1 2007-12-06 Ben Pfaff <blp@gnu.org>
5 * sys-file-reader.c (read_display_parameters): Handle variable
6 display parameters record with only 2 data items per variable.
7 Reported by Guido Gay <gay@irer.it>.
9 2007-12-04 Ben Pfaff <blp@gnu.org>
11 * identifier.c (lex_id_match_n): New function.
12 (lex_id_match): Reimplement in terms of lex_id_match_n.
14 2007-11-24 Ben Pfaff <blp@gnu.org>
16 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
19 2007-11-24 Ben Pfaff <blp@gnu.org>
21 Fix warning reported when reading back system files that include
22 very long string variables. Thanks to Guido Gay <gay@irer.it> for
25 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
26 was written where `j' was meant.
28 2007-11-18 Ben Pfaff <blp@gnu.org>
30 Properly write variables that include a range of missing values to
31 system files. Thanks to Guido Gay <gay@irer.it> for reporting
34 * sys-file-writer.c (write_variable): Correctly calculate
35 n_missing_values field when writing variables that include a range
38 2007-11-10 Ben Pfaff <blp@gnu.org>
40 Cleanups and bug fixes devised while writing up documentation.
43 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
45 * dict-class.c: New file.
46 (dict_class_from_id): Move here.
47 (dict_class_to_name): Move here.
49 * dict-class.h: New file.
50 (enum dict_class): Move here. Change from consecutive integers to
51 consecutive bits, to make testing for any of multiple values
52 easier. Add new DC_ALL constant.
54 * dictionary.c (struct dictionary): Change `case_limit' from
56 (dict_get_vars): Make final argument an enum dict_class.
57 (dict_get_vars_mutable): Ditto.
58 (dict_get_case_limit): Change return value to casenumber.
59 (dict_set_case_limit): Change final argument to a casenumber.
60 (dict_unset_split_var): Add assertion. Rephrase slightly.
61 (dict_set_label): Use xstrndup to simplify.
63 * format.c (fmt_step_width): AHEX format also needs 2-byte
65 (fmt_set_style): Simplify assertions.
67 * missing-values.c (mv_add_num_range): Rename mv_add_range.
68 Simplify implementation.
69 (mv_has_value): Simplify implementation.
70 (mv_pop_value): Remove the first value, not the last, to avoid
71 having GET followed by SAVE reverse the order of missing values.
72 (mv_peek_value): Rename mv_get_value. Simplify assertion.
73 (mv_has_range): Simplify implementation.
74 (mv_peek_range): Rename mv_get_range.
75 (can_resize_string): Removed.
76 (mv_is_resizable): Use value_is_resizable.
77 (mv_resize): Use value_resize.
79 * short-names.h (SHORT_NAME_LEN): Move here.
81 * val-type.h: New file, for definitions related to type and width
82 of abstract values. Before, these definitions were mixed among
83 those related to "union value" and those related to variables.
84 (macro SYSMIS): Move here.
85 (macro LOWEST): Move here.
86 (macro HIGHEST): Move here.
87 (macro MAX_STRING); Move here.
88 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
89 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
90 All references updated.
91 (val_type_is_valid): New function. Replaces var_type_is_valid.
92 All references updated.
93 (val_type_from_width): New function. Replaces
94 var_type_from_width. All references updated.
96 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
98 (val_labs_can_set_width): Use value_is_resizable.
99 (val_labs_add): Simply return false if the value labels set is too
100 wide, instead of having undefined behavior.
101 (val_labs_replace): Ditto.
102 (val_labs_replace): Ditto.
103 (val_labs_first): Set iterator to null if iteration is complete.
104 (val_labs_first_sorted): Ditto.
105 (val_labs_done): Become a no-op if the iterator is null.
107 * value.c (value_is_resizable): New function.
108 (value_resize): New function.
110 * variable.c (var_get_dict_class): New function.
112 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
115 2007-11-08 Ben Pfaff <blp@gnu.org>
117 * data-in.c: Make formatted data parsing locale-independent.
118 (parse_number): Use c_strtod instead of strtod, to avoid
119 locale-specific behavior.
122 2007-11-06 Ben Pfaff <blp@gnu.org>
124 Patch #6256: add support for binary, 360 file formats. Reviewed
127 * data-in.c (struct data_in): Add `encoding' member.
128 (data_in): Add `encoding' parameter, and re-encode the data passed
129 in where appropriate. Update all callers to pass it in.
130 (parse_A): Implement EBCDIC recoding wart described in manual.
131 (parse_AHEX): Implement EBCDIC recoding.
133 * data-out.c (data_out_legacy): New function.
134 (data_out): Make into a wrapper around data_out_legacy.
136 * file-handle-def.c (struct file_handle): New member `encoding'.
137 (fh_create_file): Set encoding.
138 (fh_default_properties): Set default encoding.
139 (fh_get_legacy_encoding): New function.
141 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
142 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
143 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
144 (struct fh_properties): New member `encoding'.
146 2007-11-05 Ben Pfaff <blp@gnu.org>
148 Patch #6258. Reviewed by John Darrington.
150 * file-handle-def.c (fh_lock): Add comment that TYPE should be
151 marked with N_() in the caller. Added these markings to each
152 caller too. Should make i18n easier.
153 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
155 2007-11-03 Ben Pfaff <blp@gnu.org>
157 Allow output files to overwrite input files (bug #21280).
159 * file-handle-def.c: Separate locking of files for input and for
160 output, to allow both to take place at once. Also, distinguish a
161 file handle from the identity of the underlying file, because the
162 identity of a file changes over time and the file handle can't
163 represent two different identities.
164 (struct file_handle): Remove `next', `open_cnt', `deleted',
165 `type', `open_mode', `aux', `identity' members. Change `id' from
166 char array to char *. Add `ref_cnt' member.
167 (file_handle_from_ll) New function.
168 (file_handles) Removed.
169 (named_handles) New variable.
170 (fh_init) Initialize named_handles.
171 (fh_done) Remove name from all named_handles.
172 (free_handle) Updated for modified struct file_handle.
173 (unname_handle) New function.
174 (fh_ref) New function.
175 (fh_from_file_name) Removed.
176 (fh_unref) New function.
177 (fh_unname) New function.
178 (fh_from_id) Rewritten.
179 (create_handle) Updated for modified struct file_handle.
180 (fh_inline_file) Increment the handle's ref count.
181 (fh_create_file) Updated for modified struct file_handle.
182 (fh_create_scratch) Ditto.
187 (fh_is_open) Removed.
188 (fh_get_id) Updated for modified struct file_handle.
189 (fh_get_default_handle) Increment the handle's ref count.
190 (fh_set_default_handle) Handle ref counts.
191 (struct fh_lock) New structure.
192 (locks) New static var.
193 (fh_lock) New function.
194 (fh_unlock) New function.
195 (fh_lock_get_aux) New function.
196 (fh_lock_set_aux) New function.
197 (fh_is_locked) New function.
198 (make_key) New function.
199 (free_key) New function.
200 (compare_fh_locks) New function.
201 (hash_fh_lock) New function.
203 * file-handle-def.h (enum fh_access) New enum.
205 * file-name.c: Made file_identity the same in all supported
207 (struct file_identity): New `name' member.
208 (fn_get_identity): For a file that doesn't exist, get the
209 dev/inode of its directory plus its name. If even the directory
210 doesn't exist, just use its name. Merge the Windows
211 implementation into the Unix one.
212 (fn_compare_file_identities): Rewritten. Merge the Windows
213 implementation into the Unix one.
214 (fn_hash_identity): New function.
216 * make-file.c (struct replace_file): New structure.
217 (all_files): New static var.
218 (replace_file_start): New function.
219 (replace_file_commit): New function.
220 (replace_file_abort): New function.
221 (free_replace_file): New function.
222 (unlink_replace_files): New function.
224 * por-file-reader.c (struct pfm_reader): Add `lock' member.
225 (close_reader): Unlock file.
226 (pfm_open_reader): Lock file.
228 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
230 (pfm_open_writer): Lock file and prepare for its replacement.
231 (close_writer): Unlock file.
233 * scratch-handle.h (struct scratch_handle): Add unique_id so that
234 different generations of a scratch file can be distinguished.
236 * scratch-reader.c (scratch_reader_open): Verify that the file is
239 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
240 remove scratch_handle member.
241 (scratch_writer_open): Lock handle. Prepare to replace handle
242 data, instead of doing it immediately.
243 (scratch_writer_casewriter_destroy): Replace handle data and
246 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
247 (sfm_open_reader): Lock file.
248 (close_reader): Unlock file.
250 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
252 (sfm_open_writer): Lock file and prepare for its replacement.
253 (close_writer): Unlock file.
255 2007-11-02 Ben Pfaff <blp@gnu.org>
257 * data-out.c (output_number): Use isfinite (from C99) instead of
258 gsl_isfinite, since we now have universal support for it in
260 (output_infinite): Ditto.
261 [!HAVE_ROUND] (round): Remove definition, since we now have a
262 replacement in gnulib.
264 2007-10-27 John Darrington <john@darrington.wattle.id.au>
266 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
267 wasn't initialised correctly. Closes bugs #21330 and 21397
270 2007-10-08 Ben Pfaff <blp@gnu.org>
272 Fix bug #21280. Thanks to John Darrington for review.
274 * file-name.c (create_stream): New function.
276 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
277 file handle before creating the file, to ensure that we don't
278 truncate a file that we're reading. Make code easier to read by
281 * sys-file-write.c (sfm_open_writer): Ditto.
283 2007-10-01 Ben Pfaff <blp@gnu.org>
285 Fix bug #21192. Thanks to John Darrington for review.
287 * casereader.c (casereader_read): Decrement case_cnt before
288 calling the casereader's "read" member function, so that we
289 interact properly with lazy_casereader.
291 * datasheet.c: Add regression test for above bug fix.
292 (clone_datasheet): New function.
293 (lazy_callback): New function.
294 (check_datasheet_casereader): New function.
295 (check_datasheet): Check datasheet contents are reported correctly
296 through an ordinary casereader and a lazy casereader.
297 (clone_model): Use clone_datasheet.
299 2007-09-24 Ben Pfaff <blp@gnu.org>
301 Patch #6210. Reviewed by John Darrington.
302 * settings.c: Drop "long view", which is not needed any longer.
303 (static var long_view): Removed.
304 (force_long_view): Removed.
305 (get_viewwidth): Removed.
306 (init_viewport): Removed long_view reference.
308 2007-09-19 John Darrington <john@darrington.wattle.id.au>
310 * settings.c settings.h: Changed viewport's length and width to be
311 owned by the user interface which uses the data library. This allows
312 better abstraction, and makes dynamically adjustable dimensions easier.
314 2007-09-18 Ben Pfaff <blp@gnu.org>
316 * procedure.c (proc_extract_active_file_data): New function.
318 * lazy-casereader.h: New file.
320 * lazy-casereader.c: New file.
322 * casereader.c (casereader_dynamic_cast): New function.
324 2007-09-14 Ben Pfaff <blp@gnu.org>
326 * dictionary.c (dict_clone): Copy case indexes from cloned
327 dictionary. Fixes bug #21061. Reviewed by John Darrington.
329 2007-09-13 John Darrington <john@darrington.wattle.id.au>
331 * value.c value.h (create_value): New function.
333 2007-09-12 Ben Pfaff <blp@gnu.org>
335 Make it clear that translator casereader and casewriter translate
336 functions are supposed to destroy their input case.
338 * casereader-translator.c (struct casereader_translator): Change
339 input case parameter of translate member function from const
340 struct ccase * to struct ccase *.
341 (casereader_create_translator): Ditto, for translate parameter.
343 * casewriter-translator.c (struct casewriter_translator): Ditto.
344 (casewriter_create_translator): Ditto.
346 2007-08-27 John Darrington <john@darrington.wattle.id.au>
348 * sys-file-reader.c (read_display_parameters): Force display width
349 to 8, if the sys file says 0 (like SPSS does).
351 2007-08-12 Ben Pfaff <blp@gnu.org>
353 * dictionary.c (dict_dump): New function.
355 2007-08-12 Ben Pfaff <blp@gnu.org>
357 Drop dict_compactor in favor of using the new struct case_map.
359 * dictionary.c (struct copy_map): Removed.
360 (struct dict_compactor): Removed.
361 (dict_make_compactor): Removed.
362 (dict_compactor_compact): Removed.
363 (dict_compactor_destroy): Removed.
365 * procedure.c (struct data_set): Change `compactor' member to be a
367 (proc_open): Use case_map_to_compact_dict instead of
369 (proc_casereader_read): Use case_map_execute instead of
370 dict_compactor_compact.
371 (proc_commit): Use case_map_destroy instead of
372 dict_compactor_destroy.
374 * scratch-writer.c (struct scratch_writer): Change `compactor'
375 member to be a struct case_map *.
376 (scratch_writer_open): Use case_map_to_compact_dict instead of
378 (scratch_writer_casewriter_write): Use case_map_execute instead of
379 dict_compactor_compact.
381 2007-08-12 Ben Pfaff <blp@gnu.org>
383 * automake.mk: Add case-map.c, case-map.h.
385 * case-map.c: New file.
387 * case-map.h: New file.
389 2007-08-12 Ben Pfaff <blp@gnu.org>
391 * dictionary.c (dict_compact_values): Don't delete scratch
392 variables as well as compacting case indexes. Update all callers.
393 (dict_get_compacted_value_cnt): Rename dict_count_values and
394 change interface. Update all callers.
395 (dict_get_compacted_value_cnt): Remove.
396 (dict_compacting_would_shrink): Remove.
397 (dict_compacting_would_change): Remove.
398 (dict_make_compactor): Add new parameter. Update all callers.
400 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
401 instead of dict_count_values, changing an O(N) operation into
404 2007-08-12 Ben Pfaff <blp@gnu.org>
406 * casereader.c (casereader_read): Don't require cases read by a
407 casereader to be exactly the expected size: as long as they're big
410 2007-08-12 Ben Pfaff <blp@gnu.org>
412 Make casewriters keep track of the number of `union value's in
413 each case. This is useful for two reasons: casewriter_write can
414 then check that the case being written is large enough, and later
415 recipients of the casewriter can determine the size of the case.
417 * casewriter-translator.c (casewriter_create_translator): Add
420 * casewriter.c (struct casewriter): Add value_cnt member.
421 (casewriter_write): Check that the case passed in is big enough.
422 (casewriter_get_value_cnt): New function.
423 (casewriter_create): Add value_cnt parameter.
425 2007-08-09 Ben Pfaff <blp@gnu.org>
427 Fix bug reported by Jason Stover.
428 * settings.c: Move get_termcap_viewport and in particular the
429 #include for <curses.h> to the end of the file. curses.h
430 redefines bool on some systems (e.g. OpenBSD), which causes
431 disagreement between uses of bool before its inclusion and
433 Tested by Jason Stover.
435 2007-07-29 Ben Pfaff <blp@gnu.org>
437 Provisional fix for bug #18692 and bug #20161. Reviewed by John
440 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
441 argument (never "rb" or "wb") because SUSv3 says that only those
442 modes are defined, and glibc in fact rejects other modes.
444 Open portable files with fn_open so that they can be read from
445 pipes. Fix missing fh_close call to go along with fh_open.
446 Report an error if the file close reports an error.
447 * por-file-reader.c (close_reader): New function.
448 (por_file_casereader_destroy): Use close_reader.
449 (pfm_open_reader): Open file with fn_open.
451 2007-07-28 Ben Pfaff <blp@gnu.org>
453 Make PSPP able to read all the portable files I could find on the
454 web. Thanks to John Darrington for review. Bug #17620.
455 * por-file-reader.c (struct pfm_reader): New member `line_length'.
456 (error): Print file offset in hexadecimal.
457 (warning): New function.
458 (advance): Treat lines less than 80 bytes long as padded to 80
460 (pfm_open_reader): Call read_documents if we find an "E" record.
461 (convert_format): Convert invalid formats to the default format
462 instead of aborting reading the file.
463 (read_variables): Rename duplicate variable names instead of
464 aborting reading the file.
465 (read_value_label): Allow string variables of different widths to
466 be assigned value labels in the same record. Replace duplicate
467 value labels instead of aborting.
468 (read_documents): New function.
470 * por-file-writer.c (pfm_open_writer): Call write_documents if the
471 dictionary has documents.
472 (write_documents): New function.
474 2007-07-25 Ben Pfaff <blp@gnu.org>
476 Fix bugs related to bug #17213.
478 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
479 since the former is what config.h has. Include the needed ncurses
481 (static var echo) Rename to `do_echo' because the original name is
482 the same as an ncurses identifier.
483 (get_termcap_viewport) Use error instead of msg.
485 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
486 (fn_close): Don't close stdin, stdout, stderr.
488 2007-07-26 John Darrington <john@darrington.wattle.id.au>
490 * procedure.c procedure.h: Added callbacks which get invoked whenever
491 a dataset's transformation chain changes.
493 2007-07-24 Ben Pfaff <blp@gnu.org>
496 * sys-file-writer.c (write_variable_display_parameters): Use new
497 var_default_display_width function to choose display width of
498 segments after the first one in a given variable.
499 * variable.c (var_create): Use var_default_display_width to pick
500 new variable's display width.
501 (var_default_display_width): New function.
502 Reviewed by John Darrington.
504 2007-07-24 Ben Pfaff <blp@gnu.org>
507 * por-file-writer.c (write_variables): Write weight variable.
508 Reviewed by John Darrington.
510 2007-07-23 Ben Pfaff <blp@gnu.org>
512 Improvements to system file reader and writer.
514 First, move all detailed knowledge of very long strings into
515 sys-file-private.[ch], so that this nasty stuff can be isolated.
517 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
518 (EFFECTIVE_VLS_CHUNK): New macro.
519 (min_int): New function.
520 (max_int): New function.
521 (sfm_width_to_bytes): Rewrite.
522 (sfm_width_to_octs): New function.
523 (sfm_segment_alloc_width): New function.
524 (sfm_segment_alloc_bytes): New function.
525 (sfm_segment_used_bytes): New function.
526 (sfm_segment_offset): New function.
527 (sfm_segment_effective_offset): New function.
528 (sfm_dictionary_to_sfm_vars): New function.
530 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
531 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
532 (struct sfm_var): New structure.
534 Next, improvements to the system file reader.
536 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
537 type casenumber. Added `version_major', `version_minor',
540 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
541 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
542 Change `case_cnt' to type casenumber. Removed `has_vls'.
543 (struct sfm_var): Removed.
544 (sfm_open_reader): Don't warn on wrong case size if the file was
545 written by SPSS 13, which tends to get it wrong. Use
546 sfm_dictionary_to_sfm_vars.
547 (read_header): Always output system file info.
548 (read_variable_record): Simplify code for reading missing values.
549 (read_machine_int32_info): Save version numbers from system file
550 into info struct passed as new argument.
551 (read_long_string_map): Restructured to use new sys-file-private
553 (read_value_labels): Use size_overflow_p.
554 (sys_file_casereader_read): Get rid of distinction between fast
555 and slow paths. Use information provided by sys-file-primate's
556 struct sfm_var to simplify code.
557 (skip_whole_strings): New function.
558 (read_int32): Renamed read_int. Changed return value to int.
560 (read_flt64): Renamed read_float. Changed return value to
561 double. Updated all callers.
562 (int32_to_native): Removed. Changed callers to use
564 (flt64_to_double): Removed. Changed callers to use float_convert.
566 Finally, get rid of int32, flt64 terminology and types in system
567 file writer. The former wasn't very useful since a POSIX "int"
568 can hold the whole range of int32 and we generally didn't have a
569 need for it to be exactly-32-bits, just at-least-32-bits. The
570 latter was inconvenient because we had to assume that it could be
571 different from double and thereby convert special values SYSMIS,
572 HIGHEST, LOWEST to and from it in multiple places. Instead, now
573 we just use "int" and "double" in most places, and do conversions,
574 if necessary, very close to where we do I/O. This change meant
575 that the writer code couldn't represent records in the file as C
576 structs any longer, but that's no great loss. The code actually
577 seems to be more readable without them.
579 Simplify the compression buffering code: only buffer as much as
580 necessary, which is no more than eight 8-byte units at any given
583 * sys-file-writer.c (typedef flt64): Removed.
584 (macro second_lowest_flt64): Removed.
585 (struct sysfile_header): Removed.
586 (struct sysfile_variable): Removed.
587 (struct sfm_writer): Removed `needs_translation', `has_vls',
588 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
589 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
590 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
591 compression buffering by `opcodes', `opcode_cnt', `data',
592 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
593 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
594 data writer functions instead of structures.
595 (calc_oct_idx): New function.
596 (write_header): Use simple data writer functions instead of
598 (write_format_spec): Renamed write_format. New argument.
599 (write_variable_continuation_records): New function.
600 (write_variable): Use simple data writer functions instead of
601 structures. Use write_variable_continuation_records. Write
602 entire very long string instead of requiring caller to understand
604 (write_value_labels): Use simple data writer functions instead of
606 (write_documents): Ditto.
607 (write_variable_display_parameters): Use sys-file-private
608 functions to simplify. Use simple data writer functions instead
610 (write_vls_length_table): Use simple data writer functions instead
612 (write_longvar_table): Ditto.
613 (write_rec_7_34): Break into new functions
614 write_integer_info_record, write_float_info_record. Use simple
615 data writer functions instead of structures.
616 (buf_write): Removed.
617 (append_string_max): Removed.
618 (ensure_buf_space): Removed.
619 (sys_file_casewriter_write): Get rid of the distinction between
620 fast and slow paths, which didn't seem to be too useful. Use new
621 functions write_case_uncompressed, write_case_compressed.
622 (put_instruction): Removed.
623 (put_element): Removed.
624 (write_compressed_data): Removed.
625 (close_writer): Use flush_compressed. Only write case count to
626 system file if it will fit in the field.
627 (write_case_compressed): New function.
628 (write_case_uncompressed): New function.
629 (flush_compressed): New function.
630 (put_cmp_opcode): New function.
631 (put_cmp_number): New function.
632 (write_int): New function.
633 (convert_double_to_output_format): New function.
634 (write_float): New function.
635 (write_value): New function.
636 (write_string): New function.
637 (write_bytes): New function.
638 (write_zeros): New function.
639 (write_spaces): New function.
641 Reviewed by John Darrington.
643 2007-07-22 Ben Pfaff <blp@gnu.org>
645 Don't try to write very long strings to portable files. The
646 format does not support it.
648 * por-file-writer.c (MAX_POR_WIDTH): New macro.
649 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
650 (write_format): Add arg to take width to resize format to.
651 (write_value): Limit width of value written to MAX_POR_WIDTH.
652 (write_variables): Limit width of variable and its output formats
654 Reviewed by John Darrington.
656 2007-07-22 Ben Pfaff <blp@gnu.org>
658 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
659 local variable instead of literal 5.
660 Reviewed by John Darrington.
662 2007-07-22 Ben Pfaff <blp@gnu.org>
664 Fix problems with uniqueness of short names in system files with
665 very long string variables. Now a variable may have multiple
668 * automake.mk (src_data_libdata_a_SOURCES): Add new files
669 short-names.c, short-names.h.
671 * dictionary.c (dict_clone): Clone all the short names.
672 (compare_strings): Move into short-names.c.
673 (hash_strings): Ditto.
674 (set_var_short_name_suffix): Ditto.
675 (dict_assign_short_names): Ditto, rename short_names_assign,
676 change to assign all short names.
678 * por-file-writer.c (write_variables): Use short_names_assign
679 instead of dict_assign_short_names.
681 * short-names.c: New file.
683 * short-names.h: New file.
685 * sys-file-private.c (sfm_width_to_segments): New function.
687 * sys-file-reader.c (read_long_var_name_map): Save and restore all
688 the short names, not just the first one.
690 * sys-file-writer.c (cont_var_name): Removed.
691 (sfm_open_writer): Use short_names_assign instead of
692 dict_assign_short_names. Use unique short names assigned by
693 short_names_assign instead of those generated by cont_var_name.
695 * variable.c (struct variable): Remove `short_name' member,
696 replace by `short_names' and `short_name_cnt'.
697 (var_create) Initialize new members.
698 (var_get_short_name_cnt): New function.
699 (var_get_short_name): Now takes an index argument. Changed most
701 (var_set_short_name): Ditto.
702 (var_clear_short_name): Renamed var_clear_short_names, changed to
703 clear all short names.
705 Reviewed by John Darrington.
707 2007-07-22 Ben Pfaff <blp@gnu.org>
709 * variable.c (var_set_width): Use new var_set_width function.
711 * missing-values.c (mv_n_values): Drop assertion, which was not
714 * format.c (fmt_default_for_width): New function.
715 (fmt_resize): New function.
717 Reviewed by John Darrington.
719 2007-07-18 John Darrington <john@darrington.wattle.id.au>
721 * datasheet.c (datasheet_delete_columns): Added assertion to check
722 we're not deleting outside the range of the sheet.
725 * dictionary.c dictionary.h variable.c: Added the ability for string
726 variables to be resized.
728 * vardict.h: Added some prototypes (moved from dictionary.h) as
729 these should only be called by variable.c
732 2007-07-14 John Darrington <john@darrington.wattle.id.au>
734 * sfm-reader.c: Respect case_cnt field in file header.
736 2007-07-01 John Darrington <john@darrington.wattle.id.au>
738 * transformation.c transformation.h (trns_chain_execute): Changed the
739 signature (Patch #6057)
741 2007-06-10 Ben Pfaff <blp@gnu.org>
743 * casereader-filter.c (casereader_filter_destroy): Make sure to
744 write all the remaining excluded cases to the casewriter, if any.
746 * caseinit.c (init_list_destroy): Rewrite.
747 (init_list_clear): Ditto.
749 * casegrouper.c (casegrouper_get_next_group): Always set *reader
750 to null when returning false.
752 2007-06-06 Ben Pfaff <blp@gnu.org>
754 Actually implement the new procedure code and adapt all of its
755 clients to match. Also adapt all of the other case sources and
756 sinks in the tree and their clients to use the
757 casereader/casewriter infrastructure.
759 * automake.mk: Add and remove files.
761 * any-reader.c: Change into a casereader.
762 * por-file-reader.c: Ditto.
763 * scratch-reader.c: Ditto.
764 * sys-file-reader.c: Ditto.
766 * any-writer.c: Change into a casewriter.
767 * por-file-writer.c: Ditto.
768 * scratch-writer.c: Ditto.
769 * sys-file-writer.c: Ditto.
771 * procedure.c: Change to use casereader, casewriter, caseinit, and
772 other new infrastructure.
774 * scratch-handle.c: Adapt to new infrastructure.
776 * case-sink.c: Removed, now dead code.
777 * case-sink.h: Ditto.
778 * case-source.c: Ditto.
779 * case-source.h: Ditto.
780 * casefile-factory.c: Ditto.
781 * casefile-private.h: Ditto.
784 * casefilter.c: Ditto.
785 * casefilter.h: Ditto.
788 * fastfile-factory.c: Ditto.
789 * fastfile-factory.h: Ditto.
790 * storage-stream.c: Ditto.
791 * storage-stream.h: Ditto.
793 2007-06-06 Ben Pfaff <blp@gnu.org>
797 * automake.mk: Add new files.
799 * datasheet.c: New file.
801 * datasheet.h: New file.
803 2007-06-06 Ben Pfaff <blp@gnu.org>
805 Until now, the procedure code has provided a case to the
806 case_source, which has filled in the data values that come from
807 the active file. "Left" data values that don't come from the
808 active file naturally stay the same from case to case, because the
809 procedure code keeps using that same case.
811 One of the compromises that comes with the new procedure code is
812 that the active file allocates and provides its own case, which
813 the procedure code then has to resize to provide room for any
814 other variables that should go in the case and then fill in the
815 values of "left" variables. Then, when we're done with that case,
816 we have to save the values of "left" variables to copy into the
817 next case read from the active file.
819 The caseinit code helps with this.
821 * automake.mk: Add new files.
823 * caseinit.c: New file.
825 * caseinit.h: New file.
827 2007-06-06 Ben Pfaff <blp@gnu.org>
829 * value.h (value_cnt_from_width): New function.
831 * variable.c (var_get_value_cnt): Use new function.
833 2007-06-06 Ben Pfaff <blp@gnu.org>
835 Add casegrouper, to allow cases read from a given casereader to be
836 broken into groups, each of which has its own casereader.
837 Generally cases are grouped based on having equal values for some
840 * automake.mk: Add new files.
842 * casegrouper.c: New file.
844 * casegrouper.h: New file.
846 2007-06-06 Ben Pfaff <blp@gnu.org>
848 Add interface to lexicographical ordering of cases.
850 * automake.mk: Add new files.
852 * case-ordering.c: New file.
854 * case-ordering.h: New file.
856 2007-06-06 Ben Pfaff <blp@gnu.org>
858 Add casereaders and casewriters, the basis of the new data processing
859 implementation. A casereader is a uniform interface to reading cases
860 from a data source; a casewriter is a uniform interface to writing
861 cases to a data sink.
863 * automake.mk: Add new files.
865 * casereader-filter.c: New file.
867 * casereader-provider.h: New file.
869 * casereader-translator.c: New file.
871 * casereader.c: New file.
873 * casereader.h: New file.
875 * casewriter-provider.h: New file.
877 * casewriter-translator.c: New file.
879 * casewriter.c: New file.
881 * casewriter.h: New file.
883 2007-06-06 Ben Pfaff <blp@gnu.org>
885 "casewindow" data structure that extends the deque (from libpspp)
886 of cases with the ability to dump cases to disk if we get too many
889 * automake.mk: Add new files.
891 * casewindow.c: New file.
893 * casewindow.h: New file.
895 2007-06-06 Ben Pfaff <blp@gnu.org>
897 sparse_cases data structure that augments a sparse_array of cases
898 with the ability to dump cases to disk if we get too many cases in
901 * automake.mk: Add new files.
903 * sparse-cases.c: New file.
905 * sparse-cases.h: New file.
907 2007-06-06 Ben Pfaff <blp@gnu.org>
909 Adds a low-level on-disk case array data structure.
911 * automake.mk: Add new files.
913 * case-tmpfile.c: New file.
915 * case-tmpfile.h: New file.
917 2007-06-06 Ben Pfaff <blp@gnu.org>
919 In a couple of places we calculate the maximum number of cases to
920 keep in memory based on the user-defined workspace. Enable
921 centralizing the calculation through a new function.
923 * settings.c (get_workspace_cases): New function.
925 2007-06-06 Ben Pfaff <blp@gnu.org>
927 The casenumber type is defined in transformations.h, but case.h is
928 a more sensible place. Move it.
930 * case.h (CASENUMBER_MAX): New macro.
931 (typedef casenumber): Move here, from transformations.h.
933 2007-06-03 Ben Pfaff <blp@gnu.org>
935 Slightly generalize case_to_values and case_from_values functions.
937 * case.c (case_to_values): Rename case_copy_out, change interface.
938 (case_from_values): Rename case_copy_in, change interface.
940 * fastfile.c (fastfilereader_get_next_case): Update caller.
941 (write_case_to_disk): Ditto.
943 2007-06-02 Ben Pfaff <blp@gnu.org>
945 Clean up after a forgotten part of patch #5829.
947 * casedeque.h: Remove unused file.
949 * automake.mk: Remove casedeque.h from sources.
951 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
953 * category.c: Removed redundant #include
955 2007-05-06 Ben Pfaff <blp@gnu.org>
957 Abstract the documents within a dictionary a little better.
958 Thanks to John Darrington for suggestion, initial version, and
961 * dictionary.c (struct dictionary): Change `documents' member from
962 char * to struct string.
963 (dict_clear): Destroy struct string.
964 (dict_get_documents): Convert struct string to char *.
965 (dict_set_documents): Set struct string. Pad to 80-character
967 (dict_clear_documents): New function.
968 (dict_add_document_line): New function.
969 (dict_get_document_line_cnt): New function.
970 (dict_get_document_line): New function.
972 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
974 * sys-file-reader.c (read_documents): Use new document functions.
976 2007-04-19 John Darrington <john@darrington.wattle.id.au>
978 * sys-file-reader.c: When reading a system file which has no
979 long name table, automatically create one where the long names
980 are the lower case versions of the short names.
982 2007-04-22 Ben Pfaff <blp@gnu.org>
984 * dictionary.c (dict_set_split_vars): dict_destroy expects that
985 dict_clear will free most data related to the dictionary.
986 dict_clear does a decent job, except that dict_set_split_vars on
987 some systems won't actually free the dict's "split" member.
988 Instead, it'll allocate a 1-byte region. Fix this.
990 * value.c (value_copy): New function.
991 (value_set_missing): Ditto.
993 2007-04-22 John Darrington <john@darrington.wattle.id.au>
995 * Deleted existing category.h and moved cat-routines.h into
996 category.h Encapsulated struct cat_vals better.
998 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1000 * sys-file-reader.c: When reading a system file which has no
1001 long name table, automatically create one where the long names
1002 are the lower case versions of the short names.
1004 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1006 * sys-file-reader.c: Some versions of Other Software seem to
1007 produce system files with string variables' measure set to
1008 zero. We'll assume these are supposed to be nominal variables.
1010 2007-03-30 Ben Pfaff <blp@gnu.org>
1012 * procedure.c: Adapt to new deque data structure.
1014 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1015 Ben Pfaff <blp@gnu.org>
1017 * file-name.c: Mingw compatibility fixes.
1018 (fn_search_path): Use ISSLASH instead of comparing against '/'
1020 (fn_dir_name): Use dir_name from gnulib.
1021 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1022 (fn_get_identity): Use GetFullPathName instead of canonicalize
1023 from gnulib, because the latter does not fully support
1024 Windows-style path names. Use this implementation based on the
1025 detected presence of Windows instead of the absence of Unix, since
1026 the new implementation is Windows-specific.
1027 (fn_compare_file_identities): In Windows implementation, compare
1028 names case-insensitively.
1030 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1032 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1033 <stdlib.h> to have the same effect.
1035 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1037 * por-file-reader.c: Add missing _() around messages.
1039 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1041 * make-file.c: Include "mkstemp.h", without which linking on
1044 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1046 Reduce platform dependence.
1048 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1049 Everywhere we using this, we really should have just depended on
1050 the shell to expand tildes.
1051 (fn_search_path): Simplify, given that we don't do tilde expansion
1053 (fn_normalize): Removed. Caller changed to use the canonicalize
1055 (fn_get_cwd): Removed. Only user was fn_normalize.
1056 (fn_is_absolute): Really only test for absolute names.
1057 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1059 (fn_readlink): Removed, as it was only used fn_normalize.
1060 (fn_exists): Assume the stat function is available; gnulib does.
1061 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1063 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1065 * dictionary.c (dict_create_vector_assert): New function.
1067 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1069 * file-name.c (fn_normalize): Correct name of function
1070 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1071 for pointing this out.
1073 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1075 We are using a single member in struct file_handle, the "name"
1076 field, for more than one purpose. When it begins with '"', it's a
1077 file name; otherwise, it's a token that can be used to identify
1078 it. When that assertion fires, it's because we searched for the
1079 name case-sensitively as a file name (so that there was no match),
1080 and then we try to insert it case-insensitively as a token, which
1081 fails because duplicates aren't allowed.
1083 Solution: break the two purposes into two separate fields. This
1084 fixes the problem and likely makes the code easier to read too.
1086 Fixes bug #18922. Thanks to John Darrington for bug report and
1089 * file-handle-def.c (struct file_handle): New `id' member.
1090 (fh_from_name): Rename fh_from_id. Update all callers.
1091 (create_handle): New `id' parameter. Update all callers.
1092 (fh_create_file): Ditto.
1093 (fh_get_id): New function.
1095 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1097 * case.c (case_is_null): Change return type to bool.
1099 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1103 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1104 reference counting to share data. This makes it easy for
1105 valgrind, etc. to find accesses to cases that have been destroyed
1106 but have been kept around by another user's ref-count. This often
1107 happens when the data set is small enough to find in memory; if a
1108 bigger data set that would overflow to disk were used, then data
1109 corruption would occur.
1111 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1115 * case.c (case_unshare): Make it check internally whether the
1116 ref_cnt is greater than 1, so that the callers don't have to.
1117 Update callers not to check.
1119 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1121 Before, I was thinking that I might want to get rid of reference
1122 counting at some point. Now, I'm pretty sure that it's here to
1123 stay. Thus, because we have to store the value_cnt anyway for
1124 reference-counted cases, we might as well expose it to users.
1126 * case.c (case_get_value_cnt): New function.
1127 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1128 resize case if its size actually changed.
1130 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1131 instead of peeking inside struct case directly.
1132 (casefile_append): Ditto.
1134 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1136 Get rid of the inlines for the case functions, which made the
1137 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1138 the inlines didn't speed up "make check" at all, which is not a
1139 perfect benchmark but seems indicative.)
1141 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1142 definitions. Remove some assertions on nonnull pointers that were
1143 redundant with a pointer dereference soon after in the function.
1145 (struct case_data): Move definition here from case.h.
1149 (case_data_wr): Ditto.
1151 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1153 * automake.mk: Add casedeque.h to sources.
1155 * casedeque.h: New file.
1157 * procedure.c: (struct dataset) Change lag_count, lag_head,
1158 lag_queue member into single struct casedeque member. Update all
1159 users to use the casedeque instead.
1162 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1164 * procedure.c: Simplify lagged cases interface. Updated all
1165 clients--well, the only client--to use the simplified interface.
1166 (dataset_n_lag) Removed.
1167 (dataset_set_n_lag) Removed.
1168 (dataset_need_lag) New function.
1170 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1172 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1173 whenever appropriate, but only when the dataset/dictionary is in a
1176 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1178 * dictionary.c dictionary.h : Added callbacks for change of filter and
1179 split variables. Refactored some code to ensure that callbacks get
1180 invoked when appropriate.
1182 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1183 replace_dict callback occurs when permanent_dict replaces the current
1186 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1188 * dictionary.c dictionary.h : Added callback for when the weight
1189 variable of a dictionary changes.
1191 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1193 * dictionary.c dictionary.h : Added replace_source and replace_dict
1194 callbacks, and functions to deal with them.
1196 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1198 Simplify missing value handling.
1200 * missing-values.h (enum mv_class): New type.
1201 (enum mv_type): Moved definition into missing-values.c and renamed
1202 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1203 enums. Updated all uses.
1204 (struct missing_values): Changed type of `type' from `enum
1205 mv_type' to `int' because the definition is no longer exposed.
1207 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1208 parameter. Update all callers.
1209 (mv_is_num_missing): Ditto.
1210 (mv_is_str_missing): Ditto.
1211 (mv_is_value_user_missing): Removed. Changed callers to use
1212 mv_is_value_missing.
1213 (mv_is_num_user_missing): Removed. Changed callers to use
1215 (mv_is_str_user_missing): Removed. Changed callers to use
1217 (mv_is_value_system_missing): Removed. Changed callers to use
1218 mv_is_value_missing.
1219 (mv_set_type): Removed. Changed callers to use mv_clear.
1220 (mv_clear): New function.
1222 * variable.c (var_is_value_missing): Add new enum mv_class
1223 parameter. Update all callers.
1224 (var_is_num_missing): Ditto.
1225 (var_is_str_missing): Ditto.
1226 (var_is_value_user_missing): Removed. Changed callers to use
1227 var_is_value_missing.
1228 (var_is_num_user_missing): Removed. Changed callers to use
1230 (var_is_str_user_missing): Removed. Changed callers to use
1232 (var_is_value_system_missing): Removed. Changed callers to use
1233 var_is_value_missing.
1235 * casefilter.c (struct casefilter): Use enum mv_class in place of
1237 (casefilter_variable_missing): Adapt to new member.
1238 (casefilter_create): Change signature to take enum mv_class,
1241 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1243 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1245 * case-sink.c case-sink.h procedure.c procedure.h
1246 storage-stream.c: Now uses the factory.
1248 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1250 Make it possible to pull cases from the active file with a
1251 function call, instead of requiring indirection through a callback
1254 * case-source.h (struct case_source_class): Change ->read function
1255 to return a single case, instead of calling a callback function
1256 for each case. Change ->destroy function to return an error
1259 * case-source.c (free_case_source): Pass along the value returned
1260 by the case_source ->destroy function.
1262 * procedure.c (struct write_case_data): Removed.
1263 (struct dataset): Added some members to track procedure state.
1264 (procedure): Optimize the trivial case at this level.
1265 (internal_procedure): Re-implement in terms of proc_open,
1266 proc_read, proc_close.
1267 (proc_open) New function.
1268 (proc_read) New function.
1269 (proc_close) New function.
1270 (write_case) Moved into proc_read.
1271 (close_active_file) Moved closing of data source into proc_close.
1273 * storage-source.c: Rewrote to conform with modified
1274 case_source_class interface.
1276 * transformations.c (trns_chain_execute): Added argument to allow
1277 starting execution from an arbitrary transformation. Updated
1280 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1282 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1284 * sys-file-reader.c (read_display_parameters): Don't assume that
1285 MEASURE_* and ALIGN_* have the same values found in system files.
1287 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1289 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1290 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1291 the value of n_MEASURES, which was off by 1 (at least from my
1294 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1296 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1297 callbacks which are invoked when the dictionary or its
1298 variables are changed.
1300 * missing-values.c missing-values.h value-labels.c: Tidied up
1301 consistency checks, and made some of them return false
1302 instead of assert-failing.
1304 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1306 * calendar.c (calendar_days_in_month): New function.
1308 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1310 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1311 the value label's width.
1313 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1315 * sfm-private.h: Move contents into sys-file-writer.c, which is
1316 the only remaining user. Removed Borland C++-specific directives.
1318 * sys-file-reader.c: Clean up and rewrite entire file. The
1319 rewritten version is simpler and better abstracted, and should be
1320 easier to maintain and extend. It avoids using structures to read
1321 file data, which is prone to padding variations among compilers.
1322 It should also handle non-IEEE 754 system files, although I
1323 haven't been able to find any. It has been tested against many
1324 .sav files obtained from the Web and found to produce the same
1325 results as the earlier version of the code, or in some cases
1326 improved results. It is more tolerant of format variations found
1329 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1330 member, putting an enum integer_format in its place. New member
1331 `float_format'. Changed `compressed' member to type bool.
1333 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1335 * dictionary.c (dict_delete_consecutive_vars): New function.
1337 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1339 * file-name.c (fn_search_path): Remove prefix arg that was unused
1340 by any caller. Updated all callers.
1342 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1344 * format.c (fmt_dollar_template): Use user's decimal point
1345 character. Add assertion.
1347 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1349 * format.c (fmt_dollar_template): New function, based on
1350 dollar_format_template from var-type-dialog.c.
1352 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1354 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1357 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1359 Finish converting struct variable to an opaque type. In this
1360 phase, we add remaining setter and getter functions, convert the
1361 remaining PSPP code to use them, and do a bunch of cleanup. The
1362 resulting changes are pervasive but mostly trivial, and only the
1363 notable changes are logged.
1365 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1368 * case.c (case_data): Renamed case_data_idx.
1369 (case_num): Renamed case_num_idx.
1370 (case_str): Renamed case_str_idx.
1371 (case_data_rw): Renamed case_data_rw_idx.
1373 * case.h (case_data): New function with old name and an interface
1374 that takes a variable instead of an index, which is easier to
1375 use. Updated all callers to use the new interface, or to use the
1376 new *_idx function (see above).
1379 (case_data_rw): Ditto.
1381 * category.c (cat_stored_values_destroy): Changed interface to
1382 take a struct cat_vals * instead of a struct variable *.
1384 * dictionary.c (dict_clone): Use new vector_clone function.
1385 (dict_clear) Use new var_destroy function.
1386 (add_var) New function.
1387 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1388 (dict_create_var_assert) Rewrite in terms of add_var.
1389 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1390 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1391 (dict_lookup_var) Use new var_create, var_destroy functions.
1392 (dict_contains_var) Rewrite in terms of new vardict functionality.
1393 (set_var_dict_index) New function.
1394 (set_var_case_index) New function.
1395 (reindex_vars) New function.
1396 (dict_delete_var) Rewrite in terms of new vardict functionality.
1397 (dict_reorder_var) Ditto.
1398 (dict_reorder_vars) Ditto.
1399 (rename_var) New function.
1400 (dict_rename_var) Use rename_var.
1401 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1402 (dict_get_compacted_idx_to_fv) Rename
1403 dict_get_compacted_dict_index_to_case_index, update callers.
1404 (dict_create_vector) Use new vector_create function.
1405 (dict_clear_vectors) Use new vector_destroy function.
1406 (set_var_short_name_suffix) Move here from variable.c, renamed
1407 from var_set_short_name_suffix, make static, update caller.
1409 * sys-file-private.c: New file.
1410 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1411 width_to_bytes, update callers.
1413 * sys-file-private.h: New file. Later it will supplant
1414 sfm-private.h; for now it supplements it.
1415 (macro MIN_VERY_LONG_STRING) New macro.
1416 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1418 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1419 MAX_LONG_STRING was used before.
1421 * sys-file-writer.c: Ditto.
1423 * value-labels.c: Change the paradigm here to be that a null
1424 pointer is OK for a struct val_labs * in most cases; it just
1425 represents an empty set of value labels.
1426 (val_labs_copy) A copy of a null set is a null set.
1427 (val_labs_count) A null set has 0 labels.
1428 (val_labs_replace) Change return type to void. Rewrite for
1430 (val_labs_find) A null set does not contain the value.
1431 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1432 transposed argument order, updated all callers.
1434 * value.c: New file.
1435 (value_dup) Moved here from variable.c.
1436 (compare_values) Ditto.
1439 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1440 (macro MAX_LONG_STRING) Removed, because it was only interesting
1441 for system files, not for general code.
1442 (macro MAX_VERY_LONG_STRING) Ditto.
1443 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1444 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1446 * vardict.h: New file, for an interface between variables and
1449 * variable.c: A lot of functions were moved around, for better
1451 (struct variable) Move definition here, from variable.h.
1452 (var_type_adj) Removed--makes i18n hard.
1453 (var_type_noun) Ditto.
1454 (var_create) New function.
1455 (var_clone) New function.
1456 (var_destroy) New function.
1457 (var_set_name) Assert that variable is not in a dictionary.
1458 (compare_var_names) Rename compare_vars_by_name and fix a couple
1459 of callers who thought the args were strings.
1460 (hash_var_name) Rename hash_var_by_name.
1461 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1462 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1463 (var_is_very_long_string) Removed, because it was only interesting
1464 to system file code.
1465 (var_set_missing_values) Allow the argument to be the wrong width,
1466 as long as we can resize it. Simplify callers who were doing the
1467 resizing themselves.
1468 (var_get_value_labels) New function.
1469 (var_has_value_labels) New function.
1470 (var_set_value_labels) New function.
1471 (alloc_value_labels) New function.
1472 (var_add_value_label) New function.
1473 (var_replace_value_label) New function.
1474 (var_clear_value_labels) New function.
1475 (var_lookup_value_label) New function.
1476 (var_get_value_name) Moved here from variable.c, renamed from
1477 var_get_value_name, transposed argument order, updated all
1479 (var_to_string) Moved here, from variable-label.c.
1480 (var_set_leave) New function.
1481 (var_get_leave) New function.
1482 (var_must_leave) New function.
1483 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1484 set_var_short_name_suffix.
1485 (var_get_dict_index) New function.
1486 (var_get_case_index) New function.
1487 (var_get_obs_vals) New function.
1488 (var_set_obs_vals) New function.
1489 (var_has_obs_vals) New function.
1490 (var_get_vardict) New function.
1491 (var_set_vardict) New function.
1492 (var_has_vardict) New function.
1493 (var_clear_vardict) New function.
1494 (value_dup) Moved to value.c.
1495 (compare_values) Ditto.
1498 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1499 (enum ALPHA) Rename VAR_STRING, update all users.
1501 * vector.c: New file.
1502 (struct vector) Moved here, from variable.h.
1503 (check_widths) New function.
1504 (vector_create) New function.
1505 (vector_clone) New function.
1506 (vector_destroy) New function.
1507 (vector_get_name) New function.
1508 (vector_get_var) New function.
1509 (vector_get_var_cnt) New function.
1510 (compare_vector_ptrs_by_name) New function.
1512 * vector.h: New file.
1514 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1516 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1517 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1520 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1522 * value-labels.c (destroy_atoms): New function.
1523 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1525 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1527 Thanks to Jason Stover for pointing out this problem.
1529 * data-out.c (output_number): Use gsl_finite from GSL, which is
1530 portable, instead of isfinite, which is not.
1533 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1535 * variable.c variable.h (value_dup): New function.
1537 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1539 Start converting struct variable to an opaque type. In this
1540 phase, we add a bunch of setter and getter functions and convert
1541 most of the PSPP code to use them. The resulting changes are
1542 pervasive but mostly trivial, and only the notable changes are
1545 * format.c (fmt_equal): New function.
1547 * variable.c (var_type_is_valid): New function.
1548 (measure_is_valid) Moved here, from format.c.
1549 (alignment_is_valid) Moved here, from format.c.
1550 (var_get_name) New function.
1551 (var_set_name) New function.
1552 (width_to_type) New function.
1553 (var_get_type) New function.
1554 (var_get_width) New function.
1555 (var_set_width) New function.
1556 (var_is_numeric) New function.
1557 (var_is_alpha) New function.
1558 (var_is_short_string) New function.
1559 (var_is_long_string) New function.
1560 (var_is_very_long_string) New function.
1561 (var_get_missing_values) New function.
1562 (var_set_missing_values) New function.
1563 (var_clear_missing_values) New function.
1564 (var_has_missing_values) New function.
1565 (var_is_value_missing) New function.
1566 (var_is_num_missing) New function.
1567 (var_is_str_missing) New function.
1568 (var_is_value_user_missing) New function.
1569 (var_is_num_user_missing) New function.
1570 (var_is_str_user_missing) New function.
1571 (var_is_value_system_missing) New function.
1572 (var_get_print_format) New function.
1573 (var_set_print_format) New function.
1574 (var_get_write_format) New function.
1575 (var_set_write_format) New function.
1576 (var_set_both_formats) New function.
1577 (var_get_label) New function.
1578 (var_set_label) New function.
1579 (var_clear_label) New function.
1580 (var_has_label) New function.
1581 (var_get_measure) New function.
1582 (var_set_measure) New function.
1583 (var_get_display_width) New function.
1584 (var_set_display_width) New function.
1585 (var_get_alignment) New function.
1586 (var_set_alignment) New function.
1587 (var_get_value_cnt) New function.
1588 (var_get_leave) New function.
1589 (var_get_short_name) New function.
1591 * variable.h: (struct variable) Removed "type" and "nv" members;
1592 they are now computed from "width" where needed.
1594 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1596 * missing-values.c (mv_resize): Don't write beyond end of the
1597 allocated buffer when resizing a long string.
1599 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1601 Clean up identifier code: don't require identifier enumerations to
1602 be in a particular order; make better use of string library;
1603 expose less of the internals.
1605 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1606 change interface. Updated all callers.
1607 (lex_id_match) Change interface to use struct substring, update
1609 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1610 (global array keywords[]) Make static, change form. Update all
1611 users to use lex_id_name instead.
1612 (lex_is_keyword) New function.
1613 (lex_id_to_token) Change interface to use struct substring, update
1615 (lex_id_name) New function.
1617 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1618 lex_is_keyword instead.
1619 (T_LAST_KEYWORD) Removed.
1620 (T_N_KEYWORDS) Removed.
1622 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1624 * format.c: (fmt_date_template) Distinguish characters for which a
1625 space is output and any date delimiter is allowed on input, from
1626 those for which a space is output and only a space is allowed on
1627 input. The former is represented by X, the latter by a space.
1628 Also, drop distinction between h and H, changing the former to the
1631 * data-in.c: Completely rewrite internals to conform to SPSS input
1632 formats as closely as possible.
1633 (data_in) Changed external interface by replacing the structure
1634 that was used as a single argument by a set of arguments. Updated
1636 (data_in_finite_line) Removed. Converted all callers to use plain
1638 (data_in_get_integer_format) New function.
1639 (data_in_set_integer_format) New function.
1640 (data_in_get_float_format) New function.
1641 (data_in_set_float_format) New function.
1643 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1644 (struct data_in) Removed.
1646 * data-out.c: (output_date) Drop each component from the input as
1647 it is output, to allow us to drop the distinction between h (a
1648 count of hours) and H (the hour of day) template characters.
1649 Also, handle new X template character.
1650 (output_scientific) Follow more rational rule on when to drop
1651 fraction introduced between SPSS 13 and 15. Updated test case to
1654 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1656 Fix buffer overflow reported by John Darrington.
1658 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1659 realize that DIGITS is a count of nibbles, not of bytes.
1661 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1663 * calendar.c (calendar_offset_to_gregorian) Also return the
1664 year-of-day. Change callers to new interface.
1666 * data-out.c: Completely rewrite internals to conform to SPSS
1667 output formats as completely as possible.
1668 (data_out) Change interface to put input parameters before output
1669 parameters, for consistency with the style I now prefer. Update
1671 (data_out_get_integer_format) New public function.
1672 (data_out_set_integer_format) New public function.
1673 (data_out_get_float_format) New public function.
1674 (data_out_set_float_format) New public function.
1676 * data-out.h: New file. Move prototype for data_out here, from
1679 * format.c: (fmt_step_width) Use equality comparison instead of
1680 bitwise and, for clarity.
1681 (fmt_is_string) Ditto.
1682 (fmt_input_to_output) Fix categories that are translated to F
1685 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1687 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1688 casefile-private.h: Added casefilter to assist commands with missing
1691 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1693 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1695 * settings.c: (route_errors_to_terminal) New variable.
1696 (route_errors_to_listing) New variable.
1697 (get_error_routing_to_terminal) New function.
1698 (set_error_routing_to_terminal) New function.
1699 (get_error_routing_to_listing) New function.
1700 (set_error_routing_to_listing) New function.
1702 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1704 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1706 * format.c: Completely rewrite, to achieve better abstraction.
1707 Rewrite all references to formats in other files.
1709 * format.def: Rewrite and reorganize.
1711 * settings.c: Move everything related to custom currency formats
1712 into format.[ch], changing them in form, so as to group related
1713 code and definitions better. Changed all references to use the
1715 (static var decimal) Removed.
1716 (static var grouping) Removed.
1717 (static var cc) Removed.
1718 (get_decimal) Removed.
1719 (set_decimal) Removed.
1720 (get_grouping) Removed.
1721 (set_grouping) Removed.
1725 * settings.h: (macro CC_CNT) Removed.
1726 (macro CC_WIDTH) Removed.
1727 (struct custom_currency) Removed.
1729 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1731 * data-in.c (data_in): Use switch statement instead of table, to
1732 avoid dependence on the order of the FMT_* enums.
1734 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1736 * data-out.c: (num_to_string) Removed, because it was dead code.
1738 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1740 * data-in.c (parse_trailer): Fix error message.
1742 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1744 * format.c (fmt_is_binary): New function.
1746 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1748 * procedure.c procedure.h: Encapsulated the static data into a single
1751 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1753 * casefile.c (casereader_read_xfer): Always initialize the case,
1754 even on an error condition.
1756 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1758 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1759 entire function useless.
1761 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1763 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1764 casereader method casereader_clone.
1766 * procedure.c pransformations.h: Introduced new type casenum_t
1768 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1770 * variable.c: (width_to_bytes) Rephrase code for clarify.
1772 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1774 * format.c: (fmt_type_from_string) New function.
1775 (fmt_to_string) Include decimals in output if the format has
1776 decimals, even if the format type does not. This way, we can
1777 accurately reproduce incorrect formats in user output.
1778 (check_common_specifier) Make the check for a bad format type an
1779 assertion, so we get bug reports if they show up. Fix message.
1780 Check for decimal places with a format type that doesn't allow
1782 (check_input_specifier) Remove check for FMT_X, which has been
1784 (check_output_specifier) Ditto.
1786 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1788 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1789 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1790 (enum fmt_parse_flags) Removed.
1792 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1794 * casefile.c casefile.h: Converted to an abstract base class.
1795 * casefile-private.h fastfile.c fastfile.h: New files.
1796 * automake.mk procedure.c scratch-writer.c storage-stream.c
1798 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1800 * procedure.c (internal_procedure): Create sink_case with only as
1801 many values as the compacted dictionary.
1803 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1805 Remove "debugging" code that caused plenty of false positives and
1808 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1810 * case.c: Remove all references to `this' member.
1812 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1814 Fix link error noted by Jason Stover.
1816 * storage-stream.c: Include <assert.h>.
1818 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1820 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1821 ALL) and additional underlying system file issues.
1823 Thanks to John Darrington for review.
1825 First problem: var_hash points to variables not owned by the
1826 sys-file-reader, which the caller may free or modify. Use an
1827 array of sfm_vars instead, as done earlier (e.g. CVS version
1830 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1831 members and remove all code that references it. Add vars, var_cnt
1832 members. Remove fix_specials member, which was unused.
1833 (struct sfm_var) Remove name member, which was unused.
1834 (sfm_close_reader) Free vars member instead of var_hash.
1835 (compare_var_shortnames) Removed.
1836 (hash_var_shortname) Removed.
1837 (sfm_open_reader) Fill out vars array.
1838 (compare_var_index) Removed.
1839 (sfm_read_case) Use vars instead of var_hash.
1841 Second problem: we're confused about when we actually have very
1842 long strings, causing us to choose incorrectly between slow path
1843 and fast path in sfm_read_case.
1845 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1846 have very long strings, not when we have long variable names,
1847 which is an unrelated feature.
1849 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1851 * variable.h: Move var_set and variable parsing declarations to
1852 new header, src/language/lexer/variable-parser.h. Modified lots
1853 of files to include the new header.
1855 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1857 * value-labels.c (value_to_string): When there's no value label,
1858 format the variable according to its print format, instead of
1859 always effectively using A or F format.
1861 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1863 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1866 * format.c: Removed tortological assertion.
1868 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1870 Reform string library.
1872 * file-name.c (fn_interp_vars): Change interface to take a
1873 substring as input. Updated all users.
1875 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1877 * format.c (measure_is_valid): Really return false when m >=
1880 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1882 Implement random access to casefiles, for use in GUI.
1884 * casefile.c: (struct casereader) Add `random', `file_ofs',
1885 `buffer_ofs' members.
1886 (casefile_get_random_reader) New function.
1887 (read_open_file) Break part into new function
1888 seek_and_fill_buffer().
1889 (fill_buffer) Update buffer_ofs, file_ofs.
1890 (casereader_seek) New function.
1892 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1894 * settings.c: Added call to i18n{done, init}.
1896 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1898 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1901 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1903 * casefile.c: Convert many uses of `int' to `bool'.
1905 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1907 * transformations.c (trns_chain_destroy): Destroy chain's trns
1908 member, to fix memory leak.
1910 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1912 * storage-stream.c (storage_source_decapsulate): Destroy case
1913 source to fix memory leak.
1915 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1917 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1918 case passed as argument instead of initializing the argument as a
1919 case. Fixes memory leak that showed up in
1920 tests/command/aggregate.sh with scratch files.
1922 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1924 * procedure.c (proc_done): Destroy default_dict, to fix memory
1927 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1929 Simplify procedure_with_splits().
1931 * procedure.c (struct split_aux_data): Removed case_count member.
1932 (procedure_with_splits) Don't initialize case_count.
1933 (split_procedure_case_func) Check whether prev_case is null
1934 instead of case_count.
1935 (split_procedure_end_func) Ditto.
1937 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1939 * case.c (case_move): Do nothing if dst and src are the same
1941 (case_try_create) Merge two similar cases.
1942 (case_copy) Unshare only if data must be actually copied.
1944 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1946 * data-in.c data-out.c dictionary.c sys-file-reader.c
1947 sys-file-writer.c variable.c variable.h: Reworked very long string
1948 support for better encapsulation.
1950 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1952 * value-labels.c (val_labs_can_set_width): New function.
1953 (val_labs_set_width) Clear labels if increasing width to long
1955 (val_labs_destroy) Remove unneeded test for null.
1957 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1959 * value-labels.h: Remove unneeded dependency on variable.h.
1961 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1963 Get rid of `char *c' member in union value, for cleanliness.
1965 * value.h: (union value) Remove `c' member.
1967 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1969 Make dictionary compacting functions a little more general.
1971 * sys-file-writer.c (sfm_open_writer): Use
1972 dict_compacting_would_change().
1973 (does_dict_need_translation) Removed.
1975 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1977 Make dictionary compacting functions a little more general.
1979 * dictionary.c (dict_needs_compaction): Rename
1980 dict_compacting_would_shrink(). Update all callers.
1981 (dict_compacting_would_change) New function.
1983 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1985 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1986 inverted return value (!).
1988 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1990 Continue reforming procedure execution.
1992 * procedure.c: Search and replace "vfm" by "proc". Notably:
1993 (static var vfm_source) Rename proc_source. Update all
1995 (static var vfm_sink) Rename proc_sink. Update all references.
1997 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1999 Continue reforming procedure execution. In this phase, remove
2000 PROCESS IF, which was deprecated anyway and can be easily
2001 simulated with TEMPORARY followed by SELECT IF.
2003 * procedure.c: (open_active_file) Don't call
2004 add_process_if_trns().
2005 (discard_variables) Get rid of redundant call to
2006 proc_cancel_all_transformations().
2007 (add_process_if_trns) Removed.
2008 (process_if_trns_proc) Removed.
2009 (process_if_trns_free) Removed.
2011 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2013 Continue reforming procedure execution. In this phase, add
2014 `const' to the case passed to procedure()'s callback.
2016 Updated all users of procedure() as well.
2018 * procedure.c: (struct write_case_data) Add "const" to ccase
2019 parameter for case_func member.
2020 (procedure) Add "const" to ccase parameter for proc_func
2022 (multipass_case_func) Make ccase parameter const.
2023 (internal_procedure) Add "const" to ccase parameter for case_func
2025 (split_procedure_case_func) Make ccase parameter const.
2026 (multipass_split_case_func) Make ccase parameter const.
2028 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2030 Continue reforming procedure execution. In this phase, get rid of
2031 the output code for SPLIT FILE groups in procedure.c, which really
2032 shouldn't be doing any output. Move it into the individual
2033 procedures instead. This also adds some flexibility.
2035 Updated many users of procedure_with_splits() and
2036 multipass_procedure_with_splits() to call
2037 output_split_file_values() and to deal with increased use of
2040 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2041 parameter to begin_func member.
2042 (procedure_with_splits) Add "const struct ccase *" parameter to
2043 begin_func parameter. Make ccase parameter const in proc_func
2045 (split_procedure_case_func) Don't dump split file group. Pass
2047 (dump_splits) Moved to language/dictionary/split-file.c as
2048 output_split_file_values().
2049 (struct multipass_split_aux_data) Add "const struct ccase *"
2050 parameter to split_func member.
2051 (multipass_procedure_with_splits) Add "const struct ccase *"
2052 parameter to split_func parameter.
2053 (multipass_split_case_func) Save new SPLIT FILE case before
2055 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2057 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2059 Continue reforming procedure execution. Change
2060 internal_procedure() so that it calls open_active_file() and
2061 close_active_file(), which isolates most of the actual procedure
2064 * procedure.c: (struct write_case_data) Rename `proc_func' member
2065 to `case_func' and update all references.
2066 (procedure) Rewrite as one-line wrapper around
2067 internal_procedure().
2068 (struct multipass_aux_data) New.
2069 (multipass_callback) Renamed multipass_case_func(). Use struct
2070 multipass_aux_data as auxiliary data.
2071 (multipass_end_func) New function.
2072 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2073 that uses multipass_case_func, multipass_end_func.
2074 (internal_procedure) Add `end_func' argument. Move optimization
2075 of trivial case in here. Move call to open_active_file() and
2076 close_active_file() in here. Now assert that vfm_source is
2078 (procedure_with_splits_callback) Rename
2079 split_procedure_case_func().
2080 (split_procedure_end_func) New function.
2081 (multipass_split_callback) Rename multipass_split_case_func.
2082 (multipass_split_end_func) New function.
2083 (discard_variables) No need to test for nonnull vfm_source.
2085 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2087 Continue reforming procedure execution. Get rid of unused member.
2089 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2091 (write_case) Don't increment cases_analyzed.
2093 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2095 Continue reforming procedure execution. In this phase, move
2096 procedure.c and procedure.h from src to src/data. Update
2097 makefiles and #includes accordingly.
2099 * procedure.c: Moved here from src/.
2101 * procedure.h: Moved here from src/.
2103 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2105 Continue reforming procedure execution. In this phase, get rid of
2106 many global variables, consolidating procedure execution in
2107 procedure.c. Encapsulate transformations in new "struct
2108 trns_chain". Also, change implementation of N OF CASES, FILTER,
2109 and PROCESS IF from special cases to transformations.
2111 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2114 * dictionary.c: (global variable default_dict) Move to
2117 * variable.h: (TRNS_*) Move to transformations.h.
2118 (struct transformation) Move to transformations.c.
2120 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2122 * sys-file-reader.c: Fixed invalid read problems.
2124 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2126 * storage-stream.c: Add missing function comments.
2128 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2130 Continue reforming procedure execution. In this phase, add some
2131 new, needed functionality to storage-stream.
2133 * storage-stream.c: (storage_source_decapsulate) New function.
2135 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2137 * variable.c (width_to_bytes): Declarations must precede
2138 statements for C90 compliance.
2140 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2142 * data-out.c, data-in.c, variable.c, variable.h: New functions
2143 copy_mangle and copy_demangle for reading/writing cases; emulates the
2144 way SPSS deals with strings > 255 bytes.
2146 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2147 subtype 14 needed for strings longer than 255 bytes.
2149 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2150 instead of literal values. Also fixed some constness issues.
2152 * format.h: Constness
2154 * sfm-private.h: Renamed the case_size identifier, since I discovered
2155 that SPSS's respect for this variable is very nominal.
2157 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2159 Change case limit type from int to size_t.
2161 * dictionary.c: (struct dictionary) Change type of case_limit
2163 (dict_get_case_limit) Change return type.
2164 (dict_set_case_limit) Change parameter type.
2166 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2168 * variable.h: (struct variable) Rename `reinit' member as `leave'
2169 and invert sense. Fix up all references.
2171 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2173 Continue reforming procedure execution. In this phase, break
2174 procedure.c into multiple files.
2176 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2178 * case-sink.c: New file.
2180 * case-sink.h: New file.
2182 * case-source.c: New file.
2184 * case-source.h: New file.
2186 * storage-stream.c: New file.
2188 * storage-stream.h: New file.
2190 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2192 * variable.h: (struct variable) Remove `init' member and all
2193 references to it from other files. It was initialized in several
2194 places, but nothing really ever used it for anything worthwhile.
2195 Thanks to Jason Stover for pointing out how confusing this
2198 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2200 Continue reforming error message support. In this phase, get rid
2201 of message "titles" and put the message text in `struct error'.
2202 Now `struct error' encapsulates a message more properly.
2204 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2205 Format message ourselves.
2207 * data-in.c: (vdls_error) Ditto.
2209 * por-file-reader.c: (error) Ditto.
2211 * sys-file-reader.c: (corrupt_msg) Ditto.
2213 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2215 GNU standards require "file name" instead of "filename" in
2216 documentation. It's nice for our code to follow the convention
2219 * casefile.c: (struct casefile) Rename `filename' member to
2220 `file_name'. Updated all references.
2222 * file-name.c: [!unix] (struct file_identity) Rename
2223 normalized_filename member to normalized_file_name. Updated all
2226 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2228 We don't really support anything but Unix-like environments well,
2229 so we might as well de-obfuscate by writing directory and path
2230 separators explicitly.
2232 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2234 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2235 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2237 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2240 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2242 GNU standards require "file name" instead of "filename" in
2243 documentation. It's nice for our code to follow the convention
2246 * filename.c: Rename to file-name.c.
2248 * filename.h: Rename to file-name.h. Update all inclusions.
2249 Update header guards.
2251 * automake.mk: Update file names.
2253 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2255 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2257 (fn_basename) Removed (dead code).
2258 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2259 (fn_special_p) Renamed fn_is_special(), all references updated.
2260 (fn_exists_p) Renamed fn_exists(), all references updated.
2262 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2264 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2265 Also, now it only tilde-expands file names, not paths.
2266 (fn_search_path) Tilde-expand one directory at a time.
2268 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2270 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2271 get rid of non-Unixlike version of the code, which has probably
2273 (fn_prepend_dir) Removed (dead code).
2275 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2276 (macro PATH_SEPARATOR_STRING) New.
2277 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2279 Continue reforming error message support. In this phase, we get
2280 rid of VM() and the other msg() support for "verbosity", replacing
2281 it by a new function verbose_msg().
2283 * filename.c: (fn_search_path) Use verbose_msg() instead of
2286 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2288 * sfm-private.h: Get rid of #defines after #error, which makes no
2291 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2293 Get rid of our own int32 type in favor of the standard int32_t
2296 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2299 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2301 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2303 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2305 Remove ill-considered file routines that are no longer used.
2307 * filename.c: (fn_open_ext) Removed.
2308 (fn_close_ext) Removed.
2310 * filename.h: (struct file_ext) Removed.
2312 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2314 * variable.c (var_is_valid_name): Move declarations before code
2317 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2319 * filename.ch (fn_interp_vars): Fixed small buglet.
2321 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2323 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2324 so as to modify the string inline. Thus makeing it easier to
2325 destroy the results when no longer needed.
2327 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2329 * category.c (cat_stored_values_destroy): Fixed memory leak.
2331 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2333 Add some missing frees. Thanks to John Darrington for reporting
2336 * any-writer.c (any_writer_close): Free writer.
2338 * any-reader.c (any_reader_close): Free reader.
2340 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2342 * por-file-reader.c: (error) Mark as NO_RETURN.
2344 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2346 * settings.c: Changed default value of scompress to true.
2348 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2350 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2352 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2355 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2356 here from pref.h.orig.
2358 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2360 * sys-file-reader.c: Fixed bug reading compressed files.
2362 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2364 * Numerous renames. See src/ChangeLog for details.
2366 * Moved files from src directory
projects / pspp-builds.git / blob