1 2008-02-02 Ben Pfaff <blp@gnu.org>
5 * sys-file-reader.c (read_variable_record): Allow missing values
6 to be specified on long string variables, but warn about them
7 because PSPP does not yet support them.
8 (read_extension_record): Ignore extension records 20 and 21, which
9 PSPP does not yet support.
10 (read_header): Fix error message when floating-point format cannot
13 2008-02-01 Ben Pfaff <blp@gnu.org>
15 Patch #6386. Thanks to John Darrington for review and for the
16 updates to gnumeric-reader.c.
18 * dictionary.c (make_hinted_name): New function.
19 (make_numeric_name): New function.
20 (dict_make_unique_var_name): New function.
22 * gnumeric-reader.c (devise_name): Removed.
23 (munge_name): Removed.
24 (gnumeric_open_reader): Use new function
25 dict_make_unique_var_name.
27 * short-names.c (set_var_short_name_suffix): Use new function
30 2008-01-19 John Darrington <john@darrington.wattle.id.au>
32 * settings.c settings.h: Moved static variables into a
33 single struct. Renamed functions to have a settings_ prefix.
35 2008-01-14 John Darrington <john@darrington.wattle.id.au>
37 * data-out.c (output_AHEX): Corrected number of bytes in
38 call to output_hex. Closes bug #22011
40 2008-01-02 John Darrington <john@darrington.wattle.id.au>
42 * variable.c variable.h: Replaced var_get_value_name with
43 var_append_value_name which doesn't use any static data.
44 Thanks to Ben for review.
46 2007-12-07 Ben Pfaff <blp@gnu.org>
50 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
51 made from an empty casereader produce a casegrouper with no
52 groups, instead of one with one group that has no cases.
54 * casereader.c (casereader_is_empty): New function.
56 2007-12-06 Ben Pfaff <blp@gnu.org>
60 * sys-file-reader.c (read_display_parameters): Handle variable
61 display parameters record with only 2 data items per variable.
62 Reported by Guido Gay <gay@irer.it>.
64 2007-12-04 Ben Pfaff <blp@gnu.org>
66 * identifier.c (lex_id_match_n): New function.
67 (lex_id_match): Reimplement in terms of lex_id_match_n.
69 2007-11-24 Ben Pfaff <blp@gnu.org>
71 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
74 2007-11-24 Ben Pfaff <blp@gnu.org>
76 Fix warning reported when reading back system files that include
77 very long string variables. Thanks to Guido Gay <gay@irer.it> for
80 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
81 was written where `j' was meant.
83 2007-11-18 Ben Pfaff <blp@gnu.org>
85 Properly write variables that include a range of missing values to
86 system files. Thanks to Guido Gay <gay@irer.it> for reporting
89 * sys-file-writer.c (write_variable): Correctly calculate
90 n_missing_values field when writing variables that include a range
93 2007-11-10 Ben Pfaff <blp@gnu.org>
95 Cleanups and bug fixes devised while writing up documentation.
98 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
100 * dict-class.c: New file.
101 (dict_class_from_id): Move here.
102 (dict_class_to_name): Move here.
104 * dict-class.h: New file.
105 (enum dict_class): Move here. Change from consecutive integers to
106 consecutive bits, to make testing for any of multiple values
107 easier. Add new DC_ALL constant.
109 * dictionary.c (struct dictionary): Change `case_limit' from
110 size_t to casenumber.
111 (dict_get_vars): Make final argument an enum dict_class.
112 (dict_get_vars_mutable): Ditto.
113 (dict_get_case_limit): Change return value to casenumber.
114 (dict_set_case_limit): Change final argument to a casenumber.
115 (dict_unset_split_var): Add assertion. Rephrase slightly.
116 (dict_set_label): Use xstrndup to simplify.
118 * format.c (fmt_step_width): AHEX format also needs 2-byte
120 (fmt_set_style): Simplify assertions.
122 * missing-values.c (mv_add_num_range): Rename mv_add_range.
123 Simplify implementation.
124 (mv_has_value): Simplify implementation.
125 (mv_pop_value): Remove the first value, not the last, to avoid
126 having GET followed by SAVE reverse the order of missing values.
127 (mv_peek_value): Rename mv_get_value. Simplify assertion.
128 (mv_has_range): Simplify implementation.
129 (mv_peek_range): Rename mv_get_range.
130 (can_resize_string): Removed.
131 (mv_is_resizable): Use value_is_resizable.
132 (mv_resize): Use value_resize.
134 * short-names.h (SHORT_NAME_LEN): Move here.
136 * val-type.h: New file, for definitions related to type and width
137 of abstract values. Before, these definitions were mixed among
138 those related to "union value" and those related to variables.
139 (macro SYSMIS): Move here.
140 (macro LOWEST): Move here.
141 (macro HIGHEST): Move here.
142 (macro MAX_STRING); Move here.
143 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
144 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
145 All references updated.
146 (val_type_is_valid): New function. Replaces var_type_is_valid.
147 All references updated.
148 (val_type_from_width): New function. Replaces
149 var_type_from_width. All references updated.
151 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
153 (val_labs_can_set_width): Use value_is_resizable.
154 (val_labs_add): Simply return false if the value labels set is too
155 wide, instead of having undefined behavior.
156 (val_labs_replace): Ditto.
157 (val_labs_replace): Ditto.
158 (val_labs_first): Set iterator to null if iteration is complete.
159 (val_labs_first_sorted): Ditto.
160 (val_labs_done): Become a no-op if the iterator is null.
162 * value.c (value_is_resizable): New function.
163 (value_resize): New function.
165 * variable.c (var_get_dict_class): New function.
167 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
170 2007-11-08 Ben Pfaff <blp@gnu.org>
172 * data-in.c: Make formatted data parsing locale-independent.
173 (parse_number): Use c_strtod instead of strtod, to avoid
174 locale-specific behavior.
177 2007-11-06 Ben Pfaff <blp@gnu.org>
179 Patch #6256: add support for binary, 360 file formats. Reviewed
182 * data-in.c (struct data_in): Add `encoding' member.
183 (data_in): Add `encoding' parameter, and re-encode the data passed
184 in where appropriate. Update all callers to pass it in.
185 (parse_A): Implement EBCDIC recoding wart described in manual.
186 (parse_AHEX): Implement EBCDIC recoding.
188 * data-out.c (data_out_legacy): New function.
189 (data_out): Make into a wrapper around data_out_legacy.
191 * file-handle-def.c (struct file_handle): New member `encoding'.
192 (fh_create_file): Set encoding.
193 (fh_default_properties): Set default encoding.
194 (fh_get_legacy_encoding): New function.
196 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
197 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
198 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
199 (struct fh_properties): New member `encoding'.
201 2007-11-05 Ben Pfaff <blp@gnu.org>
203 Patch #6258. Reviewed by John Darrington.
205 * file-handle-def.c (fh_lock): Add comment that TYPE should be
206 marked with N_() in the caller. Added these markings to each
207 caller too. Should make i18n easier.
208 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
210 2007-11-03 Ben Pfaff <blp@gnu.org>
212 Allow output files to overwrite input files (bug #21280).
214 * file-handle-def.c: Separate locking of files for input and for
215 output, to allow both to take place at once. Also, distinguish a
216 file handle from the identity of the underlying file, because the
217 identity of a file changes over time and the file handle can't
218 represent two different identities.
219 (struct file_handle): Remove `next', `open_cnt', `deleted',
220 `type', `open_mode', `aux', `identity' members. Change `id' from
221 char array to char *. Add `ref_cnt' member.
222 (file_handle_from_ll) New function.
223 (file_handles) Removed.
224 (named_handles) New variable.
225 (fh_init) Initialize named_handles.
226 (fh_done) Remove name from all named_handles.
227 (free_handle) Updated for modified struct file_handle.
228 (unname_handle) New function.
229 (fh_ref) New function.
230 (fh_from_file_name) Removed.
231 (fh_unref) New function.
232 (fh_unname) New function.
233 (fh_from_id) Rewritten.
234 (create_handle) Updated for modified struct file_handle.
235 (fh_inline_file) Increment the handle's ref count.
236 (fh_create_file) Updated for modified struct file_handle.
237 (fh_create_scratch) Ditto.
242 (fh_is_open) Removed.
243 (fh_get_id) Updated for modified struct file_handle.
244 (fh_get_default_handle) Increment the handle's ref count.
245 (fh_set_default_handle) Handle ref counts.
246 (struct fh_lock) New structure.
247 (locks) New static var.
248 (fh_lock) New function.
249 (fh_unlock) New function.
250 (fh_lock_get_aux) New function.
251 (fh_lock_set_aux) New function.
252 (fh_is_locked) New function.
253 (make_key) New function.
254 (free_key) New function.
255 (compare_fh_locks) New function.
256 (hash_fh_lock) New function.
258 * file-handle-def.h (enum fh_access) New enum.
260 * file-name.c: Made file_identity the same in all supported
262 (struct file_identity): New `name' member.
263 (fn_get_identity): For a file that doesn't exist, get the
264 dev/inode of its directory plus its name. If even the directory
265 doesn't exist, just use its name. Merge the Windows
266 implementation into the Unix one.
267 (fn_compare_file_identities): Rewritten. Merge the Windows
268 implementation into the Unix one.
269 (fn_hash_identity): New function.
271 * make-file.c (struct replace_file): New structure.
272 (all_files): New static var.
273 (replace_file_start): New function.
274 (replace_file_commit): New function.
275 (replace_file_abort): New function.
276 (free_replace_file): New function.
277 (unlink_replace_files): New function.
279 * por-file-reader.c (struct pfm_reader): Add `lock' member.
280 (close_reader): Unlock file.
281 (pfm_open_reader): Lock file.
283 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
285 (pfm_open_writer): Lock file and prepare for its replacement.
286 (close_writer): Unlock file.
288 * scratch-handle.h (struct scratch_handle): Add unique_id so that
289 different generations of a scratch file can be distinguished.
291 * scratch-reader.c (scratch_reader_open): Verify that the file is
294 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
295 remove scratch_handle member.
296 (scratch_writer_open): Lock handle. Prepare to replace handle
297 data, instead of doing it immediately.
298 (scratch_writer_casewriter_destroy): Replace handle data and
301 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
302 (sfm_open_reader): Lock file.
303 (close_reader): Unlock file.
305 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
307 (sfm_open_writer): Lock file and prepare for its replacement.
308 (close_writer): Unlock file.
310 2007-11-02 Ben Pfaff <blp@gnu.org>
312 * data-out.c (output_number): Use isfinite (from C99) instead of
313 gsl_isfinite, since we now have universal support for it in
315 (output_infinite): Ditto.
316 [!HAVE_ROUND] (round): Remove definition, since we now have a
317 replacement in gnulib.
319 2007-10-27 John Darrington <john@darrington.wattle.id.au>
321 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
322 wasn't initialised correctly. Closes bugs #21330 and 21397
325 2007-10-08 Ben Pfaff <blp@gnu.org>
327 Fix bug #21280. Thanks to John Darrington for review.
329 * file-name.c (create_stream): New function.
331 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
332 file handle before creating the file, to ensure that we don't
333 truncate a file that we're reading. Make code easier to read by
336 * sys-file-write.c (sfm_open_writer): Ditto.
338 2007-10-01 Ben Pfaff <blp@gnu.org>
340 Fix bug #21192. Thanks to John Darrington for review.
342 * casereader.c (casereader_read): Decrement case_cnt before
343 calling the casereader's "read" member function, so that we
344 interact properly with lazy_casereader.
346 * datasheet.c: Add regression test for above bug fix.
347 (clone_datasheet): New function.
348 (lazy_callback): New function.
349 (check_datasheet_casereader): New function.
350 (check_datasheet): Check datasheet contents are reported correctly
351 through an ordinary casereader and a lazy casereader.
352 (clone_model): Use clone_datasheet.
354 2007-09-24 Ben Pfaff <blp@gnu.org>
356 Patch #6210. Reviewed by John Darrington.
357 * settings.c: Drop "long view", which is not needed any longer.
358 (static var long_view): Removed.
359 (force_long_view): Removed.
360 (get_viewwidth): Removed.
361 (init_viewport): Removed long_view reference.
363 2007-09-19 John Darrington <john@darrington.wattle.id.au>
365 * settings.c settings.h: Changed viewport's length and width to be
366 owned by the user interface which uses the data library. This allows
367 better abstraction, and makes dynamically adjustable dimensions easier.
369 2007-09-18 Ben Pfaff <blp@gnu.org>
371 * procedure.c (proc_extract_active_file_data): New function.
373 * lazy-casereader.h: New file.
375 * lazy-casereader.c: New file.
377 * casereader.c (casereader_dynamic_cast): New function.
379 2007-09-14 Ben Pfaff <blp@gnu.org>
381 * dictionary.c (dict_clone): Copy case indexes from cloned
382 dictionary. Fixes bug #21061. Reviewed by John Darrington.
384 2007-09-13 John Darrington <john@darrington.wattle.id.au>
386 * value.c value.h (create_value): New function.
388 2007-09-12 Ben Pfaff <blp@gnu.org>
390 Make it clear that translator casereader and casewriter translate
391 functions are supposed to destroy their input case.
393 * casereader-translator.c (struct casereader_translator): Change
394 input case parameter of translate member function from const
395 struct ccase * to struct ccase *.
396 (casereader_create_translator): Ditto, for translate parameter.
398 * casewriter-translator.c (struct casewriter_translator): Ditto.
399 (casewriter_create_translator): Ditto.
401 2007-08-27 John Darrington <john@darrington.wattle.id.au>
403 * sys-file-reader.c (read_display_parameters): Force display width
404 to 8, if the sys file says 0 (like SPSS does).
406 2007-08-12 Ben Pfaff <blp@gnu.org>
408 * dictionary.c (dict_dump): New function.
410 2007-08-12 Ben Pfaff <blp@gnu.org>
412 Drop dict_compactor in favor of using the new struct case_map.
414 * dictionary.c (struct copy_map): Removed.
415 (struct dict_compactor): Removed.
416 (dict_make_compactor): Removed.
417 (dict_compactor_compact): Removed.
418 (dict_compactor_destroy): Removed.
420 * procedure.c (struct data_set): Change `compactor' member to be a
422 (proc_open): Use case_map_to_compact_dict instead of
424 (proc_casereader_read): Use case_map_execute instead of
425 dict_compactor_compact.
426 (proc_commit): Use case_map_destroy instead of
427 dict_compactor_destroy.
429 * scratch-writer.c (struct scratch_writer): Change `compactor'
430 member to be a struct case_map *.
431 (scratch_writer_open): Use case_map_to_compact_dict instead of
433 (scratch_writer_casewriter_write): Use case_map_execute instead of
434 dict_compactor_compact.
436 2007-08-12 Ben Pfaff <blp@gnu.org>
438 * automake.mk: Add case-map.c, case-map.h.
440 * case-map.c: New file.
442 * case-map.h: New file.
444 2007-08-12 Ben Pfaff <blp@gnu.org>
446 * dictionary.c (dict_compact_values): Don't delete scratch
447 variables as well as compacting case indexes. Update all callers.
448 (dict_get_compacted_value_cnt): Rename dict_count_values and
449 change interface. Update all callers.
450 (dict_get_compacted_value_cnt): Remove.
451 (dict_compacting_would_shrink): Remove.
452 (dict_compacting_would_change): Remove.
453 (dict_make_compactor): Add new parameter. Update all callers.
455 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
456 instead of dict_count_values, changing an O(N) operation into
459 2007-08-12 Ben Pfaff <blp@gnu.org>
461 * casereader.c (casereader_read): Don't require cases read by a
462 casereader to be exactly the expected size: as long as they're big
465 2007-08-12 Ben Pfaff <blp@gnu.org>
467 Make casewriters keep track of the number of `union value's in
468 each case. This is useful for two reasons: casewriter_write can
469 then check that the case being written is large enough, and later
470 recipients of the casewriter can determine the size of the case.
472 * casewriter-translator.c (casewriter_create_translator): Add
475 * casewriter.c (struct casewriter): Add value_cnt member.
476 (casewriter_write): Check that the case passed in is big enough.
477 (casewriter_get_value_cnt): New function.
478 (casewriter_create): Add value_cnt parameter.
480 2007-08-09 Ben Pfaff <blp@gnu.org>
482 Fix bug reported by Jason Stover.
483 * settings.c: Move get_termcap_viewport and in particular the
484 #include for <curses.h> to the end of the file. curses.h
485 redefines bool on some systems (e.g. OpenBSD), which causes
486 disagreement between uses of bool before its inclusion and
488 Tested by Jason Stover.
490 2007-07-29 Ben Pfaff <blp@gnu.org>
492 Provisional fix for bug #18692 and bug #20161. Reviewed by John
495 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
496 argument (never "rb" or "wb") because SUSv3 says that only those
497 modes are defined, and glibc in fact rejects other modes.
499 Open portable files with fn_open so that they can be read from
500 pipes. Fix missing fh_close call to go along with fh_open.
501 Report an error if the file close reports an error.
502 * por-file-reader.c (close_reader): New function.
503 (por_file_casereader_destroy): Use close_reader.
504 (pfm_open_reader): Open file with fn_open.
506 2007-07-28 Ben Pfaff <blp@gnu.org>
508 Make PSPP able to read all the portable files I could find on the
509 web. Thanks to John Darrington for review. Bug #17620.
510 * por-file-reader.c (struct pfm_reader): New member `line_length'.
511 (error): Print file offset in hexadecimal.
512 (warning): New function.
513 (advance): Treat lines less than 80 bytes long as padded to 80
515 (pfm_open_reader): Call read_documents if we find an "E" record.
516 (convert_format): Convert invalid formats to the default format
517 instead of aborting reading the file.
518 (read_variables): Rename duplicate variable names instead of
519 aborting reading the file.
520 (read_value_label): Allow string variables of different widths to
521 be assigned value labels in the same record. Replace duplicate
522 value labels instead of aborting.
523 (read_documents): New function.
525 * por-file-writer.c (pfm_open_writer): Call write_documents if the
526 dictionary has documents.
527 (write_documents): New function.
529 2007-07-25 Ben Pfaff <blp@gnu.org>
531 Fix bugs related to bug #17213.
533 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
534 since the former is what config.h has. Include the needed ncurses
536 (static var echo) Rename to `do_echo' because the original name is
537 the same as an ncurses identifier.
538 (get_termcap_viewport) Use error instead of msg.
540 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
541 (fn_close): Don't close stdin, stdout, stderr.
543 2007-07-26 John Darrington <john@darrington.wattle.id.au>
545 * procedure.c procedure.h: Added callbacks which get invoked whenever
546 a dataset's transformation chain changes.
548 2007-07-24 Ben Pfaff <blp@gnu.org>
551 * sys-file-writer.c (write_variable_display_parameters): Use new
552 var_default_display_width function to choose display width of
553 segments after the first one in a given variable.
554 * variable.c (var_create): Use var_default_display_width to pick
555 new variable's display width.
556 (var_default_display_width): New function.
557 Reviewed by John Darrington.
559 2007-07-24 Ben Pfaff <blp@gnu.org>
562 * por-file-writer.c (write_variables): Write weight variable.
563 Reviewed by John Darrington.
565 2007-07-23 Ben Pfaff <blp@gnu.org>
567 Improvements to system file reader and writer.
569 First, move all detailed knowledge of very long strings into
570 sys-file-private.[ch], so that this nasty stuff can be isolated.
572 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
573 (EFFECTIVE_VLS_CHUNK): New macro.
574 (min_int): New function.
575 (max_int): New function.
576 (sfm_width_to_bytes): Rewrite.
577 (sfm_width_to_octs): New function.
578 (sfm_segment_alloc_width): New function.
579 (sfm_segment_alloc_bytes): New function.
580 (sfm_segment_used_bytes): New function.
581 (sfm_segment_offset): New function.
582 (sfm_segment_effective_offset): New function.
583 (sfm_dictionary_to_sfm_vars): New function.
585 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
586 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
587 (struct sfm_var): New structure.
589 Next, improvements to the system file reader.
591 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
592 type casenumber. Added `version_major', `version_minor',
595 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
596 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
597 Change `case_cnt' to type casenumber. Removed `has_vls'.
598 (struct sfm_var): Removed.
599 (sfm_open_reader): Don't warn on wrong case size if the file was
600 written by SPSS 13, which tends to get it wrong. Use
601 sfm_dictionary_to_sfm_vars.
602 (read_header): Always output system file info.
603 (read_variable_record): Simplify code for reading missing values.
604 (read_machine_int32_info): Save version numbers from system file
605 into info struct passed as new argument.
606 (read_long_string_map): Restructured to use new sys-file-private
608 (read_value_labels): Use size_overflow_p.
609 (sys_file_casereader_read): Get rid of distinction between fast
610 and slow paths. Use information provided by sys-file-primate's
611 struct sfm_var to simplify code.
612 (skip_whole_strings): New function.
613 (read_int32): Renamed read_int. Changed return value to int.
615 (read_flt64): Renamed read_float. Changed return value to
616 double. Updated all callers.
617 (int32_to_native): Removed. Changed callers to use
619 (flt64_to_double): Removed. Changed callers to use float_convert.
621 Finally, get rid of int32, flt64 terminology and types in system
622 file writer. The former wasn't very useful since a POSIX "int"
623 can hold the whole range of int32 and we generally didn't have a
624 need for it to be exactly-32-bits, just at-least-32-bits. The
625 latter was inconvenient because we had to assume that it could be
626 different from double and thereby convert special values SYSMIS,
627 HIGHEST, LOWEST to and from it in multiple places. Instead, now
628 we just use "int" and "double" in most places, and do conversions,
629 if necessary, very close to where we do I/O. This change meant
630 that the writer code couldn't represent records in the file as C
631 structs any longer, but that's no great loss. The code actually
632 seems to be more readable without them.
634 Simplify the compression buffering code: only buffer as much as
635 necessary, which is no more than eight 8-byte units at any given
638 * sys-file-writer.c (typedef flt64): Removed.
639 (macro second_lowest_flt64): Removed.
640 (struct sysfile_header): Removed.
641 (struct sysfile_variable): Removed.
642 (struct sfm_writer): Removed `needs_translation', `has_vls',
643 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
644 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
645 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
646 compression buffering by `opcodes', `opcode_cnt', `data',
647 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
648 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
649 data writer functions instead of structures.
650 (calc_oct_idx): New function.
651 (write_header): Use simple data writer functions instead of
653 (write_format_spec): Renamed write_format. New argument.
654 (write_variable_continuation_records): New function.
655 (write_variable): Use simple data writer functions instead of
656 structures. Use write_variable_continuation_records. Write
657 entire very long string instead of requiring caller to understand
659 (write_value_labels): Use simple data writer functions instead of
661 (write_documents): Ditto.
662 (write_variable_display_parameters): Use sys-file-private
663 functions to simplify. Use simple data writer functions instead
665 (write_vls_length_table): Use simple data writer functions instead
667 (write_longvar_table): Ditto.
668 (write_rec_7_34): Break into new functions
669 write_integer_info_record, write_float_info_record. Use simple
670 data writer functions instead of structures.
671 (buf_write): Removed.
672 (append_string_max): Removed.
673 (ensure_buf_space): Removed.
674 (sys_file_casewriter_write): Get rid of the distinction between
675 fast and slow paths, which didn't seem to be too useful. Use new
676 functions write_case_uncompressed, write_case_compressed.
677 (put_instruction): Removed.
678 (put_element): Removed.
679 (write_compressed_data): Removed.
680 (close_writer): Use flush_compressed. Only write case count to
681 system file if it will fit in the field.
682 (write_case_compressed): New function.
683 (write_case_uncompressed): New function.
684 (flush_compressed): New function.
685 (put_cmp_opcode): New function.
686 (put_cmp_number): New function.
687 (write_int): New function.
688 (convert_double_to_output_format): New function.
689 (write_float): New function.
690 (write_value): New function.
691 (write_string): New function.
692 (write_bytes): New function.
693 (write_zeros): New function.
694 (write_spaces): New function.
696 Reviewed by John Darrington.
698 2007-07-22 Ben Pfaff <blp@gnu.org>
700 Don't try to write very long strings to portable files. The
701 format does not support it.
703 * por-file-writer.c (MAX_POR_WIDTH): New macro.
704 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
705 (write_format): Add arg to take width to resize format to.
706 (write_value): Limit width of value written to MAX_POR_WIDTH.
707 (write_variables): Limit width of variable and its output formats
709 Reviewed by John Darrington.
711 2007-07-22 Ben Pfaff <blp@gnu.org>
713 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
714 local variable instead of literal 5.
715 Reviewed by John Darrington.
717 2007-07-22 Ben Pfaff <blp@gnu.org>
719 Fix problems with uniqueness of short names in system files with
720 very long string variables. Now a variable may have multiple
723 * automake.mk (src_data_libdata_a_SOURCES): Add new files
724 short-names.c, short-names.h.
726 * dictionary.c (dict_clone): Clone all the short names.
727 (compare_strings): Move into short-names.c.
728 (hash_strings): Ditto.
729 (set_var_short_name_suffix): Ditto.
730 (dict_assign_short_names): Ditto, rename short_names_assign,
731 change to assign all short names.
733 * por-file-writer.c (write_variables): Use short_names_assign
734 instead of dict_assign_short_names.
736 * short-names.c: New file.
738 * short-names.h: New file.
740 * sys-file-private.c (sfm_width_to_segments): New function.
742 * sys-file-reader.c (read_long_var_name_map): Save and restore all
743 the short names, not just the first one.
745 * sys-file-writer.c (cont_var_name): Removed.
746 (sfm_open_writer): Use short_names_assign instead of
747 dict_assign_short_names. Use unique short names assigned by
748 short_names_assign instead of those generated by cont_var_name.
750 * variable.c (struct variable): Remove `short_name' member,
751 replace by `short_names' and `short_name_cnt'.
752 (var_create) Initialize new members.
753 (var_get_short_name_cnt): New function.
754 (var_get_short_name): Now takes an index argument. Changed most
756 (var_set_short_name): Ditto.
757 (var_clear_short_name): Renamed var_clear_short_names, changed to
758 clear all short names.
760 Reviewed by John Darrington.
762 2007-07-22 Ben Pfaff <blp@gnu.org>
764 * variable.c (var_set_width): Use new var_set_width function.
766 * missing-values.c (mv_n_values): Drop assertion, which was not
769 * format.c (fmt_default_for_width): New function.
770 (fmt_resize): New function.
772 Reviewed by John Darrington.
774 2007-07-18 John Darrington <john@darrington.wattle.id.au>
776 * datasheet.c (datasheet_delete_columns): Added assertion to check
777 we're not deleting outside the range of the sheet.
780 * dictionary.c dictionary.h variable.c: Added the ability for string
781 variables to be resized.
783 * vardict.h: Added some prototypes (moved from dictionary.h) as
784 these should only be called by variable.c
787 2007-07-14 John Darrington <john@darrington.wattle.id.au>
789 * sfm-reader.c: Respect case_cnt field in file header.
791 2007-07-01 John Darrington <john@darrington.wattle.id.au>
793 * transformation.c transformation.h (trns_chain_execute): Changed the
794 signature (Patch #6057)
796 2007-06-10 Ben Pfaff <blp@gnu.org>
798 * casereader-filter.c (casereader_filter_destroy): Make sure to
799 write all the remaining excluded cases to the casewriter, if any.
801 * caseinit.c (init_list_destroy): Rewrite.
802 (init_list_clear): Ditto.
804 * casegrouper.c (casegrouper_get_next_group): Always set *reader
805 to null when returning false.
807 2007-06-06 Ben Pfaff <blp@gnu.org>
809 Actually implement the new procedure code and adapt all of its
810 clients to match. Also adapt all of the other case sources and
811 sinks in the tree and their clients to use the
812 casereader/casewriter infrastructure.
814 * automake.mk: Add and remove files.
816 * any-reader.c: Change into a casereader.
817 * por-file-reader.c: Ditto.
818 * scratch-reader.c: Ditto.
819 * sys-file-reader.c: Ditto.
821 * any-writer.c: Change into a casewriter.
822 * por-file-writer.c: Ditto.
823 * scratch-writer.c: Ditto.
824 * sys-file-writer.c: Ditto.
826 * procedure.c: Change to use casereader, casewriter, caseinit, and
827 other new infrastructure.
829 * scratch-handle.c: Adapt to new infrastructure.
831 * case-sink.c: Removed, now dead code.
832 * case-sink.h: Ditto.
833 * case-source.c: Ditto.
834 * case-source.h: Ditto.
835 * casefile-factory.c: Ditto.
836 * casefile-private.h: Ditto.
839 * casefilter.c: Ditto.
840 * casefilter.h: Ditto.
843 * fastfile-factory.c: Ditto.
844 * fastfile-factory.h: Ditto.
845 * storage-stream.c: Ditto.
846 * storage-stream.h: Ditto.
848 2007-06-06 Ben Pfaff <blp@gnu.org>
852 * automake.mk: Add new files.
854 * datasheet.c: New file.
856 * datasheet.h: New file.
858 2007-06-06 Ben Pfaff <blp@gnu.org>
860 Until now, the procedure code has provided a case to the
861 case_source, which has filled in the data values that come from
862 the active file. "Left" data values that don't come from the
863 active file naturally stay the same from case to case, because the
864 procedure code keeps using that same case.
866 One of the compromises that comes with the new procedure code is
867 that the active file allocates and provides its own case, which
868 the procedure code then has to resize to provide room for any
869 other variables that should go in the case and then fill in the
870 values of "left" variables. Then, when we're done with that case,
871 we have to save the values of "left" variables to copy into the
872 next case read from the active file.
874 The caseinit code helps with this.
876 * automake.mk: Add new files.
878 * caseinit.c: New file.
880 * caseinit.h: New file.
882 2007-06-06 Ben Pfaff <blp@gnu.org>
884 * value.h (value_cnt_from_width): New function.
886 * variable.c (var_get_value_cnt): Use new function.
888 2007-06-06 Ben Pfaff <blp@gnu.org>
890 Add casegrouper, to allow cases read from a given casereader to be
891 broken into groups, each of which has its own casereader.
892 Generally cases are grouped based on having equal values for some
895 * automake.mk: Add new files.
897 * casegrouper.c: New file.
899 * casegrouper.h: New file.
901 2007-06-06 Ben Pfaff <blp@gnu.org>
903 Add interface to lexicographical ordering of cases.
905 * automake.mk: Add new files.
907 * case-ordering.c: New file.
909 * case-ordering.h: New file.
911 2007-06-06 Ben Pfaff <blp@gnu.org>
913 Add casereaders and casewriters, the basis of the new data processing
914 implementation. A casereader is a uniform interface to reading cases
915 from a data source; a casewriter is a uniform interface to writing
916 cases to a data sink.
918 * automake.mk: Add new files.
920 * casereader-filter.c: New file.
922 * casereader-provider.h: New file.
924 * casereader-translator.c: New file.
926 * casereader.c: New file.
928 * casereader.h: New file.
930 * casewriter-provider.h: New file.
932 * casewriter-translator.c: New file.
934 * casewriter.c: New file.
936 * casewriter.h: New file.
938 2007-06-06 Ben Pfaff <blp@gnu.org>
940 "casewindow" data structure that extends the deque (from libpspp)
941 of cases with the ability to dump cases to disk if we get too many
944 * automake.mk: Add new files.
946 * casewindow.c: New file.
948 * casewindow.h: New file.
950 2007-06-06 Ben Pfaff <blp@gnu.org>
952 sparse_cases data structure that augments a sparse_array of cases
953 with the ability to dump cases to disk if we get too many cases in
956 * automake.mk: Add new files.
958 * sparse-cases.c: New file.
960 * sparse-cases.h: New file.
962 2007-06-06 Ben Pfaff <blp@gnu.org>
964 Adds a low-level on-disk case array data structure.
966 * automake.mk: Add new files.
968 * case-tmpfile.c: New file.
970 * case-tmpfile.h: New file.
972 2007-06-06 Ben Pfaff <blp@gnu.org>
974 In a couple of places we calculate the maximum number of cases to
975 keep in memory based on the user-defined workspace. Enable
976 centralizing the calculation through a new function.
978 * settings.c (get_workspace_cases): New function.
980 2007-06-06 Ben Pfaff <blp@gnu.org>
982 The casenumber type is defined in transformations.h, but case.h is
983 a more sensible place. Move it.
985 * case.h (CASENUMBER_MAX): New macro.
986 (typedef casenumber): Move here, from transformations.h.
988 2007-06-03 Ben Pfaff <blp@gnu.org>
990 Slightly generalize case_to_values and case_from_values functions.
992 * case.c (case_to_values): Rename case_copy_out, change interface.
993 (case_from_values): Rename case_copy_in, change interface.
995 * fastfile.c (fastfilereader_get_next_case): Update caller.
996 (write_case_to_disk): Ditto.
998 2007-06-02 Ben Pfaff <blp@gnu.org>
1000 Clean up after a forgotten part of patch #5829.
1002 * casedeque.h: Remove unused file.
1004 * automake.mk: Remove casedeque.h from sources.
1006 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1008 * category.c: Removed redundant #include
1010 2007-05-06 Ben Pfaff <blp@gnu.org>
1012 Abstract the documents within a dictionary a little better.
1013 Thanks to John Darrington for suggestion, initial version, and
1014 review. Patch #5917.
1016 * dictionary.c (struct dictionary): Change `documents' member from
1017 char * to struct string.
1018 (dict_clear): Destroy struct string.
1019 (dict_get_documents): Convert struct string to char *.
1020 (dict_set_documents): Set struct string. Pad to 80-character
1022 (dict_clear_documents): New function.
1023 (dict_add_document_line): New function.
1024 (dict_get_document_line_cnt): New function.
1025 (dict_get_document_line): New function.
1027 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1029 * sys-file-reader.c (read_documents): Use new document functions.
1031 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1033 * sys-file-reader.c: When reading a system file which has no
1034 long name table, automatically create one where the long names
1035 are the lower case versions of the short names.
1037 2007-04-22 Ben Pfaff <blp@gnu.org>
1039 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1040 dict_clear will free most data related to the dictionary.
1041 dict_clear does a decent job, except that dict_set_split_vars on
1042 some systems won't actually free the dict's "split" member.
1043 Instead, it'll allocate a 1-byte region. Fix this.
1045 * value.c (value_copy): New function.
1046 (value_set_missing): Ditto.
1048 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1050 * Deleted existing category.h and moved cat-routines.h into
1051 category.h Encapsulated struct cat_vals better.
1053 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1055 * sys-file-reader.c: When reading a system file which has no
1056 long name table, automatically create one where the long names
1057 are the lower case versions of the short names.
1059 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1061 * sys-file-reader.c: Some versions of Other Software seem to
1062 produce system files with string variables' measure set to
1063 zero. We'll assume these are supposed to be nominal variables.
1065 2007-03-30 Ben Pfaff <blp@gnu.org>
1067 * procedure.c: Adapt to new deque data structure.
1069 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1070 Ben Pfaff <blp@gnu.org>
1072 * file-name.c: Mingw compatibility fixes.
1073 (fn_search_path): Use ISSLASH instead of comparing against '/'
1075 (fn_dir_name): Use dir_name from gnulib.
1076 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1077 (fn_get_identity): Use GetFullPathName instead of canonicalize
1078 from gnulib, because the latter does not fully support
1079 Windows-style path names. Use this implementation based on the
1080 detected presence of Windows instead of the absence of Unix, since
1081 the new implementation is Windows-specific.
1082 (fn_compare_file_identities): In Windows implementation, compare
1083 names case-insensitively.
1085 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1087 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1088 <stdlib.h> to have the same effect.
1090 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1092 * por-file-reader.c: Add missing _() around messages.
1094 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1096 * make-file.c: Include "mkstemp.h", without which linking on
1099 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1101 Reduce platform dependence.
1103 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1104 Everywhere we using this, we really should have just depended on
1105 the shell to expand tildes.
1106 (fn_search_path): Simplify, given that we don't do tilde expansion
1108 (fn_normalize): Removed. Caller changed to use the canonicalize
1110 (fn_get_cwd): Removed. Only user was fn_normalize.
1111 (fn_is_absolute): Really only test for absolute names.
1112 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1114 (fn_readlink): Removed, as it was only used fn_normalize.
1115 (fn_exists): Assume the stat function is available; gnulib does.
1116 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1118 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1120 * dictionary.c (dict_create_vector_assert): New function.
1122 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1124 * file-name.c (fn_normalize): Correct name of function
1125 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1126 for pointing this out.
1128 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1130 We are using a single member in struct file_handle, the "name"
1131 field, for more than one purpose. When it begins with '"', it's a
1132 file name; otherwise, it's a token that can be used to identify
1133 it. When that assertion fires, it's because we searched for the
1134 name case-sensitively as a file name (so that there was no match),
1135 and then we try to insert it case-insensitively as a token, which
1136 fails because duplicates aren't allowed.
1138 Solution: break the two purposes into two separate fields. This
1139 fixes the problem and likely makes the code easier to read too.
1141 Fixes bug #18922. Thanks to John Darrington for bug report and
1144 * file-handle-def.c (struct file_handle): New `id' member.
1145 (fh_from_name): Rename fh_from_id. Update all callers.
1146 (create_handle): New `id' parameter. Update all callers.
1147 (fh_create_file): Ditto.
1148 (fh_get_id): New function.
1150 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1152 * case.c (case_is_null): Change return type to bool.
1154 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1158 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1159 reference counting to share data. This makes it easy for
1160 valgrind, etc. to find accesses to cases that have been destroyed
1161 but have been kept around by another user's ref-count. This often
1162 happens when the data set is small enough to find in memory; if a
1163 bigger data set that would overflow to disk were used, then data
1164 corruption would occur.
1166 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1170 * case.c (case_unshare): Make it check internally whether the
1171 ref_cnt is greater than 1, so that the callers don't have to.
1172 Update callers not to check.
1174 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1176 Before, I was thinking that I might want to get rid of reference
1177 counting at some point. Now, I'm pretty sure that it's here to
1178 stay. Thus, because we have to store the value_cnt anyway for
1179 reference-counted cases, we might as well expose it to users.
1181 * case.c (case_get_value_cnt): New function.
1182 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1183 resize case if its size actually changed.
1185 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1186 instead of peeking inside struct case directly.
1187 (casefile_append): Ditto.
1189 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1191 Get rid of the inlines for the case functions, which made the
1192 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1193 the inlines didn't speed up "make check" at all, which is not a
1194 perfect benchmark but seems indicative.)
1196 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1197 definitions. Remove some assertions on nonnull pointers that were
1198 redundant with a pointer dereference soon after in the function.
1200 (struct case_data): Move definition here from case.h.
1204 (case_data_wr): Ditto.
1206 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1208 * automake.mk: Add casedeque.h to sources.
1210 * casedeque.h: New file.
1212 * procedure.c: (struct dataset) Change lag_count, lag_head,
1213 lag_queue member into single struct casedeque member. Update all
1214 users to use the casedeque instead.
1217 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1219 * procedure.c: Simplify lagged cases interface. Updated all
1220 clients--well, the only client--to use the simplified interface.
1221 (dataset_n_lag) Removed.
1222 (dataset_set_n_lag) Removed.
1223 (dataset_need_lag) New function.
1225 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1227 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1228 whenever appropriate, but only when the dataset/dictionary is in a
1231 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1233 * dictionary.c dictionary.h : Added callbacks for change of filter and
1234 split variables. Refactored some code to ensure that callbacks get
1235 invoked when appropriate.
1237 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1238 replace_dict callback occurs when permanent_dict replaces the current
1241 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1243 * dictionary.c dictionary.h : Added callback for when the weight
1244 variable of a dictionary changes.
1246 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1248 * dictionary.c dictionary.h : Added replace_source and replace_dict
1249 callbacks, and functions to deal with them.
1251 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1253 Simplify missing value handling.
1255 * missing-values.h (enum mv_class): New type.
1256 (enum mv_type): Moved definition into missing-values.c and renamed
1257 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1258 enums. Updated all uses.
1259 (struct missing_values): Changed type of `type' from `enum
1260 mv_type' to `int' because the definition is no longer exposed.
1262 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1263 parameter. Update all callers.
1264 (mv_is_num_missing): Ditto.
1265 (mv_is_str_missing): Ditto.
1266 (mv_is_value_user_missing): Removed. Changed callers to use
1267 mv_is_value_missing.
1268 (mv_is_num_user_missing): Removed. Changed callers to use
1270 (mv_is_str_user_missing): Removed. Changed callers to use
1272 (mv_is_value_system_missing): Removed. Changed callers to use
1273 mv_is_value_missing.
1274 (mv_set_type): Removed. Changed callers to use mv_clear.
1275 (mv_clear): New function.
1277 * variable.c (var_is_value_missing): Add new enum mv_class
1278 parameter. Update all callers.
1279 (var_is_num_missing): Ditto.
1280 (var_is_str_missing): Ditto.
1281 (var_is_value_user_missing): Removed. Changed callers to use
1282 var_is_value_missing.
1283 (var_is_num_user_missing): Removed. Changed callers to use
1285 (var_is_str_user_missing): Removed. Changed callers to use
1287 (var_is_value_system_missing): Removed. Changed callers to use
1288 var_is_value_missing.
1290 * casefilter.c (struct casefilter): Use enum mv_class in place of
1292 (casefilter_variable_missing): Adapt to new member.
1293 (casefilter_create): Change signature to take enum mv_class,
1296 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1298 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1300 * case-sink.c case-sink.h procedure.c procedure.h
1301 storage-stream.c: Now uses the factory.
1303 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1305 Make it possible to pull cases from the active file with a
1306 function call, instead of requiring indirection through a callback
1309 * case-source.h (struct case_source_class): Change ->read function
1310 to return a single case, instead of calling a callback function
1311 for each case. Change ->destroy function to return an error
1314 * case-source.c (free_case_source): Pass along the value returned
1315 by the case_source ->destroy function.
1317 * procedure.c (struct write_case_data): Removed.
1318 (struct dataset): Added some members to track procedure state.
1319 (procedure): Optimize the trivial case at this level.
1320 (internal_procedure): Re-implement in terms of proc_open,
1321 proc_read, proc_close.
1322 (proc_open) New function.
1323 (proc_read) New function.
1324 (proc_close) New function.
1325 (write_case) Moved into proc_read.
1326 (close_active_file) Moved closing of data source into proc_close.
1328 * storage-source.c: Rewrote to conform with modified
1329 case_source_class interface.
1331 * transformations.c (trns_chain_execute): Added argument to allow
1332 starting execution from an arbitrary transformation. Updated
1335 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1337 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1339 * sys-file-reader.c (read_display_parameters): Don't assume that
1340 MEASURE_* and ALIGN_* have the same values found in system files.
1342 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1344 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1345 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1346 the value of n_MEASURES, which was off by 1 (at least from my
1349 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1351 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1352 callbacks which are invoked when the dictionary or its
1353 variables are changed.
1355 * missing-values.c missing-values.h value-labels.c: Tidied up
1356 consistency checks, and made some of them return false
1357 instead of assert-failing.
1359 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1361 * calendar.c (calendar_days_in_month): New function.
1363 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1365 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1366 the value label's width.
1368 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1370 * sfm-private.h: Move contents into sys-file-writer.c, which is
1371 the only remaining user. Removed Borland C++-specific directives.
1373 * sys-file-reader.c: Clean up and rewrite entire file. The
1374 rewritten version is simpler and better abstracted, and should be
1375 easier to maintain and extend. It avoids using structures to read
1376 file data, which is prone to padding variations among compilers.
1377 It should also handle non-IEEE 754 system files, although I
1378 haven't been able to find any. It has been tested against many
1379 .sav files obtained from the Web and found to produce the same
1380 results as the earlier version of the code, or in some cases
1381 improved results. It is more tolerant of format variations found
1384 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1385 member, putting an enum integer_format in its place. New member
1386 `float_format'. Changed `compressed' member to type bool.
1388 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1390 * dictionary.c (dict_delete_consecutive_vars): New function.
1392 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1394 * file-name.c (fn_search_path): Remove prefix arg that was unused
1395 by any caller. Updated all callers.
1397 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1399 * format.c (fmt_dollar_template): Use user's decimal point
1400 character. Add assertion.
1402 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1404 * format.c (fmt_dollar_template): New function, based on
1405 dollar_format_template from var-type-dialog.c.
1407 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1409 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1412 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1414 Finish converting struct variable to an opaque type. In this
1415 phase, we add remaining setter and getter functions, convert the
1416 remaining PSPP code to use them, and do a bunch of cleanup. The
1417 resulting changes are pervasive but mostly trivial, and only the
1418 notable changes are logged.
1420 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1423 * case.c (case_data): Renamed case_data_idx.
1424 (case_num): Renamed case_num_idx.
1425 (case_str): Renamed case_str_idx.
1426 (case_data_rw): Renamed case_data_rw_idx.
1428 * case.h (case_data): New function with old name and an interface
1429 that takes a variable instead of an index, which is easier to
1430 use. Updated all callers to use the new interface, or to use the
1431 new *_idx function (see above).
1434 (case_data_rw): Ditto.
1436 * category.c (cat_stored_values_destroy): Changed interface to
1437 take a struct cat_vals * instead of a struct variable *.
1439 * dictionary.c (dict_clone): Use new vector_clone function.
1440 (dict_clear) Use new var_destroy function.
1441 (add_var) New function.
1442 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1443 (dict_create_var_assert) Rewrite in terms of add_var.
1444 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1445 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1446 (dict_lookup_var) Use new var_create, var_destroy functions.
1447 (dict_contains_var) Rewrite in terms of new vardict functionality.
1448 (set_var_dict_index) New function.
1449 (set_var_case_index) New function.
1450 (reindex_vars) New function.
1451 (dict_delete_var) Rewrite in terms of new vardict functionality.
1452 (dict_reorder_var) Ditto.
1453 (dict_reorder_vars) Ditto.
1454 (rename_var) New function.
1455 (dict_rename_var) Use rename_var.
1456 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1457 (dict_get_compacted_idx_to_fv) Rename
1458 dict_get_compacted_dict_index_to_case_index, update callers.
1459 (dict_create_vector) Use new vector_create function.
1460 (dict_clear_vectors) Use new vector_destroy function.
1461 (set_var_short_name_suffix) Move here from variable.c, renamed
1462 from var_set_short_name_suffix, make static, update caller.
1464 * sys-file-private.c: New file.
1465 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1466 width_to_bytes, update callers.
1468 * sys-file-private.h: New file. Later it will supplant
1469 sfm-private.h; for now it supplements it.
1470 (macro MIN_VERY_LONG_STRING) New macro.
1471 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1473 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1474 MAX_LONG_STRING was used before.
1476 * sys-file-writer.c: Ditto.
1478 * value-labels.c: Change the paradigm here to be that a null
1479 pointer is OK for a struct val_labs * in most cases; it just
1480 represents an empty set of value labels.
1481 (val_labs_copy) A copy of a null set is a null set.
1482 (val_labs_count) A null set has 0 labels.
1483 (val_labs_replace) Change return type to void. Rewrite for
1485 (val_labs_find) A null set does not contain the value.
1486 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1487 transposed argument order, updated all callers.
1489 * value.c: New file.
1490 (value_dup) Moved here from variable.c.
1491 (compare_values) Ditto.
1494 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1495 (macro MAX_LONG_STRING) Removed, because it was only interesting
1496 for system files, not for general code.
1497 (macro MAX_VERY_LONG_STRING) Ditto.
1498 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1499 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1501 * vardict.h: New file, for an interface between variables and
1504 * variable.c: A lot of functions were moved around, for better
1506 (struct variable) Move definition here, from variable.h.
1507 (var_type_adj) Removed--makes i18n hard.
1508 (var_type_noun) Ditto.
1509 (var_create) New function.
1510 (var_clone) New function.
1511 (var_destroy) New function.
1512 (var_set_name) Assert that variable is not in a dictionary.
1513 (compare_var_names) Rename compare_vars_by_name and fix a couple
1514 of callers who thought the args were strings.
1515 (hash_var_name) Rename hash_var_by_name.
1516 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1517 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1518 (var_is_very_long_string) Removed, because it was only interesting
1519 to system file code.
1520 (var_set_missing_values) Allow the argument to be the wrong width,
1521 as long as we can resize it. Simplify callers who were doing the
1522 resizing themselves.
1523 (var_get_value_labels) New function.
1524 (var_has_value_labels) New function.
1525 (var_set_value_labels) New function.
1526 (alloc_value_labels) New function.
1527 (var_add_value_label) New function.
1528 (var_replace_value_label) New function.
1529 (var_clear_value_labels) New function.
1530 (var_lookup_value_label) New function.
1531 (var_get_value_name) Moved here from variable.c, renamed from
1532 var_get_value_name, transposed argument order, updated all
1534 (var_to_string) Moved here, from variable-label.c.
1535 (var_set_leave) New function.
1536 (var_get_leave) New function.
1537 (var_must_leave) New function.
1538 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1539 set_var_short_name_suffix.
1540 (var_get_dict_index) New function.
1541 (var_get_case_index) New function.
1542 (var_get_obs_vals) New function.
1543 (var_set_obs_vals) New function.
1544 (var_has_obs_vals) New function.
1545 (var_get_vardict) New function.
1546 (var_set_vardict) New function.
1547 (var_has_vardict) New function.
1548 (var_clear_vardict) New function.
1549 (value_dup) Moved to value.c.
1550 (compare_values) Ditto.
1553 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1554 (enum ALPHA) Rename VAR_STRING, update all users.
1556 * vector.c: New file.
1557 (struct vector) Moved here, from variable.h.
1558 (check_widths) New function.
1559 (vector_create) New function.
1560 (vector_clone) New function.
1561 (vector_destroy) New function.
1562 (vector_get_name) New function.
1563 (vector_get_var) New function.
1564 (vector_get_var_cnt) New function.
1565 (compare_vector_ptrs_by_name) New function.
1567 * vector.h: New file.
1569 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1571 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1572 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1575 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1577 * value-labels.c (destroy_atoms): New function.
1578 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1580 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1582 Thanks to Jason Stover for pointing out this problem.
1584 * data-out.c (output_number): Use gsl_finite from GSL, which is
1585 portable, instead of isfinite, which is not.
1588 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1590 * variable.c variable.h (value_dup): New function.
1592 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1594 Start converting struct variable to an opaque type. In this
1595 phase, we add a bunch of setter and getter functions and convert
1596 most of the PSPP code to use them. The resulting changes are
1597 pervasive but mostly trivial, and only the notable changes are
1600 * format.c (fmt_equal): New function.
1602 * variable.c (var_type_is_valid): New function.
1603 (measure_is_valid) Moved here, from format.c.
1604 (alignment_is_valid) Moved here, from format.c.
1605 (var_get_name) New function.
1606 (var_set_name) New function.
1607 (width_to_type) New function.
1608 (var_get_type) New function.
1609 (var_get_width) New function.
1610 (var_set_width) New function.
1611 (var_is_numeric) New function.
1612 (var_is_alpha) New function.
1613 (var_is_short_string) New function.
1614 (var_is_long_string) New function.
1615 (var_is_very_long_string) New function.
1616 (var_get_missing_values) New function.
1617 (var_set_missing_values) New function.
1618 (var_clear_missing_values) New function.
1619 (var_has_missing_values) New function.
1620 (var_is_value_missing) New function.
1621 (var_is_num_missing) New function.
1622 (var_is_str_missing) New function.
1623 (var_is_value_user_missing) New function.
1624 (var_is_num_user_missing) New function.
1625 (var_is_str_user_missing) New function.
1626 (var_is_value_system_missing) New function.
1627 (var_get_print_format) New function.
1628 (var_set_print_format) New function.
1629 (var_get_write_format) New function.
1630 (var_set_write_format) New function.
1631 (var_set_both_formats) New function.
1632 (var_get_label) New function.
1633 (var_set_label) New function.
1634 (var_clear_label) New function.
1635 (var_has_label) New function.
1636 (var_get_measure) New function.
1637 (var_set_measure) New function.
1638 (var_get_display_width) New function.
1639 (var_set_display_width) New function.
1640 (var_get_alignment) New function.
1641 (var_set_alignment) New function.
1642 (var_get_value_cnt) New function.
1643 (var_get_leave) New function.
1644 (var_get_short_name) New function.
1646 * variable.h: (struct variable) Removed "type" and "nv" members;
1647 they are now computed from "width" where needed.
1649 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1651 * missing-values.c (mv_resize): Don't write beyond end of the
1652 allocated buffer when resizing a long string.
1654 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1656 Clean up identifier code: don't require identifier enumerations to
1657 be in a particular order; make better use of string library;
1658 expose less of the internals.
1660 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1661 change interface. Updated all callers.
1662 (lex_id_match) Change interface to use struct substring, update
1664 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1665 (global array keywords[]) Make static, change form. Update all
1666 users to use lex_id_name instead.
1667 (lex_is_keyword) New function.
1668 (lex_id_to_token) Change interface to use struct substring, update
1670 (lex_id_name) New function.
1672 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1673 lex_is_keyword instead.
1674 (T_LAST_KEYWORD) Removed.
1675 (T_N_KEYWORDS) Removed.
1677 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1679 * format.c: (fmt_date_template) Distinguish characters for which a
1680 space is output and any date delimiter is allowed on input, from
1681 those for which a space is output and only a space is allowed on
1682 input. The former is represented by X, the latter by a space.
1683 Also, drop distinction between h and H, changing the former to the
1686 * data-in.c: Completely rewrite internals to conform to SPSS input
1687 formats as closely as possible.
1688 (data_in) Changed external interface by replacing the structure
1689 that was used as a single argument by a set of arguments. Updated
1691 (data_in_finite_line) Removed. Converted all callers to use plain
1693 (data_in_get_integer_format) New function.
1694 (data_in_set_integer_format) New function.
1695 (data_in_get_float_format) New function.
1696 (data_in_set_float_format) New function.
1698 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1699 (struct data_in) Removed.
1701 * data-out.c: (output_date) Drop each component from the input as
1702 it is output, to allow us to drop the distinction between h (a
1703 count of hours) and H (the hour of day) template characters.
1704 Also, handle new X template character.
1705 (output_scientific) Follow more rational rule on when to drop
1706 fraction introduced between SPSS 13 and 15. Updated test case to
1709 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1711 Fix buffer overflow reported by John Darrington.
1713 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1714 realize that DIGITS is a count of nibbles, not of bytes.
1716 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1718 * calendar.c (calendar_offset_to_gregorian) Also return the
1719 year-of-day. Change callers to new interface.
1721 * data-out.c: Completely rewrite internals to conform to SPSS
1722 output formats as completely as possible.
1723 (data_out) Change interface to put input parameters before output
1724 parameters, for consistency with the style I now prefer. Update
1726 (data_out_get_integer_format) New public function.
1727 (data_out_set_integer_format) New public function.
1728 (data_out_get_float_format) New public function.
1729 (data_out_set_float_format) New public function.
1731 * data-out.h: New file. Move prototype for data_out here, from
1734 * format.c: (fmt_step_width) Use equality comparison instead of
1735 bitwise and, for clarity.
1736 (fmt_is_string) Ditto.
1737 (fmt_input_to_output) Fix categories that are translated to F
1740 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1742 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1743 casefile-private.h: Added casefilter to assist commands with missing
1746 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1748 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1750 * settings.c: (route_errors_to_terminal) New variable.
1751 (route_errors_to_listing) New variable.
1752 (get_error_routing_to_terminal) New function.
1753 (set_error_routing_to_terminal) New function.
1754 (get_error_routing_to_listing) New function.
1755 (set_error_routing_to_listing) New function.
1757 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1759 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1761 * format.c: Completely rewrite, to achieve better abstraction.
1762 Rewrite all references to formats in other files.
1764 * format.def: Rewrite and reorganize.
1766 * settings.c: Move everything related to custom currency formats
1767 into format.[ch], changing them in form, so as to group related
1768 code and definitions better. Changed all references to use the
1770 (static var decimal) Removed.
1771 (static var grouping) Removed.
1772 (static var cc) Removed.
1773 (get_decimal) Removed.
1774 (set_decimal) Removed.
1775 (get_grouping) Removed.
1776 (set_grouping) Removed.
1780 * settings.h: (macro CC_CNT) Removed.
1781 (macro CC_WIDTH) Removed.
1782 (struct custom_currency) Removed.
1784 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1786 * data-in.c (data_in): Use switch statement instead of table, to
1787 avoid dependence on the order of the FMT_* enums.
1789 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1791 * data-out.c: (num_to_string) Removed, because it was dead code.
1793 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1795 * data-in.c (parse_trailer): Fix error message.
1797 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1799 * format.c (fmt_is_binary): New function.
1801 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1803 * procedure.c procedure.h: Encapsulated the static data into a single
1806 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1808 * casefile.c (casereader_read_xfer): Always initialize the case,
1809 even on an error condition.
1811 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1813 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1814 entire function useless.
1816 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1818 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1819 casereader method casereader_clone.
1821 * procedure.c pransformations.h: Introduced new type casenum_t
1823 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1825 * variable.c: (width_to_bytes) Rephrase code for clarify.
1827 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1829 * format.c: (fmt_type_from_string) New function.
1830 (fmt_to_string) Include decimals in output if the format has
1831 decimals, even if the format type does not. This way, we can
1832 accurately reproduce incorrect formats in user output.
1833 (check_common_specifier) Make the check for a bad format type an
1834 assertion, so we get bug reports if they show up. Fix message.
1835 Check for decimal places with a format type that doesn't allow
1837 (check_input_specifier) Remove check for FMT_X, which has been
1839 (check_output_specifier) Ditto.
1841 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1843 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1844 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1845 (enum fmt_parse_flags) Removed.
1847 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1849 * casefile.c casefile.h: Converted to an abstract base class.
1850 * casefile-private.h fastfile.c fastfile.h: New files.
1851 * automake.mk procedure.c scratch-writer.c storage-stream.c
1853 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1855 * procedure.c (internal_procedure): Create sink_case with only as
1856 many values as the compacted dictionary.
1858 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1860 Remove "debugging" code that caused plenty of false positives and
1863 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1865 * case.c: Remove all references to `this' member.
1867 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1869 Fix link error noted by Jason Stover.
1871 * storage-stream.c: Include <assert.h>.
1873 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1875 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1876 ALL) and additional underlying system file issues.
1878 Thanks to John Darrington for review.
1880 First problem: var_hash points to variables not owned by the
1881 sys-file-reader, which the caller may free or modify. Use an
1882 array of sfm_vars instead, as done earlier (e.g. CVS version
1885 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1886 members and remove all code that references it. Add vars, var_cnt
1887 members. Remove fix_specials member, which was unused.
1888 (struct sfm_var) Remove name member, which was unused.
1889 (sfm_close_reader) Free vars member instead of var_hash.
1890 (compare_var_shortnames) Removed.
1891 (hash_var_shortname) Removed.
1892 (sfm_open_reader) Fill out vars array.
1893 (compare_var_index) Removed.
1894 (sfm_read_case) Use vars instead of var_hash.
1896 Second problem: we're confused about when we actually have very
1897 long strings, causing us to choose incorrectly between slow path
1898 and fast path in sfm_read_case.
1900 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1901 have very long strings, not when we have long variable names,
1902 which is an unrelated feature.
1904 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1906 * variable.h: Move var_set and variable parsing declarations to
1907 new header, src/language/lexer/variable-parser.h. Modified lots
1908 of files to include the new header.
1910 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1912 * value-labels.c (value_to_string): When there's no value label,
1913 format the variable according to its print format, instead of
1914 always effectively using A or F format.
1916 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1918 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1921 * format.c: Removed tortological assertion.
1923 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1925 Reform string library.
1927 * file-name.c (fn_interp_vars): Change interface to take a
1928 substring as input. Updated all users.
1930 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1932 * format.c (measure_is_valid): Really return false when m >=
1935 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1937 Implement random access to casefiles, for use in GUI.
1939 * casefile.c: (struct casereader) Add `random', `file_ofs',
1940 `buffer_ofs' members.
1941 (casefile_get_random_reader) New function.
1942 (read_open_file) Break part into new function
1943 seek_and_fill_buffer().
1944 (fill_buffer) Update buffer_ofs, file_ofs.
1945 (casereader_seek) New function.
1947 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1949 * settings.c: Added call to i18n{done, init}.
1951 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1953 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1956 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1958 * casefile.c: Convert many uses of `int' to `bool'.
1960 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1962 * transformations.c (trns_chain_destroy): Destroy chain's trns
1963 member, to fix memory leak.
1965 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1967 * storage-stream.c (storage_source_decapsulate): Destroy case
1968 source to fix memory leak.
1970 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1972 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1973 case passed as argument instead of initializing the argument as a
1974 case. Fixes memory leak that showed up in
1975 tests/command/aggregate.sh with scratch files.
1977 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1979 * procedure.c (proc_done): Destroy default_dict, to fix memory
1982 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1984 Simplify procedure_with_splits().
1986 * procedure.c (struct split_aux_data): Removed case_count member.
1987 (procedure_with_splits) Don't initialize case_count.
1988 (split_procedure_case_func) Check whether prev_case is null
1989 instead of case_count.
1990 (split_procedure_end_func) Ditto.
1992 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1994 * case.c (case_move): Do nothing if dst and src are the same
1996 (case_try_create) Merge two similar cases.
1997 (case_copy) Unshare only if data must be actually copied.
1999 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2001 * data-in.c data-out.c dictionary.c sys-file-reader.c
2002 sys-file-writer.c variable.c variable.h: Reworked very long string
2003 support for better encapsulation.
2005 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2007 * value-labels.c (val_labs_can_set_width): New function.
2008 (val_labs_set_width) Clear labels if increasing width to long
2010 (val_labs_destroy) Remove unneeded test for null.
2012 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2014 * value-labels.h: Remove unneeded dependency on variable.h.
2016 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2018 Get rid of `char *c' member in union value, for cleanliness.
2020 * value.h: (union value) Remove `c' member.
2022 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2024 Make dictionary compacting functions a little more general.
2026 * sys-file-writer.c (sfm_open_writer): Use
2027 dict_compacting_would_change().
2028 (does_dict_need_translation) Removed.
2030 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2032 Make dictionary compacting functions a little more general.
2034 * dictionary.c (dict_needs_compaction): Rename
2035 dict_compacting_would_shrink(). Update all callers.
2036 (dict_compacting_would_change) New function.
2038 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2040 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2041 inverted return value (!).
2043 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2045 Continue reforming procedure execution.
2047 * procedure.c: Search and replace "vfm" by "proc". Notably:
2048 (static var vfm_source) Rename proc_source. Update all
2050 (static var vfm_sink) Rename proc_sink. Update all references.
2052 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2054 Continue reforming procedure execution. In this phase, remove
2055 PROCESS IF, which was deprecated anyway and can be easily
2056 simulated with TEMPORARY followed by SELECT IF.
2058 * procedure.c: (open_active_file) Don't call
2059 add_process_if_trns().
2060 (discard_variables) Get rid of redundant call to
2061 proc_cancel_all_transformations().
2062 (add_process_if_trns) Removed.
2063 (process_if_trns_proc) Removed.
2064 (process_if_trns_free) Removed.
2066 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2068 Continue reforming procedure execution. In this phase, add
2069 `const' to the case passed to procedure()'s callback.
2071 Updated all users of procedure() as well.
2073 * procedure.c: (struct write_case_data) Add "const" to ccase
2074 parameter for case_func member.
2075 (procedure) Add "const" to ccase parameter for proc_func
2077 (multipass_case_func) Make ccase parameter const.
2078 (internal_procedure) Add "const" to ccase parameter for case_func
2080 (split_procedure_case_func) Make ccase parameter const.
2081 (multipass_split_case_func) Make ccase parameter const.
2083 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2085 Continue reforming procedure execution. In this phase, get rid of
2086 the output code for SPLIT FILE groups in procedure.c, which really
2087 shouldn't be doing any output. Move it into the individual
2088 procedures instead. This also adds some flexibility.
2090 Updated many users of procedure_with_splits() and
2091 multipass_procedure_with_splits() to call
2092 output_split_file_values() and to deal with increased use of
2095 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2096 parameter to begin_func member.
2097 (procedure_with_splits) Add "const struct ccase *" parameter to
2098 begin_func parameter. Make ccase parameter const in proc_func
2100 (split_procedure_case_func) Don't dump split file group. Pass
2102 (dump_splits) Moved to language/dictionary/split-file.c as
2103 output_split_file_values().
2104 (struct multipass_split_aux_data) Add "const struct ccase *"
2105 parameter to split_func member.
2106 (multipass_procedure_with_splits) Add "const struct ccase *"
2107 parameter to split_func parameter.
2108 (multipass_split_case_func) Save new SPLIT FILE case before
2110 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2112 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2114 Continue reforming procedure execution. Change
2115 internal_procedure() so that it calls open_active_file() and
2116 close_active_file(), which isolates most of the actual procedure
2119 * procedure.c: (struct write_case_data) Rename `proc_func' member
2120 to `case_func' and update all references.
2121 (procedure) Rewrite as one-line wrapper around
2122 internal_procedure().
2123 (struct multipass_aux_data) New.
2124 (multipass_callback) Renamed multipass_case_func(). Use struct
2125 multipass_aux_data as auxiliary data.
2126 (multipass_end_func) New function.
2127 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2128 that uses multipass_case_func, multipass_end_func.
2129 (internal_procedure) Add `end_func' argument. Move optimization
2130 of trivial case in here. Move call to open_active_file() and
2131 close_active_file() in here. Now assert that vfm_source is
2133 (procedure_with_splits_callback) Rename
2134 split_procedure_case_func().
2135 (split_procedure_end_func) New function.
2136 (multipass_split_callback) Rename multipass_split_case_func.
2137 (multipass_split_end_func) New function.
2138 (discard_variables) No need to test for nonnull vfm_source.
2140 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2142 Continue reforming procedure execution. Get rid of unused member.
2144 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2146 (write_case) Don't increment cases_analyzed.
2148 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2150 Continue reforming procedure execution. In this phase, move
2151 procedure.c and procedure.h from src to src/data. Update
2152 makefiles and #includes accordingly.
2154 * procedure.c: Moved here from src/.
2156 * procedure.h: Moved here from src/.
2158 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2160 Continue reforming procedure execution. In this phase, get rid of
2161 many global variables, consolidating procedure execution in
2162 procedure.c. Encapsulate transformations in new "struct
2163 trns_chain". Also, change implementation of N OF CASES, FILTER,
2164 and PROCESS IF from special cases to transformations.
2166 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2169 * dictionary.c: (global variable default_dict) Move to
2172 * variable.h: (TRNS_*) Move to transformations.h.
2173 (struct transformation) Move to transformations.c.
2175 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2177 * sys-file-reader.c: Fixed invalid read problems.
2179 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2181 * storage-stream.c: Add missing function comments.
2183 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2185 Continue reforming procedure execution. In this phase, add some
2186 new, needed functionality to storage-stream.
2188 * storage-stream.c: (storage_source_decapsulate) New function.
2190 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2192 * variable.c (width_to_bytes): Declarations must precede
2193 statements for C90 compliance.
2195 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2197 * data-out.c, data-in.c, variable.c, variable.h: New functions
2198 copy_mangle and copy_demangle for reading/writing cases; emulates the
2199 way SPSS deals with strings > 255 bytes.
2201 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2202 subtype 14 needed for strings longer than 255 bytes.
2204 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2205 instead of literal values. Also fixed some constness issues.
2207 * format.h: Constness
2209 * sfm-private.h: Renamed the case_size identifier, since I discovered
2210 that SPSS's respect for this variable is very nominal.
2212 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2214 Change case limit type from int to size_t.
2216 * dictionary.c: (struct dictionary) Change type of case_limit
2218 (dict_get_case_limit) Change return type.
2219 (dict_set_case_limit) Change parameter type.
2221 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2223 * variable.h: (struct variable) Rename `reinit' member as `leave'
2224 and invert sense. Fix up all references.
2226 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2228 Continue reforming procedure execution. In this phase, break
2229 procedure.c into multiple files.
2231 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2233 * case-sink.c: New file.
2235 * case-sink.h: New file.
2237 * case-source.c: New file.
2239 * case-source.h: New file.
2241 * storage-stream.c: New file.
2243 * storage-stream.h: New file.
2245 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2247 * variable.h: (struct variable) Remove `init' member and all
2248 references to it from other files. It was initialized in several
2249 places, but nothing really ever used it for anything worthwhile.
2250 Thanks to Jason Stover for pointing out how confusing this
2253 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2255 Continue reforming error message support. In this phase, get rid
2256 of message "titles" and put the message text in `struct error'.
2257 Now `struct error' encapsulates a message more properly.
2259 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2260 Format message ourselves.
2262 * data-in.c: (vdls_error) Ditto.
2264 * por-file-reader.c: (error) Ditto.
2266 * sys-file-reader.c: (corrupt_msg) Ditto.
2268 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2270 GNU standards require "file name" instead of "filename" in
2271 documentation. It's nice for our code to follow the convention
2274 * casefile.c: (struct casefile) Rename `filename' member to
2275 `file_name'. Updated all references.
2277 * file-name.c: [!unix] (struct file_identity) Rename
2278 normalized_filename member to normalized_file_name. Updated all
2281 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2283 We don't really support anything but Unix-like environments well,
2284 so we might as well de-obfuscate by writing directory and path
2285 separators explicitly.
2287 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2289 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2290 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2292 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2295 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2297 GNU standards require "file name" instead of "filename" in
2298 documentation. It's nice for our code to follow the convention
2301 * filename.c: Rename to file-name.c.
2303 * filename.h: Rename to file-name.h. Update all inclusions.
2304 Update header guards.
2306 * automake.mk: Update file names.
2308 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2310 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2312 (fn_basename) Removed (dead code).
2313 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2314 (fn_special_p) Renamed fn_is_special(), all references updated.
2315 (fn_exists_p) Renamed fn_exists(), all references updated.
2317 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2319 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2320 Also, now it only tilde-expands file names, not paths.
2321 (fn_search_path) Tilde-expand one directory at a time.
2323 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2325 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2326 get rid of non-Unixlike version of the code, which has probably
2328 (fn_prepend_dir) Removed (dead code).
2330 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2331 (macro PATH_SEPARATOR_STRING) New.
2332 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2334 Continue reforming error message support. In this phase, we get
2335 rid of VM() and the other msg() support for "verbosity", replacing
2336 it by a new function verbose_msg().
2338 * filename.c: (fn_search_path) Use verbose_msg() instead of
2341 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2343 * sfm-private.h: Get rid of #defines after #error, which makes no
2346 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2348 Get rid of our own int32 type in favor of the standard int32_t
2351 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2354 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2356 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2358 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2360 Remove ill-considered file routines that are no longer used.
2362 * filename.c: (fn_open_ext) Removed.
2363 (fn_close_ext) Removed.
2365 * filename.h: (struct file_ext) Removed.
2367 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2369 * variable.c (var_is_valid_name): Move declarations before code
2372 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2374 * filename.ch (fn_interp_vars): Fixed small buglet.
2376 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2378 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2379 so as to modify the string inline. Thus makeing it easier to
2380 destroy the results when no longer needed.
2382 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2384 * category.c (cat_stored_values_destroy): Fixed memory leak.
2386 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2388 Add some missing frees. Thanks to John Darrington for reporting
2391 * any-writer.c (any_writer_close): Free writer.
2393 * any-reader.c (any_reader_close): Free reader.
2395 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2397 * por-file-reader.c: (error) Mark as NO_RETURN.
2399 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2401 * settings.c: Changed default value of scompress to true.
2403 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2405 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2407 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2410 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2411 here from pref.h.orig.
2413 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2415 * sys-file-reader.c: Fixed bug reading compressed files.
2417 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2419 * Numerous renames. See src/ChangeLog for details.
2421 * Moved files from src directory