1 2008-02-06 John Darrington <john@darrington.wattle.id.au>
3 psql-reader.c psql-reader.h: Read more than one tuple at
4 once. Fix bug reading a query which returns no data. Fix bug
5 when transformation followed a reader.
6 Ask the server for the number of records in the query, for the
9 2008-02-05 John Darrington <john@darrington.wattle.id.au>
11 psql-reader.c: So yesterday they release postgresql 8.3.0
12 which has money represented with 64 bits. They must get
15 2008-02-02 John Darrington <john@darrington.wattle.id.au>
17 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
18 for reviewing this code.
20 2008-02-02 Ben Pfaff <blp@gnu.org>
24 * sys-file-reader.c (read_variable_record): Allow missing values
25 to be specified on long string variables, but warn about them
26 because PSPP does not yet support them.
27 (read_extension_record): Ignore extension records 20 and 21, which
28 PSPP does not yet support.
29 (read_header): Fix error message when floating-point format cannot
32 2008-02-01 Ben Pfaff <blp@gnu.org>
34 Patch #6386. Thanks to John Darrington for review and for the
35 updates to gnumeric-reader.c.
37 * dictionary.c (make_hinted_name): New function.
38 (make_numeric_name): New function.
39 (dict_make_unique_var_name): New function.
41 * gnumeric-reader.c (devise_name): Removed.
42 (munge_name): Removed.
43 (gnumeric_open_reader): Use new function
44 dict_make_unique_var_name.
46 * short-names.c (set_var_short_name_suffix): Use new function
49 2008-01-19 John Darrington <john@darrington.wattle.id.au>
51 * settings.c settings.h: Moved static variables into a
52 single struct. Renamed functions to have a settings_ prefix.
54 2008-01-14 John Darrington <john@darrington.wattle.id.au>
56 * data-out.c (output_AHEX): Corrected number of bytes in
57 call to output_hex. Closes bug #22011
59 2008-01-02 John Darrington <john@darrington.wattle.id.au>
61 * variable.c variable.h: Replaced var_get_value_name with
62 var_append_value_name which doesn't use any static data.
63 Thanks to Ben for review.
65 2007-12-07 Ben Pfaff <blp@gnu.org>
69 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
70 made from an empty casereader produce a casegrouper with no
71 groups, instead of one with one group that has no cases.
73 * casereader.c (casereader_is_empty): New function.
75 2007-12-06 Ben Pfaff <blp@gnu.org>
79 * sys-file-reader.c (read_display_parameters): Handle variable
80 display parameters record with only 2 data items per variable.
81 Reported by Guido Gay <gay@irer.it>.
83 2007-12-04 Ben Pfaff <blp@gnu.org>
85 * identifier.c (lex_id_match_n): New function.
86 (lex_id_match): Reimplement in terms of lex_id_match_n.
88 2007-11-24 Ben Pfaff <blp@gnu.org>
90 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
93 2007-11-24 Ben Pfaff <blp@gnu.org>
95 Fix warning reported when reading back system files that include
96 very long string variables. Thanks to Guido Gay <gay@irer.it> for
99 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
100 was written where `j' was meant.
102 2007-11-18 Ben Pfaff <blp@gnu.org>
104 Properly write variables that include a range of missing values to
105 system files. Thanks to Guido Gay <gay@irer.it> for reporting
108 * sys-file-writer.c (write_variable): Correctly calculate
109 n_missing_values field when writing variables that include a range
112 2007-11-10 Ben Pfaff <blp@gnu.org>
114 Cleanups and bug fixes devised while writing up documentation.
117 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
119 * dict-class.c: New file.
120 (dict_class_from_id): Move here.
121 (dict_class_to_name): Move here.
123 * dict-class.h: New file.
124 (enum dict_class): Move here. Change from consecutive integers to
125 consecutive bits, to make testing for any of multiple values
126 easier. Add new DC_ALL constant.
128 * dictionary.c (struct dictionary): Change `case_limit' from
129 size_t to casenumber.
130 (dict_get_vars): Make final argument an enum dict_class.
131 (dict_get_vars_mutable): Ditto.
132 (dict_get_case_limit): Change return value to casenumber.
133 (dict_set_case_limit): Change final argument to a casenumber.
134 (dict_unset_split_var): Add assertion. Rephrase slightly.
135 (dict_set_label): Use xstrndup to simplify.
137 * format.c (fmt_step_width): AHEX format also needs 2-byte
139 (fmt_set_style): Simplify assertions.
141 * missing-values.c (mv_add_num_range): Rename mv_add_range.
142 Simplify implementation.
143 (mv_has_value): Simplify implementation.
144 (mv_pop_value): Remove the first value, not the last, to avoid
145 having GET followed by SAVE reverse the order of missing values.
146 (mv_peek_value): Rename mv_get_value. Simplify assertion.
147 (mv_has_range): Simplify implementation.
148 (mv_peek_range): Rename mv_get_range.
149 (can_resize_string): Removed.
150 (mv_is_resizable): Use value_is_resizable.
151 (mv_resize): Use value_resize.
153 * short-names.h (SHORT_NAME_LEN): Move here.
155 * val-type.h: New file, for definitions related to type and width
156 of abstract values. Before, these definitions were mixed among
157 those related to "union value" and those related to variables.
158 (macro SYSMIS): Move here.
159 (macro LOWEST): Move here.
160 (macro HIGHEST): Move here.
161 (macro MAX_STRING); Move here.
162 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
163 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
164 All references updated.
165 (val_type_is_valid): New function. Replaces var_type_is_valid.
166 All references updated.
167 (val_type_from_width): New function. Replaces
168 var_type_from_width. All references updated.
170 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
172 (val_labs_can_set_width): Use value_is_resizable.
173 (val_labs_add): Simply return false if the value labels set is too
174 wide, instead of having undefined behavior.
175 (val_labs_replace): Ditto.
176 (val_labs_replace): Ditto.
177 (val_labs_first): Set iterator to null if iteration is complete.
178 (val_labs_first_sorted): Ditto.
179 (val_labs_done): Become a no-op if the iterator is null.
181 * value.c (value_is_resizable): New function.
182 (value_resize): New function.
184 * variable.c (var_get_dict_class): New function.
186 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
189 2007-11-08 Ben Pfaff <blp@gnu.org>
191 * data-in.c: Make formatted data parsing locale-independent.
192 (parse_number): Use c_strtod instead of strtod, to avoid
193 locale-specific behavior.
196 2007-11-06 Ben Pfaff <blp@gnu.org>
198 Patch #6256: add support for binary, 360 file formats. Reviewed
201 * data-in.c (struct data_in): Add `encoding' member.
202 (data_in): Add `encoding' parameter, and re-encode the data passed
203 in where appropriate. Update all callers to pass it in.
204 (parse_A): Implement EBCDIC recoding wart described in manual.
205 (parse_AHEX): Implement EBCDIC recoding.
207 * data-out.c (data_out_legacy): New function.
208 (data_out): Make into a wrapper around data_out_legacy.
210 * file-handle-def.c (struct file_handle): New member `encoding'.
211 (fh_create_file): Set encoding.
212 (fh_default_properties): Set default encoding.
213 (fh_get_legacy_encoding): New function.
215 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
216 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
217 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
218 (struct fh_properties): New member `encoding'.
220 2007-11-05 Ben Pfaff <blp@gnu.org>
222 Patch #6258. Reviewed by John Darrington.
224 * file-handle-def.c (fh_lock): Add comment that TYPE should be
225 marked with N_() in the caller. Added these markings to each
226 caller too. Should make i18n easier.
227 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
229 2007-11-03 Ben Pfaff <blp@gnu.org>
231 Allow output files to overwrite input files (bug #21280).
233 * file-handle-def.c: Separate locking of files for input and for
234 output, to allow both to take place at once. Also, distinguish a
235 file handle from the identity of the underlying file, because the
236 identity of a file changes over time and the file handle can't
237 represent two different identities.
238 (struct file_handle): Remove `next', `open_cnt', `deleted',
239 `type', `open_mode', `aux', `identity' members. Change `id' from
240 char array to char *. Add `ref_cnt' member.
241 (file_handle_from_ll) New function.
242 (file_handles) Removed.
243 (named_handles) New variable.
244 (fh_init) Initialize named_handles.
245 (fh_done) Remove name from all named_handles.
246 (free_handle) Updated for modified struct file_handle.
247 (unname_handle) New function.
248 (fh_ref) New function.
249 (fh_from_file_name) Removed.
250 (fh_unref) New function.
251 (fh_unname) New function.
252 (fh_from_id) Rewritten.
253 (create_handle) Updated for modified struct file_handle.
254 (fh_inline_file) Increment the handle's ref count.
255 (fh_create_file) Updated for modified struct file_handle.
256 (fh_create_scratch) Ditto.
261 (fh_is_open) Removed.
262 (fh_get_id) Updated for modified struct file_handle.
263 (fh_get_default_handle) Increment the handle's ref count.
264 (fh_set_default_handle) Handle ref counts.
265 (struct fh_lock) New structure.
266 (locks) New static var.
267 (fh_lock) New function.
268 (fh_unlock) New function.
269 (fh_lock_get_aux) New function.
270 (fh_lock_set_aux) New function.
271 (fh_is_locked) New function.
272 (make_key) New function.
273 (free_key) New function.
274 (compare_fh_locks) New function.
275 (hash_fh_lock) New function.
277 * file-handle-def.h (enum fh_access) New enum.
279 * file-name.c: Made file_identity the same in all supported
281 (struct file_identity): New `name' member.
282 (fn_get_identity): For a file that doesn't exist, get the
283 dev/inode of its directory plus its name. If even the directory
284 doesn't exist, just use its name. Merge the Windows
285 implementation into the Unix one.
286 (fn_compare_file_identities): Rewritten. Merge the Windows
287 implementation into the Unix one.
288 (fn_hash_identity): New function.
290 * make-file.c (struct replace_file): New structure.
291 (all_files): New static var.
292 (replace_file_start): New function.
293 (replace_file_commit): New function.
294 (replace_file_abort): New function.
295 (free_replace_file): New function.
296 (unlink_replace_files): New function.
298 * por-file-reader.c (struct pfm_reader): Add `lock' member.
299 (close_reader): Unlock file.
300 (pfm_open_reader): Lock file.
302 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
304 (pfm_open_writer): Lock file and prepare for its replacement.
305 (close_writer): Unlock file.
307 * scratch-handle.h (struct scratch_handle): Add unique_id so that
308 different generations of a scratch file can be distinguished.
310 * scratch-reader.c (scratch_reader_open): Verify that the file is
313 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
314 remove scratch_handle member.
315 (scratch_writer_open): Lock handle. Prepare to replace handle
316 data, instead of doing it immediately.
317 (scratch_writer_casewriter_destroy): Replace handle data and
320 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
321 (sfm_open_reader): Lock file.
322 (close_reader): Unlock file.
324 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
326 (sfm_open_writer): Lock file and prepare for its replacement.
327 (close_writer): Unlock file.
329 2007-11-02 Ben Pfaff <blp@gnu.org>
331 * data-out.c (output_number): Use isfinite (from C99) instead of
332 gsl_isfinite, since we now have universal support for it in
334 (output_infinite): Ditto.
335 [!HAVE_ROUND] (round): Remove definition, since we now have a
336 replacement in gnulib.
338 2007-10-27 John Darrington <john@darrington.wattle.id.au>
340 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
341 wasn't initialised correctly. Closes bugs #21330 and 21397
344 2007-10-08 Ben Pfaff <blp@gnu.org>
346 Fix bug #21280. Thanks to John Darrington for review.
348 * file-name.c (create_stream): New function.
350 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
351 file handle before creating the file, to ensure that we don't
352 truncate a file that we're reading. Make code easier to read by
355 * sys-file-write.c (sfm_open_writer): Ditto.
357 2007-10-01 Ben Pfaff <blp@gnu.org>
359 Fix bug #21192. Thanks to John Darrington for review.
361 * casereader.c (casereader_read): Decrement case_cnt before
362 calling the casereader's "read" member function, so that we
363 interact properly with lazy_casereader.
365 * datasheet.c: Add regression test for above bug fix.
366 (clone_datasheet): New function.
367 (lazy_callback): New function.
368 (check_datasheet_casereader): New function.
369 (check_datasheet): Check datasheet contents are reported correctly
370 through an ordinary casereader and a lazy casereader.
371 (clone_model): Use clone_datasheet.
373 2007-09-24 Ben Pfaff <blp@gnu.org>
375 Patch #6210. Reviewed by John Darrington.
376 * settings.c: Drop "long view", which is not needed any longer.
377 (static var long_view): Removed.
378 (force_long_view): Removed.
379 (get_viewwidth): Removed.
380 (init_viewport): Removed long_view reference.
382 2007-09-19 John Darrington <john@darrington.wattle.id.au>
384 * settings.c settings.h: Changed viewport's length and width to be
385 owned by the user interface which uses the data library. This allows
386 better abstraction, and makes dynamically adjustable dimensions easier.
388 2007-09-18 Ben Pfaff <blp@gnu.org>
390 * procedure.c (proc_extract_active_file_data): New function.
392 * lazy-casereader.h: New file.
394 * lazy-casereader.c: New file.
396 * casereader.c (casereader_dynamic_cast): New function.
398 2007-09-14 Ben Pfaff <blp@gnu.org>
400 * dictionary.c (dict_clone): Copy case indexes from cloned
401 dictionary. Fixes bug #21061. Reviewed by John Darrington.
403 2007-09-13 John Darrington <john@darrington.wattle.id.au>
405 * value.c value.h (create_value): New function.
407 2007-09-12 Ben Pfaff <blp@gnu.org>
409 Make it clear that translator casereader and casewriter translate
410 functions are supposed to destroy their input case.
412 * casereader-translator.c (struct casereader_translator): Change
413 input case parameter of translate member function from const
414 struct ccase * to struct ccase *.
415 (casereader_create_translator): Ditto, for translate parameter.
417 * casewriter-translator.c (struct casewriter_translator): Ditto.
418 (casewriter_create_translator): Ditto.
420 2007-08-27 John Darrington <john@darrington.wattle.id.au>
422 * sys-file-reader.c (read_display_parameters): Force display width
423 to 8, if the sys file says 0 (like SPSS does).
425 2007-08-12 Ben Pfaff <blp@gnu.org>
427 * dictionary.c (dict_dump): New function.
429 2007-08-12 Ben Pfaff <blp@gnu.org>
431 Drop dict_compactor in favor of using the new struct case_map.
433 * dictionary.c (struct copy_map): Removed.
434 (struct dict_compactor): Removed.
435 (dict_make_compactor): Removed.
436 (dict_compactor_compact): Removed.
437 (dict_compactor_destroy): Removed.
439 * procedure.c (struct data_set): Change `compactor' member to be a
441 (proc_open): Use case_map_to_compact_dict instead of
443 (proc_casereader_read): Use case_map_execute instead of
444 dict_compactor_compact.
445 (proc_commit): Use case_map_destroy instead of
446 dict_compactor_destroy.
448 * scratch-writer.c (struct scratch_writer): Change `compactor'
449 member to be a struct case_map *.
450 (scratch_writer_open): Use case_map_to_compact_dict instead of
452 (scratch_writer_casewriter_write): Use case_map_execute instead of
453 dict_compactor_compact.
455 2007-08-12 Ben Pfaff <blp@gnu.org>
457 * automake.mk: Add case-map.c, case-map.h.
459 * case-map.c: New file.
461 * case-map.h: New file.
463 2007-08-12 Ben Pfaff <blp@gnu.org>
465 * dictionary.c (dict_compact_values): Don't delete scratch
466 variables as well as compacting case indexes. Update all callers.
467 (dict_get_compacted_value_cnt): Rename dict_count_values and
468 change interface. Update all callers.
469 (dict_get_compacted_value_cnt): Remove.
470 (dict_compacting_would_shrink): Remove.
471 (dict_compacting_would_change): Remove.
472 (dict_make_compactor): Add new parameter. Update all callers.
474 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
475 instead of dict_count_values, changing an O(N) operation into
478 2007-08-12 Ben Pfaff <blp@gnu.org>
480 * casereader.c (casereader_read): Don't require cases read by a
481 casereader to be exactly the expected size: as long as they're big
484 2007-08-12 Ben Pfaff <blp@gnu.org>
486 Make casewriters keep track of the number of `union value's in
487 each case. This is useful for two reasons: casewriter_write can
488 then check that the case being written is large enough, and later
489 recipients of the casewriter can determine the size of the case.
491 * casewriter-translator.c (casewriter_create_translator): Add
494 * casewriter.c (struct casewriter): Add value_cnt member.
495 (casewriter_write): Check that the case passed in is big enough.
496 (casewriter_get_value_cnt): New function.
497 (casewriter_create): Add value_cnt parameter.
499 2007-08-09 Ben Pfaff <blp@gnu.org>
501 Fix bug reported by Jason Stover.
502 * settings.c: Move get_termcap_viewport and in particular the
503 #include for <curses.h> to the end of the file. curses.h
504 redefines bool on some systems (e.g. OpenBSD), which causes
505 disagreement between uses of bool before its inclusion and
507 Tested by Jason Stover.
509 2007-07-29 Ben Pfaff <blp@gnu.org>
511 Provisional fix for bug #18692 and bug #20161. Reviewed by John
514 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
515 argument (never "rb" or "wb") because SUSv3 says that only those
516 modes are defined, and glibc in fact rejects other modes.
518 Open portable files with fn_open so that they can be read from
519 pipes. Fix missing fh_close call to go along with fh_open.
520 Report an error if the file close reports an error.
521 * por-file-reader.c (close_reader): New function.
522 (por_file_casereader_destroy): Use close_reader.
523 (pfm_open_reader): Open file with fn_open.
525 2007-07-28 Ben Pfaff <blp@gnu.org>
527 Make PSPP able to read all the portable files I could find on the
528 web. Thanks to John Darrington for review. Bug #17620.
529 * por-file-reader.c (struct pfm_reader): New member `line_length'.
530 (error): Print file offset in hexadecimal.
531 (warning): New function.
532 (advance): Treat lines less than 80 bytes long as padded to 80
534 (pfm_open_reader): Call read_documents if we find an "E" record.
535 (convert_format): Convert invalid formats to the default format
536 instead of aborting reading the file.
537 (read_variables): Rename duplicate variable names instead of
538 aborting reading the file.
539 (read_value_label): Allow string variables of different widths to
540 be assigned value labels in the same record. Replace duplicate
541 value labels instead of aborting.
542 (read_documents): New function.
544 * por-file-writer.c (pfm_open_writer): Call write_documents if the
545 dictionary has documents.
546 (write_documents): New function.
548 2007-07-25 Ben Pfaff <blp@gnu.org>
550 Fix bugs related to bug #17213.
552 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
553 since the former is what config.h has. Include the needed ncurses
555 (static var echo) Rename to `do_echo' because the original name is
556 the same as an ncurses identifier.
557 (get_termcap_viewport) Use error instead of msg.
559 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
560 (fn_close): Don't close stdin, stdout, stderr.
562 2007-07-26 John Darrington <john@darrington.wattle.id.au>
564 * procedure.c procedure.h: Added callbacks which get invoked whenever
565 a dataset's transformation chain changes.
567 2007-07-24 Ben Pfaff <blp@gnu.org>
570 * sys-file-writer.c (write_variable_display_parameters): Use new
571 var_default_display_width function to choose display width of
572 segments after the first one in a given variable.
573 * variable.c (var_create): Use var_default_display_width to pick
574 new variable's display width.
575 (var_default_display_width): New function.
576 Reviewed by John Darrington.
578 2007-07-24 Ben Pfaff <blp@gnu.org>
581 * por-file-writer.c (write_variables): Write weight variable.
582 Reviewed by John Darrington.
584 2007-07-23 Ben Pfaff <blp@gnu.org>
586 Improvements to system file reader and writer.
588 First, move all detailed knowledge of very long strings into
589 sys-file-private.[ch], so that this nasty stuff can be isolated.
591 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
592 (EFFECTIVE_VLS_CHUNK): New macro.
593 (min_int): New function.
594 (max_int): New function.
595 (sfm_width_to_bytes): Rewrite.
596 (sfm_width_to_octs): New function.
597 (sfm_segment_alloc_width): New function.
598 (sfm_segment_alloc_bytes): New function.
599 (sfm_segment_used_bytes): New function.
600 (sfm_segment_offset): New function.
601 (sfm_segment_effective_offset): New function.
602 (sfm_dictionary_to_sfm_vars): New function.
604 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
605 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
606 (struct sfm_var): New structure.
608 Next, improvements to the system file reader.
610 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
611 type casenumber. Added `version_major', `version_minor',
614 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
615 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
616 Change `case_cnt' to type casenumber. Removed `has_vls'.
617 (struct sfm_var): Removed.
618 (sfm_open_reader): Don't warn on wrong case size if the file was
619 written by SPSS 13, which tends to get it wrong. Use
620 sfm_dictionary_to_sfm_vars.
621 (read_header): Always output system file info.
622 (read_variable_record): Simplify code for reading missing values.
623 (read_machine_int32_info): Save version numbers from system file
624 into info struct passed as new argument.
625 (read_long_string_map): Restructured to use new sys-file-private
627 (read_value_labels): Use size_overflow_p.
628 (sys_file_casereader_read): Get rid of distinction between fast
629 and slow paths. Use information provided by sys-file-primate's
630 struct sfm_var to simplify code.
631 (skip_whole_strings): New function.
632 (read_int32): Renamed read_int. Changed return value to int.
634 (read_flt64): Renamed read_float. Changed return value to
635 double. Updated all callers.
636 (int32_to_native): Removed. Changed callers to use
638 (flt64_to_double): Removed. Changed callers to use float_convert.
640 Finally, get rid of int32, flt64 terminology and types in system
641 file writer. The former wasn't very useful since a POSIX "int"
642 can hold the whole range of int32 and we generally didn't have a
643 need for it to be exactly-32-bits, just at-least-32-bits. The
644 latter was inconvenient because we had to assume that it could be
645 different from double and thereby convert special values SYSMIS,
646 HIGHEST, LOWEST to and from it in multiple places. Instead, now
647 we just use "int" and "double" in most places, and do conversions,
648 if necessary, very close to where we do I/O. This change meant
649 that the writer code couldn't represent records in the file as C
650 structs any longer, but that's no great loss. The code actually
651 seems to be more readable without them.
653 Simplify the compression buffering code: only buffer as much as
654 necessary, which is no more than eight 8-byte units at any given
657 * sys-file-writer.c (typedef flt64): Removed.
658 (macro second_lowest_flt64): Removed.
659 (struct sysfile_header): Removed.
660 (struct sysfile_variable): Removed.
661 (struct sfm_writer): Removed `needs_translation', `has_vls',
662 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
663 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
664 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
665 compression buffering by `opcodes', `opcode_cnt', `data',
666 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
667 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
668 data writer functions instead of structures.
669 (calc_oct_idx): New function.
670 (write_header): Use simple data writer functions instead of
672 (write_format_spec): Renamed write_format. New argument.
673 (write_variable_continuation_records): New function.
674 (write_variable): Use simple data writer functions instead of
675 structures. Use write_variable_continuation_records. Write
676 entire very long string instead of requiring caller to understand
678 (write_value_labels): Use simple data writer functions instead of
680 (write_documents): Ditto.
681 (write_variable_display_parameters): Use sys-file-private
682 functions to simplify. Use simple data writer functions instead
684 (write_vls_length_table): Use simple data writer functions instead
686 (write_longvar_table): Ditto.
687 (write_rec_7_34): Break into new functions
688 write_integer_info_record, write_float_info_record. Use simple
689 data writer functions instead of structures.
690 (buf_write): Removed.
691 (append_string_max): Removed.
692 (ensure_buf_space): Removed.
693 (sys_file_casewriter_write): Get rid of the distinction between
694 fast and slow paths, which didn't seem to be too useful. Use new
695 functions write_case_uncompressed, write_case_compressed.
696 (put_instruction): Removed.
697 (put_element): Removed.
698 (write_compressed_data): Removed.
699 (close_writer): Use flush_compressed. Only write case count to
700 system file if it will fit in the field.
701 (write_case_compressed): New function.
702 (write_case_uncompressed): New function.
703 (flush_compressed): New function.
704 (put_cmp_opcode): New function.
705 (put_cmp_number): New function.
706 (write_int): New function.
707 (convert_double_to_output_format): New function.
708 (write_float): New function.
709 (write_value): New function.
710 (write_string): New function.
711 (write_bytes): New function.
712 (write_zeros): New function.
713 (write_spaces): New function.
715 Reviewed by John Darrington.
717 2007-07-22 Ben Pfaff <blp@gnu.org>
719 Don't try to write very long strings to portable files. The
720 format does not support it.
722 * por-file-writer.c (MAX_POR_WIDTH): New macro.
723 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
724 (write_format): Add arg to take width to resize format to.
725 (write_value): Limit width of value written to MAX_POR_WIDTH.
726 (write_variables): Limit width of variable and its output formats
728 Reviewed by John Darrington.
730 2007-07-22 Ben Pfaff <blp@gnu.org>
732 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
733 local variable instead of literal 5.
734 Reviewed by John Darrington.
736 2007-07-22 Ben Pfaff <blp@gnu.org>
738 Fix problems with uniqueness of short names in system files with
739 very long string variables. Now a variable may have multiple
742 * automake.mk (src_data_libdata_a_SOURCES): Add new files
743 short-names.c, short-names.h.
745 * dictionary.c (dict_clone): Clone all the short names.
746 (compare_strings): Move into short-names.c.
747 (hash_strings): Ditto.
748 (set_var_short_name_suffix): Ditto.
749 (dict_assign_short_names): Ditto, rename short_names_assign,
750 change to assign all short names.
752 * por-file-writer.c (write_variables): Use short_names_assign
753 instead of dict_assign_short_names.
755 * short-names.c: New file.
757 * short-names.h: New file.
759 * sys-file-private.c (sfm_width_to_segments): New function.
761 * sys-file-reader.c (read_long_var_name_map): Save and restore all
762 the short names, not just the first one.
764 * sys-file-writer.c (cont_var_name): Removed.
765 (sfm_open_writer): Use short_names_assign instead of
766 dict_assign_short_names. Use unique short names assigned by
767 short_names_assign instead of those generated by cont_var_name.
769 * variable.c (struct variable): Remove `short_name' member,
770 replace by `short_names' and `short_name_cnt'.
771 (var_create) Initialize new members.
772 (var_get_short_name_cnt): New function.
773 (var_get_short_name): Now takes an index argument. Changed most
775 (var_set_short_name): Ditto.
776 (var_clear_short_name): Renamed var_clear_short_names, changed to
777 clear all short names.
779 Reviewed by John Darrington.
781 2007-07-22 Ben Pfaff <blp@gnu.org>
783 * variable.c (var_set_width): Use new var_set_width function.
785 * missing-values.c (mv_n_values): Drop assertion, which was not
788 * format.c (fmt_default_for_width): New function.
789 (fmt_resize): New function.
791 Reviewed by John Darrington.
793 2007-07-18 John Darrington <john@darrington.wattle.id.au>
795 * datasheet.c (datasheet_delete_columns): Added assertion to check
796 we're not deleting outside the range of the sheet.
799 * dictionary.c dictionary.h variable.c: Added the ability for string
800 variables to be resized.
802 * vardict.h: Added some prototypes (moved from dictionary.h) as
803 these should only be called by variable.c
806 2007-07-14 John Darrington <john@darrington.wattle.id.au>
808 * sfm-reader.c: Respect case_cnt field in file header.
810 2007-07-01 John Darrington <john@darrington.wattle.id.au>
812 * transformation.c transformation.h (trns_chain_execute): Changed the
813 signature (Patch #6057)
815 2007-06-10 Ben Pfaff <blp@gnu.org>
817 * casereader-filter.c (casereader_filter_destroy): Make sure to
818 write all the remaining excluded cases to the casewriter, if any.
820 * caseinit.c (init_list_destroy): Rewrite.
821 (init_list_clear): Ditto.
823 * casegrouper.c (casegrouper_get_next_group): Always set *reader
824 to null when returning false.
826 2007-06-06 Ben Pfaff <blp@gnu.org>
828 Actually implement the new procedure code and adapt all of its
829 clients to match. Also adapt all of the other case sources and
830 sinks in the tree and their clients to use the
831 casereader/casewriter infrastructure.
833 * automake.mk: Add and remove files.
835 * any-reader.c: Change into a casereader.
836 * por-file-reader.c: Ditto.
837 * scratch-reader.c: Ditto.
838 * sys-file-reader.c: Ditto.
840 * any-writer.c: Change into a casewriter.
841 * por-file-writer.c: Ditto.
842 * scratch-writer.c: Ditto.
843 * sys-file-writer.c: Ditto.
845 * procedure.c: Change to use casereader, casewriter, caseinit, and
846 other new infrastructure.
848 * scratch-handle.c: Adapt to new infrastructure.
850 * case-sink.c: Removed, now dead code.
851 * case-sink.h: Ditto.
852 * case-source.c: Ditto.
853 * case-source.h: Ditto.
854 * casefile-factory.c: Ditto.
855 * casefile-private.h: Ditto.
858 * casefilter.c: Ditto.
859 * casefilter.h: Ditto.
862 * fastfile-factory.c: Ditto.
863 * fastfile-factory.h: Ditto.
864 * storage-stream.c: Ditto.
865 * storage-stream.h: Ditto.
867 2007-06-06 Ben Pfaff <blp@gnu.org>
871 * automake.mk: Add new files.
873 * datasheet.c: New file.
875 * datasheet.h: New file.
877 2007-06-06 Ben Pfaff <blp@gnu.org>
879 Until now, the procedure code has provided a case to the
880 case_source, which has filled in the data values that come from
881 the active file. "Left" data values that don't come from the
882 active file naturally stay the same from case to case, because the
883 procedure code keeps using that same case.
885 One of the compromises that comes with the new procedure code is
886 that the active file allocates and provides its own case, which
887 the procedure code then has to resize to provide room for any
888 other variables that should go in the case and then fill in the
889 values of "left" variables. Then, when we're done with that case,
890 we have to save the values of "left" variables to copy into the
891 next case read from the active file.
893 The caseinit code helps with this.
895 * automake.mk: Add new files.
897 * caseinit.c: New file.
899 * caseinit.h: New file.
901 2007-06-06 Ben Pfaff <blp@gnu.org>
903 * value.h (value_cnt_from_width): New function.
905 * variable.c (var_get_value_cnt): Use new function.
907 2007-06-06 Ben Pfaff <blp@gnu.org>
909 Add casegrouper, to allow cases read from a given casereader to be
910 broken into groups, each of which has its own casereader.
911 Generally cases are grouped based on having equal values for some
914 * automake.mk: Add new files.
916 * casegrouper.c: New file.
918 * casegrouper.h: New file.
920 2007-06-06 Ben Pfaff <blp@gnu.org>
922 Add interface to lexicographical ordering of cases.
924 * automake.mk: Add new files.
926 * case-ordering.c: New file.
928 * case-ordering.h: New file.
930 2007-06-06 Ben Pfaff <blp@gnu.org>
932 Add casereaders and casewriters, the basis of the new data processing
933 implementation. A casereader is a uniform interface to reading cases
934 from a data source; a casewriter is a uniform interface to writing
935 cases to a data sink.
937 * automake.mk: Add new files.
939 * casereader-filter.c: New file.
941 * casereader-provider.h: New file.
943 * casereader-translator.c: New file.
945 * casereader.c: New file.
947 * casereader.h: New file.
949 * casewriter-provider.h: New file.
951 * casewriter-translator.c: New file.
953 * casewriter.c: New file.
955 * casewriter.h: New file.
957 2007-06-06 Ben Pfaff <blp@gnu.org>
959 "casewindow" data structure that extends the deque (from libpspp)
960 of cases with the ability to dump cases to disk if we get too many
963 * automake.mk: Add new files.
965 * casewindow.c: New file.
967 * casewindow.h: New file.
969 2007-06-06 Ben Pfaff <blp@gnu.org>
971 sparse_cases data structure that augments a sparse_array of cases
972 with the ability to dump cases to disk if we get too many cases in
975 * automake.mk: Add new files.
977 * sparse-cases.c: New file.
979 * sparse-cases.h: New file.
981 2007-06-06 Ben Pfaff <blp@gnu.org>
983 Adds a low-level on-disk case array data structure.
985 * automake.mk: Add new files.
987 * case-tmpfile.c: New file.
989 * case-tmpfile.h: New file.
991 2007-06-06 Ben Pfaff <blp@gnu.org>
993 In a couple of places we calculate the maximum number of cases to
994 keep in memory based on the user-defined workspace. Enable
995 centralizing the calculation through a new function.
997 * settings.c (get_workspace_cases): New function.
999 2007-06-06 Ben Pfaff <blp@gnu.org>
1001 The casenumber type is defined in transformations.h, but case.h is
1002 a more sensible place. Move it.
1004 * case.h (CASENUMBER_MAX): New macro.
1005 (typedef casenumber): Move here, from transformations.h.
1007 2007-06-03 Ben Pfaff <blp@gnu.org>
1009 Slightly generalize case_to_values and case_from_values functions.
1011 * case.c (case_to_values): Rename case_copy_out, change interface.
1012 (case_from_values): Rename case_copy_in, change interface.
1014 * fastfile.c (fastfilereader_get_next_case): Update caller.
1015 (write_case_to_disk): Ditto.
1017 2007-06-02 Ben Pfaff <blp@gnu.org>
1019 Clean up after a forgotten part of patch #5829.
1021 * casedeque.h: Remove unused file.
1023 * automake.mk: Remove casedeque.h from sources.
1025 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1027 * category.c: Removed redundant #include
1029 2007-05-06 Ben Pfaff <blp@gnu.org>
1031 Abstract the documents within a dictionary a little better.
1032 Thanks to John Darrington for suggestion, initial version, and
1033 review. Patch #5917.
1035 * dictionary.c (struct dictionary): Change `documents' member from
1036 char * to struct string.
1037 (dict_clear): Destroy struct string.
1038 (dict_get_documents): Convert struct string to char *.
1039 (dict_set_documents): Set struct string. Pad to 80-character
1041 (dict_clear_documents): New function.
1042 (dict_add_document_line): New function.
1043 (dict_get_document_line_cnt): New function.
1044 (dict_get_document_line): New function.
1046 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1048 * sys-file-reader.c (read_documents): Use new document functions.
1050 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1052 * sys-file-reader.c: When reading a system file which has no
1053 long name table, automatically create one where the long names
1054 are the lower case versions of the short names.
1056 2007-04-22 Ben Pfaff <blp@gnu.org>
1058 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1059 dict_clear will free most data related to the dictionary.
1060 dict_clear does a decent job, except that dict_set_split_vars on
1061 some systems won't actually free the dict's "split" member.
1062 Instead, it'll allocate a 1-byte region. Fix this.
1064 * value.c (value_copy): New function.
1065 (value_set_missing): Ditto.
1067 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1069 * Deleted existing category.h and moved cat-routines.h into
1070 category.h Encapsulated struct cat_vals better.
1072 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1074 * sys-file-reader.c: When reading a system file which has no
1075 long name table, automatically create one where the long names
1076 are the lower case versions of the short names.
1078 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1080 * sys-file-reader.c: Some versions of Other Software seem to
1081 produce system files with string variables' measure set to
1082 zero. We'll assume these are supposed to be nominal variables.
1084 2007-03-30 Ben Pfaff <blp@gnu.org>
1086 * procedure.c: Adapt to new deque data structure.
1088 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1089 Ben Pfaff <blp@gnu.org>
1091 * file-name.c: Mingw compatibility fixes.
1092 (fn_search_path): Use ISSLASH instead of comparing against '/'
1094 (fn_dir_name): Use dir_name from gnulib.
1095 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1096 (fn_get_identity): Use GetFullPathName instead of canonicalize
1097 from gnulib, because the latter does not fully support
1098 Windows-style path names. Use this implementation based on the
1099 detected presence of Windows instead of the absence of Unix, since
1100 the new implementation is Windows-specific.
1101 (fn_compare_file_identities): In Windows implementation, compare
1102 names case-insensitively.
1104 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1106 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1107 <stdlib.h> to have the same effect.
1109 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1111 * por-file-reader.c: Add missing _() around messages.
1113 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1115 * make-file.c: Include "mkstemp.h", without which linking on
1118 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1120 Reduce platform dependence.
1122 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1123 Everywhere we using this, we really should have just depended on
1124 the shell to expand tildes.
1125 (fn_search_path): Simplify, given that we don't do tilde expansion
1127 (fn_normalize): Removed. Caller changed to use the canonicalize
1129 (fn_get_cwd): Removed. Only user was fn_normalize.
1130 (fn_is_absolute): Really only test for absolute names.
1131 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1133 (fn_readlink): Removed, as it was only used fn_normalize.
1134 (fn_exists): Assume the stat function is available; gnulib does.
1135 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1137 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1139 * dictionary.c (dict_create_vector_assert): New function.
1141 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1143 * file-name.c (fn_normalize): Correct name of function
1144 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1145 for pointing this out.
1147 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1149 We are using a single member in struct file_handle, the "name"
1150 field, for more than one purpose. When it begins with '"', it's a
1151 file name; otherwise, it's a token that can be used to identify
1152 it. When that assertion fires, it's because we searched for the
1153 name case-sensitively as a file name (so that there was no match),
1154 and then we try to insert it case-insensitively as a token, which
1155 fails because duplicates aren't allowed.
1157 Solution: break the two purposes into two separate fields. This
1158 fixes the problem and likely makes the code easier to read too.
1160 Fixes bug #18922. Thanks to John Darrington for bug report and
1163 * file-handle-def.c (struct file_handle): New `id' member.
1164 (fh_from_name): Rename fh_from_id. Update all callers.
1165 (create_handle): New `id' parameter. Update all callers.
1166 (fh_create_file): Ditto.
1167 (fh_get_id): New function.
1169 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1171 * case.c (case_is_null): Change return type to bool.
1173 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1177 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1178 reference counting to share data. This makes it easy for
1179 valgrind, etc. to find accesses to cases that have been destroyed
1180 but have been kept around by another user's ref-count. This often
1181 happens when the data set is small enough to find in memory; if a
1182 bigger data set that would overflow to disk were used, then data
1183 corruption would occur.
1185 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1189 * case.c (case_unshare): Make it check internally whether the
1190 ref_cnt is greater than 1, so that the callers don't have to.
1191 Update callers not to check.
1193 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1195 Before, I was thinking that I might want to get rid of reference
1196 counting at some point. Now, I'm pretty sure that it's here to
1197 stay. Thus, because we have to store the value_cnt anyway for
1198 reference-counted cases, we might as well expose it to users.
1200 * case.c (case_get_value_cnt): New function.
1201 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1202 resize case if its size actually changed.
1204 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1205 instead of peeking inside struct case directly.
1206 (casefile_append): Ditto.
1208 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1210 Get rid of the inlines for the case functions, which made the
1211 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1212 the inlines didn't speed up "make check" at all, which is not a
1213 perfect benchmark but seems indicative.)
1215 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1216 definitions. Remove some assertions on nonnull pointers that were
1217 redundant with a pointer dereference soon after in the function.
1219 (struct case_data): Move definition here from case.h.
1223 (case_data_wr): Ditto.
1225 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1227 * automake.mk: Add casedeque.h to sources.
1229 * casedeque.h: New file.
1231 * procedure.c: (struct dataset) Change lag_count, lag_head,
1232 lag_queue member into single struct casedeque member. Update all
1233 users to use the casedeque instead.
1236 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1238 * procedure.c: Simplify lagged cases interface. Updated all
1239 clients--well, the only client--to use the simplified interface.
1240 (dataset_n_lag) Removed.
1241 (dataset_set_n_lag) Removed.
1242 (dataset_need_lag) New function.
1244 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1246 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1247 whenever appropriate, but only when the dataset/dictionary is in a
1250 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1252 * dictionary.c dictionary.h : Added callbacks for change of filter and
1253 split variables. Refactored some code to ensure that callbacks get
1254 invoked when appropriate.
1256 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1257 replace_dict callback occurs when permanent_dict replaces the current
1260 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1262 * dictionary.c dictionary.h : Added callback for when the weight
1263 variable of a dictionary changes.
1265 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1267 * dictionary.c dictionary.h : Added replace_source and replace_dict
1268 callbacks, and functions to deal with them.
1270 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1272 Simplify missing value handling.
1274 * missing-values.h (enum mv_class): New type.
1275 (enum mv_type): Moved definition into missing-values.c and renamed
1276 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1277 enums. Updated all uses.
1278 (struct missing_values): Changed type of `type' from `enum
1279 mv_type' to `int' because the definition is no longer exposed.
1281 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1282 parameter. Update all callers.
1283 (mv_is_num_missing): Ditto.
1284 (mv_is_str_missing): Ditto.
1285 (mv_is_value_user_missing): Removed. Changed callers to use
1286 mv_is_value_missing.
1287 (mv_is_num_user_missing): Removed. Changed callers to use
1289 (mv_is_str_user_missing): Removed. Changed callers to use
1291 (mv_is_value_system_missing): Removed. Changed callers to use
1292 mv_is_value_missing.
1293 (mv_set_type): Removed. Changed callers to use mv_clear.
1294 (mv_clear): New function.
1296 * variable.c (var_is_value_missing): Add new enum mv_class
1297 parameter. Update all callers.
1298 (var_is_num_missing): Ditto.
1299 (var_is_str_missing): Ditto.
1300 (var_is_value_user_missing): Removed. Changed callers to use
1301 var_is_value_missing.
1302 (var_is_num_user_missing): Removed. Changed callers to use
1304 (var_is_str_user_missing): Removed. Changed callers to use
1306 (var_is_value_system_missing): Removed. Changed callers to use
1307 var_is_value_missing.
1309 * casefilter.c (struct casefilter): Use enum mv_class in place of
1311 (casefilter_variable_missing): Adapt to new member.
1312 (casefilter_create): Change signature to take enum mv_class,
1315 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1317 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1319 * case-sink.c case-sink.h procedure.c procedure.h
1320 storage-stream.c: Now uses the factory.
1322 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1324 Make it possible to pull cases from the active file with a
1325 function call, instead of requiring indirection through a callback
1328 * case-source.h (struct case_source_class): Change ->read function
1329 to return a single case, instead of calling a callback function
1330 for each case. Change ->destroy function to return an error
1333 * case-source.c (free_case_source): Pass along the value returned
1334 by the case_source ->destroy function.
1336 * procedure.c (struct write_case_data): Removed.
1337 (struct dataset): Added some members to track procedure state.
1338 (procedure): Optimize the trivial case at this level.
1339 (internal_procedure): Re-implement in terms of proc_open,
1340 proc_read, proc_close.
1341 (proc_open) New function.
1342 (proc_read) New function.
1343 (proc_close) New function.
1344 (write_case) Moved into proc_read.
1345 (close_active_file) Moved closing of data source into proc_close.
1347 * storage-source.c: Rewrote to conform with modified
1348 case_source_class interface.
1350 * transformations.c (trns_chain_execute): Added argument to allow
1351 starting execution from an arbitrary transformation. Updated
1354 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1356 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1358 * sys-file-reader.c (read_display_parameters): Don't assume that
1359 MEASURE_* and ALIGN_* have the same values found in system files.
1361 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1363 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1364 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1365 the value of n_MEASURES, which was off by 1 (at least from my
1368 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1370 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1371 callbacks which are invoked when the dictionary or its
1372 variables are changed.
1374 * missing-values.c missing-values.h value-labels.c: Tidied up
1375 consistency checks, and made some of them return false
1376 instead of assert-failing.
1378 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1380 * calendar.c (calendar_days_in_month): New function.
1382 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1384 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1385 the value label's width.
1387 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1389 * sfm-private.h: Move contents into sys-file-writer.c, which is
1390 the only remaining user. Removed Borland C++-specific directives.
1392 * sys-file-reader.c: Clean up and rewrite entire file. The
1393 rewritten version is simpler and better abstracted, and should be
1394 easier to maintain and extend. It avoids using structures to read
1395 file data, which is prone to padding variations among compilers.
1396 It should also handle non-IEEE 754 system files, although I
1397 haven't been able to find any. It has been tested against many
1398 .sav files obtained from the Web and found to produce the same
1399 results as the earlier version of the code, or in some cases
1400 improved results. It is more tolerant of format variations found
1403 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1404 member, putting an enum integer_format in its place. New member
1405 `float_format'. Changed `compressed' member to type bool.
1407 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1409 * dictionary.c (dict_delete_consecutive_vars): New function.
1411 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1413 * file-name.c (fn_search_path): Remove prefix arg that was unused
1414 by any caller. Updated all callers.
1416 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1418 * format.c (fmt_dollar_template): Use user's decimal point
1419 character. Add assertion.
1421 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1423 * format.c (fmt_dollar_template): New function, based on
1424 dollar_format_template from var-type-dialog.c.
1426 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1428 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1431 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1433 Finish converting struct variable to an opaque type. In this
1434 phase, we add remaining setter and getter functions, convert the
1435 remaining PSPP code to use them, and do a bunch of cleanup. The
1436 resulting changes are pervasive but mostly trivial, and only the
1437 notable changes are logged.
1439 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1442 * case.c (case_data): Renamed case_data_idx.
1443 (case_num): Renamed case_num_idx.
1444 (case_str): Renamed case_str_idx.
1445 (case_data_rw): Renamed case_data_rw_idx.
1447 * case.h (case_data): New function with old name and an interface
1448 that takes a variable instead of an index, which is easier to
1449 use. Updated all callers to use the new interface, or to use the
1450 new *_idx function (see above).
1453 (case_data_rw): Ditto.
1455 * category.c (cat_stored_values_destroy): Changed interface to
1456 take a struct cat_vals * instead of a struct variable *.
1458 * dictionary.c (dict_clone): Use new vector_clone function.
1459 (dict_clear) Use new var_destroy function.
1460 (add_var) New function.
1461 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1462 (dict_create_var_assert) Rewrite in terms of add_var.
1463 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1464 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1465 (dict_lookup_var) Use new var_create, var_destroy functions.
1466 (dict_contains_var) Rewrite in terms of new vardict functionality.
1467 (set_var_dict_index) New function.
1468 (set_var_case_index) New function.
1469 (reindex_vars) New function.
1470 (dict_delete_var) Rewrite in terms of new vardict functionality.
1471 (dict_reorder_var) Ditto.
1472 (dict_reorder_vars) Ditto.
1473 (rename_var) New function.
1474 (dict_rename_var) Use rename_var.
1475 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1476 (dict_get_compacted_idx_to_fv) Rename
1477 dict_get_compacted_dict_index_to_case_index, update callers.
1478 (dict_create_vector) Use new vector_create function.
1479 (dict_clear_vectors) Use new vector_destroy function.
1480 (set_var_short_name_suffix) Move here from variable.c, renamed
1481 from var_set_short_name_suffix, make static, update caller.
1483 * sys-file-private.c: New file.
1484 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1485 width_to_bytes, update callers.
1487 * sys-file-private.h: New file. Later it will supplant
1488 sfm-private.h; for now it supplements it.
1489 (macro MIN_VERY_LONG_STRING) New macro.
1490 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1492 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1493 MAX_LONG_STRING was used before.
1495 * sys-file-writer.c: Ditto.
1497 * value-labels.c: Change the paradigm here to be that a null
1498 pointer is OK for a struct val_labs * in most cases; it just
1499 represents an empty set of value labels.
1500 (val_labs_copy) A copy of a null set is a null set.
1501 (val_labs_count) A null set has 0 labels.
1502 (val_labs_replace) Change return type to void. Rewrite for
1504 (val_labs_find) A null set does not contain the value.
1505 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1506 transposed argument order, updated all callers.
1508 * value.c: New file.
1509 (value_dup) Moved here from variable.c.
1510 (compare_values) Ditto.
1513 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1514 (macro MAX_LONG_STRING) Removed, because it was only interesting
1515 for system files, not for general code.
1516 (macro MAX_VERY_LONG_STRING) Ditto.
1517 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1518 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1520 * vardict.h: New file, for an interface between variables and
1523 * variable.c: A lot of functions were moved around, for better
1525 (struct variable) Move definition here, from variable.h.
1526 (var_type_adj) Removed--makes i18n hard.
1527 (var_type_noun) Ditto.
1528 (var_create) New function.
1529 (var_clone) New function.
1530 (var_destroy) New function.
1531 (var_set_name) Assert that variable is not in a dictionary.
1532 (compare_var_names) Rename compare_vars_by_name and fix a couple
1533 of callers who thought the args were strings.
1534 (hash_var_name) Rename hash_var_by_name.
1535 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1536 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1537 (var_is_very_long_string) Removed, because it was only interesting
1538 to system file code.
1539 (var_set_missing_values) Allow the argument to be the wrong width,
1540 as long as we can resize it. Simplify callers who were doing the
1541 resizing themselves.
1542 (var_get_value_labels) New function.
1543 (var_has_value_labels) New function.
1544 (var_set_value_labels) New function.
1545 (alloc_value_labels) New function.
1546 (var_add_value_label) New function.
1547 (var_replace_value_label) New function.
1548 (var_clear_value_labels) New function.
1549 (var_lookup_value_label) New function.
1550 (var_get_value_name) Moved here from variable.c, renamed from
1551 var_get_value_name, transposed argument order, updated all
1553 (var_to_string) Moved here, from variable-label.c.
1554 (var_set_leave) New function.
1555 (var_get_leave) New function.
1556 (var_must_leave) New function.
1557 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1558 set_var_short_name_suffix.
1559 (var_get_dict_index) New function.
1560 (var_get_case_index) New function.
1561 (var_get_obs_vals) New function.
1562 (var_set_obs_vals) New function.
1563 (var_has_obs_vals) New function.
1564 (var_get_vardict) New function.
1565 (var_set_vardict) New function.
1566 (var_has_vardict) New function.
1567 (var_clear_vardict) New function.
1568 (value_dup) Moved to value.c.
1569 (compare_values) Ditto.
1572 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1573 (enum ALPHA) Rename VAR_STRING, update all users.
1575 * vector.c: New file.
1576 (struct vector) Moved here, from variable.h.
1577 (check_widths) New function.
1578 (vector_create) New function.
1579 (vector_clone) New function.
1580 (vector_destroy) New function.
1581 (vector_get_name) New function.
1582 (vector_get_var) New function.
1583 (vector_get_var_cnt) New function.
1584 (compare_vector_ptrs_by_name) New function.
1586 * vector.h: New file.
1588 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1590 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1591 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1594 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1596 * value-labels.c (destroy_atoms): New function.
1597 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1599 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1601 Thanks to Jason Stover for pointing out this problem.
1603 * data-out.c (output_number): Use gsl_finite from GSL, which is
1604 portable, instead of isfinite, which is not.
1607 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1609 * variable.c variable.h (value_dup): New function.
1611 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1613 Start converting struct variable to an opaque type. In this
1614 phase, we add a bunch of setter and getter functions and convert
1615 most of the PSPP code to use them. The resulting changes are
1616 pervasive but mostly trivial, and only the notable changes are
1619 * format.c (fmt_equal): New function.
1621 * variable.c (var_type_is_valid): New function.
1622 (measure_is_valid) Moved here, from format.c.
1623 (alignment_is_valid) Moved here, from format.c.
1624 (var_get_name) New function.
1625 (var_set_name) New function.
1626 (width_to_type) New function.
1627 (var_get_type) New function.
1628 (var_get_width) New function.
1629 (var_set_width) New function.
1630 (var_is_numeric) New function.
1631 (var_is_alpha) New function.
1632 (var_is_short_string) New function.
1633 (var_is_long_string) New function.
1634 (var_is_very_long_string) New function.
1635 (var_get_missing_values) New function.
1636 (var_set_missing_values) New function.
1637 (var_clear_missing_values) New function.
1638 (var_has_missing_values) New function.
1639 (var_is_value_missing) New function.
1640 (var_is_num_missing) New function.
1641 (var_is_str_missing) New function.
1642 (var_is_value_user_missing) New function.
1643 (var_is_num_user_missing) New function.
1644 (var_is_str_user_missing) New function.
1645 (var_is_value_system_missing) New function.
1646 (var_get_print_format) New function.
1647 (var_set_print_format) New function.
1648 (var_get_write_format) New function.
1649 (var_set_write_format) New function.
1650 (var_set_both_formats) New function.
1651 (var_get_label) New function.
1652 (var_set_label) New function.
1653 (var_clear_label) New function.
1654 (var_has_label) New function.
1655 (var_get_measure) New function.
1656 (var_set_measure) New function.
1657 (var_get_display_width) New function.
1658 (var_set_display_width) New function.
1659 (var_get_alignment) New function.
1660 (var_set_alignment) New function.
1661 (var_get_value_cnt) New function.
1662 (var_get_leave) New function.
1663 (var_get_short_name) New function.
1665 * variable.h: (struct variable) Removed "type" and "nv" members;
1666 they are now computed from "width" where needed.
1668 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1670 * missing-values.c (mv_resize): Don't write beyond end of the
1671 allocated buffer when resizing a long string.
1673 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1675 Clean up identifier code: don't require identifier enumerations to
1676 be in a particular order; make better use of string library;
1677 expose less of the internals.
1679 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1680 change interface. Updated all callers.
1681 (lex_id_match) Change interface to use struct substring, update
1683 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1684 (global array keywords[]) Make static, change form. Update all
1685 users to use lex_id_name instead.
1686 (lex_is_keyword) New function.
1687 (lex_id_to_token) Change interface to use struct substring, update
1689 (lex_id_name) New function.
1691 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1692 lex_is_keyword instead.
1693 (T_LAST_KEYWORD) Removed.
1694 (T_N_KEYWORDS) Removed.
1696 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1698 * format.c: (fmt_date_template) Distinguish characters for which a
1699 space is output and any date delimiter is allowed on input, from
1700 those for which a space is output and only a space is allowed on
1701 input. The former is represented by X, the latter by a space.
1702 Also, drop distinction between h and H, changing the former to the
1705 * data-in.c: Completely rewrite internals to conform to SPSS input
1706 formats as closely as possible.
1707 (data_in) Changed external interface by replacing the structure
1708 that was used as a single argument by a set of arguments. Updated
1710 (data_in_finite_line) Removed. Converted all callers to use plain
1712 (data_in_get_integer_format) New function.
1713 (data_in_set_integer_format) New function.
1714 (data_in_get_float_format) New function.
1715 (data_in_set_float_format) New function.
1717 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1718 (struct data_in) Removed.
1720 * data-out.c: (output_date) Drop each component from the input as
1721 it is output, to allow us to drop the distinction between h (a
1722 count of hours) and H (the hour of day) template characters.
1723 Also, handle new X template character.
1724 (output_scientific) Follow more rational rule on when to drop
1725 fraction introduced between SPSS 13 and 15. Updated test case to
1728 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1730 Fix buffer overflow reported by John Darrington.
1732 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1733 realize that DIGITS is a count of nibbles, not of bytes.
1735 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1737 * calendar.c (calendar_offset_to_gregorian) Also return the
1738 year-of-day. Change callers to new interface.
1740 * data-out.c: Completely rewrite internals to conform to SPSS
1741 output formats as completely as possible.
1742 (data_out) Change interface to put input parameters before output
1743 parameters, for consistency with the style I now prefer. Update
1745 (data_out_get_integer_format) New public function.
1746 (data_out_set_integer_format) New public function.
1747 (data_out_get_float_format) New public function.
1748 (data_out_set_float_format) New public function.
1750 * data-out.h: New file. Move prototype for data_out here, from
1753 * format.c: (fmt_step_width) Use equality comparison instead of
1754 bitwise and, for clarity.
1755 (fmt_is_string) Ditto.
1756 (fmt_input_to_output) Fix categories that are translated to F
1759 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1761 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1762 casefile-private.h: Added casefilter to assist commands with missing
1765 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1767 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1769 * settings.c: (route_errors_to_terminal) New variable.
1770 (route_errors_to_listing) New variable.
1771 (get_error_routing_to_terminal) New function.
1772 (set_error_routing_to_terminal) New function.
1773 (get_error_routing_to_listing) New function.
1774 (set_error_routing_to_listing) New function.
1776 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1778 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1780 * format.c: Completely rewrite, to achieve better abstraction.
1781 Rewrite all references to formats in other files.
1783 * format.def: Rewrite and reorganize.
1785 * settings.c: Move everything related to custom currency formats
1786 into format.[ch], changing them in form, so as to group related
1787 code and definitions better. Changed all references to use the
1789 (static var decimal) Removed.
1790 (static var grouping) Removed.
1791 (static var cc) Removed.
1792 (get_decimal) Removed.
1793 (set_decimal) Removed.
1794 (get_grouping) Removed.
1795 (set_grouping) Removed.
1799 * settings.h: (macro CC_CNT) Removed.
1800 (macro CC_WIDTH) Removed.
1801 (struct custom_currency) Removed.
1803 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1805 * data-in.c (data_in): Use switch statement instead of table, to
1806 avoid dependence on the order of the FMT_* enums.
1808 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1810 * data-out.c: (num_to_string) Removed, because it was dead code.
1812 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1814 * data-in.c (parse_trailer): Fix error message.
1816 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1818 * format.c (fmt_is_binary): New function.
1820 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1822 * procedure.c procedure.h: Encapsulated the static data into a single
1825 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1827 * casefile.c (casereader_read_xfer): Always initialize the case,
1828 even on an error condition.
1830 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1832 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1833 entire function useless.
1835 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1837 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1838 casereader method casereader_clone.
1840 * procedure.c pransformations.h: Introduced new type casenum_t
1842 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1844 * variable.c: (width_to_bytes) Rephrase code for clarify.
1846 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1848 * format.c: (fmt_type_from_string) New function.
1849 (fmt_to_string) Include decimals in output if the format has
1850 decimals, even if the format type does not. This way, we can
1851 accurately reproduce incorrect formats in user output.
1852 (check_common_specifier) Make the check for a bad format type an
1853 assertion, so we get bug reports if they show up. Fix message.
1854 Check for decimal places with a format type that doesn't allow
1856 (check_input_specifier) Remove check for FMT_X, which has been
1858 (check_output_specifier) Ditto.
1860 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1862 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1863 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1864 (enum fmt_parse_flags) Removed.
1866 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1868 * casefile.c casefile.h: Converted to an abstract base class.
1869 * casefile-private.h fastfile.c fastfile.h: New files.
1870 * automake.mk procedure.c scratch-writer.c storage-stream.c
1872 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1874 * procedure.c (internal_procedure): Create sink_case with only as
1875 many values as the compacted dictionary.
1877 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1879 Remove "debugging" code that caused plenty of false positives and
1882 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1884 * case.c: Remove all references to `this' member.
1886 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1888 Fix link error noted by Jason Stover.
1890 * storage-stream.c: Include <assert.h>.
1892 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1894 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1895 ALL) and additional underlying system file issues.
1897 Thanks to John Darrington for review.
1899 First problem: var_hash points to variables not owned by the
1900 sys-file-reader, which the caller may free or modify. Use an
1901 array of sfm_vars instead, as done earlier (e.g. CVS version
1904 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1905 members and remove all code that references it. Add vars, var_cnt
1906 members. Remove fix_specials member, which was unused.
1907 (struct sfm_var) Remove name member, which was unused.
1908 (sfm_close_reader) Free vars member instead of var_hash.
1909 (compare_var_shortnames) Removed.
1910 (hash_var_shortname) Removed.
1911 (sfm_open_reader) Fill out vars array.
1912 (compare_var_index) Removed.
1913 (sfm_read_case) Use vars instead of var_hash.
1915 Second problem: we're confused about when we actually have very
1916 long strings, causing us to choose incorrectly between slow path
1917 and fast path in sfm_read_case.
1919 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1920 have very long strings, not when we have long variable names,
1921 which is an unrelated feature.
1923 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1925 * variable.h: Move var_set and variable parsing declarations to
1926 new header, src/language/lexer/variable-parser.h. Modified lots
1927 of files to include the new header.
1929 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1931 * value-labels.c (value_to_string): When there's no value label,
1932 format the variable according to its print format, instead of
1933 always effectively using A or F format.
1935 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1937 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1940 * format.c: Removed tortological assertion.
1942 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1944 Reform string library.
1946 * file-name.c (fn_interp_vars): Change interface to take a
1947 substring as input. Updated all users.
1949 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1951 * format.c (measure_is_valid): Really return false when m >=
1954 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1956 Implement random access to casefiles, for use in GUI.
1958 * casefile.c: (struct casereader) Add `random', `file_ofs',
1959 `buffer_ofs' members.
1960 (casefile_get_random_reader) New function.
1961 (read_open_file) Break part into new function
1962 seek_and_fill_buffer().
1963 (fill_buffer) Update buffer_ofs, file_ofs.
1964 (casereader_seek) New function.
1966 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1968 * settings.c: Added call to i18n{done, init}.
1970 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1972 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1975 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1977 * casefile.c: Convert many uses of `int' to `bool'.
1979 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1981 * transformations.c (trns_chain_destroy): Destroy chain's trns
1982 member, to fix memory leak.
1984 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1986 * storage-stream.c (storage_source_decapsulate): Destroy case
1987 source to fix memory leak.
1989 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1991 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1992 case passed as argument instead of initializing the argument as a
1993 case. Fixes memory leak that showed up in
1994 tests/command/aggregate.sh with scratch files.
1996 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1998 * procedure.c (proc_done): Destroy default_dict, to fix memory
2001 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2003 Simplify procedure_with_splits().
2005 * procedure.c (struct split_aux_data): Removed case_count member.
2006 (procedure_with_splits) Don't initialize case_count.
2007 (split_procedure_case_func) Check whether prev_case is null
2008 instead of case_count.
2009 (split_procedure_end_func) Ditto.
2011 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2013 * case.c (case_move): Do nothing if dst and src are the same
2015 (case_try_create) Merge two similar cases.
2016 (case_copy) Unshare only if data must be actually copied.
2018 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2020 * data-in.c data-out.c dictionary.c sys-file-reader.c
2021 sys-file-writer.c variable.c variable.h: Reworked very long string
2022 support for better encapsulation.
2024 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2026 * value-labels.c (val_labs_can_set_width): New function.
2027 (val_labs_set_width) Clear labels if increasing width to long
2029 (val_labs_destroy) Remove unneeded test for null.
2031 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2033 * value-labels.h: Remove unneeded dependency on variable.h.
2035 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2037 Get rid of `char *c' member in union value, for cleanliness.
2039 * value.h: (union value) Remove `c' member.
2041 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2043 Make dictionary compacting functions a little more general.
2045 * sys-file-writer.c (sfm_open_writer): Use
2046 dict_compacting_would_change().
2047 (does_dict_need_translation) Removed.
2049 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2051 Make dictionary compacting functions a little more general.
2053 * dictionary.c (dict_needs_compaction): Rename
2054 dict_compacting_would_shrink(). Update all callers.
2055 (dict_compacting_would_change) New function.
2057 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2059 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2060 inverted return value (!).
2062 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2064 Continue reforming procedure execution.
2066 * procedure.c: Search and replace "vfm" by "proc". Notably:
2067 (static var vfm_source) Rename proc_source. Update all
2069 (static var vfm_sink) Rename proc_sink. Update all references.
2071 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2073 Continue reforming procedure execution. In this phase, remove
2074 PROCESS IF, which was deprecated anyway and can be easily
2075 simulated with TEMPORARY followed by SELECT IF.
2077 * procedure.c: (open_active_file) Don't call
2078 add_process_if_trns().
2079 (discard_variables) Get rid of redundant call to
2080 proc_cancel_all_transformations().
2081 (add_process_if_trns) Removed.
2082 (process_if_trns_proc) Removed.
2083 (process_if_trns_free) Removed.
2085 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2087 Continue reforming procedure execution. In this phase, add
2088 `const' to the case passed to procedure()'s callback.
2090 Updated all users of procedure() as well.
2092 * procedure.c: (struct write_case_data) Add "const" to ccase
2093 parameter for case_func member.
2094 (procedure) Add "const" to ccase parameter for proc_func
2096 (multipass_case_func) Make ccase parameter const.
2097 (internal_procedure) Add "const" to ccase parameter for case_func
2099 (split_procedure_case_func) Make ccase parameter const.
2100 (multipass_split_case_func) Make ccase parameter const.
2102 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2104 Continue reforming procedure execution. In this phase, get rid of
2105 the output code for SPLIT FILE groups in procedure.c, which really
2106 shouldn't be doing any output. Move it into the individual
2107 procedures instead. This also adds some flexibility.
2109 Updated many users of procedure_with_splits() and
2110 multipass_procedure_with_splits() to call
2111 output_split_file_values() and to deal with increased use of
2114 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2115 parameter to begin_func member.
2116 (procedure_with_splits) Add "const struct ccase *" parameter to
2117 begin_func parameter. Make ccase parameter const in proc_func
2119 (split_procedure_case_func) Don't dump split file group. Pass
2121 (dump_splits) Moved to language/dictionary/split-file.c as
2122 output_split_file_values().
2123 (struct multipass_split_aux_data) Add "const struct ccase *"
2124 parameter to split_func member.
2125 (multipass_procedure_with_splits) Add "const struct ccase *"
2126 parameter to split_func parameter.
2127 (multipass_split_case_func) Save new SPLIT FILE case before
2129 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2131 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2133 Continue reforming procedure execution. Change
2134 internal_procedure() so that it calls open_active_file() and
2135 close_active_file(), which isolates most of the actual procedure
2138 * procedure.c: (struct write_case_data) Rename `proc_func' member
2139 to `case_func' and update all references.
2140 (procedure) Rewrite as one-line wrapper around
2141 internal_procedure().
2142 (struct multipass_aux_data) New.
2143 (multipass_callback) Renamed multipass_case_func(). Use struct
2144 multipass_aux_data as auxiliary data.
2145 (multipass_end_func) New function.
2146 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2147 that uses multipass_case_func, multipass_end_func.
2148 (internal_procedure) Add `end_func' argument. Move optimization
2149 of trivial case in here. Move call to open_active_file() and
2150 close_active_file() in here. Now assert that vfm_source is
2152 (procedure_with_splits_callback) Rename
2153 split_procedure_case_func().
2154 (split_procedure_end_func) New function.
2155 (multipass_split_callback) Rename multipass_split_case_func.
2156 (multipass_split_end_func) New function.
2157 (discard_variables) No need to test for nonnull vfm_source.
2159 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2161 Continue reforming procedure execution. Get rid of unused member.
2163 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2165 (write_case) Don't increment cases_analyzed.
2167 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2169 Continue reforming procedure execution. In this phase, move
2170 procedure.c and procedure.h from src to src/data. Update
2171 makefiles and #includes accordingly.
2173 * procedure.c: Moved here from src/.
2175 * procedure.h: Moved here from src/.
2177 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2179 Continue reforming procedure execution. In this phase, get rid of
2180 many global variables, consolidating procedure execution in
2181 procedure.c. Encapsulate transformations in new "struct
2182 trns_chain". Also, change implementation of N OF CASES, FILTER,
2183 and PROCESS IF from special cases to transformations.
2185 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2188 * dictionary.c: (global variable default_dict) Move to
2191 * variable.h: (TRNS_*) Move to transformations.h.
2192 (struct transformation) Move to transformations.c.
2194 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2196 * sys-file-reader.c: Fixed invalid read problems.
2198 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2200 * storage-stream.c: Add missing function comments.
2202 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2204 Continue reforming procedure execution. In this phase, add some
2205 new, needed functionality to storage-stream.
2207 * storage-stream.c: (storage_source_decapsulate) New function.
2209 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2211 * variable.c (width_to_bytes): Declarations must precede
2212 statements for C90 compliance.
2214 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2216 * data-out.c, data-in.c, variable.c, variable.h: New functions
2217 copy_mangle and copy_demangle for reading/writing cases; emulates the
2218 way SPSS deals with strings > 255 bytes.
2220 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2221 subtype 14 needed for strings longer than 255 bytes.
2223 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2224 instead of literal values. Also fixed some constness issues.
2226 * format.h: Constness
2228 * sfm-private.h: Renamed the case_size identifier, since I discovered
2229 that SPSS's respect for this variable is very nominal.
2231 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2233 Change case limit type from int to size_t.
2235 * dictionary.c: (struct dictionary) Change type of case_limit
2237 (dict_get_case_limit) Change return type.
2238 (dict_set_case_limit) Change parameter type.
2240 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2242 * variable.h: (struct variable) Rename `reinit' member as `leave'
2243 and invert sense. Fix up all references.
2245 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2247 Continue reforming procedure execution. In this phase, break
2248 procedure.c into multiple files.
2250 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2252 * case-sink.c: New file.
2254 * case-sink.h: New file.
2256 * case-source.c: New file.
2258 * case-source.h: New file.
2260 * storage-stream.c: New file.
2262 * storage-stream.h: New file.
2264 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2266 * variable.h: (struct variable) Remove `init' member and all
2267 references to it from other files. It was initialized in several
2268 places, but nothing really ever used it for anything worthwhile.
2269 Thanks to Jason Stover for pointing out how confusing this
2272 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2274 Continue reforming error message support. In this phase, get rid
2275 of message "titles" and put the message text in `struct error'.
2276 Now `struct error' encapsulates a message more properly.
2278 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2279 Format message ourselves.
2281 * data-in.c: (vdls_error) Ditto.
2283 * por-file-reader.c: (error) Ditto.
2285 * sys-file-reader.c: (corrupt_msg) Ditto.
2287 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2289 GNU standards require "file name" instead of "filename" in
2290 documentation. It's nice for our code to follow the convention
2293 * casefile.c: (struct casefile) Rename `filename' member to
2294 `file_name'. Updated all references.
2296 * file-name.c: [!unix] (struct file_identity) Rename
2297 normalized_filename member to normalized_file_name. Updated all
2300 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2302 We don't really support anything but Unix-like environments well,
2303 so we might as well de-obfuscate by writing directory and path
2304 separators explicitly.
2306 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2308 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2309 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2311 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2314 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2316 GNU standards require "file name" instead of "filename" in
2317 documentation. It's nice for our code to follow the convention
2320 * filename.c: Rename to file-name.c.
2322 * filename.h: Rename to file-name.h. Update all inclusions.
2323 Update header guards.
2325 * automake.mk: Update file names.
2327 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2329 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2331 (fn_basename) Removed (dead code).
2332 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2333 (fn_special_p) Renamed fn_is_special(), all references updated.
2334 (fn_exists_p) Renamed fn_exists(), all references updated.
2336 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2338 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2339 Also, now it only tilde-expands file names, not paths.
2340 (fn_search_path) Tilde-expand one directory at a time.
2342 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2344 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2345 get rid of non-Unixlike version of the code, which has probably
2347 (fn_prepend_dir) Removed (dead code).
2349 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2350 (macro PATH_SEPARATOR_STRING) New.
2351 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2353 Continue reforming error message support. In this phase, we get
2354 rid of VM() and the other msg() support for "verbosity", replacing
2355 it by a new function verbose_msg().
2357 * filename.c: (fn_search_path) Use verbose_msg() instead of
2360 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2362 * sfm-private.h: Get rid of #defines after #error, which makes no
2365 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2367 Get rid of our own int32 type in favor of the standard int32_t
2370 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2373 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2375 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2377 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2379 Remove ill-considered file routines that are no longer used.
2381 * filename.c: (fn_open_ext) Removed.
2382 (fn_close_ext) Removed.
2384 * filename.h: (struct file_ext) Removed.
2386 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2388 * variable.c (var_is_valid_name): Move declarations before code
2391 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2393 * filename.ch (fn_interp_vars): Fixed small buglet.
2395 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2397 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2398 so as to modify the string inline. Thus makeing it easier to
2399 destroy the results when no longer needed.
2401 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2403 * category.c (cat_stored_values_destroy): Fixed memory leak.
2405 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2407 Add some missing frees. Thanks to John Darrington for reporting
2410 * any-writer.c (any_writer_close): Free writer.
2412 * any-reader.c (any_reader_close): Free reader.
2414 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2416 * por-file-reader.c: (error) Mark as NO_RETURN.
2418 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2420 * settings.c: Changed default value of scompress to true.
2422 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2424 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2426 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2429 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2430 here from pref.h.orig.
2432 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2434 * sys-file-reader.c: Fixed bug reading compressed files.
2436 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2438 * Numerous renames. See src/ChangeLog for details.
2440 * Moved files from src directory