1 2008-01-14 John Darrington <john@darrington.wattle.id.au>
3 * data-out.c (output_AHEX): Corrected number of bytes in
4 call to output_hex. Closes bug #22011
6 2008-01-02 John Darrington <john@darrington.wattle.id.au>
8 * variable.c variable.h: Replaced var_get_value_name with
9 var_append_value_name which doesn't use any static data.
10 Thanks to Ben for review.
12 2007-12-07 Ben Pfaff <blp@gnu.org>
16 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
17 made from an empty casereader produce a casegrouper with no
18 groups, instead of one with one group that has no cases.
20 * casereader.c (casereader_is_empty): New function.
22 2007-12-06 Ben Pfaff <blp@gnu.org>
26 * sys-file-reader.c (read_display_parameters): Handle variable
27 display parameters record with only 2 data items per variable.
28 Reported by Guido Gay <gay@irer.it>.
30 2007-12-04 Ben Pfaff <blp@gnu.org>
32 * identifier.c (lex_id_match_n): New function.
33 (lex_id_match): Reimplement in terms of lex_id_match_n.
35 2007-11-24 Ben Pfaff <blp@gnu.org>
37 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
40 2007-11-24 Ben Pfaff <blp@gnu.org>
42 Fix warning reported when reading back system files that include
43 very long string variables. Thanks to Guido Gay <gay@irer.it> for
46 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
47 was written where `j' was meant.
49 2007-11-18 Ben Pfaff <blp@gnu.org>
51 Properly write variables that include a range of missing values to
52 system files. Thanks to Guido Gay <gay@irer.it> for reporting
55 * sys-file-writer.c (write_variable): Correctly calculate
56 n_missing_values field when writing variables that include a range
59 2007-11-10 Ben Pfaff <blp@gnu.org>
61 Cleanups and bug fixes devised while writing up documentation.
64 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
66 * dict-class.c: New file.
67 (dict_class_from_id): Move here.
68 (dict_class_to_name): Move here.
70 * dict-class.h: New file.
71 (enum dict_class): Move here. Change from consecutive integers to
72 consecutive bits, to make testing for any of multiple values
73 easier. Add new DC_ALL constant.
75 * dictionary.c (struct dictionary): Change `case_limit' from
77 (dict_get_vars): Make final argument an enum dict_class.
78 (dict_get_vars_mutable): Ditto.
79 (dict_get_case_limit): Change return value to casenumber.
80 (dict_set_case_limit): Change final argument to a casenumber.
81 (dict_unset_split_var): Add assertion. Rephrase slightly.
82 (dict_set_label): Use xstrndup to simplify.
84 * format.c (fmt_step_width): AHEX format also needs 2-byte
86 (fmt_set_style): Simplify assertions.
88 * missing-values.c (mv_add_num_range): Rename mv_add_range.
89 Simplify implementation.
90 (mv_has_value): Simplify implementation.
91 (mv_pop_value): Remove the first value, not the last, to avoid
92 having GET followed by SAVE reverse the order of missing values.
93 (mv_peek_value): Rename mv_get_value. Simplify assertion.
94 (mv_has_range): Simplify implementation.
95 (mv_peek_range): Rename mv_get_range.
96 (can_resize_string): Removed.
97 (mv_is_resizable): Use value_is_resizable.
98 (mv_resize): Use value_resize.
100 * short-names.h (SHORT_NAME_LEN): Move here.
102 * val-type.h: New file, for definitions related to type and width
103 of abstract values. Before, these definitions were mixed among
104 those related to "union value" and those related to variables.
105 (macro SYSMIS): Move here.
106 (macro LOWEST): Move here.
107 (macro HIGHEST): Move here.
108 (macro MAX_STRING); Move here.
109 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
110 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
111 All references updated.
112 (val_type_is_valid): New function. Replaces var_type_is_valid.
113 All references updated.
114 (val_type_from_width): New function. Replaces
115 var_type_from_width. All references updated.
117 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
119 (val_labs_can_set_width): Use value_is_resizable.
120 (val_labs_add): Simply return false if the value labels set is too
121 wide, instead of having undefined behavior.
122 (val_labs_replace): Ditto.
123 (val_labs_replace): Ditto.
124 (val_labs_first): Set iterator to null if iteration is complete.
125 (val_labs_first_sorted): Ditto.
126 (val_labs_done): Become a no-op if the iterator is null.
128 * value.c (value_is_resizable): New function.
129 (value_resize): New function.
131 * variable.c (var_get_dict_class): New function.
133 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
136 2007-11-08 Ben Pfaff <blp@gnu.org>
138 * data-in.c: Make formatted data parsing locale-independent.
139 (parse_number): Use c_strtod instead of strtod, to avoid
140 locale-specific behavior.
143 2007-11-06 Ben Pfaff <blp@gnu.org>
145 Patch #6256: add support for binary, 360 file formats. Reviewed
148 * data-in.c (struct data_in): Add `encoding' member.
149 (data_in): Add `encoding' parameter, and re-encode the data passed
150 in where appropriate. Update all callers to pass it in.
151 (parse_A): Implement EBCDIC recoding wart described in manual.
152 (parse_AHEX): Implement EBCDIC recoding.
154 * data-out.c (data_out_legacy): New function.
155 (data_out): Make into a wrapper around data_out_legacy.
157 * file-handle-def.c (struct file_handle): New member `encoding'.
158 (fh_create_file): Set encoding.
159 (fh_default_properties): Set default encoding.
160 (fh_get_legacy_encoding): New function.
162 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
163 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
164 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
165 (struct fh_properties): New member `encoding'.
167 2007-11-05 Ben Pfaff <blp@gnu.org>
169 Patch #6258. Reviewed by John Darrington.
171 * file-handle-def.c (fh_lock): Add comment that TYPE should be
172 marked with N_() in the caller. Added these markings to each
173 caller too. Should make i18n easier.
174 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
176 2007-11-03 Ben Pfaff <blp@gnu.org>
178 Allow output files to overwrite input files (bug #21280).
180 * file-handle-def.c: Separate locking of files for input and for
181 output, to allow both to take place at once. Also, distinguish a
182 file handle from the identity of the underlying file, because the
183 identity of a file changes over time and the file handle can't
184 represent two different identities.
185 (struct file_handle): Remove `next', `open_cnt', `deleted',
186 `type', `open_mode', `aux', `identity' members. Change `id' from
187 char array to char *. Add `ref_cnt' member.
188 (file_handle_from_ll) New function.
189 (file_handles) Removed.
190 (named_handles) New variable.
191 (fh_init) Initialize named_handles.
192 (fh_done) Remove name from all named_handles.
193 (free_handle) Updated for modified struct file_handle.
194 (unname_handle) New function.
195 (fh_ref) New function.
196 (fh_from_file_name) Removed.
197 (fh_unref) New function.
198 (fh_unname) New function.
199 (fh_from_id) Rewritten.
200 (create_handle) Updated for modified struct file_handle.
201 (fh_inline_file) Increment the handle's ref count.
202 (fh_create_file) Updated for modified struct file_handle.
203 (fh_create_scratch) Ditto.
208 (fh_is_open) Removed.
209 (fh_get_id) Updated for modified struct file_handle.
210 (fh_get_default_handle) Increment the handle's ref count.
211 (fh_set_default_handle) Handle ref counts.
212 (struct fh_lock) New structure.
213 (locks) New static var.
214 (fh_lock) New function.
215 (fh_unlock) New function.
216 (fh_lock_get_aux) New function.
217 (fh_lock_set_aux) New function.
218 (fh_is_locked) New function.
219 (make_key) New function.
220 (free_key) New function.
221 (compare_fh_locks) New function.
222 (hash_fh_lock) New function.
224 * file-handle-def.h (enum fh_access) New enum.
226 * file-name.c: Made file_identity the same in all supported
228 (struct file_identity): New `name' member.
229 (fn_get_identity): For a file that doesn't exist, get the
230 dev/inode of its directory plus its name. If even the directory
231 doesn't exist, just use its name. Merge the Windows
232 implementation into the Unix one.
233 (fn_compare_file_identities): Rewritten. Merge the Windows
234 implementation into the Unix one.
235 (fn_hash_identity): New function.
237 * make-file.c (struct replace_file): New structure.
238 (all_files): New static var.
239 (replace_file_start): New function.
240 (replace_file_commit): New function.
241 (replace_file_abort): New function.
242 (free_replace_file): New function.
243 (unlink_replace_files): New function.
245 * por-file-reader.c (struct pfm_reader): Add `lock' member.
246 (close_reader): Unlock file.
247 (pfm_open_reader): Lock file.
249 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
251 (pfm_open_writer): Lock file and prepare for its replacement.
252 (close_writer): Unlock file.
254 * scratch-handle.h (struct scratch_handle): Add unique_id so that
255 different generations of a scratch file can be distinguished.
257 * scratch-reader.c (scratch_reader_open): Verify that the file is
260 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
261 remove scratch_handle member.
262 (scratch_writer_open): Lock handle. Prepare to replace handle
263 data, instead of doing it immediately.
264 (scratch_writer_casewriter_destroy): Replace handle data and
267 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
268 (sfm_open_reader): Lock file.
269 (close_reader): Unlock file.
271 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
273 (sfm_open_writer): Lock file and prepare for its replacement.
274 (close_writer): Unlock file.
276 2007-11-02 Ben Pfaff <blp@gnu.org>
278 * data-out.c (output_number): Use isfinite (from C99) instead of
279 gsl_isfinite, since we now have universal support for it in
281 (output_infinite): Ditto.
282 [!HAVE_ROUND] (round): Remove definition, since we now have a
283 replacement in gnulib.
285 2007-10-27 John Darrington <john@darrington.wattle.id.au>
287 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
288 wasn't initialised correctly. Closes bugs #21330 and 21397
291 2007-10-08 Ben Pfaff <blp@gnu.org>
293 Fix bug #21280. Thanks to John Darrington for review.
295 * file-name.c (create_stream): New function.
297 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
298 file handle before creating the file, to ensure that we don't
299 truncate a file that we're reading. Make code easier to read by
302 * sys-file-write.c (sfm_open_writer): Ditto.
304 2007-10-01 Ben Pfaff <blp@gnu.org>
306 Fix bug #21192. Thanks to John Darrington for review.
308 * casereader.c (casereader_read): Decrement case_cnt before
309 calling the casereader's "read" member function, so that we
310 interact properly with lazy_casereader.
312 * datasheet.c: Add regression test for above bug fix.
313 (clone_datasheet): New function.
314 (lazy_callback): New function.
315 (check_datasheet_casereader): New function.
316 (check_datasheet): Check datasheet contents are reported correctly
317 through an ordinary casereader and a lazy casereader.
318 (clone_model): Use clone_datasheet.
320 2007-09-24 Ben Pfaff <blp@gnu.org>
322 Patch #6210. Reviewed by John Darrington.
323 * settings.c: Drop "long view", which is not needed any longer.
324 (static var long_view): Removed.
325 (force_long_view): Removed.
326 (get_viewwidth): Removed.
327 (init_viewport): Removed long_view reference.
329 2007-09-19 John Darrington <john@darrington.wattle.id.au>
331 * settings.c settings.h: Changed viewport's length and width to be
332 owned by the user interface which uses the data library. This allows
333 better abstraction, and makes dynamically adjustable dimensions easier.
335 2007-09-18 Ben Pfaff <blp@gnu.org>
337 * procedure.c (proc_extract_active_file_data): New function.
339 * lazy-casereader.h: New file.
341 * lazy-casereader.c: New file.
343 * casereader.c (casereader_dynamic_cast): New function.
345 2007-09-14 Ben Pfaff <blp@gnu.org>
347 * dictionary.c (dict_clone): Copy case indexes from cloned
348 dictionary. Fixes bug #21061. Reviewed by John Darrington.
350 2007-09-13 John Darrington <john@darrington.wattle.id.au>
352 * value.c value.h (create_value): New function.
354 2007-09-12 Ben Pfaff <blp@gnu.org>
356 Make it clear that translator casereader and casewriter translate
357 functions are supposed to destroy their input case.
359 * casereader-translator.c (struct casereader_translator): Change
360 input case parameter of translate member function from const
361 struct ccase * to struct ccase *.
362 (casereader_create_translator): Ditto, for translate parameter.
364 * casewriter-translator.c (struct casewriter_translator): Ditto.
365 (casewriter_create_translator): Ditto.
367 2007-08-27 John Darrington <john@darrington.wattle.id.au>
369 * sys-file-reader.c (read_display_parameters): Force display width
370 to 8, if the sys file says 0 (like SPSS does).
372 2007-08-12 Ben Pfaff <blp@gnu.org>
374 * dictionary.c (dict_dump): New function.
376 2007-08-12 Ben Pfaff <blp@gnu.org>
378 Drop dict_compactor in favor of using the new struct case_map.
380 * dictionary.c (struct copy_map): Removed.
381 (struct dict_compactor): Removed.
382 (dict_make_compactor): Removed.
383 (dict_compactor_compact): Removed.
384 (dict_compactor_destroy): Removed.
386 * procedure.c (struct data_set): Change `compactor' member to be a
388 (proc_open): Use case_map_to_compact_dict instead of
390 (proc_casereader_read): Use case_map_execute instead of
391 dict_compactor_compact.
392 (proc_commit): Use case_map_destroy instead of
393 dict_compactor_destroy.
395 * scratch-writer.c (struct scratch_writer): Change `compactor'
396 member to be a struct case_map *.
397 (scratch_writer_open): Use case_map_to_compact_dict instead of
399 (scratch_writer_casewriter_write): Use case_map_execute instead of
400 dict_compactor_compact.
402 2007-08-12 Ben Pfaff <blp@gnu.org>
404 * automake.mk: Add case-map.c, case-map.h.
406 * case-map.c: New file.
408 * case-map.h: New file.
410 2007-08-12 Ben Pfaff <blp@gnu.org>
412 * dictionary.c (dict_compact_values): Don't delete scratch
413 variables as well as compacting case indexes. Update all callers.
414 (dict_get_compacted_value_cnt): Rename dict_count_values and
415 change interface. Update all callers.
416 (dict_get_compacted_value_cnt): Remove.
417 (dict_compacting_would_shrink): Remove.
418 (dict_compacting_would_change): Remove.
419 (dict_make_compactor): Add new parameter. Update all callers.
421 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
422 instead of dict_count_values, changing an O(N) operation into
425 2007-08-12 Ben Pfaff <blp@gnu.org>
427 * casereader.c (casereader_read): Don't require cases read by a
428 casereader to be exactly the expected size: as long as they're big
431 2007-08-12 Ben Pfaff <blp@gnu.org>
433 Make casewriters keep track of the number of `union value's in
434 each case. This is useful for two reasons: casewriter_write can
435 then check that the case being written is large enough, and later
436 recipients of the casewriter can determine the size of the case.
438 * casewriter-translator.c (casewriter_create_translator): Add
441 * casewriter.c (struct casewriter): Add value_cnt member.
442 (casewriter_write): Check that the case passed in is big enough.
443 (casewriter_get_value_cnt): New function.
444 (casewriter_create): Add value_cnt parameter.
446 2007-08-09 Ben Pfaff <blp@gnu.org>
448 Fix bug reported by Jason Stover.
449 * settings.c: Move get_termcap_viewport and in particular the
450 #include for <curses.h> to the end of the file. curses.h
451 redefines bool on some systems (e.g. OpenBSD), which causes
452 disagreement between uses of bool before its inclusion and
454 Tested by Jason Stover.
456 2007-07-29 Ben Pfaff <blp@gnu.org>
458 Provisional fix for bug #18692 and bug #20161. Reviewed by John
461 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
462 argument (never "rb" or "wb") because SUSv3 says that only those
463 modes are defined, and glibc in fact rejects other modes.
465 Open portable files with fn_open so that they can be read from
466 pipes. Fix missing fh_close call to go along with fh_open.
467 Report an error if the file close reports an error.
468 * por-file-reader.c (close_reader): New function.
469 (por_file_casereader_destroy): Use close_reader.
470 (pfm_open_reader): Open file with fn_open.
472 2007-07-28 Ben Pfaff <blp@gnu.org>
474 Make PSPP able to read all the portable files I could find on the
475 web. Thanks to John Darrington for review. Bug #17620.
476 * por-file-reader.c (struct pfm_reader): New member `line_length'.
477 (error): Print file offset in hexadecimal.
478 (warning): New function.
479 (advance): Treat lines less than 80 bytes long as padded to 80
481 (pfm_open_reader): Call read_documents if we find an "E" record.
482 (convert_format): Convert invalid formats to the default format
483 instead of aborting reading the file.
484 (read_variables): Rename duplicate variable names instead of
485 aborting reading the file.
486 (read_value_label): Allow string variables of different widths to
487 be assigned value labels in the same record. Replace duplicate
488 value labels instead of aborting.
489 (read_documents): New function.
491 * por-file-writer.c (pfm_open_writer): Call write_documents if the
492 dictionary has documents.
493 (write_documents): New function.
495 2007-07-25 Ben Pfaff <blp@gnu.org>
497 Fix bugs related to bug #17213.
499 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
500 since the former is what config.h has. Include the needed ncurses
502 (static var echo) Rename to `do_echo' because the original name is
503 the same as an ncurses identifier.
504 (get_termcap_viewport) Use error instead of msg.
506 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
507 (fn_close): Don't close stdin, stdout, stderr.
509 2007-07-26 John Darrington <john@darrington.wattle.id.au>
511 * procedure.c procedure.h: Added callbacks which get invoked whenever
512 a dataset's transformation chain changes.
514 2007-07-24 Ben Pfaff <blp@gnu.org>
517 * sys-file-writer.c (write_variable_display_parameters): Use new
518 var_default_display_width function to choose display width of
519 segments after the first one in a given variable.
520 * variable.c (var_create): Use var_default_display_width to pick
521 new variable's display width.
522 (var_default_display_width): New function.
523 Reviewed by John Darrington.
525 2007-07-24 Ben Pfaff <blp@gnu.org>
528 * por-file-writer.c (write_variables): Write weight variable.
529 Reviewed by John Darrington.
531 2007-07-23 Ben Pfaff <blp@gnu.org>
533 Improvements to system file reader and writer.
535 First, move all detailed knowledge of very long strings into
536 sys-file-private.[ch], so that this nasty stuff can be isolated.
538 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
539 (EFFECTIVE_VLS_CHUNK): New macro.
540 (min_int): New function.
541 (max_int): New function.
542 (sfm_width_to_bytes): Rewrite.
543 (sfm_width_to_octs): New function.
544 (sfm_segment_alloc_width): New function.
545 (sfm_segment_alloc_bytes): New function.
546 (sfm_segment_used_bytes): New function.
547 (sfm_segment_offset): New function.
548 (sfm_segment_effective_offset): New function.
549 (sfm_dictionary_to_sfm_vars): New function.
551 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
552 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
553 (struct sfm_var): New structure.
555 Next, improvements to the system file reader.
557 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
558 type casenumber. Added `version_major', `version_minor',
561 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
562 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
563 Change `case_cnt' to type casenumber. Removed `has_vls'.
564 (struct sfm_var): Removed.
565 (sfm_open_reader): Don't warn on wrong case size if the file was
566 written by SPSS 13, which tends to get it wrong. Use
567 sfm_dictionary_to_sfm_vars.
568 (read_header): Always output system file info.
569 (read_variable_record): Simplify code for reading missing values.
570 (read_machine_int32_info): Save version numbers from system file
571 into info struct passed as new argument.
572 (read_long_string_map): Restructured to use new sys-file-private
574 (read_value_labels): Use size_overflow_p.
575 (sys_file_casereader_read): Get rid of distinction between fast
576 and slow paths. Use information provided by sys-file-primate's
577 struct sfm_var to simplify code.
578 (skip_whole_strings): New function.
579 (read_int32): Renamed read_int. Changed return value to int.
581 (read_flt64): Renamed read_float. Changed return value to
582 double. Updated all callers.
583 (int32_to_native): Removed. Changed callers to use
585 (flt64_to_double): Removed. Changed callers to use float_convert.
587 Finally, get rid of int32, flt64 terminology and types in system
588 file writer. The former wasn't very useful since a POSIX "int"
589 can hold the whole range of int32 and we generally didn't have a
590 need for it to be exactly-32-bits, just at-least-32-bits. The
591 latter was inconvenient because we had to assume that it could be
592 different from double and thereby convert special values SYSMIS,
593 HIGHEST, LOWEST to and from it in multiple places. Instead, now
594 we just use "int" and "double" in most places, and do conversions,
595 if necessary, very close to where we do I/O. This change meant
596 that the writer code couldn't represent records in the file as C
597 structs any longer, but that's no great loss. The code actually
598 seems to be more readable without them.
600 Simplify the compression buffering code: only buffer as much as
601 necessary, which is no more than eight 8-byte units at any given
604 * sys-file-writer.c (typedef flt64): Removed.
605 (macro second_lowest_flt64): Removed.
606 (struct sysfile_header): Removed.
607 (struct sysfile_variable): Removed.
608 (struct sfm_writer): Removed `needs_translation', `has_vls',
609 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
610 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
611 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
612 compression buffering by `opcodes', `opcode_cnt', `data',
613 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
614 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
615 data writer functions instead of structures.
616 (calc_oct_idx): New function.
617 (write_header): Use simple data writer functions instead of
619 (write_format_spec): Renamed write_format. New argument.
620 (write_variable_continuation_records): New function.
621 (write_variable): Use simple data writer functions instead of
622 structures. Use write_variable_continuation_records. Write
623 entire very long string instead of requiring caller to understand
625 (write_value_labels): Use simple data writer functions instead of
627 (write_documents): Ditto.
628 (write_variable_display_parameters): Use sys-file-private
629 functions to simplify. Use simple data writer functions instead
631 (write_vls_length_table): Use simple data writer functions instead
633 (write_longvar_table): Ditto.
634 (write_rec_7_34): Break into new functions
635 write_integer_info_record, write_float_info_record. Use simple
636 data writer functions instead of structures.
637 (buf_write): Removed.
638 (append_string_max): Removed.
639 (ensure_buf_space): Removed.
640 (sys_file_casewriter_write): Get rid of the distinction between
641 fast and slow paths, which didn't seem to be too useful. Use new
642 functions write_case_uncompressed, write_case_compressed.
643 (put_instruction): Removed.
644 (put_element): Removed.
645 (write_compressed_data): Removed.
646 (close_writer): Use flush_compressed. Only write case count to
647 system file if it will fit in the field.
648 (write_case_compressed): New function.
649 (write_case_uncompressed): New function.
650 (flush_compressed): New function.
651 (put_cmp_opcode): New function.
652 (put_cmp_number): New function.
653 (write_int): New function.
654 (convert_double_to_output_format): New function.
655 (write_float): New function.
656 (write_value): New function.
657 (write_string): New function.
658 (write_bytes): New function.
659 (write_zeros): New function.
660 (write_spaces): New function.
662 Reviewed by John Darrington.
664 2007-07-22 Ben Pfaff <blp@gnu.org>
666 Don't try to write very long strings to portable files. The
667 format does not support it.
669 * por-file-writer.c (MAX_POR_WIDTH): New macro.
670 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
671 (write_format): Add arg to take width to resize format to.
672 (write_value): Limit width of value written to MAX_POR_WIDTH.
673 (write_variables): Limit width of variable and its output formats
675 Reviewed by John Darrington.
677 2007-07-22 Ben Pfaff <blp@gnu.org>
679 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
680 local variable instead of literal 5.
681 Reviewed by John Darrington.
683 2007-07-22 Ben Pfaff <blp@gnu.org>
685 Fix problems with uniqueness of short names in system files with
686 very long string variables. Now a variable may have multiple
689 * automake.mk (src_data_libdata_a_SOURCES): Add new files
690 short-names.c, short-names.h.
692 * dictionary.c (dict_clone): Clone all the short names.
693 (compare_strings): Move into short-names.c.
694 (hash_strings): Ditto.
695 (set_var_short_name_suffix): Ditto.
696 (dict_assign_short_names): Ditto, rename short_names_assign,
697 change to assign all short names.
699 * por-file-writer.c (write_variables): Use short_names_assign
700 instead of dict_assign_short_names.
702 * short-names.c: New file.
704 * short-names.h: New file.
706 * sys-file-private.c (sfm_width_to_segments): New function.
708 * sys-file-reader.c (read_long_var_name_map): Save and restore all
709 the short names, not just the first one.
711 * sys-file-writer.c (cont_var_name): Removed.
712 (sfm_open_writer): Use short_names_assign instead of
713 dict_assign_short_names. Use unique short names assigned by
714 short_names_assign instead of those generated by cont_var_name.
716 * variable.c (struct variable): Remove `short_name' member,
717 replace by `short_names' and `short_name_cnt'.
718 (var_create) Initialize new members.
719 (var_get_short_name_cnt): New function.
720 (var_get_short_name): Now takes an index argument. Changed most
722 (var_set_short_name): Ditto.
723 (var_clear_short_name): Renamed var_clear_short_names, changed to
724 clear all short names.
726 Reviewed by John Darrington.
728 2007-07-22 Ben Pfaff <blp@gnu.org>
730 * variable.c (var_set_width): Use new var_set_width function.
732 * missing-values.c (mv_n_values): Drop assertion, which was not
735 * format.c (fmt_default_for_width): New function.
736 (fmt_resize): New function.
738 Reviewed by John Darrington.
740 2007-07-18 John Darrington <john@darrington.wattle.id.au>
742 * datasheet.c (datasheet_delete_columns): Added assertion to check
743 we're not deleting outside the range of the sheet.
746 * dictionary.c dictionary.h variable.c: Added the ability for string
747 variables to be resized.
749 * vardict.h: Added some prototypes (moved from dictionary.h) as
750 these should only be called by variable.c
753 2007-07-14 John Darrington <john@darrington.wattle.id.au>
755 * sfm-reader.c: Respect case_cnt field in file header.
757 2007-07-01 John Darrington <john@darrington.wattle.id.au>
759 * transformation.c transformation.h (trns_chain_execute): Changed the
760 signature (Patch #6057)
762 2007-06-10 Ben Pfaff <blp@gnu.org>
764 * casereader-filter.c (casereader_filter_destroy): Make sure to
765 write all the remaining excluded cases to the casewriter, if any.
767 * caseinit.c (init_list_destroy): Rewrite.
768 (init_list_clear): Ditto.
770 * casegrouper.c (casegrouper_get_next_group): Always set *reader
771 to null when returning false.
773 2007-06-06 Ben Pfaff <blp@gnu.org>
775 Actually implement the new procedure code and adapt all of its
776 clients to match. Also adapt all of the other case sources and
777 sinks in the tree and their clients to use the
778 casereader/casewriter infrastructure.
780 * automake.mk: Add and remove files.
782 * any-reader.c: Change into a casereader.
783 * por-file-reader.c: Ditto.
784 * scratch-reader.c: Ditto.
785 * sys-file-reader.c: Ditto.
787 * any-writer.c: Change into a casewriter.
788 * por-file-writer.c: Ditto.
789 * scratch-writer.c: Ditto.
790 * sys-file-writer.c: Ditto.
792 * procedure.c: Change to use casereader, casewriter, caseinit, and
793 other new infrastructure.
795 * scratch-handle.c: Adapt to new infrastructure.
797 * case-sink.c: Removed, now dead code.
798 * case-sink.h: Ditto.
799 * case-source.c: Ditto.
800 * case-source.h: Ditto.
801 * casefile-factory.c: Ditto.
802 * casefile-private.h: Ditto.
805 * casefilter.c: Ditto.
806 * casefilter.h: Ditto.
809 * fastfile-factory.c: Ditto.
810 * fastfile-factory.h: Ditto.
811 * storage-stream.c: Ditto.
812 * storage-stream.h: Ditto.
814 2007-06-06 Ben Pfaff <blp@gnu.org>
818 * automake.mk: Add new files.
820 * datasheet.c: New file.
822 * datasheet.h: New file.
824 2007-06-06 Ben Pfaff <blp@gnu.org>
826 Until now, the procedure code has provided a case to the
827 case_source, which has filled in the data values that come from
828 the active file. "Left" data values that don't come from the
829 active file naturally stay the same from case to case, because the
830 procedure code keeps using that same case.
832 One of the compromises that comes with the new procedure code is
833 that the active file allocates and provides its own case, which
834 the procedure code then has to resize to provide room for any
835 other variables that should go in the case and then fill in the
836 values of "left" variables. Then, when we're done with that case,
837 we have to save the values of "left" variables to copy into the
838 next case read from the active file.
840 The caseinit code helps with this.
842 * automake.mk: Add new files.
844 * caseinit.c: New file.
846 * caseinit.h: New file.
848 2007-06-06 Ben Pfaff <blp@gnu.org>
850 * value.h (value_cnt_from_width): New function.
852 * variable.c (var_get_value_cnt): Use new function.
854 2007-06-06 Ben Pfaff <blp@gnu.org>
856 Add casegrouper, to allow cases read from a given casereader to be
857 broken into groups, each of which has its own casereader.
858 Generally cases are grouped based on having equal values for some
861 * automake.mk: Add new files.
863 * casegrouper.c: New file.
865 * casegrouper.h: New file.
867 2007-06-06 Ben Pfaff <blp@gnu.org>
869 Add interface to lexicographical ordering of cases.
871 * automake.mk: Add new files.
873 * case-ordering.c: New file.
875 * case-ordering.h: New file.
877 2007-06-06 Ben Pfaff <blp@gnu.org>
879 Add casereaders and casewriters, the basis of the new data processing
880 implementation. A casereader is a uniform interface to reading cases
881 from a data source; a casewriter is a uniform interface to writing
882 cases to a data sink.
884 * automake.mk: Add new files.
886 * casereader-filter.c: New file.
888 * casereader-provider.h: New file.
890 * casereader-translator.c: New file.
892 * casereader.c: New file.
894 * casereader.h: New file.
896 * casewriter-provider.h: New file.
898 * casewriter-translator.c: New file.
900 * casewriter.c: New file.
902 * casewriter.h: New file.
904 2007-06-06 Ben Pfaff <blp@gnu.org>
906 "casewindow" data structure that extends the deque (from libpspp)
907 of cases with the ability to dump cases to disk if we get too many
910 * automake.mk: Add new files.
912 * casewindow.c: New file.
914 * casewindow.h: New file.
916 2007-06-06 Ben Pfaff <blp@gnu.org>
918 sparse_cases data structure that augments a sparse_array of cases
919 with the ability to dump cases to disk if we get too many cases in
922 * automake.mk: Add new files.
924 * sparse-cases.c: New file.
926 * sparse-cases.h: New file.
928 2007-06-06 Ben Pfaff <blp@gnu.org>
930 Adds a low-level on-disk case array data structure.
932 * automake.mk: Add new files.
934 * case-tmpfile.c: New file.
936 * case-tmpfile.h: New file.
938 2007-06-06 Ben Pfaff <blp@gnu.org>
940 In a couple of places we calculate the maximum number of cases to
941 keep in memory based on the user-defined workspace. Enable
942 centralizing the calculation through a new function.
944 * settings.c (get_workspace_cases): New function.
946 2007-06-06 Ben Pfaff <blp@gnu.org>
948 The casenumber type is defined in transformations.h, but case.h is
949 a more sensible place. Move it.
951 * case.h (CASENUMBER_MAX): New macro.
952 (typedef casenumber): Move here, from transformations.h.
954 2007-06-03 Ben Pfaff <blp@gnu.org>
956 Slightly generalize case_to_values and case_from_values functions.
958 * case.c (case_to_values): Rename case_copy_out, change interface.
959 (case_from_values): Rename case_copy_in, change interface.
961 * fastfile.c (fastfilereader_get_next_case): Update caller.
962 (write_case_to_disk): Ditto.
964 2007-06-02 Ben Pfaff <blp@gnu.org>
966 Clean up after a forgotten part of patch #5829.
968 * casedeque.h: Remove unused file.
970 * automake.mk: Remove casedeque.h from sources.
972 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
974 * category.c: Removed redundant #include
976 2007-05-06 Ben Pfaff <blp@gnu.org>
978 Abstract the documents within a dictionary a little better.
979 Thanks to John Darrington for suggestion, initial version, and
982 * dictionary.c (struct dictionary): Change `documents' member from
983 char * to struct string.
984 (dict_clear): Destroy struct string.
985 (dict_get_documents): Convert struct string to char *.
986 (dict_set_documents): Set struct string. Pad to 80-character
988 (dict_clear_documents): New function.
989 (dict_add_document_line): New function.
990 (dict_get_document_line_cnt): New function.
991 (dict_get_document_line): New function.
993 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
995 * sys-file-reader.c (read_documents): Use new document functions.
997 2007-04-19 John Darrington <john@darrington.wattle.id.au>
999 * sys-file-reader.c: When reading a system file which has no
1000 long name table, automatically create one where the long names
1001 are the lower case versions of the short names.
1003 2007-04-22 Ben Pfaff <blp@gnu.org>
1005 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1006 dict_clear will free most data related to the dictionary.
1007 dict_clear does a decent job, except that dict_set_split_vars on
1008 some systems won't actually free the dict's "split" member.
1009 Instead, it'll allocate a 1-byte region. Fix this.
1011 * value.c (value_copy): New function.
1012 (value_set_missing): Ditto.
1014 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1016 * Deleted existing category.h and moved cat-routines.h into
1017 category.h Encapsulated struct cat_vals better.
1019 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1021 * sys-file-reader.c: When reading a system file which has no
1022 long name table, automatically create one where the long names
1023 are the lower case versions of the short names.
1025 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1027 * sys-file-reader.c: Some versions of Other Software seem to
1028 produce system files with string variables' measure set to
1029 zero. We'll assume these are supposed to be nominal variables.
1031 2007-03-30 Ben Pfaff <blp@gnu.org>
1033 * procedure.c: Adapt to new deque data structure.
1035 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1036 Ben Pfaff <blp@gnu.org>
1038 * file-name.c: Mingw compatibility fixes.
1039 (fn_search_path): Use ISSLASH instead of comparing against '/'
1041 (fn_dir_name): Use dir_name from gnulib.
1042 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1043 (fn_get_identity): Use GetFullPathName instead of canonicalize
1044 from gnulib, because the latter does not fully support
1045 Windows-style path names. Use this implementation based on the
1046 detected presence of Windows instead of the absence of Unix, since
1047 the new implementation is Windows-specific.
1048 (fn_compare_file_identities): In Windows implementation, compare
1049 names case-insensitively.
1051 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1053 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1054 <stdlib.h> to have the same effect.
1056 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1058 * por-file-reader.c: Add missing _() around messages.
1060 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1062 * make-file.c: Include "mkstemp.h", without which linking on
1065 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1067 Reduce platform dependence.
1069 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1070 Everywhere we using this, we really should have just depended on
1071 the shell to expand tildes.
1072 (fn_search_path): Simplify, given that we don't do tilde expansion
1074 (fn_normalize): Removed. Caller changed to use the canonicalize
1076 (fn_get_cwd): Removed. Only user was fn_normalize.
1077 (fn_is_absolute): Really only test for absolute names.
1078 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1080 (fn_readlink): Removed, as it was only used fn_normalize.
1081 (fn_exists): Assume the stat function is available; gnulib does.
1082 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1084 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1086 * dictionary.c (dict_create_vector_assert): New function.
1088 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1090 * file-name.c (fn_normalize): Correct name of function
1091 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1092 for pointing this out.
1094 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1096 We are using a single member in struct file_handle, the "name"
1097 field, for more than one purpose. When it begins with '"', it's a
1098 file name; otherwise, it's a token that can be used to identify
1099 it. When that assertion fires, it's because we searched for the
1100 name case-sensitively as a file name (so that there was no match),
1101 and then we try to insert it case-insensitively as a token, which
1102 fails because duplicates aren't allowed.
1104 Solution: break the two purposes into two separate fields. This
1105 fixes the problem and likely makes the code easier to read too.
1107 Fixes bug #18922. Thanks to John Darrington for bug report and
1110 * file-handle-def.c (struct file_handle): New `id' member.
1111 (fh_from_name): Rename fh_from_id. Update all callers.
1112 (create_handle): New `id' parameter. Update all callers.
1113 (fh_create_file): Ditto.
1114 (fh_get_id): New function.
1116 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1118 * case.c (case_is_null): Change return type to bool.
1120 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1124 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1125 reference counting to share data. This makes it easy for
1126 valgrind, etc. to find accesses to cases that have been destroyed
1127 but have been kept around by another user's ref-count. This often
1128 happens when the data set is small enough to find in memory; if a
1129 bigger data set that would overflow to disk were used, then data
1130 corruption would occur.
1132 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1136 * case.c (case_unshare): Make it check internally whether the
1137 ref_cnt is greater than 1, so that the callers don't have to.
1138 Update callers not to check.
1140 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1142 Before, I was thinking that I might want to get rid of reference
1143 counting at some point. Now, I'm pretty sure that it's here to
1144 stay. Thus, because we have to store the value_cnt anyway for
1145 reference-counted cases, we might as well expose it to users.
1147 * case.c (case_get_value_cnt): New function.
1148 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1149 resize case if its size actually changed.
1151 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1152 instead of peeking inside struct case directly.
1153 (casefile_append): Ditto.
1155 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1157 Get rid of the inlines for the case functions, which made the
1158 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1159 the inlines didn't speed up "make check" at all, which is not a
1160 perfect benchmark but seems indicative.)
1162 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1163 definitions. Remove some assertions on nonnull pointers that were
1164 redundant with a pointer dereference soon after in the function.
1166 (struct case_data): Move definition here from case.h.
1170 (case_data_wr): Ditto.
1172 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1174 * automake.mk: Add casedeque.h to sources.
1176 * casedeque.h: New file.
1178 * procedure.c: (struct dataset) Change lag_count, lag_head,
1179 lag_queue member into single struct casedeque member. Update all
1180 users to use the casedeque instead.
1183 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1185 * procedure.c: Simplify lagged cases interface. Updated all
1186 clients--well, the only client--to use the simplified interface.
1187 (dataset_n_lag) Removed.
1188 (dataset_set_n_lag) Removed.
1189 (dataset_need_lag) New function.
1191 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1193 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1194 whenever appropriate, but only when the dataset/dictionary is in a
1197 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1199 * dictionary.c dictionary.h : Added callbacks for change of filter and
1200 split variables. Refactored some code to ensure that callbacks get
1201 invoked when appropriate.
1203 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1204 replace_dict callback occurs when permanent_dict replaces the current
1207 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1209 * dictionary.c dictionary.h : Added callback for when the weight
1210 variable of a dictionary changes.
1212 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1214 * dictionary.c dictionary.h : Added replace_source and replace_dict
1215 callbacks, and functions to deal with them.
1217 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1219 Simplify missing value handling.
1221 * missing-values.h (enum mv_class): New type.
1222 (enum mv_type): Moved definition into missing-values.c and renamed
1223 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1224 enums. Updated all uses.
1225 (struct missing_values): Changed type of `type' from `enum
1226 mv_type' to `int' because the definition is no longer exposed.
1228 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1229 parameter. Update all callers.
1230 (mv_is_num_missing): Ditto.
1231 (mv_is_str_missing): Ditto.
1232 (mv_is_value_user_missing): Removed. Changed callers to use
1233 mv_is_value_missing.
1234 (mv_is_num_user_missing): Removed. Changed callers to use
1236 (mv_is_str_user_missing): Removed. Changed callers to use
1238 (mv_is_value_system_missing): Removed. Changed callers to use
1239 mv_is_value_missing.
1240 (mv_set_type): Removed. Changed callers to use mv_clear.
1241 (mv_clear): New function.
1243 * variable.c (var_is_value_missing): Add new enum mv_class
1244 parameter. Update all callers.
1245 (var_is_num_missing): Ditto.
1246 (var_is_str_missing): Ditto.
1247 (var_is_value_user_missing): Removed. Changed callers to use
1248 var_is_value_missing.
1249 (var_is_num_user_missing): Removed. Changed callers to use
1251 (var_is_str_user_missing): Removed. Changed callers to use
1253 (var_is_value_system_missing): Removed. Changed callers to use
1254 var_is_value_missing.
1256 * casefilter.c (struct casefilter): Use enum mv_class in place of
1258 (casefilter_variable_missing): Adapt to new member.
1259 (casefilter_create): Change signature to take enum mv_class,
1262 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1264 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1266 * case-sink.c case-sink.h procedure.c procedure.h
1267 storage-stream.c: Now uses the factory.
1269 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1271 Make it possible to pull cases from the active file with a
1272 function call, instead of requiring indirection through a callback
1275 * case-source.h (struct case_source_class): Change ->read function
1276 to return a single case, instead of calling a callback function
1277 for each case. Change ->destroy function to return an error
1280 * case-source.c (free_case_source): Pass along the value returned
1281 by the case_source ->destroy function.
1283 * procedure.c (struct write_case_data): Removed.
1284 (struct dataset): Added some members to track procedure state.
1285 (procedure): Optimize the trivial case at this level.
1286 (internal_procedure): Re-implement in terms of proc_open,
1287 proc_read, proc_close.
1288 (proc_open) New function.
1289 (proc_read) New function.
1290 (proc_close) New function.
1291 (write_case) Moved into proc_read.
1292 (close_active_file) Moved closing of data source into proc_close.
1294 * storage-source.c: Rewrote to conform with modified
1295 case_source_class interface.
1297 * transformations.c (trns_chain_execute): Added argument to allow
1298 starting execution from an arbitrary transformation. Updated
1301 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1303 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1305 * sys-file-reader.c (read_display_parameters): Don't assume that
1306 MEASURE_* and ALIGN_* have the same values found in system files.
1308 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1310 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1311 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1312 the value of n_MEASURES, which was off by 1 (at least from my
1315 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1317 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1318 callbacks which are invoked when the dictionary or its
1319 variables are changed.
1321 * missing-values.c missing-values.h value-labels.c: Tidied up
1322 consistency checks, and made some of them return false
1323 instead of assert-failing.
1325 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1327 * calendar.c (calendar_days_in_month): New function.
1329 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1331 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1332 the value label's width.
1334 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1336 * sfm-private.h: Move contents into sys-file-writer.c, which is
1337 the only remaining user. Removed Borland C++-specific directives.
1339 * sys-file-reader.c: Clean up and rewrite entire file. The
1340 rewritten version is simpler and better abstracted, and should be
1341 easier to maintain and extend. It avoids using structures to read
1342 file data, which is prone to padding variations among compilers.
1343 It should also handle non-IEEE 754 system files, although I
1344 haven't been able to find any. It has been tested against many
1345 .sav files obtained from the Web and found to produce the same
1346 results as the earlier version of the code, or in some cases
1347 improved results. It is more tolerant of format variations found
1350 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1351 member, putting an enum integer_format in its place. New member
1352 `float_format'. Changed `compressed' member to type bool.
1354 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1356 * dictionary.c (dict_delete_consecutive_vars): New function.
1358 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1360 * file-name.c (fn_search_path): Remove prefix arg that was unused
1361 by any caller. Updated all callers.
1363 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1365 * format.c (fmt_dollar_template): Use user's decimal point
1366 character. Add assertion.
1368 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1370 * format.c (fmt_dollar_template): New function, based on
1371 dollar_format_template from var-type-dialog.c.
1373 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1375 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1378 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1380 Finish converting struct variable to an opaque type. In this
1381 phase, we add remaining setter and getter functions, convert the
1382 remaining PSPP code to use them, and do a bunch of cleanup. The
1383 resulting changes are pervasive but mostly trivial, and only the
1384 notable changes are logged.
1386 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1389 * case.c (case_data): Renamed case_data_idx.
1390 (case_num): Renamed case_num_idx.
1391 (case_str): Renamed case_str_idx.
1392 (case_data_rw): Renamed case_data_rw_idx.
1394 * case.h (case_data): New function with old name and an interface
1395 that takes a variable instead of an index, which is easier to
1396 use. Updated all callers to use the new interface, or to use the
1397 new *_idx function (see above).
1400 (case_data_rw): Ditto.
1402 * category.c (cat_stored_values_destroy): Changed interface to
1403 take a struct cat_vals * instead of a struct variable *.
1405 * dictionary.c (dict_clone): Use new vector_clone function.
1406 (dict_clear) Use new var_destroy function.
1407 (add_var) New function.
1408 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1409 (dict_create_var_assert) Rewrite in terms of add_var.
1410 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1411 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1412 (dict_lookup_var) Use new var_create, var_destroy functions.
1413 (dict_contains_var) Rewrite in terms of new vardict functionality.
1414 (set_var_dict_index) New function.
1415 (set_var_case_index) New function.
1416 (reindex_vars) New function.
1417 (dict_delete_var) Rewrite in terms of new vardict functionality.
1418 (dict_reorder_var) Ditto.
1419 (dict_reorder_vars) Ditto.
1420 (rename_var) New function.
1421 (dict_rename_var) Use rename_var.
1422 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1423 (dict_get_compacted_idx_to_fv) Rename
1424 dict_get_compacted_dict_index_to_case_index, update callers.
1425 (dict_create_vector) Use new vector_create function.
1426 (dict_clear_vectors) Use new vector_destroy function.
1427 (set_var_short_name_suffix) Move here from variable.c, renamed
1428 from var_set_short_name_suffix, make static, update caller.
1430 * sys-file-private.c: New file.
1431 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1432 width_to_bytes, update callers.
1434 * sys-file-private.h: New file. Later it will supplant
1435 sfm-private.h; for now it supplements it.
1436 (macro MIN_VERY_LONG_STRING) New macro.
1437 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1439 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1440 MAX_LONG_STRING was used before.
1442 * sys-file-writer.c: Ditto.
1444 * value-labels.c: Change the paradigm here to be that a null
1445 pointer is OK for a struct val_labs * in most cases; it just
1446 represents an empty set of value labels.
1447 (val_labs_copy) A copy of a null set is a null set.
1448 (val_labs_count) A null set has 0 labels.
1449 (val_labs_replace) Change return type to void. Rewrite for
1451 (val_labs_find) A null set does not contain the value.
1452 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1453 transposed argument order, updated all callers.
1455 * value.c: New file.
1456 (value_dup) Moved here from variable.c.
1457 (compare_values) Ditto.
1460 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1461 (macro MAX_LONG_STRING) Removed, because it was only interesting
1462 for system files, not for general code.
1463 (macro MAX_VERY_LONG_STRING) Ditto.
1464 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1465 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1467 * vardict.h: New file, for an interface between variables and
1470 * variable.c: A lot of functions were moved around, for better
1472 (struct variable) Move definition here, from variable.h.
1473 (var_type_adj) Removed--makes i18n hard.
1474 (var_type_noun) Ditto.
1475 (var_create) New function.
1476 (var_clone) New function.
1477 (var_destroy) New function.
1478 (var_set_name) Assert that variable is not in a dictionary.
1479 (compare_var_names) Rename compare_vars_by_name and fix a couple
1480 of callers who thought the args were strings.
1481 (hash_var_name) Rename hash_var_by_name.
1482 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1483 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1484 (var_is_very_long_string) Removed, because it was only interesting
1485 to system file code.
1486 (var_set_missing_values) Allow the argument to be the wrong width,
1487 as long as we can resize it. Simplify callers who were doing the
1488 resizing themselves.
1489 (var_get_value_labels) New function.
1490 (var_has_value_labels) New function.
1491 (var_set_value_labels) New function.
1492 (alloc_value_labels) New function.
1493 (var_add_value_label) New function.
1494 (var_replace_value_label) New function.
1495 (var_clear_value_labels) New function.
1496 (var_lookup_value_label) New function.
1497 (var_get_value_name) Moved here from variable.c, renamed from
1498 var_get_value_name, transposed argument order, updated all
1500 (var_to_string) Moved here, from variable-label.c.
1501 (var_set_leave) New function.
1502 (var_get_leave) New function.
1503 (var_must_leave) New function.
1504 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1505 set_var_short_name_suffix.
1506 (var_get_dict_index) New function.
1507 (var_get_case_index) New function.
1508 (var_get_obs_vals) New function.
1509 (var_set_obs_vals) New function.
1510 (var_has_obs_vals) New function.
1511 (var_get_vardict) New function.
1512 (var_set_vardict) New function.
1513 (var_has_vardict) New function.
1514 (var_clear_vardict) New function.
1515 (value_dup) Moved to value.c.
1516 (compare_values) Ditto.
1519 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1520 (enum ALPHA) Rename VAR_STRING, update all users.
1522 * vector.c: New file.
1523 (struct vector) Moved here, from variable.h.
1524 (check_widths) New function.
1525 (vector_create) New function.
1526 (vector_clone) New function.
1527 (vector_destroy) New function.
1528 (vector_get_name) New function.
1529 (vector_get_var) New function.
1530 (vector_get_var_cnt) New function.
1531 (compare_vector_ptrs_by_name) New function.
1533 * vector.h: New file.
1535 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1537 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1538 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1541 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1543 * value-labels.c (destroy_atoms): New function.
1544 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1546 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1548 Thanks to Jason Stover for pointing out this problem.
1550 * data-out.c (output_number): Use gsl_finite from GSL, which is
1551 portable, instead of isfinite, which is not.
1554 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1556 * variable.c variable.h (value_dup): New function.
1558 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1560 Start converting struct variable to an opaque type. In this
1561 phase, we add a bunch of setter and getter functions and convert
1562 most of the PSPP code to use them. The resulting changes are
1563 pervasive but mostly trivial, and only the notable changes are
1566 * format.c (fmt_equal): New function.
1568 * variable.c (var_type_is_valid): New function.
1569 (measure_is_valid) Moved here, from format.c.
1570 (alignment_is_valid) Moved here, from format.c.
1571 (var_get_name) New function.
1572 (var_set_name) New function.
1573 (width_to_type) New function.
1574 (var_get_type) New function.
1575 (var_get_width) New function.
1576 (var_set_width) New function.
1577 (var_is_numeric) New function.
1578 (var_is_alpha) New function.
1579 (var_is_short_string) New function.
1580 (var_is_long_string) New function.
1581 (var_is_very_long_string) New function.
1582 (var_get_missing_values) New function.
1583 (var_set_missing_values) New function.
1584 (var_clear_missing_values) New function.
1585 (var_has_missing_values) New function.
1586 (var_is_value_missing) New function.
1587 (var_is_num_missing) New function.
1588 (var_is_str_missing) New function.
1589 (var_is_value_user_missing) New function.
1590 (var_is_num_user_missing) New function.
1591 (var_is_str_user_missing) New function.
1592 (var_is_value_system_missing) New function.
1593 (var_get_print_format) New function.
1594 (var_set_print_format) New function.
1595 (var_get_write_format) New function.
1596 (var_set_write_format) New function.
1597 (var_set_both_formats) New function.
1598 (var_get_label) New function.
1599 (var_set_label) New function.
1600 (var_clear_label) New function.
1601 (var_has_label) New function.
1602 (var_get_measure) New function.
1603 (var_set_measure) New function.
1604 (var_get_display_width) New function.
1605 (var_set_display_width) New function.
1606 (var_get_alignment) New function.
1607 (var_set_alignment) New function.
1608 (var_get_value_cnt) New function.
1609 (var_get_leave) New function.
1610 (var_get_short_name) New function.
1612 * variable.h: (struct variable) Removed "type" and "nv" members;
1613 they are now computed from "width" where needed.
1615 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1617 * missing-values.c (mv_resize): Don't write beyond end of the
1618 allocated buffer when resizing a long string.
1620 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1622 Clean up identifier code: don't require identifier enumerations to
1623 be in a particular order; make better use of string library;
1624 expose less of the internals.
1626 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1627 change interface. Updated all callers.
1628 (lex_id_match) Change interface to use struct substring, update
1630 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1631 (global array keywords[]) Make static, change form. Update all
1632 users to use lex_id_name instead.
1633 (lex_is_keyword) New function.
1634 (lex_id_to_token) Change interface to use struct substring, update
1636 (lex_id_name) New function.
1638 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1639 lex_is_keyword instead.
1640 (T_LAST_KEYWORD) Removed.
1641 (T_N_KEYWORDS) Removed.
1643 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1645 * format.c: (fmt_date_template) Distinguish characters for which a
1646 space is output and any date delimiter is allowed on input, from
1647 those for which a space is output and only a space is allowed on
1648 input. The former is represented by X, the latter by a space.
1649 Also, drop distinction between h and H, changing the former to the
1652 * data-in.c: Completely rewrite internals to conform to SPSS input
1653 formats as closely as possible.
1654 (data_in) Changed external interface by replacing the structure
1655 that was used as a single argument by a set of arguments. Updated
1657 (data_in_finite_line) Removed. Converted all callers to use plain
1659 (data_in_get_integer_format) New function.
1660 (data_in_set_integer_format) New function.
1661 (data_in_get_float_format) New function.
1662 (data_in_set_float_format) New function.
1664 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1665 (struct data_in) Removed.
1667 * data-out.c: (output_date) Drop each component from the input as
1668 it is output, to allow us to drop the distinction between h (a
1669 count of hours) and H (the hour of day) template characters.
1670 Also, handle new X template character.
1671 (output_scientific) Follow more rational rule on when to drop
1672 fraction introduced between SPSS 13 and 15. Updated test case to
1675 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1677 Fix buffer overflow reported by John Darrington.
1679 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1680 realize that DIGITS is a count of nibbles, not of bytes.
1682 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1684 * calendar.c (calendar_offset_to_gregorian) Also return the
1685 year-of-day. Change callers to new interface.
1687 * data-out.c: Completely rewrite internals to conform to SPSS
1688 output formats as completely as possible.
1689 (data_out) Change interface to put input parameters before output
1690 parameters, for consistency with the style I now prefer. Update
1692 (data_out_get_integer_format) New public function.
1693 (data_out_set_integer_format) New public function.
1694 (data_out_get_float_format) New public function.
1695 (data_out_set_float_format) New public function.
1697 * data-out.h: New file. Move prototype for data_out here, from
1700 * format.c: (fmt_step_width) Use equality comparison instead of
1701 bitwise and, for clarity.
1702 (fmt_is_string) Ditto.
1703 (fmt_input_to_output) Fix categories that are translated to F
1706 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1708 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1709 casefile-private.h: Added casefilter to assist commands with missing
1712 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1714 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1716 * settings.c: (route_errors_to_terminal) New variable.
1717 (route_errors_to_listing) New variable.
1718 (get_error_routing_to_terminal) New function.
1719 (set_error_routing_to_terminal) New function.
1720 (get_error_routing_to_listing) New function.
1721 (set_error_routing_to_listing) New function.
1723 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1725 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1727 * format.c: Completely rewrite, to achieve better abstraction.
1728 Rewrite all references to formats in other files.
1730 * format.def: Rewrite and reorganize.
1732 * settings.c: Move everything related to custom currency formats
1733 into format.[ch], changing them in form, so as to group related
1734 code and definitions better. Changed all references to use the
1736 (static var decimal) Removed.
1737 (static var grouping) Removed.
1738 (static var cc) Removed.
1739 (get_decimal) Removed.
1740 (set_decimal) Removed.
1741 (get_grouping) Removed.
1742 (set_grouping) Removed.
1746 * settings.h: (macro CC_CNT) Removed.
1747 (macro CC_WIDTH) Removed.
1748 (struct custom_currency) Removed.
1750 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1752 * data-in.c (data_in): Use switch statement instead of table, to
1753 avoid dependence on the order of the FMT_* enums.
1755 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1757 * data-out.c: (num_to_string) Removed, because it was dead code.
1759 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1761 * data-in.c (parse_trailer): Fix error message.
1763 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1765 * format.c (fmt_is_binary): New function.
1767 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1769 * procedure.c procedure.h: Encapsulated the static data into a single
1772 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1774 * casefile.c (casereader_read_xfer): Always initialize the case,
1775 even on an error condition.
1777 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1779 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1780 entire function useless.
1782 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1784 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1785 casereader method casereader_clone.
1787 * procedure.c pransformations.h: Introduced new type casenum_t
1789 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1791 * variable.c: (width_to_bytes) Rephrase code for clarify.
1793 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1795 * format.c: (fmt_type_from_string) New function.
1796 (fmt_to_string) Include decimals in output if the format has
1797 decimals, even if the format type does not. This way, we can
1798 accurately reproduce incorrect formats in user output.
1799 (check_common_specifier) Make the check for a bad format type an
1800 assertion, so we get bug reports if they show up. Fix message.
1801 Check for decimal places with a format type that doesn't allow
1803 (check_input_specifier) Remove check for FMT_X, which has been
1805 (check_output_specifier) Ditto.
1807 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1809 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1810 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1811 (enum fmt_parse_flags) Removed.
1813 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1815 * casefile.c casefile.h: Converted to an abstract base class.
1816 * casefile-private.h fastfile.c fastfile.h: New files.
1817 * automake.mk procedure.c scratch-writer.c storage-stream.c
1819 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1821 * procedure.c (internal_procedure): Create sink_case with only as
1822 many values as the compacted dictionary.
1824 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1826 Remove "debugging" code that caused plenty of false positives and
1829 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1831 * case.c: Remove all references to `this' member.
1833 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1835 Fix link error noted by Jason Stover.
1837 * storage-stream.c: Include <assert.h>.
1839 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1841 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1842 ALL) and additional underlying system file issues.
1844 Thanks to John Darrington for review.
1846 First problem: var_hash points to variables not owned by the
1847 sys-file-reader, which the caller may free or modify. Use an
1848 array of sfm_vars instead, as done earlier (e.g. CVS version
1851 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1852 members and remove all code that references it. Add vars, var_cnt
1853 members. Remove fix_specials member, which was unused.
1854 (struct sfm_var) Remove name member, which was unused.
1855 (sfm_close_reader) Free vars member instead of var_hash.
1856 (compare_var_shortnames) Removed.
1857 (hash_var_shortname) Removed.
1858 (sfm_open_reader) Fill out vars array.
1859 (compare_var_index) Removed.
1860 (sfm_read_case) Use vars instead of var_hash.
1862 Second problem: we're confused about when we actually have very
1863 long strings, causing us to choose incorrectly between slow path
1864 and fast path in sfm_read_case.
1866 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1867 have very long strings, not when we have long variable names,
1868 which is an unrelated feature.
1870 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1872 * variable.h: Move var_set and variable parsing declarations to
1873 new header, src/language/lexer/variable-parser.h. Modified lots
1874 of files to include the new header.
1876 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1878 * value-labels.c (value_to_string): When there's no value label,
1879 format the variable according to its print format, instead of
1880 always effectively using A or F format.
1882 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1884 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1887 * format.c: Removed tortological assertion.
1889 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1891 Reform string library.
1893 * file-name.c (fn_interp_vars): Change interface to take a
1894 substring as input. Updated all users.
1896 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1898 * format.c (measure_is_valid): Really return false when m >=
1901 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1903 Implement random access to casefiles, for use in GUI.
1905 * casefile.c: (struct casereader) Add `random', `file_ofs',
1906 `buffer_ofs' members.
1907 (casefile_get_random_reader) New function.
1908 (read_open_file) Break part into new function
1909 seek_and_fill_buffer().
1910 (fill_buffer) Update buffer_ofs, file_ofs.
1911 (casereader_seek) New function.
1913 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1915 * settings.c: Added call to i18n{done, init}.
1917 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1919 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1922 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1924 * casefile.c: Convert many uses of `int' to `bool'.
1926 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1928 * transformations.c (trns_chain_destroy): Destroy chain's trns
1929 member, to fix memory leak.
1931 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1933 * storage-stream.c (storage_source_decapsulate): Destroy case
1934 source to fix memory leak.
1936 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1938 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1939 case passed as argument instead of initializing the argument as a
1940 case. Fixes memory leak that showed up in
1941 tests/command/aggregate.sh with scratch files.
1943 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1945 * procedure.c (proc_done): Destroy default_dict, to fix memory
1948 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1950 Simplify procedure_with_splits().
1952 * procedure.c (struct split_aux_data): Removed case_count member.
1953 (procedure_with_splits) Don't initialize case_count.
1954 (split_procedure_case_func) Check whether prev_case is null
1955 instead of case_count.
1956 (split_procedure_end_func) Ditto.
1958 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1960 * case.c (case_move): Do nothing if dst and src are the same
1962 (case_try_create) Merge two similar cases.
1963 (case_copy) Unshare only if data must be actually copied.
1965 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1967 * data-in.c data-out.c dictionary.c sys-file-reader.c
1968 sys-file-writer.c variable.c variable.h: Reworked very long string
1969 support for better encapsulation.
1971 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1973 * value-labels.c (val_labs_can_set_width): New function.
1974 (val_labs_set_width) Clear labels if increasing width to long
1976 (val_labs_destroy) Remove unneeded test for null.
1978 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1980 * value-labels.h: Remove unneeded dependency on variable.h.
1982 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1984 Get rid of `char *c' member in union value, for cleanliness.
1986 * value.h: (union value) Remove `c' member.
1988 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1990 Make dictionary compacting functions a little more general.
1992 * sys-file-writer.c (sfm_open_writer): Use
1993 dict_compacting_would_change().
1994 (does_dict_need_translation) Removed.
1996 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1998 Make dictionary compacting functions a little more general.
2000 * dictionary.c (dict_needs_compaction): Rename
2001 dict_compacting_would_shrink(). Update all callers.
2002 (dict_compacting_would_change) New function.
2004 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2006 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2007 inverted return value (!).
2009 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2011 Continue reforming procedure execution.
2013 * procedure.c: Search and replace "vfm" by "proc". Notably:
2014 (static var vfm_source) Rename proc_source. Update all
2016 (static var vfm_sink) Rename proc_sink. Update all references.
2018 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2020 Continue reforming procedure execution. In this phase, remove
2021 PROCESS IF, which was deprecated anyway and can be easily
2022 simulated with TEMPORARY followed by SELECT IF.
2024 * procedure.c: (open_active_file) Don't call
2025 add_process_if_trns().
2026 (discard_variables) Get rid of redundant call to
2027 proc_cancel_all_transformations().
2028 (add_process_if_trns) Removed.
2029 (process_if_trns_proc) Removed.
2030 (process_if_trns_free) Removed.
2032 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2034 Continue reforming procedure execution. In this phase, add
2035 `const' to the case passed to procedure()'s callback.
2037 Updated all users of procedure() as well.
2039 * procedure.c: (struct write_case_data) Add "const" to ccase
2040 parameter for case_func member.
2041 (procedure) Add "const" to ccase parameter for proc_func
2043 (multipass_case_func) Make ccase parameter const.
2044 (internal_procedure) Add "const" to ccase parameter for case_func
2046 (split_procedure_case_func) Make ccase parameter const.
2047 (multipass_split_case_func) Make ccase parameter const.
2049 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2051 Continue reforming procedure execution. In this phase, get rid of
2052 the output code for SPLIT FILE groups in procedure.c, which really
2053 shouldn't be doing any output. Move it into the individual
2054 procedures instead. This also adds some flexibility.
2056 Updated many users of procedure_with_splits() and
2057 multipass_procedure_with_splits() to call
2058 output_split_file_values() and to deal with increased use of
2061 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2062 parameter to begin_func member.
2063 (procedure_with_splits) Add "const struct ccase *" parameter to
2064 begin_func parameter. Make ccase parameter const in proc_func
2066 (split_procedure_case_func) Don't dump split file group. Pass
2068 (dump_splits) Moved to language/dictionary/split-file.c as
2069 output_split_file_values().
2070 (struct multipass_split_aux_data) Add "const struct ccase *"
2071 parameter to split_func member.
2072 (multipass_procedure_with_splits) Add "const struct ccase *"
2073 parameter to split_func parameter.
2074 (multipass_split_case_func) Save new SPLIT FILE case before
2076 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2078 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2080 Continue reforming procedure execution. Change
2081 internal_procedure() so that it calls open_active_file() and
2082 close_active_file(), which isolates most of the actual procedure
2085 * procedure.c: (struct write_case_data) Rename `proc_func' member
2086 to `case_func' and update all references.
2087 (procedure) Rewrite as one-line wrapper around
2088 internal_procedure().
2089 (struct multipass_aux_data) New.
2090 (multipass_callback) Renamed multipass_case_func(). Use struct
2091 multipass_aux_data as auxiliary data.
2092 (multipass_end_func) New function.
2093 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2094 that uses multipass_case_func, multipass_end_func.
2095 (internal_procedure) Add `end_func' argument. Move optimization
2096 of trivial case in here. Move call to open_active_file() and
2097 close_active_file() in here. Now assert that vfm_source is
2099 (procedure_with_splits_callback) Rename
2100 split_procedure_case_func().
2101 (split_procedure_end_func) New function.
2102 (multipass_split_callback) Rename multipass_split_case_func.
2103 (multipass_split_end_func) New function.
2104 (discard_variables) No need to test for nonnull vfm_source.
2106 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2108 Continue reforming procedure execution. Get rid of unused member.
2110 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2112 (write_case) Don't increment cases_analyzed.
2114 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2116 Continue reforming procedure execution. In this phase, move
2117 procedure.c and procedure.h from src to src/data. Update
2118 makefiles and #includes accordingly.
2120 * procedure.c: Moved here from src/.
2122 * procedure.h: Moved here from src/.
2124 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2126 Continue reforming procedure execution. In this phase, get rid of
2127 many global variables, consolidating procedure execution in
2128 procedure.c. Encapsulate transformations in new "struct
2129 trns_chain". Also, change implementation of N OF CASES, FILTER,
2130 and PROCESS IF from special cases to transformations.
2132 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2135 * dictionary.c: (global variable default_dict) Move to
2138 * variable.h: (TRNS_*) Move to transformations.h.
2139 (struct transformation) Move to transformations.c.
2141 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2143 * sys-file-reader.c: Fixed invalid read problems.
2145 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2147 * storage-stream.c: Add missing function comments.
2149 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2151 Continue reforming procedure execution. In this phase, add some
2152 new, needed functionality to storage-stream.
2154 * storage-stream.c: (storage_source_decapsulate) New function.
2156 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2158 * variable.c (width_to_bytes): Declarations must precede
2159 statements for C90 compliance.
2161 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2163 * data-out.c, data-in.c, variable.c, variable.h: New functions
2164 copy_mangle and copy_demangle for reading/writing cases; emulates the
2165 way SPSS deals with strings > 255 bytes.
2167 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2168 subtype 14 needed for strings longer than 255 bytes.
2170 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2171 instead of literal values. Also fixed some constness issues.
2173 * format.h: Constness
2175 * sfm-private.h: Renamed the case_size identifier, since I discovered
2176 that SPSS's respect for this variable is very nominal.
2178 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2180 Change case limit type from int to size_t.
2182 * dictionary.c: (struct dictionary) Change type of case_limit
2184 (dict_get_case_limit) Change return type.
2185 (dict_set_case_limit) Change parameter type.
2187 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2189 * variable.h: (struct variable) Rename `reinit' member as `leave'
2190 and invert sense. Fix up all references.
2192 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2194 Continue reforming procedure execution. In this phase, break
2195 procedure.c into multiple files.
2197 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2199 * case-sink.c: New file.
2201 * case-sink.h: New file.
2203 * case-source.c: New file.
2205 * case-source.h: New file.
2207 * storage-stream.c: New file.
2209 * storage-stream.h: New file.
2211 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2213 * variable.h: (struct variable) Remove `init' member and all
2214 references to it from other files. It was initialized in several
2215 places, but nothing really ever used it for anything worthwhile.
2216 Thanks to Jason Stover for pointing out how confusing this
2219 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2221 Continue reforming error message support. In this phase, get rid
2222 of message "titles" and put the message text in `struct error'.
2223 Now `struct error' encapsulates a message more properly.
2225 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2226 Format message ourselves.
2228 * data-in.c: (vdls_error) Ditto.
2230 * por-file-reader.c: (error) Ditto.
2232 * sys-file-reader.c: (corrupt_msg) Ditto.
2234 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2236 GNU standards require "file name" instead of "filename" in
2237 documentation. It's nice for our code to follow the convention
2240 * casefile.c: (struct casefile) Rename `filename' member to
2241 `file_name'. Updated all references.
2243 * file-name.c: [!unix] (struct file_identity) Rename
2244 normalized_filename member to normalized_file_name. Updated all
2247 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2249 We don't really support anything but Unix-like environments well,
2250 so we might as well de-obfuscate by writing directory and path
2251 separators explicitly.
2253 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2255 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2256 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2258 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2261 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2263 GNU standards require "file name" instead of "filename" in
2264 documentation. It's nice for our code to follow the convention
2267 * filename.c: Rename to file-name.c.
2269 * filename.h: Rename to file-name.h. Update all inclusions.
2270 Update header guards.
2272 * automake.mk: Update file names.
2274 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2276 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2278 (fn_basename) Removed (dead code).
2279 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2280 (fn_special_p) Renamed fn_is_special(), all references updated.
2281 (fn_exists_p) Renamed fn_exists(), all references updated.
2283 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2285 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2286 Also, now it only tilde-expands file names, not paths.
2287 (fn_search_path) Tilde-expand one directory at a time.
2289 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2291 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2292 get rid of non-Unixlike version of the code, which has probably
2294 (fn_prepend_dir) Removed (dead code).
2296 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2297 (macro PATH_SEPARATOR_STRING) New.
2298 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2300 Continue reforming error message support. In this phase, we get
2301 rid of VM() and the other msg() support for "verbosity", replacing
2302 it by a new function verbose_msg().
2304 * filename.c: (fn_search_path) Use verbose_msg() instead of
2307 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2309 * sfm-private.h: Get rid of #defines after #error, which makes no
2312 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2314 Get rid of our own int32 type in favor of the standard int32_t
2317 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2320 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2322 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2324 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2326 Remove ill-considered file routines that are no longer used.
2328 * filename.c: (fn_open_ext) Removed.
2329 (fn_close_ext) Removed.
2331 * filename.h: (struct file_ext) Removed.
2333 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2335 * variable.c (var_is_valid_name): Move declarations before code
2338 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2340 * filename.ch (fn_interp_vars): Fixed small buglet.
2342 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2344 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2345 so as to modify the string inline. Thus makeing it easier to
2346 destroy the results when no longer needed.
2348 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2350 * category.c (cat_stored_values_destroy): Fixed memory leak.
2352 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2354 Add some missing frees. Thanks to John Darrington for reporting
2357 * any-writer.c (any_writer_close): Free writer.
2359 * any-reader.c (any_reader_close): Free reader.
2361 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2363 * por-file-reader.c: (error) Mark as NO_RETURN.
2365 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2367 * settings.c: Changed default value of scompress to true.
2369 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2371 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2373 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2376 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2377 here from pref.h.orig.
2379 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2381 * sys-file-reader.c: Fixed bug reading compressed files.
2383 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2385 * Numerous renames. See src/ChangeLog for details.
2387 * Moved files from src directory
projects / pspp-builds.git / blob