Updated all callers.
1408 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1410 * format.c (fmt_dollar_template): Use user's decimal point
1411 character. Add assertion.
1413 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1415 * format.c (fmt_dollar_template): New function, based on
1416 dollar_format_template from var-type-dialog.c.
1418 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1420 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1423 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1425 Finish converting struct variable to an opaque type. In this
1426 phase, we add remaining setter and getter functions, convert the
1427 remaining PSPP code to use them, and do a bunch of cleanup. The
1428 resulting changes are pervasive but mostly trivial, and only the
1429 notable changes are logged.
1431 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1434 * case.c (case_data): Renamed case_data_idx.
1435 (case_num): Renamed case_num_idx.
1436 (case_str): Renamed case_str_idx.
1437 (case_data_rw): Renamed case_data_rw_idx.
1439 * case.h (case_data): New function with old name and an interface
1440 that takes a variable instead of an index, which is easier to
1441 use. Updated all callers to use the new interface, or to use the
1442 new *_idx function (see above).
1445 (case_data_rw): Ditto.
1447 * category.c (cat_stored_values_destroy): Changed interface to
1448 take a struct cat_vals * instead of a struct variable *.
1450 * dictionary.c (dict_clone): Use new vector_clone function.
1451 (dict_clear) Use new var_destroy function.
1452 (add_var) New function.
1453 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1454 (dict_create_var_assert) Rewrite in terms of add_var.
1455 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1456 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1457 (dict_lookup_var) Use new var_create, var_destroy functions.
1458 (dict_contains_var) Rewrite in terms of new vardict functionality.
1459 (set_var_dict_index) New function.
1460 (set_var_case_index) New function.
1461 (reindex_vars) New function.
1462 (dict_delete_var) Rewrite in terms of new vardict functionality.
1463 (dict_reorder_var) Ditto.
1464 (dict_reorder_vars) Ditto.
1465 (rename_var) New function.
1466 (dict_rename_var) Use rename_var.
1467 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1468 (dict_get_compacted_idx_to_fv) Rename
1469 dict_get_compacted_dict_index_to_case_index, update callers.
1470 (dict_create_vector) Use new vector_create function.
1471 (dict_clear_vectors) Use new vector_destroy function.
1472 (set_var_short_name_suffix) Move here from variable.c, renamed
1473 from var_set_short_name_suffix, make static, update caller.
1475 * sys-file-private.c: New file.
1476 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1477 width_to_bytes, update callers.
1479 * sys-file-private.h: New file. Later it will supplant
1480 sfm-private.h; for now it supplements it.
1481 (macro MIN_VERY_LONG_STRING) New macro.
1482 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1484 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1485 MAX_LONG_STRING was used before.
1487 * sys-file-writer.c: Ditto.
1489 * value-labels.c: Change the paradigm here to be that a null
1490 pointer is OK for a struct val_labs * in most cases; it just
1491 represents an empty set of value labels.
1492 (val_labs_copy) A copy of a null set is a null set.
1493 (val_labs_count) A null set has 0 labels.
1494 (val_labs_replace) Change return type to void. Rewrite for
1496 (val_labs_find) A null set does not contain the value.
1497 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1498 transposed argument order, updated all callers.
1500 * value.c: New file.
1501 (value_dup) Moved here from variable.c.
1502 (compare_values) Ditto.
1505 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1506 (macro MAX_LONG_STRING) Removed, because it was only interesting
1507 for system files, not for general code.
1508 (macro MAX_VERY_LONG_STRING) Ditto.
1509 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1510 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1512 * vardict.h: New file, for an interface between variables and
1515 * variable.c: A lot of functions were moved around, for better
1517 (struct variable) Move definition here, from variable.h.
1518 (var_type_adj) Removed--makes i18n hard.
1519 (var_type_noun) Ditto.
1520 (var_create) New function.
1521 (var_clone) New function.
1522 (var_destroy) New function.
1523 (var_set_name) Assert that variable is not in a dictionary.
1524 (compare_var_names) Rename compare_vars_by_name and fix a couple
1525 of callers who thought the args were strings.
1526 (hash_var_name) Rename hash_var_by_name.
1527 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1528 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1529 (var_is_very_long_string) Removed, because it was only interesting
1530 to system file code.
1531 (var_set_missing_values) Allow the argument to be the wrong width,
1532 as long as we can resize it. Simplify callers who were doing the
1533 resizing themselves.
1534 (var_get_value_labels) New function.
1535 (var_has_value_labels) New function.
1536 (var_set_value_labels) New function.
1537 (alloc_value_labels) New function.
1538 (var_add_value_label) New function.
1539 (var_replace_value_label) New function.
1540 (var_clear_value_labels) New function.
1541 (var_lookup_value_label) New function.
1542 (var_get_value_name) Moved here from variable.c, renamed from
1543 var_get_value_name, transposed argument order, updated all
1545 (var_to_string) Moved here, from variable-label.c.
1546 (var_set_leave) New function.
1547 (var_get_leave) New function.
1548 (var_must_leave) New function.
1549 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1550 set_var_short_name_suffix.
1551 (var_get_dict_index) New function.
1552 (var_get_case_index) New function.
1553 (var_get_obs_vals) New function.
1554 (var_set_obs_vals) New function.
1555 (var_has_obs_vals) New function.
1556 (var_get_vardict) New function.
1557 (var_set_vardict) New function.
1558 (var_has_vardict) New function.
1559 (var_clear_vardict) New function.
1560 (value_dup) Moved to value.c.
1561 (compare_values) Ditto.
1564 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1565 (enum ALPHA) Rename VAR_STRING, update all users.
1567 * vector.c: New file.
1568 (struct vector) Moved here, from variable.h.
1569 (check_widths) New function.
1570 (vector_create) New function.
1571 (vector_clone) New function.
1572 (vector_destroy) New function.
1573 (vector_get_name) New function.
1574 (vector_get_var) New function.
1575 (vector_get_var_cnt) New function.
1576 (compare_vector_ptrs_by_name) New function.
1578 * vector.h: New file.
1580 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1582 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1583 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1586 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1588 * value-labels.c (destroy_atoms): New function.
1589 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1591 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1593 Thanks to Jason Stover for pointing out this problem.
1595 * data-out.c (output_number): Use gsl_finite from GSL, which is
1596 portable, instead of isfinite, which is not.
1599 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1601 * variable.c variable.h (value_dup): New function.
1603 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1605 Start converting struct variable to an opaque type. In this
1606 phase, we add a bunch of setter and getter functions and convert
1607 most of the PSPP code to use them. The resulting changes are
1608 pervasive but mostly trivial, and only the notable changes are
1611 * format.c (fmt_equal): New function.
1613 * variable.c (var_type_is_valid): New function.
1614 (measure_is_valid) Moved here, from format.c.
1615 (alignment_is_valid) Moved here, from format.c.
1616 (var_get_name) New function.
1617 (var_set_name) New function.
1618 (width_to_type) New function.
1619 (var_get_type) New function.
1620 (var_get_width) New function.
1621 (var_set_width) New function.
1622 (var_is_numeric) New function.
1623 (var_is_alpha) New function.
1624 (var_is_short_string) New function.
1625 (var_is_long_string) New function.
1626 (var_is_very_long_string) New function.
1627 (var_get_missing_values) New function.
1628 (var_set_missing_values) New function.
1629 (var_clear_missing_values) New function.
1630 (var_has_missing_values) New function.
1631 (var_is_value_missing) New function.
1632 (var_is_num_missing) New function.
1633 (var_is_str_missing) New function.
1634 (var_is_value_user_missing) New function.
1635 (var_is_num_user_missing) New function.
1636 (var_is_str_user_missing) New function.
1637 (var_is_value_system_missing) New function.
1638 (var_get_print_format) New function.
1639 (var_set_print_format) New function.
1640 (var_get_write_format) New function.
1641 (var_set_write_format) New function.
1642 (var_set_both_formats) New function.
1643 (var_get_label) New function.
1644 (var_set_label) New function.
1645 (var_clear_label) New function.
1646 (var_has_label) New function.
1647 (var_get_measure) New function.
1648 (var_set_measure) New function.
1649 (var_get_display_width) New function.
1650 (var_set_display_width) New function.
1651 (var_get_alignment) New function.
1652 (var_set_alignment) New function.
1653 (var_get_value_cnt) New function.
1654 (var_get_leave) New function.
1655 (var_get_short_name) New function.
1657 * variable.h: (struct variable) Removed "type" and "nv" members;
1658 they are now computed from "width" where needed.
1660 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1662 * missing-values.c (mv_resize): Don't write beyond end of the
1663 allocated buffer when resizing a long string.
1665 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1667 Clean up identifier code: don't require identifier enumerations to
1668 be in a particular order; make better use of string library;
1669 expose less of the internals.
1671 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1672 change interface. Updated all callers.
1673 (lex_id_match) Change interface to use struct substring, update
1675 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1676 (global array keywords[]) Make static, change form. Update all
1677 users to use lex_id_name instead.
1678 (lex_is_keyword) New function.
1679 (lex_id_to_token) Change interface to use struct substring, update
1681 (lex_id_name) New function.
1683 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1684 lex_is_keyword instead.
1685 (T_LAST_KEYWORD) Removed.
1686 (T_N_KEYWORDS) Removed.
1688 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1690 * format.c: (fmt_date_template) Distinguish characters for which a
1691 space is output and any date delimiter is allowed on input, from
1692 those for which a space is output and only a space is allowed on
1693 input. The former is represented by X, the latter by a space.
1694 Also, drop distinction between h and H, changing the former to the
1697 * data-in.c: Completely rewrite internals to conform to SPSS input
1698 formats as closely as possible.
1699 (data_in) Changed external interface by replacing the structure
1700 that was used as a single argument by a set of arguments. Updated
1702 (data_in_finite_line) Removed. Converted all callers to use plain
1704 (data_in_get_integer_format) New function.
1705 (data_in_set_integer_format) New function.
1706 (data_in_get_float_format) New function.
1707 (data_in_set_float_format) New function.
1709 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1710 (struct data_in) Removed.
1712 * data-out.c: (output_date) Drop each component from the input as
1713 it is output, to allow us to drop the distinction between h (a
1714 count of hours) and H (the hour of day) template characters.
1715 Also, handle new X template character.
1716 (output_scientific) Follow more rational rule on when to drop
1717 fraction introduced between SPSS 13 and 15. Updated test case to
1720 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1722 Fix buffer overflow reported by John Darrington.
1724 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1725 realize that DIGITS is a count of nibbles, not of bytes.
1727 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1729 * calendar.c (calendar_offset_to_gregorian) Also return the
1730 year-of-day. Change callers to new interface.
1732 * data-out.c: Completely rewrite internals to conform to SPSS
1733 output formats as completely as possible.
1734 (data_out) Change interface to put input parameters before output
1735 parameters, for consistency with the style I now prefer. Update
1737 (data_out_get_integer_format) New public function.
1738 (data_out_set_integer_format) New public function.
1739 (data_out_get_float_format) New public function.
1740 (data_out_set_float_format) New public function.
1742 * data-out.h: New file. Move prototype for data_out here, from
1745 * format.c: (fmt_step_width) Use equality comparison instead of
1746 bitwise and, for clarity.
1747 (fmt_is_string) Ditto.
1748 (fmt_input_to_output) Fix categories that are translated to F
1751 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1753 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1754 casefile-private.h: Added casefilter to assist commands with missing
1757 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1759 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1761 * settings.c: (route_errors_to_terminal) New variable.
1762 (route_errors_to_listing) New variable.
1763 (get_error_routing_to_terminal) New function.
1764 (set_error_routing_to_terminal) New function.
1765 (get_error_routing_to_listing) New function.
1766 (set_error_routing_to_listing) New function.
1768 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1770 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1772 * format.c: Completely rewrite, to achieve better abstraction.
1773 Rewrite all references to formats in other files.
1775 * format.def: Rewrite and reorganize.
1777 * settings.c: Move everything related to custom currency formats
1778 into format.[ch], changing them in form, so as to group related
1779 code and definitions better. Changed all references to use the
1781 (static var decimal) Removed.
1782 (static var grouping) Removed.
1783 (static var cc) Removed.
1784 (get_decimal) Removed.
1785 (set_decimal) Removed.
1786 (get_grouping) Removed.
1787 (set_grouping) Removed.
1791 * settings.h: (macro CC_CNT) Removed.
1792 (macro CC_WIDTH) Removed.
1793 (struct custom_currency) Removed.
1795 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1797 * data-in.c (data_in): Use switch statement instead of table, to
1798 avoid dependence on the order of the FMT_* enums.
1800 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1802 * data-out.c: (num_to_string) Removed, because it was dead code.
1804 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1806 * data-in.c (parse_trailer): Fix error message.
1808 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1810 * format.c (fmt_is_binary): New function.
1812 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1814 * procedure.c procedure.h: Encapsulated the static data into a single
1817 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1819 * casefile.c (casereader_read_xfer): Always initialize the case,
1820 even on an error condition.
1822 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1824 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1825 entire function useless.
1827 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1829 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1830 casereader method casereader_clone.
1832 * procedure.c pransformations.h: Introduced new type casenum_t
1834 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1836 * variable.c: (width_to_bytes) Rephrase code for clarify.
1838 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1840 * format.c: (fmt_type_from_string) New function.
1841 (fmt_to_string) Include decimals in output if the format has
1842 decimals, even if the format type does not. This way, we can
1843 accurately reproduce incorrect formats in user output.
1844 (check_common_specifier) Make the check for a bad format type an
1845 assertion, so we get bug reports if they show up. Fix message.
1846 Check for decimal places with a format type that doesn't allow
1848 (check_input_specifier) Remove check for FMT_X, which has been
1850 (check_output_specifier) Ditto.
1852 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1854 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1855 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1856 (enum fmt_parse_flags) Removed.
1858 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1860 * casefile.c casefile.h: Converted to an abstract base class.
1861 * casefile-private.h fastfile.c fastfile.h: New files.
1862 * automake.mk procedure.c scratch-writer.c storage-stream.c
1864 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1866 * procedure.c (internal_procedure): Create sink_case with only as
1867 many values as the compacted dictionary.
1869 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1871 Remove "debugging" code that caused plenty of false positives and
1874 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1876 * case.c: Remove all references to `this' member.
1878 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1880 Fix link error noted by Jason Stover.
1882 * storage-stream.c: Include <assert.h>.
1884 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1886 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1887 ALL) and additional underlying system file issues.
1889 Thanks to John Darrington for review.
1891 First problem: var_hash points to variables not owned by the
1892 sys-file-reader, which the caller may free or modify. Use an
1893 array of sfm_vars instead, as done earlier (e.g. CVS version
1896 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1897 members and remove all code that references it. Add vars, var_cnt
1898 members. Remove fix_specials member, which was unused.
1899 (struct sfm_var) Remove name member, which was unused.
1900 (sfm_close_reader) Free vars member instead of var_hash.
1901 (compare_var_shortnames) Removed.
1902 (hash_var_shortname) Removed.
1903 (sfm_open_reader) Fill out vars array.
1904 (compare_var_index) Removed.
1905 (sfm_read_case) Use vars instead of var_hash.
1907 Second problem: we're confused about when we actually have very
1908 long strings, causing us to choose incorrectly between slow path
1909 and fast path in sfm_read_case.
1911 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1912 have very long strings, not when we have long variable names,
1913 which is an unrelated feature.
1915 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1917 * variable.h: Move var_set and variable parsing declarations to
1918 new header, src/language/lexer/variable-parser.h. Modified lots
1919 of files to include the new header.
1921 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1923 * value-labels.c (value_to_string): When there's no value label,
1924 format the variable according to its print format, instead of
1925 always effectively using A or F format.
1927 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1929 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1932 * format.c: Removed tortological assertion.
1934 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1936 Reform string library.
1938 * file-name.c (fn_interp_vars): Change interface to take a
1939 substring as input. Updated all users.
1941 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1943 * format.c (measure_is_valid): Really return false when m >=
1946 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1948 Implement random access to casefiles, for use in GUI.
1950 * casefile.c: (struct casereader) Add `random', `file_ofs',
1951 `buffer_ofs' members.
1952 (casefile_get_random_reader) New function.
1953 (read_open_file) Break part into new function
1954 seek_and_fill_buffer().
1955 (fill_buffer) Update buffer_ofs, file_ofs.
1956 (casereader_seek) New function.
1958 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1960 * settings.c: Added call to i18n{done, init}.
1962 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1964 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1967 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1969 * casefile.c: Convert many uses of `int' to `bool'.
1971 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1973 * transformations.c (trns_chain_destroy): Destroy chain's trns
1974 member, to fix memory leak.
1976 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1978 * storage-stream.c (storage_source_decapsulate): Destroy case
1979 source to fix memory leak.
1981 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1983 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1984 case passed as argument instead of initializing the argument as a
1985 case. Fixes memory leak that showed up in
1986 tests/command/aggregate.sh with scratch files.
1988 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1990 * procedure.c (proc_done): Destroy default_dict, to fix memory
1993 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1995 Simplify procedure_with_splits().
1997 * procedure.c (struct split_aux_data): Removed case_count member.
1998 (procedure_with_splits) Don't initialize case_count.
1999 (split_procedure_case_func) Check whether prev_case is null
2000 instead of case_count.
2001 (split_procedure_end_func) Ditto.
2003 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2005 * case.c (case_move): Do nothing if dst and src are the same
2007 (case_try_create) Merge two similar cases.
2008 (case_copy) Unshare only if data must be actually copied.
2010 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2012 * data-in.c data-out.c dictionary.c sys-file-reader.c
2013 sys-file-writer.c variable.c variable.h: Reworked very long string
2014 support for better encapsulation.
2016 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2018 * value-labels.c (val_labs_can_set_width): New function.
2019 (val_labs_set_width) Clear labels if increasing width to long
2021 (val_labs_destroy) Remove unneeded test for null.
2023 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2025 * value-labels.h: Remove unneeded dependency on variable.h.
2027 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2029 Get rid of `char *c' member in union value, for cleanliness.
2031 * value.h: (union value) Remove `c' member.
2033 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2035 Make dictionary compacting functions a little more general.
2037 * sys-file-writer.c (sfm_open_writer): Use
2038 dict_compacting_would_change().
2039 (does_dict_need_translation) Removed.
2041 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2043 Make dictionary compacting functions a little more general.
2045 * dictionary.c (dict_needs_compaction): Rename
2046 dict_compacting_would_shrink(). Update all callers.
2047 (dict_compacting_would_change) New function.
2049 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2051 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2052 inverted return value (!).
2054 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2056 Continue reforming procedure execution.
2058 * procedure.c: Search and replace "vfm" by "proc". Notably:
2059 (static var vfm_source) Rename proc_source. Update all
2061 (static var vfm_sink) Rename proc_sink. Update all references.
2063 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2065 Continue reforming procedure execution. In this phase, remove
2066 PROCESS IF, which was deprecated anyway and can be easily
2067 simulated with TEMPORARY followed by SELECT IF.
2069 * procedure.c: (open_active_file) Don't call
2070 add_process_if_trns().
2071 (discard_variables) Get rid of redundant call to
2072 proc_cancel_all_transformations().
2073 (add_process_if_trns) Removed.
2074 (process_if_trns_proc) Removed.
2075 (process_if_trns_free) Removed.
2077 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2079 Continue reforming procedure execution. In this phase, add
2080 `const' to the case passed to procedure()'s callback.
2082 Updated all users of procedure() as well.
2084 * procedure.c: (struct write_case_data) Add "const" to ccase
2085 parameter for case_func member.
2086 (procedure) Add "const" to ccase parameter for proc_func
2088 (multipass_case_func) Make ccase parameter const.
2089 (internal_procedure) Add "const" to ccase parameter for case_func
2091 (split_procedure_case_func) Make ccase parameter const.
2092 (multipass_split_case_func) Make ccase parameter const.
2094 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2096 Continue reforming procedure execution. In this phase, get rid of
2097 the output code for SPLIT FILE groups in procedure.c, which really
2098 shouldn't be doing any output. Move it into the individual
2099 procedures instead. This also adds some flexibility.
2101 Updated many users of procedure_with_splits() and
2102 multipass_procedure_with_splits() to call
2103 output_split_file_values() and to deal with increased use of
2106 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2107 parameter to begin_func member.
2108 (procedure_with_splits) Add "const struct ccase *" parameter to
2109 begin_func parameter. Make ccase parameter const in proc_func
2111 (split_procedure_case_func) Don't dump split file group. Pass
2113 (dump_splits) Moved to language/dictionary/split-file.c as
2114 output_split_file_values().
2115 (struct multipass_split_aux_data) Add "const struct ccase *"
2116 parameter to split_func member.
2117 (multipass_procedure_with_splits) Add "const struct ccase *"
2118 parameter to split_func parameter.
2119 (multipass_split_case_func) Save new SPLIT FILE case before
2121 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2123 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2125 Continue reforming procedure execution. Change
2126 internal_procedure() so that it calls open_active_file() and
2127 close_active_file(), which isolates most of the actual procedure
2130 * procedure.c: (struct write_case_data) Rename `proc_func' member
2131 to `case_func' and update all references.
2132 (procedure) Rewrite as one-line wrapper around
2133 internal_procedure().
2134 (struct multipass_aux_data) New.
2135 (multipass_callback) Renamed multipass_case_func(). Use struct
2136 multipass_aux_data as auxiliary data.
2137 (multipass_end_func) New function.
2138 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2139 that uses multipass_case_func, multipass_end_func.
2140 (internal_procedure) Add `end_func' argument. Move optimization
2141 of trivial case in here. Move call to open_active_file() and
2142 close_active_file() in here. Now assert that vfm_source is
2144 (procedure_with_splits_callback) Rename
2145 split_procedure_case_func().
2146 (split_procedure_end_func) New function.
2147 (multipass_split_callback) Rename multipass_split_case_func.
2148 (multipass_split_end_func) New function.
2149 (discard_variables) No need to test for nonnull vfm_source.
2151 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2153 Continue reforming procedure execution. Get rid of unused member.
2155 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2157 (write_case) Don't increment cases_analyzed.
2159 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2161 Continue reforming procedure execution. In this phase, move
2162 procedure.c and procedure.h from src to src/data. Update
2163 makefiles and #includes accordingly.
2165 * procedure.c: Moved here from src/.
2167 * procedure.h: Moved here from src/.
2169 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2171 Continue reforming procedure execution. In this phase, get rid of
2172 many global variables, consolidating procedure execution in
2173 procedure.c. Encapsulate transformations in new "struct
2174 trns_chain". Also, change implementation of N OF CASES, FILTER,
2175 and PROCESS IF from special cases to transformations.
2177 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2180 * dictionary.c: (global variable default_dict) Move to
2183 * variable.h: (TRNS_*) Move to transformations.h.
2184 (struct transformation) Move to transformations.c.
2186 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2188 * sys-file-reader.c: Fixed invalid read problems.
2190 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2192 * storage-stream.c: Add missing function comments.
2194 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2196 Continue reforming procedure execution. In this phase, add some
2197 new, needed functionality to storage-stream.
2199 * storage-stream.c: (storage_source_decapsulate) New function.
2201 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2203 * variable.c (width_to_bytes): Declarations must precede
2204 statements for C90 compliance.
2206 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2208 * data-out.c, data-in.c, variable.c, variable.h: New functions
2209 copy_mangle and copy_demangle for reading/writing cases; emulates the
2210 way SPSS deals with strings > 255 bytes.
2212 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2213 subtype 14 needed for strings longer than 255 bytes.
2215 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2216 instead of literal values. Also fixed some constness issues.
2218 * format.h: Constness
2220 * sfm-private.h: Renamed the case_size identifier, since I discovered
2221 that SPSS's respect for this variable is very nominal.
2223 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2225 Change case limit type from int to size_t.
2227 * dictionary.c: (struct dictionary) Change type of case_limit
2229 (dict_get_case_limit) Change return type.
2230 (dict_set_case_limit) Change parameter type.
2232 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2234 * variable.h: (struct variable) Rename `reinit' member as `leave'
2235 and invert sense. Fix up all references.
2237 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2239 Continue reforming procedure execution. In this phase, break
2240 procedure.c into multiple files.
2242 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2244 * case-sink.c: New file.
2246 * case-sink.h: New file.
2248 * case-source.c: New file.
2250 * case-source.h: New file.
2252 * storage-stream.c: New file.
2254 * storage-stream.h: New file.
2256 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2258 * variable.h: (struct variable) Remove `init' member and all
2259 references to it from other files. It was initialized in several
2260 places, but nothing really ever used it for anything worthwhile.
2261 Thanks to Jason Stover for pointing out how confusing this
2264 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2266 Continue reforming error message support. In this phase, get rid
2267 of message "titles" and put the message text in `struct error'.
2268 Now `struct error' encapsulates a message more properly.
2270 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2271 Format message ourselves.
2273 * data-in.c: (vdls_error) Ditto.
2275 * por-file-reader.c: (error) Ditto.
2277 * sys-file-reader.c: (corrupt_msg) Ditto.
2279 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2281 GNU standards require "file name" instead of "filename" in
2282 documentation. It's nice for our code to follow the convention
2285 * casefile.c: (struct casefile) Rename `filename' member to
2286 `file_name'. Updated all references.
2288 * file-name.c: [!unix] (struct file_identity) Rename
2289 normalized_filename member to normalized_file_name. Updated all
2292 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2294 We don't really support anything but Unix-like environments well,
2295 so we might as well de-obfuscate by writing directory and path
2296 separators explicitly.
2298 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2300 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2301 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2303 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2306 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2308 GNU standards require "file name" instead of "filename" in
2309 documentation. It's nice for our code to follow the convention
2312 * filename.c: Rename to file-name.c.
2314 * filename.h: Rename to file-name.h. Update all inclusions.
2315 Update header guards.
2317 * automake.mk: Update file names.
2319 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2321 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2323 (fn_basename) Removed (dead code).
2324 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2325 (fn_special_p) Renamed fn_is_special(), all references updated.
2326 (fn_exists_p) Renamed fn_exists(), all references updated.
2328 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2330 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2331 Also, now it only tilde-expands file names, not paths.
2332 (fn_search_path) Tilde-expand one directory at a time.
2334 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2336 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2337 get rid of non-Unixlike version of the code, which has probably
2339 (fn_prepend_dir) Removed (dead code).
2341 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2342 (macro PATH_SEPARATOR_STRING) New.
2343 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2345 Continue reforming error message support. In this phase, we get
2346 rid of VM() and the other msg() support for "verbosity", replacing
2347 it by a new function verbose_msg().
2349 * filename.c: (fn_search_path) Use verbose_msg() instead of
2352 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2354 * sfm-private.h: Get rid of #defines after #error, which makes no
2357 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2359 Get rid of our own int32 type in favor of the standard int32_t
2362 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2365 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2367 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2369 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2371 Remove ill-considered file routines that are no longer used.
2373 * filename.c: (fn_open_ext) Removed.
2374 (fn_close_ext) Removed.
2376 * filename.h: (struct file_ext) Removed.
2378 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2380 * variable.c (var_is_valid_name): Move declarations before code
2383 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2385 * filename.ch (fn_interp_vars): Fixed small buglet.
2387 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2389 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2390 so as to modify the string inline. Thus makeing it easier to
2391 destroy the results when no longer needed.
2393 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2395 * category.c (cat_stored_values_destroy): Fixed memory leak.
2397 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2399 Add some missing frees. Thanks to John Darrington for reporting
2402 * any-writer.c (any_writer_close): Free writer.
2404 * any-reader.c (any_reader_close): Free reader.
2406 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2408 * por-file-reader.c: (error) Mark as NO_RETURN.
2410 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2412 * settings.c: Changed default value of scompress to true.
2414 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2416 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2418 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2421 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2422 here from pref.h.orig.
2424 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2426 * sys-file-reader.c: Fixed bug reading compressed files.
2428 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2430 * Numerous renames. See src/ChangeLog for details.
2432 * Moved files from src directory