1 2007-12-07 Ben Pfaff <blp@gnu.org>
5 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
6 made from an empty casereader produce a casegrouper with no
7 groups, instead of one with one group that has no cases.
9 * casereader.c (casereader_is_empty): New function.
11 2007-12-06 Ben Pfaff <blp@gnu.org>
15 * sys-file-reader.c (read_display_parameters): Handle variable
16 display parameters record with only 2 data items per variable.
17 Reported by Guido Gay <gay@irer.it>.
19 2007-12-04 Ben Pfaff <blp@gnu.org>
21 * identifier.c (lex_id_match_n): New function.
22 (lex_id_match): Reimplement in terms of lex_id_match_n.
24 2007-11-24 Ben Pfaff <blp@gnu.org>
26 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
29 2007-11-24 Ben Pfaff <blp@gnu.org>
31 Fix warning reported when reading back system files that include
32 very long string variables. Thanks to Guido Gay <gay@irer.it> for
35 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
36 was written where `j' was meant.
38 2007-11-18 Ben Pfaff <blp@gnu.org>
40 Properly write variables that include a range of missing values to
41 system files. Thanks to Guido Gay <gay@irer.it> for reporting
44 * sys-file-writer.c (write_variable): Correctly calculate
45 n_missing_values field when writing variables that include a range
48 2007-11-10 Ben Pfaff <blp@gnu.org>
50 Cleanups and bug fixes devised while writing up documentation.
53 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
55 * dict-class.c: New file.
56 (dict_class_from_id): Move here.
57 (dict_class_to_name): Move here.
59 * dict-class.h: New file.
60 (enum dict_class): Move here. Change from consecutive integers to
61 consecutive bits, to make testing for any of multiple values
62 easier. Add new DC_ALL constant.
64 * dictionary.c (struct dictionary): Change `case_limit' from
66 (dict_get_vars): Make final argument an enum dict_class.
67 (dict_get_vars_mutable): Ditto.
68 (dict_get_case_limit): Change return value to casenumber.
69 (dict_set_case_limit): Change final argument to a casenumber.
70 (dict_unset_split_var): Add assertion. Rephrase slightly.
71 (dict_set_label): Use xstrndup to simplify.
73 * format.c (fmt_step_width): AHEX format also needs 2-byte
75 (fmt_set_style): Simplify assertions.
77 * missing-values.c (mv_add_num_range): Rename mv_add_range.
78 Simplify implementation.
79 (mv_has_value): Simplify implementation.
80 (mv_pop_value): Remove the first value, not the last, to avoid
81 having GET followed by SAVE reverse the order of missing values.
82 (mv_peek_value): Rename mv_get_value. Simplify assertion.
83 (mv_has_range): Simplify implementation.
84 (mv_peek_range): Rename mv_get_range.
85 (can_resize_string): Removed.
86 (mv_is_resizable): Use value_is_resizable.
87 (mv_resize): Use value_resize.
89 * short-names.h (SHORT_NAME_LEN): Move here.
91 * val-type.h: New file, for definitions related to type and width
92 of abstract values. Before, these definitions were mixed among
93 those related to "union value" and those related to variables.
94 (macro SYSMIS): Move here.
95 (macro LOWEST): Move here.
96 (macro HIGHEST): Move here.
97 (macro MAX_STRING); Move here.
98 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
99 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
100 All references updated.
101 (val_type_is_valid): New function. Replaces var_type_is_valid.
102 All references updated.
103 (val_type_from_width): New function. Replaces
104 var_type_from_width. All references updated.
106 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
108 (val_labs_can_set_width): Use value_is_resizable.
109 (val_labs_add): Simply return false if the value labels set is too
110 wide, instead of having undefined behavior.
111 (val_labs_replace): Ditto.
112 (val_labs_replace): Ditto.
113 (val_labs_first): Set iterator to null if iteration is complete.
114 (val_labs_first_sorted): Ditto.
115 (val_labs_done): Become a no-op if the iterator is null.
117 * value.c (value_is_resizable): New function.
118 (value_resize): New function.
120 * variable.c (var_get_dict_class): New function.
122 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
125 2007-11-08 Ben Pfaff <blp@gnu.org>
127 * data-in.c: Make formatted data parsing locale-independent.
128 (parse_number): Use c_strtod instead of strtod, to avoid
129 locale-specific behavior.
132 2007-11-06 Ben Pfaff <blp@gnu.org>
134 Patch #6256: add support for binary, 360 file formats. Reviewed
137 * data-in.c (struct data_in): Add `encoding' member.
138 (data_in): Add `encoding' parameter, and re-encode the data passed
139 in where appropriate. Update all callers to pass it in.
140 (parse_A): Implement EBCDIC recoding wart described in manual.
141 (parse_AHEX): Implement EBCDIC recoding.
143 * data-out.c (data_out_legacy): New function.
144 (data_out): Make into a wrapper around data_out_legacy.
146 * file-handle-def.c (struct file_handle): New member `encoding'.
147 (fh_create_file): Set encoding.
148 (fh_default_properties): Set default encoding.
149 (fh_get_legacy_encoding): New function.
151 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
152 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
153 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
154 (struct fh_properties): New member `encoding'.
156 2007-11-05 Ben Pfaff <blp@gnu.org>
158 Patch #6258. Reviewed by John Darrington.
160 * file-handle-def.c (fh_lock): Add comment that TYPE should be
161 marked with N_() in the caller. Added these markings to each
162 caller too. Should make i18n easier.
163 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
165 2007-11-03 Ben Pfaff <blp@gnu.org>
167 Allow output files to overwrite input files (bug #21280).
169 * file-handle-def.c: Separate locking of files for input and for
170 output, to allow both to take place at once. Also, distinguish a
171 file handle from the identity of the underlying file, because the
172 identity of a file changes over time and the file handle can't
173 represent two different identities.
174 (struct file_handle): Remove `next', `open_cnt', `deleted',
175 `type', `open_mode', `aux', `identity' members. Change `id' from
176 char array to char *. Add `ref_cnt' member.
177 (file_handle_from_ll) New function.
178 (file_handles) Removed.
179 (named_handles) New variable.
180 (fh_init) Initialize named_handles.
181 (fh_done) Remove name from all named_handles.
182 (free_handle) Updated for modified struct file_handle.
183 (unname_handle) New function.
184 (fh_ref) New function.
185 (fh_from_file_name) Removed.
186 (fh_unref) New function.
187 (fh_unname) New function.
188 (fh_from_id) Rewritten.
189 (create_handle) Updated for modified struct file_handle.
190 (fh_inline_file) Increment the handle's ref count.
191 (fh_create_file) Updated for modified struct file_handle.
192 (fh_create_scratch) Ditto.
197 (fh_is_open) Removed.
198 (fh_get_id) Updated for modified struct file_handle.
199 (fh_get_default_handle) Increment the handle's ref count.
200 (fh_set_default_handle) Handle ref counts.
201 (struct fh_lock) New structure.
202 (locks) New static var.
203 (fh_lock) New function.
204 (fh_unlock) New function.
205 (fh_lock_get_aux) New function.
206 (fh_lock_set_aux) New function.
207 (fh_is_locked) New function.
208 (make_key) New function.
209 (free_key) New function.
210 (compare_fh_locks) New function.
211 (hash_fh_lock) New function.
213 * file-handle-def.h (enum fh_access) New enum.
215 * file-name.c: Made file_identity the same in all supported
217 (struct file_identity): New `name' member.
218 (fn_get_identity): For a file that doesn't exist, get the
219 dev/inode of its directory plus its name. If even the directory
220 doesn't exist, just use its name. Merge the Windows
221 implementation into the Unix one.
222 (fn_compare_file_identities): Rewritten. Merge the Windows
223 implementation into the Unix one.
224 (fn_hash_identity): New function.
226 * make-file.c (struct replace_file): New structure.
227 (all_files): New static var.
228 (replace_file_start): New function.
229 (replace_file_commit): New function.
230 (replace_file_abort): New function.
231 (free_replace_file): New function.
232 (unlink_replace_files): New function.
234 * por-file-reader.c (struct pfm_reader): Add `lock' member.
235 (close_reader): Unlock file.
236 (pfm_open_reader): Lock file.
238 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
240 (pfm_open_writer): Lock file and prepare for its replacement.
241 (close_writer): Unlock file.
243 * scratch-handle.h (struct scratch_handle): Add unique_id so that
244 different generations of a scratch file can be distinguished.
246 * scratch-reader.c (scratch_reader_open): Verify that the file is
249 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
250 remove scratch_handle member.
251 (scratch_writer_open): Lock handle. Prepare to replace handle
252 data, instead of doing it immediately.
253 (scratch_writer_casewriter_destroy): Replace handle data and
256 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
257 (sfm_open_reader): Lock file.
258 (close_reader): Unlock file.
260 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
262 (sfm_open_writer): Lock file and prepare for its replacement.
263 (close_writer): Unlock file.
265 2007-11-02 Ben Pfaff <blp@gnu.org>
267 * data-out.c (output_number): Use isfinite (from C99) instead of
268 gsl_isfinite, since we now have universal support for it in
270 (output_infinite): Ditto.
271 [!HAVE_ROUND] (round): Remove definition, since we now have a
272 replacement in gnulib.
274 2007-10-27 John Darrington <john@darrington.wattle.id.au>
276 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
277 wasn't initialised correctly. Closes bugs #21330 and 21397
280 2007-10-08 Ben Pfaff <blp@gnu.org>
282 Fix bug #21280. Thanks to John Darrington for review.
284 * file-name.c (create_stream): New function.
286 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
287 file handle before creating the file, to ensure that we don't
288 truncate a file that we're reading. Make code easier to read by
291 * sys-file-write.c (sfm_open_writer): Ditto.
293 2007-10-01 Ben Pfaff <blp@gnu.org>
295 Fix bug #21192. Thanks to John Darrington for review.
297 * casereader.c (casereader_read): Decrement case_cnt before
298 calling the casereader's "read" member function, so that we
299 interact properly with lazy_casereader.
301 * datasheet.c: Add regression test for above bug fix.
302 (clone_datasheet): New function.
303 (lazy_callback): New function.
304 (check_datasheet_casereader): New function.
305 (check_datasheet): Check datasheet contents are reported correctly
306 through an ordinary casereader and a lazy casereader.
307 (clone_model): Use clone_datasheet.
309 2007-09-24 Ben Pfaff <blp@gnu.org>
311 Patch #6210. Reviewed by John Darrington.
312 * settings.c: Drop "long view", which is not needed any longer.
313 (static var long_view): Removed.
314 (force_long_view): Removed.
315 (get_viewwidth): Removed.
316 (init_viewport): Removed long_view reference.
318 2007-09-19 John Darrington <john@darrington.wattle.id.au>
320 * settings.c settings.h: Changed viewport's length and width to be
321 owned by the user interface which uses the data library. This allows
322 better abstraction, and makes dynamically adjustable dimensions easier.
324 2007-09-18 Ben Pfaff <blp@gnu.org>
326 * procedure.c (proc_extract_active_file_data): New function.
328 * lazy-casereader.h: New file.
330 * lazy-casereader.c: New file.
332 * casereader.c (casereader_dynamic_cast): New function.
334 2007-09-14 Ben Pfaff <blp@gnu.org>
336 * dictionary.c (dict_clone): Copy case indexes from cloned
337 dictionary. Fixes bug #21061. Reviewed by John Darrington.
339 2007-09-13 John Darrington <john@darrington.wattle.id.au>
341 * value.c value.h (create_value): New function.
343 2007-09-12 Ben Pfaff <blp@gnu.org>
345 Make it clear that translator casereader and casewriter translate
346 functions are supposed to destroy their input case.
348 * casereader-translator.c (struct casereader_translator): Change
349 input case parameter of translate member function from const
350 struct ccase * to struct ccase *.
351 (casereader_create_translator): Ditto, for translate parameter.
353 * casewriter-translator.c (struct casewriter_translator): Ditto.
354 (casewriter_create_translator): Ditto.
356 2007-08-27 John Darrington <john@darrington.wattle.id.au>
358 * sys-file-reader.c (read_display_parameters): Force display width
359 to 8, if the sys file says 0 (like SPSS does).
361 2007-08-12 Ben Pfaff <blp@gnu.org>
363 * dictionary.c (dict_dump): New function.
365 2007-08-12 Ben Pfaff <blp@gnu.org>
367 Drop dict_compactor in favor of using the new struct case_map.
369 * dictionary.c (struct copy_map): Removed.
370 (struct dict_compactor): Removed.
371 (dict_make_compactor): Removed.
372 (dict_compactor_compact): Removed.
373 (dict_compactor_destroy): Removed.
375 * procedure.c (struct data_set): Change `compactor' member to be a
377 (proc_open): Use case_map_to_compact_dict instead of
379 (proc_casereader_read): Use case_map_execute instead of
380 dict_compactor_compact.
381 (proc_commit): Use case_map_destroy instead of
382 dict_compactor_destroy.
384 * scratch-writer.c (struct scratch_writer): Change `compactor'
385 member to be a struct case_map *.
386 (scratch_writer_open): Use case_map_to_compact_dict instead of
388 (scratch_writer_casewriter_write): Use case_map_execute instead of
389 dict_compactor_compact.
391 2007-08-12 Ben Pfaff <blp@gnu.org>
393 * automake.mk: Add case-map.c, case-map.h.
395 * case-map.c: New file.
397 * case-map.h: New file.
399 2007-08-12 Ben Pfaff <blp@gnu.org>
401 * dictionary.c (dict_compact_values): Don't delete scratch
402 variables as well as compacting case indexes. Update all callers.
403 (dict_get_compacted_value_cnt): Rename dict_count_values and
404 change interface. Update all callers.
405 (dict_get_compacted_value_cnt): Remove.
406 (dict_compacting_would_shrink): Remove.
407 (dict_compacting_would_change): Remove.
408 (dict_make_compactor): Add new parameter. Update all callers.
410 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
411 instead of dict_count_values, changing an O(N) operation into
414 2007-08-12 Ben Pfaff <blp@gnu.org>
416 * casereader.c (casereader_read): Don't require cases read by a
417 casereader to be exactly the expected size: as long as they're big
420 2007-08-12 Ben Pfaff <blp@gnu.org>
422 Make casewriters keep track of the number of `union value's in
423 each case. This is useful for two reasons: casewriter_write can
424 then check that the case being written is large enough, and later
425 recipients of the casewriter can determine the size of the case.
427 * casewriter-translator.c (casewriter_create_translator): Add
430 * casewriter.c (struct casewriter): Add value_cnt member.
431 (casewriter_write): Check that the case passed in is big enough.
432 (casewriter_get_value_cnt): New function.
433 (casewriter_create): Add value_cnt parameter.
435 2007-08-09 Ben Pfaff <blp@gnu.org>
437 Fix bug reported by Jason Stover.
438 * settings.c: Move get_termcap_viewport and in particular the
439 #include for <curses.h> to the end of the file. curses.h
440 redefines bool on some systems (e.g. OpenBSD), which causes
441 disagreement between uses of bool before its inclusion and
443 Tested by Jason Stover.
445 2007-07-29 Ben Pfaff <blp@gnu.org>
447 Provisional fix for bug #18692 and bug #20161. Reviewed by John
450 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
451 argument (never "rb" or "wb") because SUSv3 says that only those
452 modes are defined, and glibc in fact rejects other modes.
454 Open portable files with fn_open so that they can be read from
455 pipes. Fix missing fh_close call to go along with fh_open.
456 Report an error if the file close reports an error.
457 * por-file-reader.c (close_reader): New function.
458 (por_file_casereader_destroy): Use close_reader.
459 (pfm_open_reader): Open file with fn_open.
461 2007-07-28 Ben Pfaff <blp@gnu.org>
463 Make PSPP able to read all the portable files I could find on the
464 web. Thanks to John Darrington for review. Bug #17620.
465 * por-file-reader.c (struct pfm_reader): New member `line_length'.
466 (error): Print file offset in hexadecimal.
467 (warning): New function.
468 (advance): Treat lines less than 80 bytes long as padded to 80
470 (pfm_open_reader): Call read_documents if we find an "E" record.
471 (convert_format): Convert invalid formats to the default format
472 instead of aborting reading the file.
473 (read_variables): Rename duplicate variable names instead of
474 aborting reading the file.
475 (read_value_label): Allow string variables of different widths to
476 be assigned value labels in the same record. Replace duplicate
477 value labels instead of aborting.
478 (read_documents): New function.
480 * por-file-writer.c (pfm_open_writer): Call write_documents if the
481 dictionary has documents.
482 (write_documents): New function.
484 2007-07-25 Ben Pfaff <blp@gnu.org>
486 Fix bugs related to bug #17213.
488 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
489 since the former is what config.h has. Include the needed ncurses
491 (static var echo) Rename to `do_echo' because the original name is
492 the same as an ncurses identifier.
493 (get_termcap_viewport) Use error instead of msg.
495 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
496 (fn_close): Don't close stdin, stdout, stderr.
498 2007-07-26 John Darrington <john@darrington.wattle.id.au>
500 * procedure.c procedure.h: Added callbacks which get invoked whenever
501 a dataset's transformation chain changes.
503 2007-07-24 Ben Pfaff <blp@gnu.org>
506 * sys-file-writer.c (write_variable_display_parameters): Use new
507 var_default_display_width function to choose display width of
508 segments after the first one in a given variable.
509 * variable.c (var_create): Use var_default_display_width to pick
510 new variable's display width.
511 (var_default_display_width): New function.
512 Reviewed by John Darrington.
514 2007-07-24 Ben Pfaff <blp@gnu.org>
517 * por-file-writer.c (write_variables): Write weight variable.
518 Reviewed by John Darrington.
520 2007-07-23 Ben Pfaff <blp@gnu.org>
522 Improvements to system file reader and writer.
524 First, move all detailed knowledge of very long strings into
525 sys-file-private.[ch], so that this nasty stuff can be isolated.
527 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
528 (EFFECTIVE_VLS_CHUNK): New macro.
529 (min_int): New function.
530 (max_int): New function.
531 (sfm_width_to_bytes): Rewrite.
532 (sfm_width_to_octs): New function.
533 (sfm_segment_alloc_width): New function.
534 (sfm_segment_alloc_bytes): New function.
535 (sfm_segment_used_bytes): New function.
536 (sfm_segment_offset): New function.
537 (sfm_segment_effective_offset): New function.
538 (sfm_dictionary_to_sfm_vars): New function.
540 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
541 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
542 (struct sfm_var): New structure.
544 Next, improvements to the system file reader.
546 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
547 type casenumber. Added `version_major', `version_minor',
550 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
551 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
552 Change `case_cnt' to type casenumber. Removed `has_vls'.
553 (struct sfm_var): Removed.
554 (sfm_open_reader): Don't warn on wrong case size if the file was
555 written by SPSS 13, which tends to get it wrong. Use
556 sfm_dictionary_to_sfm_vars.
557 (read_header): Always output system file info.
558 (read_variable_record): Simplify code for reading missing values.
559 (read_machine_int32_info): Save version numbers from system file
560 into info struct passed as new argument.
561 (read_long_string_map): Restructured to use new sys-file-private
563 (read_value_labels): Use size_overflow_p.
564 (sys_file_casereader_read): Get rid of distinction between fast
565 and slow paths. Use information provided by sys-file-primate's
566 struct sfm_var to simplify code.
567 (skip_whole_strings): New function.
568 (read_int32): Renamed read_int. Changed return value to int.
570 (read_flt64): Renamed read_float. Changed return value to
571 double. Updated all callers.
572 (int32_to_native): Removed. Changed callers to use
574 (flt64_to_double): Removed. Changed callers to use float_convert.
576 Finally, get rid of int32, flt64 terminology and types in system
577 file writer. The former wasn't very useful since a POSIX "int"
578 can hold the whole range of int32 and we generally didn't have a
579 need for it to be exactly-32-bits, just at-least-32-bits. The
580 latter was inconvenient because we had to assume that it could be
581 different from double and thereby convert special values SYSMIS,
582 HIGHEST, LOWEST to and from it in multiple places. Instead, now
583 we just use "int" and "double" in most places, and do conversions,
584 if necessary, very close to where we do I/O. This change meant
585 that the writer code couldn't represent records in the file as C
586 structs any longer, but that's no great loss. The code actually
587 seems to be more readable without them.
589 Simplify the compression buffering code: only buffer as much as
590 necessary, which is no more than eight 8-byte units at any given
593 * sys-file-writer.c (typedef flt64): Removed.
594 (macro second_lowest_flt64): Removed.
595 (struct sysfile_header): Removed.
596 (struct sysfile_variable): Removed.
597 (struct sfm_writer): Removed `needs_translation', `has_vls',
598 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
599 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
600 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
601 compression buffering by `opcodes', `opcode_cnt', `data',
602 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
603 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
604 data writer functions instead of structures.
605 (calc_oct_idx): New function.
606 (write_header): Use simple data writer functions instead of
608 (write_format_spec): Renamed write_format. New argument.
609 (write_variable_continuation_records): New function.
610 (write_variable): Use simple data writer functions instead of
611 structures. Use write_variable_continuation_records. Write
612 entire very long string instead of requiring caller to understand
614 (write_value_labels): Use simple data writer functions instead of
616 (write_documents): Ditto.
617 (write_variable_display_parameters): Use sys-file-private
618 functions to simplify. Use simple data writer functions instead
620 (write_vls_length_table): Use simple data writer functions instead
622 (write_longvar_table): Ditto.
623 (write_rec_7_34): Break into new functions
624 write_integer_info_record, write_float_info_record. Use simple
625 data writer functions instead of structures.
626 (buf_write): Removed.
627 (append_string_max): Removed.
628 (ensure_buf_space): Removed.
629 (sys_file_casewriter_write): Get rid of the distinction between
630 fast and slow paths, which didn't seem to be too useful. Use new
631 functions write_case_uncompressed, write_case_compressed.
632 (put_instruction): Removed.
633 (put_element): Removed.
634 (write_compressed_data): Removed.
635 (close_writer): Use flush_compressed. Only write case count to
636 system file if it will fit in the field.
637 (write_case_compressed): New function.
638 (write_case_uncompressed): New function.
639 (flush_compressed): New function.
640 (put_cmp_opcode): New function.
641 (put_cmp_number): New function.
642 (write_int): New function.
643 (convert_double_to_output_format): New function.
644 (write_float): New function.
645 (write_value): New function.
646 (write_string): New function.
647 (write_bytes): New function.
648 (write_zeros): New function.
649 (write_spaces): New function.
651 Reviewed by John Darrington.
653 2007-07-22 Ben Pfaff <blp@gnu.org>
655 Don't try to write very long strings to portable files. The
656 format does not support it.
658 * por-file-writer.c (MAX_POR_WIDTH): New macro.
659 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
660 (write_format): Add arg to take width to resize format to.
661 (write_value): Limit width of value written to MAX_POR_WIDTH.
662 (write_variables): Limit width of variable and its output formats
664 Reviewed by John Darrington.
666 2007-07-22 Ben Pfaff <blp@gnu.org>
668 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
669 local variable instead of literal 5.
670 Reviewed by John Darrington.
672 2007-07-22 Ben Pfaff <blp@gnu.org>
674 Fix problems with uniqueness of short names in system files with
675 very long string variables. Now a variable may have multiple
678 * automake.mk (src_data_libdata_a_SOURCES): Add new files
679 short-names.c, short-names.h.
681 * dictionary.c (dict_clone): Clone all the short names.
682 (compare_strings): Move into short-names.c.
683 (hash_strings): Ditto.
684 (set_var_short_name_suffix): Ditto.
685 (dict_assign_short_names): Ditto, rename short_names_assign,
686 change to assign all short names.
688 * por-file-writer.c (write_variables): Use short_names_assign
689 instead of dict_assign_short_names.
691 * short-names.c: New file.
693 * short-names.h: New file.
695 * sys-file-private.c (sfm_width_to_segments): New function.
697 * sys-file-reader.c (read_long_var_name_map): Save and restore all
698 the short names, not just the first one.
700 * sys-file-writer.c (cont_var_name): Removed.
701 (sfm_open_writer): Use short_names_assign instead of
702 dict_assign_short_names. Use unique short names assigned by
703 short_names_assign instead of those generated by cont_var_name.
705 * variable.c (struct variable): Remove `short_name' member,
706 replace by `short_names' and `short_name_cnt'.
707 (var_create) Initialize new members.
708 (var_get_short_name_cnt): New function.
709 (var_get_short_name): Now takes an index argument. Changed most
711 (var_set_short_name): Ditto.
712 (var_clear_short_name): Renamed var_clear_short_names, changed to
713 clear all short names.
715 Reviewed by John Darrington.
717 2007-07-22 Ben Pfaff <blp@gnu.org>
719 * variable.c (var_set_width): Use new var_set_width function.
721 * missing-values.c (mv_n_values): Drop assertion, which was not
724 * format.c (fmt_default_for_width): New function.
725 (fmt_resize): New function.
727 Reviewed by John Darrington.
729 2007-07-18 John Darrington <john@darrington.wattle.id.au>
731 * datasheet.c (datasheet_delete_columns): Added assertion to check
732 we're not deleting outside the range of the sheet.
735 * dictionary.c dictionary.h variable.c: Added the ability for string
736 variables to be resized.
738 * vardict.h: Added some prototypes (moved from dictionary.h) as
739 these should only be called by variable.c
742 2007-07-14 John Darrington <john@darrington.wattle.id.au>
744 * sfm-reader.c: Respect case_cnt field in file header.
746 2007-07-01 John Darrington <john@darrington.wattle.id.au>
748 * transformation.c transformation.h (trns_chain_execute): Changed the
749 signature (Patch #6057)
751 2007-06-10 Ben Pfaff <blp@gnu.org>
753 * casereader-filter.c (casereader_filter_destroy): Make sure to
754 write all the remaining excluded cases to the casewriter, if any.
756 * caseinit.c (init_list_destroy): Rewrite.
757 (init_list_clear): Ditto.
759 * casegrouper.c (casegrouper_get_next_group): Always set *reader
760 to null when returning false.
762 2007-06-06 Ben Pfaff <blp@gnu.org>
764 Actually implement the new procedure code and adapt all of its
765 clients to match. Also adapt all of the other case sources and
766 sinks in the tree and their clients to use the
767 casereader/casewriter infrastructure.
769 * automake.mk: Add and remove files.
771 * any-reader.c: Change into a casereader.
772 * por-file-reader.c: Ditto.
773 * scratch-reader.c: Ditto.
774 * sys-file-reader.c: Ditto.
776 * any-writer.c: Change into a casewriter.
777 * por-file-writer.c: Ditto.
778 * scratch-writer.c: Ditto.
779 * sys-file-writer.c: Ditto.
781 * procedure.c: Change to use casereader, casewriter, caseinit, and
782 other new infrastructure.
784 * scratch-handle.c: Adapt to new infrastructure.
786 * case-sink.c: Removed, now dead code.
787 * case-sink.h: Ditto.
788 * case-source.c: Ditto.
789 * case-source.h: Ditto.
790 * casefile-factory.c: Ditto.
791 * casefile-private.h: Ditto.
794 * casefilter.c: Ditto.
795 * casefilter.h: Ditto.
798 * fastfile-factory.c: Ditto.
799 * fastfile-factory.h: Ditto.
800 * storage-stream.c: Ditto.
801 * storage-stream.h: Ditto.
803 2007-06-06 Ben Pfaff <blp@gnu.org>
807 * automake.mk: Add new files.
809 * datasheet.c: New file.
811 * datasheet.h: New file.
813 2007-06-06 Ben Pfaff <blp@gnu.org>
815 Until now, the procedure code has provided a case to the
816 case_source, which has filled in the data values that come from
817 the active file. "Left" data values that don't come from the
818 active file naturally stay the same from case to case, because the
819 procedure code keeps using that same case.
821 One of the compromises that comes with the new procedure code is
822 that the active file allocates and provides its own case, which
823 the procedure code then has to resize to provide room for any
824 other variables that should go in the case and then fill in the
825 values of "left" variables. Then, when we're done with that case,
826 we have to save the values of "left" variables to copy into the
827 next case read from the active file.
829 The caseinit code helps with this.
831 * automake.mk: Add new files.
833 * caseinit.c: New file.
835 * caseinit.h: New file.
837 2007-06-06 Ben Pfaff <blp@gnu.org>
839 * value.h (value_cnt_from_width): New function.
841 * variable.c (var_get_value_cnt): Use new function.
843 2007-06-06 Ben Pfaff <blp@gnu.org>
845 Add casegrouper, to allow cases read from a given casereader to be
846 broken into groups, each of which has its own casereader.
847 Generally cases are grouped based on having equal values for some
850 * automake.mk: Add new files.
852 * casegrouper.c: New file.
854 * casegrouper.h: New file.
856 2007-06-06 Ben Pfaff <blp@gnu.org>
858 Add interface to lexicographical ordering of cases.
860 * automake.mk: Add new files.
862 * case-ordering.c: New file.
864 * case-ordering.h: New file.
866 2007-06-06 Ben Pfaff <blp@gnu.org>
868 Add casereaders and casewriters, the basis of the new data processing
869 implementation. A casereader is a uniform interface to reading cases
870 from a data source; a casewriter is a uniform interface to writing
871 cases to a data sink.
873 * automake.mk: Add new files.
875 * casereader-filter.c: New file.
877 * casereader-provider.h: New file.
879 * casereader-translator.c: New file.
881 * casereader.c: New file.
883 * casereader.h: New file.
885 * casewriter-provider.h: New file.
887 * casewriter-translator.c: New file.
889 * casewriter.c: New file.
891 * casewriter.h: New file.
893 2007-06-06 Ben Pfaff <blp@gnu.org>
895 "casewindow" data structure that extends the deque (from libpspp)
896 of cases with the ability to dump cases to disk if we get too many
899 * automake.mk: Add new files.
901 * casewindow.c: New file.
903 * casewindow.h: New file.
905 2007-06-06 Ben Pfaff <blp@gnu.org>
907 sparse_cases data structure that augments a sparse_array of cases
908 with the ability to dump cases to disk if we get too many cases in
911 * automake.mk: Add new files.
913 * sparse-cases.c: New file.
915 * sparse-cases.h: New file.
917 2007-06-06 Ben Pfaff <blp@gnu.org>
919 Adds a low-level on-disk case array data structure.
921 * automake.mk: Add new files.
923 * case-tmpfile.c: New file.
925 * case-tmpfile.h: New file.
927 2007-06-06 Ben Pfaff <blp@gnu.org>
929 In a couple of places we calculate the maximum number of cases to
930 keep in memory based on the user-defined workspace. Enable
931 centralizing the calculation through a new function.
933 * settings.c (get_workspace_cases): New function.
935 2007-06-06 Ben Pfaff <blp@gnu.org>
937 The casenumber type is defined in transformations.h, but case.h is
938 a more sensible place. Move it.
940 * case.h (CASENUMBER_MAX): New macro.
941 (typedef casenumber): Move here, from transformations.h.
943 2007-06-03 Ben Pfaff <blp@gnu.org>
945 Slightly generalize case_to_values and case_from_values functions.
947 * case.c (case_to_values): Rename case_copy_out, change interface.
948 (case_from_values): Rename case_copy_in, change interface.
950 * fastfile.c (fastfilereader_get_next_case): Update caller.
951 (write_case_to_disk): Ditto.
953 2007-06-02 Ben Pfaff <blp@gnu.org>
955 Clean up after a forgotten part of patch #5829.
957 * casedeque.h: Remove unused file.
959 * automake.mk: Remove casedeque.h from sources.
961 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
963 * category.c: Removed redundant #include
965 2007-05-06 Ben Pfaff <blp@gnu.org>
967 Abstract the documents within a dictionary a little better.
968 Thanks to John Darrington for suggestion, initial version, and
971 * dictionary.c (struct dictionary): Change `documents' member from
972 char * to struct string.
973 (dict_clear): Destroy struct string.
974 (dict_get_documents): Convert struct string to char *.
975 (dict_set_documents): Set struct string. Pad to 80-character
977 (dict_clear_documents): New function.
978 (dict_add_document_line): New function.
979 (dict_get_document_line_cnt): New function.
980 (dict_get_document_line): New function.
982 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
984 * sys-file-reader.c (read_documents): Use new document functions.
986 2007-04-19 John Darrington <john@darrington.wattle.id.au>
988 * sys-file-reader.c: When reading a system file which has no
989 long name table, automatically create one where the long names
990 are the lower case versions of the short names.
992 2007-04-22 Ben Pfaff <blp@gnu.org>
994 * dictionary.c (dict_set_split_vars): dict_destroy expects that
995 dict_clear will free most data related to the dictionary.
996 dict_clear does a decent job, except that dict_set_split_vars on
997 some systems won't actually free the dict's "split" member.
998 Instead, it'll allocate a 1-byte region. Fix this.
1000 * value.c (value_copy): New function.
1001 (value_set_missing): Ditto.
1003 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1005 * Deleted existing category.h and moved cat-routines.h into
1006 category.h Encapsulated struct cat_vals better.
1008 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1010 * sys-file-reader.c: When reading a system file which has no
1011 long name table, automatically create one where the long names
1012 are the lower case versions of the short names.
1014 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1016 * sys-file-reader.c: Some versions of Other Software seem to
1017 produce system files with string variables' measure set to
1018 zero. We'll assume these are supposed to be nominal variables.
1020 2007-03-30 Ben Pfaff <blp@gnu.org>
1022 * procedure.c: Adapt to new deque data structure.
1024 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1025 Ben Pfaff <blp@gnu.org>
1027 * file-name.c: Mingw compatibility fixes.
1028 (fn_search_path): Use ISSLASH instead of comparing against '/'
1030 (fn_dir_name): Use dir_name from gnulib.
1031 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1032 (fn_get_identity): Use GetFullPathName instead of canonicalize
1033 from gnulib, because the latter does not fully support
1034 Windows-style path names. Use this implementation based on the
1035 detected presence of Windows instead of the absence of Unix, since
1036 the new implementation is Windows-specific.
1037 (fn_compare_file_identities): In Windows implementation, compare
1038 names case-insensitively.
1040 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1042 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1043 <stdlib.h> to have the same effect.
1045 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1047 * por-file-reader.c: Add missing _() around messages.
1049 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1051 * make-file.c: Include "mkstemp.h", without which linking on
1054 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1056 Reduce platform dependence.
1058 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1059 Everywhere we using this, we really should have just depended on
1060 the shell to expand tildes.
1061 (fn_search_path): Simplify, given that we don't do tilde expansion
1063 (fn_normalize): Removed. Caller changed to use the canonicalize
1065 (fn_get_cwd): Removed. Only user was fn_normalize.
1066 (fn_is_absolute): Really only test for absolute names.
1067 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1069 (fn_readlink): Removed, as it was only used fn_normalize.
1070 (fn_exists): Assume the stat function is available; gnulib does.
1071 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1073 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1075 * dictionary.c (dict_create_vector_assert): New function.
1077 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1079 * file-name.c (fn_normalize): Correct name of function
1080 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1081 for pointing this out.
1083 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1085 We are using a single member in struct file_handle, the "name"
1086 field, for more than one purpose. When it begins with '"', it's a
1087 file name; otherwise, it's a token that can be used to identify
1088 it. When that assertion fires, it's because we searched for the
1089 name case-sensitively as a file name (so that there was no match),
1090 and then we try to insert it case-insensitively as a token, which
1091 fails because duplicates aren't allowed.
1093 Solution: break the two purposes into two separate fields. This
1094 fixes the problem and likely makes the code easier to read too.
1096 Fixes bug #18922. Thanks to John Darrington for bug report and
1099 * file-handle-def.c (struct file_handle): New `id' member.
1100 (fh_from_name): Rename fh_from_id. Update all callers.
1101 (create_handle): New `id' parameter. Update all callers.
1102 (fh_create_file): Ditto.
1103 (fh_get_id): New function.
1105 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1107 * case.c (case_is_null): Change return type to bool.
1109 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1113 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1114 reference counting to share data. This makes it easy for
1115 valgrind, etc. to find accesses to cases that have been destroyed
1116 but have been kept around by another user's ref-count. This often
1117 happens when the data set is small enough to find in memory; if a
1118 bigger data set that would overflow to disk were used, then data
1119 corruption would occur.
1121 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1125 * case.c (case_unshare): Make it check internally whether the
1126 ref_cnt is greater than 1, so that the callers don't have to.
1127 Update callers not to check.
1129 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1131 Before, I was thinking that I might want to get rid of reference
1132 counting at some point. Now, I'm pretty sure that it's here to
1133 stay. Thus, because we have to store the value_cnt anyway for
1134 reference-counted cases, we might as well expose it to users.
1136 * case.c (case_get_value_cnt): New function.
1137 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1138 resize case if its size actually changed.
1140 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1141 instead of peeking inside struct case directly.
1142 (casefile_append): Ditto.
1144 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1146 Get rid of the inlines for the case functions, which made the
1147 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1148 the inlines didn't speed up "make check" at all, which is not a
1149 perfect benchmark but seems indicative.)
1151 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1152 definitions. Remove some assertions on nonnull pointers that were
1153 redundant with a pointer dereference soon after in the function.
1155 (struct case_data): Move definition here from case.h.
1159 (case_data_wr): Ditto.
1161 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1163 * automake.mk: Add casedeque.h to sources.
1165 * casedeque.h: New file.
1167 * procedure.c: (struct dataset) Change lag_count, lag_head,
1168 lag_queue member into single struct casedeque member. Update all
1169 users to use the casedeque instead.
1172 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1174 * procedure.c: Simplify lagged cases interface. Updated all
1175 clients--well, the only client--to use the simplified interface.
1176 (dataset_n_lag) Removed.
1177 (dataset_set_n_lag) Removed.
1178 (dataset_need_lag) New function.
1180 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1182 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1183 whenever appropriate, but only when the dataset/dictionary is in a
1186 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1188 * dictionary.c dictionary.h : Added callbacks for change of filter and
1189 split variables. Refactored some code to ensure that callbacks get
1190 invoked when appropriate.
1192 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1193 replace_dict callback occurs when permanent_dict replaces the current
1196 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1198 * dictionary.c dictionary.h : Added callback for when the weight
1199 variable of a dictionary changes.
1201 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1203 * dictionary.c dictionary.h : Added replace_source and replace_dict
1204 callbacks, and functions to deal with them.
1206 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1208 Simplify missing value handling.
1210 * missing-values.h (enum mv_class): New type.
1211 (enum mv_type): Moved definition into missing-values.c and renamed
1212 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1213 enums. Updated all uses.
1214 (struct missing_values): Changed type of `type' from `enum
1215 mv_type' to `int' because the definition is no longer exposed.
1217 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1218 parameter. Update all callers.
1219 (mv_is_num_missing): Ditto.
1220 (mv_is_str_missing): Ditto.
1221 (mv_is_value_user_missing): Removed. Changed callers to use
1222 mv_is_value_missing.
1223 (mv_is_num_user_missing): Removed. Changed callers to use
1225 (mv_is_str_user_missing): Removed. Changed callers to use
1227 (mv_is_value_system_missing): Removed. Changed callers to use
1228 mv_is_value_missing.
1229 (mv_set_type): Removed. Changed callers to use mv_clear.
1230 (mv_clear): New function.
1232 * variable.c (var_is_value_missing): Add new enum mv_class
1233 parameter. Update all callers.
1234 (var_is_num_missing): Ditto.
1235 (var_is_str_missing): Ditto.
1236 (var_is_value_user_missing): Removed. Changed callers to use
1237 var_is_value_missing.
1238 (var_is_num_user_missing): Removed. Changed callers to use
1240 (var_is_str_user_missing): Removed. Changed callers to use
1242 (var_is_value_system_missing): Removed. Changed callers to use
1243 var_is_value_missing.
1245 * casefilter.c (struct casefilter): Use enum mv_class in place of
1247 (casefilter_variable_missing): Adapt to new member.
1248 (casefilter_create): Change signature to take enum mv_class,
1251 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1253 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1255 * case-sink.c case-sink.h procedure.c procedure.h
1256 storage-stream.c: Now uses the factory.
1258 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1260 Make it possible to pull cases from the active file with a
1261 function call, instead of requiring indirection through a callback
1264 * case-source.h (struct case_source_class): Change ->read function
1265 to return a single case, instead of calling a callback function
1266 for each case. Change ->destroy function to return an error
1269 * case-source.c (free_case_source): Pass along the value returned
1270 by the case_source ->destroy function.
1272 * procedure.c (struct write_case_data): Removed.
1273 (struct dataset): Added some members to track procedure state.
1274 (procedure): Optimize the trivial case at this level.
1275 (internal_procedure): Re-implement in terms of proc_open,
1276 proc_read, proc_close.
1277 (proc_open) New function.
1278 (proc_read) New function.
1279 (proc_close) New function.
1280 (write_case) Moved into proc_read.
1281 (close_active_file) Moved closing of data source into proc_close.
1283 * storage-source.c: Rewrote to conform with modified
1284 case_source_class interface.
1286 * transformations.c (trns_chain_execute): Added argument to allow
1287 starting execution from an arbitrary transformation. Updated
1290 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1292 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1294 * sys-file-reader.c (read_display_parameters): Don't assume that
1295 MEASURE_* and ALIGN_* have the same values found in system files.
1297 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1299 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1300 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1301 the value of n_MEASURES, which was off by 1 (at least from my
1304 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1306 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1307 callbacks which are invoked when the dictionary or its
1308 variables are changed.
1310 * missing-values.c missing-values.h value-labels.c: Tidied up
1311 consistency checks, and made some of them return false
1312 instead of assert-failing.
1314 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1316 * calendar.c (calendar_days_in_month): New function.
1318 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1320 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1321 the value label's width.
1323 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1325 * sfm-private.h: Move contents into sys-file-writer.c, which is
1326 the only remaining user. Removed Borland C++-specific directives.
1328 * sys-file-reader.c: Clean up and rewrite entire file. The
1329 rewritten version is simpler and better abstracted, and should be
1330 easier to maintain and extend. It avoids using structures to read
1331 file data, which is prone to padding variations among compilers.
1332 It should also handle non-IEEE 754 system files, although I
1333 haven't been able to find any. It has been tested against many
1334 .sav files obtained from the Web and found to produce the same
1335 results as the earlier version of the code, or in some cases
1336 improved results. It is more tolerant of format variations found
1339 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1340 member, putting an enum integer_format in its place. New member
1341 `float_format'. Changed `compressed' member to type bool.
1343 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1345 * dictionary.c (dict_delete_consecutive_vars): New function.
1347 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1349 * file-name.c (fn_search_path): Remove prefix arg that was unused
1350 by any caller. Updated all callers.
1352 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1354 * format.c (fmt_dollar_template): Use user's decimal point
1355 character. Add assertion.
1357 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1359 * format.c (fmt_dollar_template): New function, based on
1360 dollar_format_template from var-type-dialog.c.
1362 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1364 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1367 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1369 Finish converting struct variable to an opaque type. In this
1370 phase, we add remaining setter and getter functions, convert the
1371 remaining PSPP code to use them, and do a bunch of cleanup. The
1372 resulting changes are pervasive but mostly trivial, and only the
1373 notable changes are logged.
1375 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1378 * case.c (case_data): Renamed case_data_idx.
1379 (case_num): Renamed case_num_idx.
1380 (case_str): Renamed case_str_idx.
1381 (case_data_rw): Renamed case_data_rw_idx.
1383 * case.h (case_data): New function with old name and an interface
1384 that takes a variable instead of an index, which is easier to
1385 use. Updated all callers to use the new interface, or to use the
1386 new *_idx function (see above).
1389 (case_data_rw): Ditto.
1391 * category.c (cat_stored_values_destroy): Changed interface to
1392 take a struct cat_vals * instead of a struct variable *.
1394 * dictionary.c (dict_clone): Use new vector_clone function.
1395 (dict_clear) Use new var_destroy function.
1396 (add_var) New function.
1397 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1398 (dict_create_var_assert) Rewrite in terms of add_var.
1399 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1400 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1401 (dict_lookup_var) Use new var_create, var_destroy functions.
1402 (dict_contains_var) Rewrite in terms of new vardict functionality.
1403 (set_var_dict_index) New function.
1404 (set_var_case_index) New function.
1405 (reindex_vars) New function.
1406 (dict_delete_var) Rewrite in terms of new vardict functionality.
1407 (dict_reorder_var) Ditto.
1408 (dict_reorder_vars) Ditto.
1409 (rename_var) New function.
1410 (dict_rename_var) Use rename_var.
1411 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1412 (dict_get_compacted_idx_to_fv) Rename
1413 dict_get_compacted_dict_index_to_case_index, update callers.
1414 (dict_create_vector) Use new vector_create function.
1415 (dict_clear_vectors) Use new vector_destroy function.
1416 (set_var_short_name_suffix) Move here from variable.c, renamed
1417 from var_set_short_name_suffix, make static, update caller.
1419 * sys-file-private.c: New file.
1420 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1421 width_to_bytes, update callers.
1423 * sys-file-private.h: New file. Later it will supplant
1424 sfm-private.h; for now it supplements it.
1425 (macro MIN_VERY_LONG_STRING) New macro.
1426 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1428 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1429 MAX_LONG_STRING was used before.
1431 * sys-file-writer.c: Ditto.
1433 * value-labels.c: Change the paradigm here to be that a null
1434 pointer is OK for a struct val_labs * in most cases; it just
1435 represents an empty set of value labels.
1436 (val_labs_copy) A copy of a null set is a null set.
1437 (val_labs_count) A null set has 0 labels.
1438 (val_labs_replace) Change return type to void. Rewrite for
1440 (val_labs_find) A null set does not contain the value.
1441 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1442 transposed argument order, updated all callers.
1444 * value.c: New file.
1445 (value_dup) Moved here from variable.c.
1446 (compare_values) Ditto.
1449 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1450 (macro MAX_LONG_STRING) Removed, because it was only interesting
1451 for system files, not for general code.
1452 (macro MAX_VERY_LONG_STRING) Ditto.
1453 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1454 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1456 * vardict.h: New file, for an interface between variables and
1459 * variable.c: A lot of functions were moved around, for better
1461 (struct variable) Move definition here, from variable.h.
1462 (var_type_adj) Removed--makes i18n hard.
1463 (var_type_noun) Ditto.
1464 (var_create) New function.
1465 (var_clone) New function.
1466 (var_destroy) New function.
1467 (var_set_name) Assert that variable is not in a dictionary.
1468 (compare_var_names) Rename compare_vars_by_name and fix a couple
1469 of callers who thought the args were strings.
1470 (hash_var_name) Rename hash_var_by_name.
1471 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1472 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1473 (var_is_very_long_string) Removed, because it was only interesting
1474 to system file code.
1475 (var_set_missing_values) Allow the argument to be the wrong width,
1476 as long as we can resize it. Simplify callers who were doing the
1477 resizing themselves.
1478 (var_get_value_labels) New function.
1479 (var_has_value_labels) New function.
1480 (var_set_value_labels) New function.
1481 (alloc_value_labels) New function.
1482 (var_add_value_label) New function.
1483 (var_replace_value_label) New function.
1484 (var_clear_value_labels) New function.
1485 (var_lookup_value_label) New function.
1486 (var_get_value_name) Moved here from variable.c, renamed from
1487 var_get_value_name, transposed argument order, updated all
1489 (var_to_string) Moved here, from variable-label.c.
1490 (var_set_leave) New function.
1491 (var_get_leave) New function.
1492 (var_must_leave) New function.
1493 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1494 set_var_short_name_suffix.
1495 (var_get_dict_index) New function.
1496 (var_get_case_index) New function.
1497 (var_get_obs_vals) New function.
1498 (var_set_obs_vals) New function.
1499 (var_has_obs_vals) New function.
1500 (var_get_vardict) New function.
1501 (var_set_vardict) New function.
1502 (var_has_vardict) New function.
1503 (var_clear_vardict) New function.
1504 (value_dup) Moved to value.c.
1505 (compare_values) Ditto.
1508 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1509 (enum ALPHA) Rename VAR_STRING, update all users.
1511 * vector.c: New file.
1512 (struct vector) Moved here, from variable.h.
1513 (check_widths) New function.
1514 (vector_create) New function.
1515 (vector_clone) New function.
1516 (vector_destroy) New function.
1517 (vector_get_name) New function.
1518 (vector_get_var) New function.
1519 (vector_get_var_cnt) New function.
1520 (compare_vector_ptrs_by_name) New function.
1522 * vector.h: New file.
1524 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1526 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1527 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1530 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1532 * value-labels.c (destroy_atoms): New function.
1533 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1535 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1537 Thanks to Jason Stover for pointing out this problem.
1539 * data-out.c (output_number): Use gsl_finite from GSL, which is
1540 portable, instead of isfinite, which is not.
1543 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1545 * variable.c variable.h (value_dup): New function.
1547 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1549 Start converting struct variable to an opaque type. In this
1550 phase, we add a bunch of setter and getter functions and convert
1551 most of the PSPP code to use them. The resulting changes are
1552 pervasive but mostly trivial, and only the notable changes are
1555 * format.c (fmt_equal): New function.
1557 * variable.c (var_type_is_valid): New function.
1558 (measure_is_valid) Moved here, from format.c.
1559 (alignment_is_valid) Moved here, from format.c.
1560 (var_get_name) New function.
1561 (var_set_name) New function.
1562 (width_to_type) New function.
1563 (var_get_type) New function.
1564 (var_get_width) New function.
1565 (var_set_width) New function.
1566 (var_is_numeric) New function.
1567 (var_is_alpha) New function.
1568 (var_is_short_string) New function.
1569 (var_is_long_string) New function.
1570 (var_is_very_long_string) New function.
1571 (var_get_missing_values) New function.
1572 (var_set_missing_values) New function.
1573 (var_clear_missing_values) New function.
1574 (var_has_missing_values) New function.
1575 (var_is_value_missing) New function.
1576 (var_is_num_missing) New function.
1577 (var_is_str_missing) New function.
1578 (var_is_value_user_missing) New function.
1579 (var_is_num_user_missing) New function.
1580 (var_is_str_user_missing) New function.
1581 (var_is_value_system_missing) New function.
1582 (var_get_print_format) New function.
1583 (var_set_print_format) New function.
1584 (var_get_write_format) New function.
1585 (var_set_write_format) New function.
1586 (var_set_both_formats) New function.
1587 (var_get_label) New function.
1588 (var_set_label) New function.
1589 (var_clear_label) New function.
1590 (var_has_label) New function.
1591 (var_get_measure) New function.
1592 (var_set_measure) New function.
1593 (var_get_display_width) New function.
1594 (var_set_display_width) New function.
1595 (var_get_alignment) New function.
1596 (var_set_alignment) New function.
1597 (var_get_value_cnt) New function.
1598 (var_get_leave) New function.
1599 (var_get_short_name) New function.
1601 * variable.h: (struct variable) Removed "type" and "nv" members;
1602 they are now computed from "width" where needed.
1604 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1606 * missing-values.c (mv_resize): Don't write beyond end of the
1607 allocated buffer when resizing a long string.
1609 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1611 Clean up identifier code: don't require identifier enumerations to
1612 be in a particular order; make better use of string library;
1613 expose less of the internals.
1615 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1616 change interface. Updated all callers.
1617 (lex_id_match) Change interface to use struct substring, update
1619 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1620 (global array keywords[]) Make static, change form. Update all
1621 users to use lex_id_name instead.
1622 (lex_is_keyword) New function.
1623 (lex_id_to_token) Change interface to use struct substring, update
1625 (lex_id_name) New function.
1627 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1628 lex_is_keyword instead.
1629 (T_LAST_KEYWORD) Removed.
1630 (T_N_KEYWORDS) Removed.
1632 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1634 * format.c: (fmt_date_template) Distinguish characters for which a
1635 space is output and any date delimiter is allowed on input, from
1636 those for which a space is output and only a space is allowed on
1637 input. The former is represented by X, the latter by a space.
1638 Also, drop distinction between h and H, changing the former to the
1641 * data-in.c: Completely rewrite internals to conform to SPSS input
1642 formats as closely as possible.
1643 (data_in) Changed external interface by replacing the structure
1644 that was used as a single argument by a set of arguments. Updated
1646 (data_in_finite_line) Removed. Converted all callers to use plain
1648 (data_in_get_integer_format) New function.
1649 (data_in_set_integer_format) New function.
1650 (data_in_get_float_format) New function.
1651 (data_in_set_float_format) New function.
1653 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1654 (struct data_in) Removed.
1656 * data-out.c: (output_date) Drop each component from the input as
1657 it is output, to allow us to drop the distinction between h (a
1658 count of hours) and H (the hour of day) template characters.
1659 Also, handle new X template character.
1660 (output_scientific) Follow more rational rule on when to drop
1661 fraction introduced between SPSS 13 and 15. Updated test case to
1664 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1666 Fix buffer overflow reported by John Darrington.
1668 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1669 realize that DIGITS is a count of nibbles, not of bytes.
1671 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1673 * calendar.c (calendar_offset_to_gregorian) Also return the
1674 year-of-day. Change callers to new interface.
1676 * data-out.c: Completely rewrite internals to conform to SPSS
1677 output formats as completely as possible.
1678 (data_out) Change interface to put input parameters before output
1679 parameters, for consistency with the style I now prefer. Update
1681 (data_out_get_integer_format) New public function.
1682 (data_out_set_integer_format) New public function.
1683 (data_out_get_float_format) New public function.
1684 (data_out_set_float_format) New public function.
1686 * data-out.h: New file. Move prototype for data_out here, from
1689 * format.c: (fmt_step_width) Use equality comparison instead of
1690 bitwise and, for clarity.
1691 (fmt_is_string) Ditto.
1692 (fmt_input_to_output) Fix categories that are translated to F
1695 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1697 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1698 casefile-private.h: Added casefilter to assist commands with missing
1701 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1703 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1705 * settings.c: (route_errors_to_terminal) New variable.
1706 (route_errors_to_listing) New variable.
1707 (get_error_routing_to_terminal) New function.
1708 (set_error_routing_to_terminal) New function.
1709 (get_error_routing_to_listing) New function.
1710 (set_error_routing_to_listing) New function.
1712 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1714 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1716 * format.c: Completely rewrite, to achieve better abstraction.
1717 Rewrite all references to formats in other files.
1719 * format.def: Rewrite and reorganize.
1721 * settings.c: Move everything related to custom currency formats
1722 into format.[ch], changing them in form, so as to group related
1723 code and definitions better. Changed all references to use the
1725 (static var decimal) Removed.
1726 (static var grouping) Removed.
1727 (static var cc) Removed.
1728 (get_decimal) Removed.
1729 (set_decimal) Removed.
1730 (get_grouping) Removed.
1731 (set_grouping) Removed.
1735 * settings.h: (macro CC_CNT) Removed.
1736 (macro CC_WIDTH) Removed.
1737 (struct custom_currency) Removed.
1739 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1741 * data-in.c (data_in): Use switch statement instead of table, to
1742 avoid dependence on the order of the FMT_* enums.
1744 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1746 * data-out.c: (num_to_string) Removed, because it was dead code.
1748 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1750 * data-in.c (parse_trailer): Fix error message.
1752 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1754 * format.c (fmt_is_binary): New function.
1756 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1758 * procedure.c procedure.h: Encapsulated the static data into a single
1761 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1763 * casefile.c (casereader_read_xfer): Always initialize the case,
1764 even on an error condition.
1766 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1768 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1769 entire function useless.
1771 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1773 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1774 casereader method casereader_clone.
1776 * procedure.c pransformations.h: Introduced new type casenum_t
1778 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1780 * variable.c: (width_to_bytes) Rephrase code for clarify.
1782 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1784 * format.c: (fmt_type_from_string) New function.
1785 (fmt_to_string) Include decimals in output if the format has
1786 decimals, even if the format type does not. This way, we can
1787 accurately reproduce incorrect formats in user output.
1788 (check_common_specifier) Make the check for a bad format type an
1789 assertion, so we get bug reports if they show up. Fix message.
1790 Check for decimal places with a format type that doesn't allow
1792 (check_input_specifier) Remove check for FMT_X, which has been
1794 (check_output_specifier) Ditto.
1796 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1798 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1799 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1800 (enum fmt_parse_flags) Removed.
1802 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1804 * casefile.c casefile.h: Converted to an abstract base class.
1805 * casefile-private.h fastfile.c fastfile.h: New files.
1806 * automake.mk procedure.c scratch-writer.c storage-stream.c
1808 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1810 * procedure.c (internal_procedure): Create sink_case with only as
1811 many values as the compacted dictionary.
1813 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1815 Remove "debugging" code that caused plenty of false positives and
1818 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1820 * case.c: Remove all references to `this' member.
1822 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1824 Fix link error noted by Jason Stover.
1826 * storage-stream.c: Include <assert.h>.
1828 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1830 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1831 ALL) and additional underlying system file issues.
1833 Thanks to John Darrington for review.
1835 First problem: var_hash points to variables not owned by the
1836 sys-file-reader, which the caller may free or modify. Use an
1837 array of sfm_vars instead, as done earlier (e.g. CVS version
1840 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1841 members and remove all code that references it. Add vars, var_cnt
1842 members. Remove fix_specials member, which was unused.
1843 (struct sfm_var) Remove name member, which was unused.
1844 (sfm_close_reader) Free vars member instead of var_hash.
1845 (compare_var_shortnames) Removed.
1846 (hash_var_shortname) Removed.
1847 (sfm_open_reader) Fill out vars array.
1848 (compare_var_index) Removed.
1849 (sfm_read_case) Use vars instead of var_hash.
1851 Second problem: we're confused about when we actually have very
1852 long strings, causing us to choose incorrectly between slow path
1853 and fast path in sfm_read_case.
1855 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1856 have very long strings, not when we have long variable names,
1857 which is an unrelated feature.
1859 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1861 * variable.h: Move var_set and variable parsing declarations to
1862 new header, src/language/lexer/variable-parser.h. Modified lots
1863 of files to include the new header.
1865 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1867 * value-labels.c (value_to_string): When there's no value label,
1868 format the variable according to its print format, instead of
1869 always effectively using A or F format.
1871 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1873 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1876 * format.c: Removed tortological assertion.
1878 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1880 Reform string library.
1882 * file-name.c (fn_interp_vars): Change interface to take a
1883 substring as input. Updated all users.
1885 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1887 * format.c (measure_is_valid): Really return false when m >=
1890 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1892 Implement random access to casefiles, for use in GUI.
1894 * casefile.c: (struct casereader) Add `random', `file_ofs',
1895 `buffer_ofs' members.
1896 (casefile_get_random_reader) New function.
1897 (read_open_file) Break part into new function
1898 seek_and_fill_buffer().
1899 (fill_buffer) Update buffer_ofs, file_ofs.
1900 (casereader_seek) New function.
1902 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1904 * settings.c: Added call to i18n{done, init}.
1906 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1908 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1911 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1913 * casefile.c: Convert many uses of `int' to `bool'.
1915 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1917 * transformations.c (trns_chain_destroy): Destroy chain's trns
1918 member, to fix memory leak.
1920 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1922 * storage-stream.c (storage_source_decapsulate): Destroy case
1923 source to fix memory leak.
1925 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1927 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1928 case passed as argument instead of initializing the argument as a
1929 case. Fixes memory leak that showed up in
1930 tests/command/aggregate.sh with scratch files.
1932 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1934 * procedure.c (proc_done): Destroy default_dict, to fix memory
1937 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1939 Simplify procedure_with_splits().
1941 * procedure.c (struct split_aux_data): Removed case_count member.
1942 (procedure_with_splits) Don't initialize case_count.
1943 (split_procedure_case_func) Check whether prev_case is null
1944 instead of case_count.
1945 (split_procedure_end_func) Ditto.
1947 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1949 * case.c (case_move): Do nothing if dst and src are the same
1951 (case_try_create) Merge two similar cases.
1952 (case_copy) Unshare only if data must be actually copied.
1954 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1956 * data-in.c data-out.c dictionary.c sys-file-reader.c
1957 sys-file-writer.c variable.c variable.h: Reworked very long string
1958 support for better encapsulation.
1960 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1962 * value-labels.c (val_labs_can_set_width): New function.
1963 (val_labs_set_width) Clear labels if increasing width to long
1965 (val_labs_destroy) Remove unneeded test for null.
1967 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1969 * value-labels.h: Remove unneeded dependency on variable.h.
1971 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1973 Get rid of `char *c' member in union value, for cleanliness.
1975 * value.h: (union value) Remove `c' member.
1977 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1979 Make dictionary compacting functions a little more general.
1981 * sys-file-writer.c (sfm_open_writer): Use
1982 dict_compacting_would_change().
1983 (does_dict_need_translation) Removed.
1985 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1987 Make dictionary compacting functions a little more general.
1989 * dictionary.c (dict_needs_compaction): Rename
1990 dict_compacting_would_shrink(). Update all callers.
1991 (dict_compacting_would_change) New function.
1993 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1995 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1996 inverted return value (!).
1998 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2000 Continue reforming procedure execution.
2002 * procedure.c: Search and replace "vfm" by "proc". Notably:
2003 (static var vfm_source) Rename proc_source. Update all
2005 (static var vfm_sink) Rename proc_sink. Update all references.
2007 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2009 Continue reforming procedure execution. In this phase, remove
2010 PROCESS IF, which was deprecated anyway and can be easily
2011 simulated with TEMPORARY followed by SELECT IF.
2013 * procedure.c: (open_active_file) Don't call
2014 add_process_if_trns().
2015 (discard_variables) Get rid of redundant call to
2016 proc_cancel_all_transformations().
2017 (add_process_if_trns) Removed.
2018 (process_if_trns_proc) Removed.
2019 (process_if_trns_free) Removed.
2021 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2023 Continue reforming procedure execution. In this phase, add
2024 `const' to the case passed to procedure()'s callback.
2026 Updated all users of procedure() as well.
2028 * procedure.c: (struct write_case_data) Add "const" to ccase
2029 parameter for case_func member.
2030 (procedure) Add "const" to ccase parameter for proc_func
2032 (multipass_case_func) Make ccase parameter const.
2033 (internal_procedure) Add "const" to ccase parameter for case_func
2035 (split_procedure_case_func) Make ccase parameter const.
2036 (multipass_split_case_func) Make ccase parameter const.
2038 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2040 Continue reforming procedure execution. In this phase, get rid of
2041 the output code for SPLIT FILE groups in procedure.c, which really
2042 shouldn't be doing any output. Move it into the individual
2043 procedures instead. This also adds some flexibility.
2045 Updated many users of procedure_with_splits() and
2046 multipass_procedure_with_splits() to call
2047 output_split_file_values() and to deal with increased use of
2050 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2051 parameter to begin_func member.
2052 (procedure_with_splits) Add "const struct ccase *" parameter to
2053 begin_func parameter. Make ccase parameter const in proc_func
2055 (split_procedure_case_func) Don't dump split file group. Pass
2057 (dump_splits) Moved to language/dictionary/split-file.c as
2058 output_split_file_values().
2059 (struct multipass_split_aux_data) Add "const struct ccase *"
2060 parameter to split_func member.
2061 (multipass_procedure_with_splits) Add "const struct ccase *"
2062 parameter to split_func parameter.
2063 (multipass_split_case_func) Save new SPLIT FILE case before
2065 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2067 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2069 Continue reforming procedure execution. Change
2070 internal_procedure() so that it calls open_active_file() and
2071 close_active_file(), which isolates most of the actual procedure
2074 * procedure.c: (struct write_case_data) Rename `proc_func' member
2075 to `case_func' and update all references.
2076 (procedure) Rewrite as one-line wrapper around
2077 internal_procedure().
2078 (struct multipass_aux_data) New.
2079 (multipass_callback) Renamed multipass_case_func(). Use struct
2080 multipass_aux_data as auxiliary data.
2081 (multipass_end_func) New function.
2082 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2083 that uses multipass_case_func, multipass_end_func.
2084 (internal_procedure) Add `end_func' argument. Move optimization
2085 of trivial case in here. Move call to open_active_file() and
2086 close_active_file() in here. Now assert that vfm_source is
2088 (procedure_with_splits_callback) Rename
2089 split_procedure_case_func().
2090 (split_procedure_end_func) New function.
2091 (multipass_split_callback) Rename multipass_split_case_func.
2092 (multipass_split_end_func) New function.
2093 (discard_variables) No need to test for nonnull vfm_source.
2095 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2097 Continue reforming procedure execution. Get rid of unused member.
2099 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2101 (write_case) Don't increment cases_analyzed.
2103 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2105 Continue reforming procedure execution. In this phase, move
2106 procedure.c and procedure.h from src to src/data. Update
2107 makefiles and #includes accordingly.
2109 * procedure.c: Moved here from src/.
2111 * procedure.h: Moved here from src/.
2113 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2115 Continue reforming procedure execution. In this phase, get rid of
2116 many global variables, consolidating procedure execution in
2117 procedure.c. Encapsulate transformations in new "struct
2118 trns_chain". Also, change implementation of N OF CASES, FILTER,
2119 and PROCESS IF from special cases to transformations.
2121 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2124 * dictionary.c: (global variable default_dict) Move to
2127 * variable.h: (TRNS_*) Move to transformations.h.
2128 (struct transformation) Move to transformations.c.
2130 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2132 * sys-file-reader.c: Fixed invalid read problems.
2134 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2136 * storage-stream.c: Add missing function comments.
2138 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2140 Continue reforming procedure execution. In this phase, add some
2141 new, needed functionality to storage-stream.
2143 * storage-stream.c: (storage_source_decapsulate) New function.
2145 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2147 * variable.c (width_to_bytes): Declarations must precede
2148 statements for C90 compliance.
2150 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2152 * data-out.c, data-in.c, variable.c, variable.h: New functions
2153 copy_mangle and copy_demangle for reading/writing cases; emulates the
2154 way SPSS deals with strings > 255 bytes.
2156 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2157 subtype 14 needed for strings longer than 255 bytes.
2159 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2160 instead of literal values. Also fixed some constness issues.
2162 * format.h: Constness
2164 * sfm-private.h: Renamed the case_size identifier, since I discovered
2165 that SPSS's respect for this variable is very nominal.
2167 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2169 Change case limit type from int to size_t.
2171 * dictionary.c: (struct dictionary) Change type of case_limit
2173 (dict_get_case_limit) Change return type.
2174 (dict_set_case_limit) Change parameter type.
2176 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2178 * variable.h: (struct variable) Rename `reinit' member as `leave'
2179 and invert sense. Fix up all references.
2181 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2183 Continue reforming procedure execution. In this phase, break
2184 procedure.c into multiple files.
2186 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2188 * case-sink.c: New file.
2190 * case-sink.h: New file.
2192 * case-source.c: New file.
2194 * case-source.h: New file.
2196 * storage-stream.c: New file.
2198 * storage-stream.h: New file.
2200 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2202 * variable.h: (struct variable) Remove `init' member and all
2203 references to it from other files. It was initialized in several
2204 places, but nothing really ever used it for anything worthwhile.
2205 Thanks to Jason Stover for pointing out how confusing this
2208 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2210 Continue reforming error message support. In this phase, get rid
2211 of message "titles" and put the message text in `struct error'.
2212 Now `struct error' encapsulates a message more properly.
2214 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2215 Format message ourselves.
2217 * data-in.c: (vdls_error) Ditto.
2219 * por-file-reader.c: (error) Ditto.
2221 * sys-file-reader.c: (corrupt_msg) Ditto.
2223 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2225 GNU standards require "file name" instead of "filename" in
2226 documentation. It's nice for our code to follow the convention
2229 * casefile.c: (struct casefile) Rename `filename' member to
2230 `file_name'. Updated all references.
2232 * file-name.c: [!unix] (struct file_identity) Rename
2233 normalized_filename member to normalized_file_name. Updated all
2236 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2238 We don't really support anything but Unix-like environments well,
2239 so we might as well de-obfuscate by writing directory and path
2240 separators explicitly.
2242 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2244 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2245 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2247 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2250 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2252 GNU standards require "file name" instead of "filename" in
2253 documentation. It's nice for our code to follow the convention
2256 * filename.c: Rename to file-name.c.
2258 * filename.h: Rename to file-name.h. Update all inclusions.
2259 Update header guards.
2261 * automake.mk: Update file names.
2263 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2265 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2267 (fn_basename) Removed (dead code).
2268 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2269 (fn_special_p) Renamed fn_is_special(), all references updated.
2270 (fn_exists_p) Renamed fn_exists(), all references updated.
2272 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2274 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2275 Also, now it only tilde-expands file names, not paths.
2276 (fn_search_path) Tilde-expand one directory at a time.
2278 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2280 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2281 get rid of non-Unixlike version of the code, which has probably
2283 (fn_prepend_dir) Removed (dead code).
2285 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2286 (macro PATH_SEPARATOR_STRING) New.
2287 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2289 Continue reforming error message support. In this phase, we get
2290 rid of VM() and the other msg() support for "verbosity", replacing
2291 it by a new function verbose_msg().
2293 * filename.c: (fn_search_path) Use verbose_msg() instead of
2296 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2298 * sfm-private.h: Get rid of #defines after #error, which makes no
2301 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2303 Get rid of our own int32 type in favor of the standard int32_t
2306 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2309 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2311 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2313 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2315 Remove ill-considered file routines that are no longer used.
2317 * filename.c: (fn_open_ext) Removed.
2318 (fn_close_ext) Removed.
2320 * filename.h: (struct file_ext) Removed.
2322 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2324 * variable.c (var_is_valid_name): Move declarations before code
2327 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2329 * filename.ch (fn_interp_vars): Fixed small buglet.
2331 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2333 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2334 so as to modify the string inline. Thus makeing it easier to
2335 destroy the results when no longer needed.
2337 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2339 * category.c (cat_stored_values_destroy): Fixed memory leak.
2341 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2343 Add some missing frees. Thanks to John Darrington for reporting
2346 * any-writer.c (any_writer_close): Free writer.
2348 * any-reader.c (any_reader_close): Free reader.
2350 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2352 * por-file-reader.c: (error) Mark as NO_RETURN.
2354 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2356 * settings.c: Changed default value of scompress to true.
2358 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2360 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2362 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2365 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2366 here from pref.h.orig.
2368 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2370 * sys-file-reader.c: Fixed bug reading compressed files.
2372 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2374 * Numerous renames. See src/ChangeLog for details.
2376 * Moved files from src directory