1 2008-02-18 Ben Pfaff <blp@gnu.org>
3 Patch #6426. Reviewed by John Darrington.
5 * format.c (min_width): Renamed fmt_min_width and made public.
6 Updated all references.
7 (max_width): Renamed fmt_max_width and made public. Updated all
9 (max_decimals): Renamed fmt_max_decimals and made public. Updated
12 2008-02-09 Ben Pfaff <blp@gnu.org>
14 Add a couple of extensions to GET DATA TYPE=TXT. Patch #6412.
15 Thanks to John Darrington for review.
17 * data-in.c (data_in): Add new argument to designate the last
18 column of the data field being parsed, for use in error messages.
21 * data-parser (struct data_parser): New member `quote_escape'.
22 (data_parser_create): Initialize quote_escape.
23 (data_parser_set_quotes): New function.
24 (cut_field): Support escaped quotes.
25 (parse_delimited_span): Ditto.
26 (parse_delimited_no_span): Ditto.
28 * get-data.c (parse_get_txt): Support ESCAPE extension subcommand
29 in enhanced mode. Only support multiple quote characters in
32 2008-02-06 John Darrington <john@darrington.wattle.id.au>
34 psql-reader.c psql-reader.h: Read more than one tuple at
35 once. Fix bug reading a query which returns no data. Fix bug
36 when transformation followed a reader.
37 Ask the server for the number of records in the query, for the
40 2008-02-05 John Darrington <john@darrington.wattle.id.au>
42 psql-reader.c: So yesterday they release postgresql 8.3.0
43 which has money represented with 64 bits. They must get
46 2008-02-02 John Darrington <john@darrington.wattle.id.au>
48 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
49 for reviewing this code.
51 2008-02-02 Ben Pfaff <blp@gnu.org>
55 * sys-file-reader.c (read_variable_record): Allow missing values
56 to be specified on long string variables, but warn about them
57 because PSPP does not yet support them.
58 (read_extension_record): Ignore extension records 20 and 21, which
59 PSPP does not yet support.
60 (read_header): Fix error message when floating-point format cannot
63 2008-02-01 Ben Pfaff <blp@gnu.org>
65 Patch #6386. Thanks to John Darrington for review and for the
66 updates to gnumeric-reader.c.
68 * dictionary.c (make_hinted_name): New function.
69 (make_numeric_name): New function.
70 (dict_make_unique_var_name): New function.
72 * gnumeric-reader.c (devise_name): Removed.
73 (munge_name): Removed.
74 (gnumeric_open_reader): Use new function
75 dict_make_unique_var_name.
77 * short-names.c (set_var_short_name_suffix): Use new function
80 2008-01-19 John Darrington <john@darrington.wattle.id.au>
82 * settings.c settings.h: Moved static variables into a
83 single struct. Renamed functions to have a settings_ prefix.
85 2008-01-14 John Darrington <john@darrington.wattle.id.au>
87 * data-out.c (output_AHEX): Corrected number of bytes in
88 call to output_hex. Closes bug #22011
90 2008-01-02 John Darrington <john@darrington.wattle.id.au>
92 * variable.c variable.h: Replaced var_get_value_name with
93 var_append_value_name which doesn't use any static data.
94 Thanks to Ben for review.
96 2007-12-07 Ben Pfaff <blp@gnu.org>
100 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
101 made from an empty casereader produce a casegrouper with no
102 groups, instead of one with one group that has no cases.
104 * casereader.c (casereader_is_empty): New function.
106 2007-12-06 Ben Pfaff <blp@gnu.org>
110 * sys-file-reader.c (read_display_parameters): Handle variable
111 display parameters record with only 2 data items per variable.
112 Reported by Guido Gay <gay@irer.it>.
114 2007-12-04 Ben Pfaff <blp@gnu.org>
116 * identifier.c (lex_id_match_n): New function.
117 (lex_id_match): Reimplement in terms of lex_id_match_n.
119 2007-11-24 Ben Pfaff <blp@gnu.org>
121 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
124 2007-11-24 Ben Pfaff <blp@gnu.org>
126 Fix warning reported when reading back system files that include
127 very long string variables. Thanks to Guido Gay <gay@irer.it> for
130 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
131 was written where `j' was meant.
133 2007-11-18 Ben Pfaff <blp@gnu.org>
135 Properly write variables that include a range of missing values to
136 system files. Thanks to Guido Gay <gay@irer.it> for reporting
139 * sys-file-writer.c (write_variable): Correctly calculate
140 n_missing_values field when writing variables that include a range
143 2007-11-10 Ben Pfaff <blp@gnu.org>
145 Cleanups and bug fixes devised while writing up documentation.
148 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
150 * dict-class.c: New file.
151 (dict_class_from_id): Move here.
152 (dict_class_to_name): Move here.
154 * dict-class.h: New file.
155 (enum dict_class): Move here. Change from consecutive integers to
156 consecutive bits, to make testing for any of multiple values
157 easier. Add new DC_ALL constant.
159 * dictionary.c (struct dictionary): Change `case_limit' from
160 size_t to casenumber.
161 (dict_get_vars): Make final argument an enum dict_class.
162 (dict_get_vars_mutable): Ditto.
163 (dict_get_case_limit): Change return value to casenumber.
164 (dict_set_case_limit): Change final argument to a casenumber.
165 (dict_unset_split_var): Add assertion. Rephrase slightly.
166 (dict_set_label): Use xstrndup to simplify.
168 * format.c (fmt_step_width): AHEX format also needs 2-byte
170 (fmt_set_style): Simplify assertions.
172 * missing-values.c (mv_add_num_range): Rename mv_add_range.
173 Simplify implementation.
174 (mv_has_value): Simplify implementation.
175 (mv_pop_value): Remove the first value, not the last, to avoid
176 having GET followed by SAVE reverse the order of missing values.
177 (mv_peek_value): Rename mv_get_value. Simplify assertion.
178 (mv_has_range): Simplify implementation.
179 (mv_peek_range): Rename mv_get_range.
180 (can_resize_string): Removed.
181 (mv_is_resizable): Use value_is_resizable.
182 (mv_resize): Use value_resize.
184 * short-names.h (SHORT_NAME_LEN): Move here.
186 * val-type.h: New file, for definitions related to type and width
187 of abstract values. Before, these definitions were mixed among
188 those related to "union value" and those related to variables.
189 (macro SYSMIS): Move here.
190 (macro LOWEST): Move here.
191 (macro HIGHEST): Move here.
192 (macro MAX_STRING); Move here.
193 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
194 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
195 All references updated.
196 (val_type_is_valid): New function. Replaces var_type_is_valid.
197 All references updated.
198 (val_type_from_width): New function. Replaces
199 var_type_from_width. All references updated.
201 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
203 (val_labs_can_set_width): Use value_is_resizable.
204 (val_labs_add): Simply return false if the value labels set is too
205 wide, instead of having undefined behavior.
206 (val_labs_replace): Ditto.
207 (val_labs_replace): Ditto.
208 (val_labs_first): Set iterator to null if iteration is complete.
209 (val_labs_first_sorted): Ditto.
210 (val_labs_done): Become a no-op if the iterator is null.
212 * value.c (value_is_resizable): New function.
213 (value_resize): New function.
215 * variable.c (var_get_dict_class): New function.
217 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
220 2007-11-08 Ben Pfaff <blp@gnu.org>
222 * data-in.c: Make formatted data parsing locale-independent.
223 (parse_number): Use c_strtod instead of strtod, to avoid
224 locale-specific behavior.
227 2007-11-06 Ben Pfaff <blp@gnu.org>
229 Patch #6256: add support for binary, 360 file formats. Reviewed
232 * data-in.c (struct data_in): Add `encoding' member.
233 (data_in): Add `encoding' parameter, and re-encode the data passed
234 in where appropriate. Update all callers to pass it in.
235 (parse_A): Implement EBCDIC recoding wart described in manual.
236 (parse_AHEX): Implement EBCDIC recoding.
238 * data-out.c (data_out_legacy): New function.
239 (data_out): Make into a wrapper around data_out_legacy.
241 * file-handle-def.c (struct file_handle): New member `encoding'.
242 (fh_create_file): Set encoding.
243 (fh_default_properties): Set default encoding.
244 (fh_get_legacy_encoding): New function.
246 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
247 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
248 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
249 (struct fh_properties): New member `encoding'.
251 2007-11-05 Ben Pfaff <blp@gnu.org>
253 Patch #6258. Reviewed by John Darrington.
255 * file-handle-def.c (fh_lock): Add comment that TYPE should be
256 marked with N_() in the caller. Added these markings to each
257 caller too. Should make i18n easier.
258 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
260 2007-11-03 Ben Pfaff <blp@gnu.org>
262 Allow output files to overwrite input files (bug #21280).
264 * file-handle-def.c: Separate locking of files for input and for
265 output, to allow both to take place at once. Also, distinguish a
266 file handle from the identity of the underlying file, because the
267 identity of a file changes over time and the file handle can't
268 represent two different identities.
269 (struct file_handle): Remove `next', `open_cnt', `deleted',
270 `type', `open_mode', `aux', `identity' members. Change `id' from
271 char array to char *. Add `ref_cnt' member.
272 (file_handle_from_ll) New function.
273 (file_handles) Removed.
274 (named_handles) New variable.
275 (fh_init) Initialize named_handles.
276 (fh_done) Remove name from all named_handles.
277 (free_handle) Updated for modified struct file_handle.
278 (unname_handle) New function.
279 (fh_ref) New function.
280 (fh_from_file_name) Removed.
281 (fh_unref) New function.
282 (fh_unname) New function.
283 (fh_from_id) Rewritten.
284 (create_handle) Updated for modified struct file_handle.
285 (fh_inline_file) Increment the handle's ref count.
286 (fh_create_file) Updated for modified struct file_handle.
287 (fh_create_scratch) Ditto.
292 (fh_is_open) Removed.
293 (fh_get_id) Updated for modified struct file_handle.
294 (fh_get_default_handle) Increment the handle's ref count.
295 (fh_set_default_handle) Handle ref counts.
296 (struct fh_lock) New structure.
297 (locks) New static var.
298 (fh_lock) New function.
299 (fh_unlock) New function.
300 (fh_lock_get_aux) New function.
301 (fh_lock_set_aux) New function.
302 (fh_is_locked) New function.
303 (make_key) New function.
304 (free_key) New function.
305 (compare_fh_locks) New function.
306 (hash_fh_lock) New function.
308 * file-handle-def.h (enum fh_access) New enum.
310 * file-name.c: Made file_identity the same in all supported
312 (struct file_identity): New `name' member.
313 (fn_get_identity): For a file that doesn't exist, get the
314 dev/inode of its directory plus its name. If even the directory
315 doesn't exist, just use its name. Merge the Windows
316 implementation into the Unix one.
317 (fn_compare_file_identities): Rewritten. Merge the Windows
318 implementation into the Unix one.
319 (fn_hash_identity): New function.
321 * make-file.c (struct replace_file): New structure.
322 (all_files): New static var.
323 (replace_file_start): New function.
324 (replace_file_commit): New function.
325 (replace_file_abort): New function.
326 (free_replace_file): New function.
327 (unlink_replace_files): New function.
329 * por-file-reader.c (struct pfm_reader): Add `lock' member.
330 (close_reader): Unlock file.
331 (pfm_open_reader): Lock file.
333 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
335 (pfm_open_writer): Lock file and prepare for its replacement.
336 (close_writer): Unlock file.
338 * scratch-handle.h (struct scratch_handle): Add unique_id so that
339 different generations of a scratch file can be distinguished.
341 * scratch-reader.c (scratch_reader_open): Verify that the file is
344 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
345 remove scratch_handle member.
346 (scratch_writer_open): Lock handle. Prepare to replace handle
347 data, instead of doing it immediately.
348 (scratch_writer_casewriter_destroy): Replace handle data and
351 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
352 (sfm_open_reader): Lock file.
353 (close_reader): Unlock file.
355 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
357 (sfm_open_writer): Lock file and prepare for its replacement.
358 (close_writer): Unlock file.
360 2007-11-02 Ben Pfaff <blp@gnu.org>
362 * data-out.c (output_number): Use isfinite (from C99) instead of
363 gsl_isfinite, since we now have universal support for it in
365 (output_infinite): Ditto.
366 [!HAVE_ROUND] (round): Remove definition, since we now have a
367 replacement in gnulib.
369 2007-10-27 John Darrington <john@darrington.wattle.id.au>
371 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
372 wasn't initialised correctly. Closes bugs #21330 and 21397
375 2007-10-08 Ben Pfaff <blp@gnu.org>
377 Fix bug #21280. Thanks to John Darrington for review.
379 * file-name.c (create_stream): New function.
381 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
382 file handle before creating the file, to ensure that we don't
383 truncate a file that we're reading. Make code easier to read by
386 * sys-file-write.c (sfm_open_writer): Ditto.
388 2007-10-01 Ben Pfaff <blp@gnu.org>
390 Fix bug #21192. Thanks to John Darrington for review.
392 * casereader.c (casereader_read): Decrement case_cnt before
393 calling the casereader's "read" member function, so that we
394 interact properly with lazy_casereader.
396 * datasheet.c: Add regression test for above bug fix.
397 (clone_datasheet): New function.
398 (lazy_callback): New function.
399 (check_datasheet_casereader): New function.
400 (check_datasheet): Check datasheet contents are reported correctly
401 through an ordinary casereader and a lazy casereader.
402 (clone_model): Use clone_datasheet.
404 2007-09-24 Ben Pfaff <blp@gnu.org>
406 Patch #6210. Reviewed by John Darrington.
407 * settings.c: Drop "long view", which is not needed any longer.
408 (static var long_view): Removed.
409 (force_long_view): Removed.
410 (get_viewwidth): Removed.
411 (init_viewport): Removed long_view reference.
413 2007-09-19 John Darrington <john@darrington.wattle.id.au>
415 * settings.c settings.h: Changed viewport's length and width to be
416 owned by the user interface which uses the data library. This allows
417 better abstraction, and makes dynamically adjustable dimensions easier.
419 2007-09-18 Ben Pfaff <blp@gnu.org>
421 * procedure.c (proc_extract_active_file_data): New function.
423 * lazy-casereader.h: New file.
425 * lazy-casereader.c: New file.
427 * casereader.c (casereader_dynamic_cast): New function.
429 2007-09-14 Ben Pfaff <blp@gnu.org>
431 * dictionary.c (dict_clone): Copy case indexes from cloned
432 dictionary. Fixes bug #21061. Reviewed by John Darrington.
434 2007-09-13 John Darrington <john@darrington.wattle.id.au>
436 * value.c value.h (create_value): New function.
438 2007-09-12 Ben Pfaff <blp@gnu.org>
440 Make it clear that translator casereader and casewriter translate
441 functions are supposed to destroy their input case.
443 * casereader-translator.c (struct casereader_translator): Change
444 input case parameter of translate member function from const
445 struct ccase * to struct ccase *.
446 (casereader_create_translator): Ditto, for translate parameter.
448 * casewriter-translator.c (struct casewriter_translator): Ditto.
449 (casewriter_create_translator): Ditto.
451 2007-08-27 John Darrington <john@darrington.wattle.id.au>
453 * sys-file-reader.c (read_display_parameters): Force display width
454 to 8, if the sys file says 0 (like SPSS does).
456 2007-08-12 Ben Pfaff <blp@gnu.org>
458 * dictionary.c (dict_dump): New function.
460 2007-08-12 Ben Pfaff <blp@gnu.org>
462 Drop dict_compactor in favor of using the new struct case_map.
464 * dictionary.c (struct copy_map): Removed.
465 (struct dict_compactor): Removed.
466 (dict_make_compactor): Removed.
467 (dict_compactor_compact): Removed.
468 (dict_compactor_destroy): Removed.
470 * procedure.c (struct data_set): Change `compactor' member to be a
472 (proc_open): Use case_map_to_compact_dict instead of
474 (proc_casereader_read): Use case_map_execute instead of
475 dict_compactor_compact.
476 (proc_commit): Use case_map_destroy instead of
477 dict_compactor_destroy.
479 * scratch-writer.c (struct scratch_writer): Change `compactor'
480 member to be a struct case_map *.
481 (scratch_writer_open): Use case_map_to_compact_dict instead of
483 (scratch_writer_casewriter_write): Use case_map_execute instead of
484 dict_compactor_compact.
486 2007-08-12 Ben Pfaff <blp@gnu.org>
488 * automake.mk: Add case-map.c, case-map.h.
490 * case-map.c: New file.
492 * case-map.h: New file.
494 2007-08-12 Ben Pfaff <blp@gnu.org>
496 * dictionary.c (dict_compact_values): Don't delete scratch
497 variables as well as compacting case indexes. Update all callers.
498 (dict_get_compacted_value_cnt): Rename dict_count_values and
499 change interface. Update all callers.
500 (dict_get_compacted_value_cnt): Remove.
501 (dict_compacting_would_shrink): Remove.
502 (dict_compacting_would_change): Remove.
503 (dict_make_compactor): Add new parameter. Update all callers.
505 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
506 instead of dict_count_values, changing an O(N) operation into
509 2007-08-12 Ben Pfaff <blp@gnu.org>
511 * casereader.c (casereader_read): Don't require cases read by a
512 casereader to be exactly the expected size: as long as they're big
515 2007-08-12 Ben Pfaff <blp@gnu.org>
517 Make casewriters keep track of the number of `union value's in
518 each case. This is useful for two reasons: casewriter_write can
519 then check that the case being written is large enough, and later
520 recipients of the casewriter can determine the size of the case.
522 * casewriter-translator.c (casewriter_create_translator): Add
525 * casewriter.c (struct casewriter): Add value_cnt member.
526 (casewriter_write): Check that the case passed in is big enough.
527 (casewriter_get_value_cnt): New function.
528 (casewriter_create): Add value_cnt parameter.
530 2007-08-09 Ben Pfaff <blp@gnu.org>
532 Fix bug reported by Jason Stover.
533 * settings.c: Move get_termcap_viewport and in particular the
534 #include for <curses.h> to the end of the file. curses.h
535 redefines bool on some systems (e.g. OpenBSD), which causes
536 disagreement between uses of bool before its inclusion and
538 Tested by Jason Stover.
540 2007-07-29 Ben Pfaff <blp@gnu.org>
542 Provisional fix for bug #18692 and bug #20161. Reviewed by John
545 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
546 argument (never "rb" or "wb") because SUSv3 says that only those
547 modes are defined, and glibc in fact rejects other modes.
549 Open portable files with fn_open so that they can be read from
550 pipes. Fix missing fh_close call to go along with fh_open.
551 Report an error if the file close reports an error.
552 * por-file-reader.c (close_reader): New function.
553 (por_file_casereader_destroy): Use close_reader.
554 (pfm_open_reader): Open file with fn_open.
556 2007-07-28 Ben Pfaff <blp@gnu.org>
558 Make PSPP able to read all the portable files I could find on the
559 web. Thanks to John Darrington for review. Bug #17620.
560 * por-file-reader.c (struct pfm_reader): New member `line_length'.
561 (error): Print file offset in hexadecimal.
562 (warning): New function.
563 (advance): Treat lines less than 80 bytes long as padded to 80
565 (pfm_open_reader): Call read_documents if we find an "E" record.
566 (convert_format): Convert invalid formats to the default format
567 instead of aborting reading the file.
568 (read_variables): Rename duplicate variable names instead of
569 aborting reading the file.
570 (read_value_label): Allow string variables of different widths to
571 be assigned value labels in the same record. Replace duplicate
572 value labels instead of aborting.
573 (read_documents): New function.
575 * por-file-writer.c (pfm_open_writer): Call write_documents if the
576 dictionary has documents.
577 (write_documents): New function.
579 2007-07-25 Ben Pfaff <blp@gnu.org>
581 Fix bugs related to bug #17213.
583 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
584 since the former is what config.h has. Include the needed ncurses
586 (static var echo) Rename to `do_echo' because the original name is
587 the same as an ncurses identifier.
588 (get_termcap_viewport) Use error instead of msg.
590 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
591 (fn_close): Don't close stdin, stdout, stderr.
593 2007-07-26 John Darrington <john@darrington.wattle.id.au>
595 * procedure.c procedure.h: Added callbacks which get invoked whenever
596 a dataset's transformation chain changes.
598 2007-07-24 Ben Pfaff <blp@gnu.org>
601 * sys-file-writer.c (write_variable_display_parameters): Use new
602 var_default_display_width function to choose display width of
603 segments after the first one in a given variable.
604 * variable.c (var_create): Use var_default_display_width to pick
605 new variable's display width.
606 (var_default_display_width): New function.
607 Reviewed by John Darrington.
609 2007-07-24 Ben Pfaff <blp@gnu.org>
612 * por-file-writer.c (write_variables): Write weight variable.
613 Reviewed by John Darrington.
615 2007-07-23 Ben Pfaff <blp@gnu.org>
617 Improvements to system file reader and writer.
619 First, move all detailed knowledge of very long strings into
620 sys-file-private.[ch], so that this nasty stuff can be isolated.
622 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
623 (EFFECTIVE_VLS_CHUNK): New macro.
624 (min_int): New function.
625 (max_int): New function.
626 (sfm_width_to_bytes): Rewrite.
627 (sfm_width_to_octs): New function.
628 (sfm_segment_alloc_width): New function.
629 (sfm_segment_alloc_bytes): New function.
630 (sfm_segment_used_bytes): New function.
631 (sfm_segment_offset): New function.
632 (sfm_segment_effective_offset): New function.
633 (sfm_dictionary_to_sfm_vars): New function.
635 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
636 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
637 (struct sfm_var): New structure.
639 Next, improvements to the system file reader.
641 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
642 type casenumber. Added `version_major', `version_minor',
645 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
646 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
647 Change `case_cnt' to type casenumber. Removed `has_vls'.
648 (struct sfm_var): Removed.
649 (sfm_open_reader): Don't warn on wrong case size if the file was
650 written by SPSS 13, which tends to get it wrong. Use
651 sfm_dictionary_to_sfm_vars.
652 (read_header): Always output system file info.
653 (read_variable_record): Simplify code for reading missing values.
654 (read_machine_int32_info): Save version numbers from system file
655 into info struct passed as new argument.
656 (read_long_string_map): Restructured to use new sys-file-private
658 (read_value_labels): Use size_overflow_p.
659 (sys_file_casereader_read): Get rid of distinction between fast
660 and slow paths. Use information provided by sys-file-primate's
661 struct sfm_var to simplify code.
662 (skip_whole_strings): New function.
663 (read_int32): Renamed read_int. Changed return value to int.
665 (read_flt64): Renamed read_float. Changed return value to
666 double. Updated all callers.
667 (int32_to_native): Removed. Changed callers to use
669 (flt64_to_double): Removed. Changed callers to use float_convert.
671 Finally, get rid of int32, flt64 terminology and types in system
672 file writer. The former wasn't very useful since a POSIX "int"
673 can hold the whole range of int32 and we generally didn't have a
674 need for it to be exactly-32-bits, just at-least-32-bits. The
675 latter was inconvenient because we had to assume that it could be
676 different from double and thereby convert special values SYSMIS,
677 HIGHEST, LOWEST to and from it in multiple places. Instead, now
678 we just use "int" and "double" in most places, and do conversions,
679 if necessary, very close to where we do I/O. This change meant
680 that the writer code couldn't represent records in the file as C
681 structs any longer, but that's no great loss. The code actually
682 seems to be more readable without them.
684 Simplify the compression buffering code: only buffer as much as
685 necessary, which is no more than eight 8-byte units at any given
688 * sys-file-writer.c (typedef flt64): Removed.
689 (macro second_lowest_flt64): Removed.
690 (struct sysfile_header): Removed.
691 (struct sysfile_variable): Removed.
692 (struct sfm_writer): Removed `needs_translation', `has_vls',
693 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
694 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
695 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
696 compression buffering by `opcodes', `opcode_cnt', `data',
697 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
698 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
699 data writer functions instead of structures.
700 (calc_oct_idx): New function.
701 (write_header): Use simple data writer functions instead of
703 (write_format_spec): Renamed write_format. New argument.
704 (write_variable_continuation_records): New function.
705 (write_variable): Use simple data writer functions instead of
706 structures. Use write_variable_continuation_records. Write
707 entire very long string instead of requiring caller to understand
709 (write_value_labels): Use simple data writer functions instead of
711 (write_documents): Ditto.
712 (write_variable_display_parameters): Use sys-file-private
713 functions to simplify. Use simple data writer functions instead
715 (write_vls_length_table): Use simple data writer functions instead
717 (write_longvar_table): Ditto.
718 (write_rec_7_34): Break into new functions
719 write_integer_info_record, write_float_info_record. Use simple
720 data writer functions instead of structures.
721 (buf_write): Removed.
722 (append_string_max): Removed.
723 (ensure_buf_space): Removed.
724 (sys_file_casewriter_write): Get rid of the distinction between
725 fast and slow paths, which didn't seem to be too useful. Use new
726 functions write_case_uncompressed, write_case_compressed.
727 (put_instruction): Removed.
728 (put_element): Removed.
729 (write_compressed_data): Removed.
730 (close_writer): Use flush_compressed. Only write case count to
731 system file if it will fit in the field.
732 (write_case_compressed): New function.
733 (write_case_uncompressed): New function.
734 (flush_compressed): New function.
735 (put_cmp_opcode): New function.
736 (put_cmp_number): New function.
737 (write_int): New function.
738 (convert_double_to_output_format): New function.
739 (write_float): New function.
740 (write_value): New function.
741 (write_string): New function.
742 (write_bytes): New function.
743 (write_zeros): New function.
744 (write_spaces): New function.
746 Reviewed by John Darrington.
748 2007-07-22 Ben Pfaff <blp@gnu.org>
750 Don't try to write very long strings to portable files. The
751 format does not support it.
753 * por-file-writer.c (MAX_POR_WIDTH): New macro.
754 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
755 (write_format): Add arg to take width to resize format to.
756 (write_value): Limit width of value written to MAX_POR_WIDTH.
757 (write_variables): Limit width of variable and its output formats
759 Reviewed by John Darrington.
761 2007-07-22 Ben Pfaff <blp@gnu.org>
763 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
764 local variable instead of literal 5.
765 Reviewed by John Darrington.
767 2007-07-22 Ben Pfaff <blp@gnu.org>
769 Fix problems with uniqueness of short names in system files with
770 very long string variables. Now a variable may have multiple
773 * automake.mk (src_data_libdata_a_SOURCES): Add new files
774 short-names.c, short-names.h.
776 * dictionary.c (dict_clone): Clone all the short names.
777 (compare_strings): Move into short-names.c.
778 (hash_strings): Ditto.
779 (set_var_short_name_suffix): Ditto.
780 (dict_assign_short_names): Ditto, rename short_names_assign,
781 change to assign all short names.
783 * por-file-writer.c (write_variables): Use short_names_assign
784 instead of dict_assign_short_names.
786 * short-names.c: New file.
788 * short-names.h: New file.
790 * sys-file-private.c (sfm_width_to_segments): New function.
792 * sys-file-reader.c (read_long_var_name_map): Save and restore all
793 the short names, not just the first one.
795 * sys-file-writer.c (cont_var_name): Removed.
796 (sfm_open_writer): Use short_names_assign instead of
797 dict_assign_short_names. Use unique short names assigned by
798 short_names_assign instead of those generated by cont_var_name.
800 * variable.c (struct variable): Remove `short_name' member,
801 replace by `short_names' and `short_name_cnt'.
802 (var_create) Initialize new members.
803 (var_get_short_name_cnt): New function.
804 (var_get_short_name): Now takes an index argument. Changed most
806 (var_set_short_name): Ditto.
807 (var_clear_short_name): Renamed var_clear_short_names, changed to
808 clear all short names.
810 Reviewed by John Darrington.
812 2007-07-22 Ben Pfaff <blp@gnu.org>
814 * variable.c (var_set_width): Use new var_set_width function.
816 * missing-values.c (mv_n_values): Drop assertion, which was not
819 * format.c (fmt_default_for_width): New function.
820 (fmt_resize): New function.
822 Reviewed by John Darrington.
824 2007-07-18 John Darrington <john@darrington.wattle.id.au>
826 * datasheet.c (datasheet_delete_columns): Added assertion to check
827 we're not deleting outside the range of the sheet.
830 * dictionary.c dictionary.h variable.c: Added the ability for string
831 variables to be resized.
833 * vardict.h: Added some prototypes (moved from dictionary.h) as
834 these should only be called by variable.c
837 2007-07-14 John Darrington <john@darrington.wattle.id.au>
839 * sfm-reader.c: Respect case_cnt field in file header.
841 2007-07-01 John Darrington <john@darrington.wattle.id.au>
843 * transformation.c transformation.h (trns_chain_execute): Changed the
844 signature (Patch #6057)
846 2007-06-10 Ben Pfaff <blp@gnu.org>
848 * casereader-filter.c (casereader_filter_destroy): Make sure to
849 write all the remaining excluded cases to the casewriter, if any.
851 * caseinit.c (init_list_destroy): Rewrite.
852 (init_list_clear): Ditto.
854 * casegrouper.c (casegrouper_get_next_group): Always set *reader
855 to null when returning false.
857 2007-06-06 Ben Pfaff <blp@gnu.org>
859 Actually implement the new procedure code and adapt all of its
860 clients to match. Also adapt all of the other case sources and
861 sinks in the tree and their clients to use the
862 casereader/casewriter infrastructure.
864 * automake.mk: Add and remove files.
866 * any-reader.c: Change into a casereader.
867 * por-file-reader.c: Ditto.
868 * scratch-reader.c: Ditto.
869 * sys-file-reader.c: Ditto.
871 * any-writer.c: Change into a casewriter.
872 * por-file-writer.c: Ditto.
873 * scratch-writer.c: Ditto.
874 * sys-file-writer.c: Ditto.
876 * procedure.c: Change to use casereader, casewriter, caseinit, and
877 other new infrastructure.
879 * scratch-handle.c: Adapt to new infrastructure.
881 * case-sink.c: Removed, now dead code.
882 * case-sink.h: Ditto.
883 * case-source.c: Ditto.
884 * case-source.h: Ditto.
885 * casefile-factory.c: Ditto.
886 * casefile-private.h: Ditto.
889 * casefilter.c: Ditto.
890 * casefilter.h: Ditto.
893 * fastfile-factory.c: Ditto.
894 * fastfile-factory.h: Ditto.
895 * storage-stream.c: Ditto.
896 * storage-stream.h: Ditto.
898 2007-06-06 Ben Pfaff <blp@gnu.org>
902 * automake.mk: Add new files.
904 * datasheet.c: New file.
906 * datasheet.h: New file.
908 2007-06-06 Ben Pfaff <blp@gnu.org>
910 Until now, the procedure code has provided a case to the
911 case_source, which has filled in the data values that come from
912 the active file. "Left" data values that don't come from the
913 active file naturally stay the same from case to case, because the
914 procedure code keeps using that same case.
916 One of the compromises that comes with the new procedure code is
917 that the active file allocates and provides its own case, which
918 the procedure code then has to resize to provide room for any
919 other variables that should go in the case and then fill in the
920 values of "left" variables. Then, when we're done with that case,
921 we have to save the values of "left" variables to copy into the
922 next case read from the active file.
924 The caseinit code helps with this.
926 * automake.mk: Add new files.
928 * caseinit.c: New file.
930 * caseinit.h: New file.
932 2007-06-06 Ben Pfaff <blp@gnu.org>
934 * value.h (value_cnt_from_width): New function.
936 * variable.c (var_get_value_cnt): Use new function.
938 2007-06-06 Ben Pfaff <blp@gnu.org>
940 Add casegrouper, to allow cases read from a given casereader to be
941 broken into groups, each of which has its own casereader.
942 Generally cases are grouped based on having equal values for some
945 * automake.mk: Add new files.
947 * casegrouper.c: New file.
949 * casegrouper.h: New file.
951 2007-06-06 Ben Pfaff <blp@gnu.org>
953 Add interface to lexicographical ordering of cases.
955 * automake.mk: Add new files.
957 * case-ordering.c: New file.
959 * case-ordering.h: New file.
961 2007-06-06 Ben Pfaff <blp@gnu.org>
963 Add casereaders and casewriters, the basis of the new data processing
964 implementation. A casereader is a uniform interface to reading cases
965 from a data source; a casewriter is a uniform interface to writing
966 cases to a data sink.
968 * automake.mk: Add new files.
970 * casereader-filter.c: New file.
972 * casereader-provider.h: New file.
974 * casereader-translator.c: New file.
976 * casereader.c: New file.
978 * casereader.h: New file.
980 * casewriter-provider.h: New file.
982 * casewriter-translator.c: New file.
984 * casewriter.c: New file.
986 * casewriter.h: New file.
988 2007-06-06 Ben Pfaff <blp@gnu.org>
990 "casewindow" data structure that extends the deque (from libpspp)
991 of cases with the ability to dump cases to disk if we get too many
994 * automake.mk: Add new files.
996 * casewindow.c: New file.
998 * casewindow.h: New file.
1000 2007-06-06 Ben Pfaff <blp@gnu.org>
1002 sparse_cases data structure that augments a sparse_array of cases
1003 with the ability to dump cases to disk if we get too many cases in
1006 * automake.mk: Add new files.
1008 * sparse-cases.c: New file.
1010 * sparse-cases.h: New file.
1012 2007-06-06 Ben Pfaff <blp@gnu.org>
1014 Adds a low-level on-disk case array data structure.
1016 * automake.mk: Add new files.
1018 * case-tmpfile.c: New file.
1020 * case-tmpfile.h: New file.
1022 2007-06-06 Ben Pfaff <blp@gnu.org>
1024 In a couple of places we calculate the maximum number of cases to
1025 keep in memory based on the user-defined workspace. Enable
1026 centralizing the calculation through a new function.
1028 * settings.c (get_workspace_cases): New function.
1030 2007-06-06 Ben Pfaff <blp@gnu.org>
1032 The casenumber type is defined in transformations.h, but case.h is
1033 a more sensible place. Move it.
1035 * case.h (CASENUMBER_MAX): New macro.
1036 (typedef casenumber): Move here, from transformations.h.
1038 2007-06-03 Ben Pfaff <blp@gnu.org>
1040 Slightly generalize case_to_values and case_from_values functions.
1042 * case.c (case_to_values): Rename case_copy_out, change interface.
1043 (case_from_values): Rename case_copy_in, change interface.
1045 * fastfile.c (fastfilereader_get_next_case): Update caller.
1046 (write_case_to_disk): Ditto.
1048 2007-06-02 Ben Pfaff <blp@gnu.org>
1050 Clean up after a forgotten part of patch #5829.
1052 * casedeque.h: Remove unused file.
1054 * automake.mk: Remove casedeque.h from sources.
1056 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1058 * category.c: Removed redundant #include
1060 2007-05-06 Ben Pfaff <blp@gnu.org>
1062 Abstract the documents within a dictionary a little better.
1063 Thanks to John Darrington for suggestion, initial version, and
1064 review. Patch #5917.
1066 * dictionary.c (struct dictionary): Change `documents' member from
1067 char * to struct string.
1068 (dict_clear): Destroy struct string.
1069 (dict_get_documents): Convert struct string to char *.
1070 (dict_set_documents): Set struct string. Pad to 80-character
1072 (dict_clear_documents): New function.
1073 (dict_add_document_line): New function.
1074 (dict_get_document_line_cnt): New function.
1075 (dict_get_document_line): New function.
1077 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1079 * sys-file-reader.c (read_documents): Use new document functions.
1081 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1083 * sys-file-reader.c: When reading a system file which has no
1084 long name table, automatically create one where the long names
1085 are the lower case versions of the short names.
1087 2007-04-22 Ben Pfaff <blp@gnu.org>
1089 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1090 dict_clear will free most data related to the dictionary.
1091 dict_clear does a decent job, except that dict_set_split_vars on
1092 some systems won't actually free the dict's "split" member.
1093 Instead, it'll allocate a 1-byte region. Fix this.
1095 * value.c (value_copy): New function.
1096 (value_set_missing): Ditto.
1098 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1100 * Deleted existing category.h and moved cat-routines.h into
1101 category.h Encapsulated struct cat_vals better.
1103 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1105 * sys-file-reader.c: When reading a system file which has no
1106 long name table, automatically create one where the long names
1107 are the lower case versions of the short names.
1109 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1111 * sys-file-reader.c: Some versions of Other Software seem to
1112 produce system files with string variables' measure set to
1113 zero. We'll assume these are supposed to be nominal variables.
1115 2007-03-30 Ben Pfaff <blp@gnu.org>
1117 * procedure.c: Adapt to new deque data structure.
1119 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1120 Ben Pfaff <blp@gnu.org>
1122 * file-name.c: Mingw compatibility fixes.
1123 (fn_search_path): Use ISSLASH instead of comparing against '/'
1125 (fn_dir_name): Use dir_name from gnulib.
1126 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1127 (fn_get_identity): Use GetFullPathName instead of canonicalize
1128 from gnulib, because the latter does not fully support
1129 Windows-style path names. Use this implementation based on the
1130 detected presence of Windows instead of the absence of Unix, since
1131 the new implementation is Windows-specific.
1132 (fn_compare_file_identities): In Windows implementation, compare
1133 names case-insensitively.
1135 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1137 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1138 <stdlib.h> to have the same effect.
1140 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1142 * por-file-reader.c: Add missing _() around messages.
1144 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1146 * make-file.c: Include "mkstemp.h", without which linking on
1149 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1151 Reduce platform dependence.
1153 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1154 Everywhere we using this, we really should have just depended on
1155 the shell to expand tildes.
1156 (fn_search_path): Simplify, given that we don't do tilde expansion
1158 (fn_normalize): Removed. Caller changed to use the canonicalize
1160 (fn_get_cwd): Removed. Only user was fn_normalize.
1161 (fn_is_absolute): Really only test for absolute names.
1162 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1164 (fn_readlink): Removed, as it was only used fn_normalize.
1165 (fn_exists): Assume the stat function is available; gnulib does.
1166 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1168 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1170 * dictionary.c (dict_create_vector_assert): New function.
1172 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1174 * file-name.c (fn_normalize): Correct name of function
1175 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1176 for pointing this out.
1178 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1180 We are using a single member in struct file_handle, the "name"
1181 field, for more than one purpose. When it begins with '"', it's a
1182 file name; otherwise, it's a token that can be used to identify
1183 it. When that assertion fires, it's because we searched for the
1184 name case-sensitively as a file name (so that there was no match),
1185 and then we try to insert it case-insensitively as a token, which
1186 fails because duplicates aren't allowed.
1188 Solution: break the two purposes into two separate fields. This
1189 fixes the problem and likely makes the code easier to read too.
1191 Fixes bug #18922. Thanks to John Darrington for bug report and
1194 * file-handle-def.c (struct file_handle): New `id' member.
1195 (fh_from_name): Rename fh_from_id. Update all callers.
1196 (create_handle): New `id' parameter. Update all callers.
1197 (fh_create_file): Ditto.
1198 (fh_get_id): New function.
1200 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1202 * case.c (case_is_null): Change return type to bool.
1204 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1208 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1209 reference counting to share data. This makes it easy for
1210 valgrind, etc. to find accesses to cases that have been destroyed
1211 but have been kept around by another user's ref-count. This often
1212 happens when the data set is small enough to find in memory; if a
1213 bigger data set that would overflow to disk were used, then data
1214 corruption would occur.
1216 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1220 * case.c (case_unshare): Make it check internally whether the
1221 ref_cnt is greater than 1, so that the callers don't have to.
1222 Update callers not to check.
1224 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1226 Before, I was thinking that I might want to get rid of reference
1227 counting at some point. Now, I'm pretty sure that it's here to
1228 stay. Thus, because we have to store the value_cnt anyway for
1229 reference-counted cases, we might as well expose it to users.
1231 * case.c (case_get_value_cnt): New function.
1232 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1233 resize case if its size actually changed.
1235 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1236 instead of peeking inside struct case directly.
1237 (casefile_append): Ditto.
1239 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1241 Get rid of the inlines for the case functions, which made the
1242 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1243 the inlines didn't speed up "make check" at all, which is not a
1244 perfect benchmark but seems indicative.)
1246 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1247 definitions. Remove some assertions on nonnull pointers that were
1248 redundant with a pointer dereference soon after in the function.
1250 (struct case_data): Move definition here from case.h.
1254 (case_data_wr): Ditto.
1256 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1258 * automake.mk: Add casedeque.h to sources.
1260 * casedeque.h: New file.
1262 * procedure.c: (struct dataset) Change lag_count, lag_head,
1263 lag_queue member into single struct casedeque member. Update all
1264 users to use the casedeque instead.
1267 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1269 * procedure.c: Simplify lagged cases interface. Updated all
1270 clients--well, the only client--to use the simplified interface.
1271 (dataset_n_lag) Removed.
1272 (dataset_set_n_lag) Removed.
1273 (dataset_need_lag) New function.
1275 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1277 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1278 whenever appropriate, but only when the dataset/dictionary is in a
1281 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1283 * dictionary.c dictionary.h : Added callbacks for change of filter and
1284 split variables. Refactored some code to ensure that callbacks get
1285 invoked when appropriate.
1287 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1288 replace_dict callback occurs when permanent_dict replaces the current
1291 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1293 * dictionary.c dictionary.h : Added callback for when the weight
1294 variable of a dictionary changes.
1296 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1298 * dictionary.c dictionary.h : Added replace_source and replace_dict
1299 callbacks, and functions to deal with them.
1301 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1303 Simplify missing value handling.
1305 * missing-values.h (enum mv_class): New type.
1306 (enum mv_type): Moved definition into missing-values.c and renamed
1307 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1308 enums. Updated all uses.
1309 (struct missing_values): Changed type of `type' from `enum
1310 mv_type' to `int' because the definition is no longer exposed.
1312 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1313 parameter. Update all callers.
1314 (mv_is_num_missing): Ditto.
1315 (mv_is_str_missing): Ditto.
1316 (mv_is_value_user_missing): Removed. Changed callers to use
1317 mv_is_value_missing.
1318 (mv_is_num_user_missing): Removed. Changed callers to use
1320 (mv_is_str_user_missing): Removed. Changed callers to use
1322 (mv_is_value_system_missing): Removed. Changed callers to use
1323 mv_is_value_missing.
1324 (mv_set_type): Removed. Changed callers to use mv_clear.
1325 (mv_clear): New function.
1327 * variable.c (var_is_value_missing): Add new enum mv_class
1328 parameter. Update all callers.
1329 (var_is_num_missing): Ditto.
1330 (var_is_str_missing): Ditto.
1331 (var_is_value_user_missing): Removed. Changed callers to use
1332 var_is_value_missing.
1333 (var_is_num_user_missing): Removed. Changed callers to use
1335 (var_is_str_user_missing): Removed. Changed callers to use
1337 (var_is_value_system_missing): Removed. Changed callers to use
1338 var_is_value_missing.
1340 * casefilter.c (struct casefilter): Use enum mv_class in place of
1342 (casefilter_variable_missing): Adapt to new member.
1343 (casefilter_create): Change signature to take enum mv_class,
1346 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1348 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1350 * case-sink.c case-sink.h procedure.c procedure.h
1351 storage-stream.c: Now uses the factory.
1353 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1355 Make it possible to pull cases from the active file with a
1356 function call, instead of requiring indirection through a callback
1359 * case-source.h (struct case_source_class): Change ->read function
1360 to return a single case, instead of calling a callback function
1361 for each case. Change ->destroy function to return an error
1364 * case-source.c (free_case_source): Pass along the value returned
1365 by the case_source ->destroy function.
1367 * procedure.c (struct write_case_data): Removed.
1368 (struct dataset): Added some members to track procedure state.
1369 (procedure): Optimize the trivial case at this level.
1370 (internal_procedure): Re-implement in terms of proc_open,
1371 proc_read, proc_close.
1372 (proc_open) New function.
1373 (proc_read) New function.
1374 (proc_close) New function.
1375 (write_case) Moved into proc_read.
1376 (close_active_file) Moved closing of data source into proc_close.
1378 * storage-source.c: Rewrote to conform with modified
1379 case_source_class interface.
1381 * transformations.c (trns_chain_execute): Added argument to allow
1382 starting execution from an arbitrary transformation. Updated
1385 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1387 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1389 * sys-file-reader.c (read_display_parameters): Don't assume that
1390 MEASURE_* and ALIGN_* have the same values found in system files.
1392 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1394 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1395 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1396 the value of n_MEASURES, which was off by 1 (at least from my
1399 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1401 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1402 callbacks which are invoked when the dictionary or its
1403 variables are changed.
1405 * missing-values.c missing-values.h value-labels.c: Tidied up
1406 consistency checks, and made some of them return false
1407 instead of assert-failing.
1409 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1411 * calendar.c (calendar_days_in_month): New function.
1413 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1415 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1416 the value label's width.
1418 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1420 * sfm-private.h: Move contents into sys-file-writer.c, which is
1421 the only remaining user. Removed Borland C++-specific directives.
1423 * sys-file-reader.c: Clean up and rewrite entire file. The
1424 rewritten version is simpler and better abstracted, and should be
1425 easier to maintain and extend. It avoids using structures to read
1426 file data, which is prone to padding variations among compilers.
1427 It should also handle non-IEEE 754 system files, although I
1428 haven't been able to find any. It has been tested against many
1429 .sav files obtained from the Web and found to produce the same
1430 results as the earlier version of the code, or in some cases
1431 improved results. It is more tolerant of format variations found
1434 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1435 member, putting an enum integer_format in its place. New member
1436 `float_format'. Changed `compressed' member to type bool.
1438 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1440 * dictionary.c (dict_delete_consecutive_vars): New function.
1442 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1444 * file-name.c (fn_search_path): Remove prefix arg that was unused
1445 by any caller. Updated all callers.
1447 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1449 * format.c (fmt_dollar_template): Use user's decimal point
1450 character. Add assertion.
1452 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1454 * format.c (fmt_dollar_template): New function, based on
1455 dollar_format_template from var-type-dialog.c.
1457 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1459 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1462 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1464 Finish converting struct variable to an opaque type. In this
1465 phase, we add remaining setter and getter functions, convert the
1466 remaining PSPP code to use them, and do a bunch of cleanup. The
1467 resulting changes are pervasive but mostly trivial, and only the
1468 notable changes are logged.
1470 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1473 * case.c (case_data): Renamed case_data_idx.
1474 (case_num): Renamed case_num_idx.
1475 (case_str): Renamed case_str_idx.
1476 (case_data_rw): Renamed case_data_rw_idx.
1478 * case.h (case_data): New function with old name and an interface
1479 that takes a variable instead of an index, which is easier to
1480 use. Updated all callers to use the new interface, or to use the
1481 new *_idx function (see above).
1484 (case_data_rw): Ditto.
1486 * category.c (cat_stored_values_destroy): Changed interface to
1487 take a struct cat_vals * instead of a struct variable *.
1489 * dictionary.c (dict_clone): Use new vector_clone function.
1490 (dict_clear) Use new var_destroy function.
1491 (add_var) New function.
1492 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1493 (dict_create_var_assert) Rewrite in terms of add_var.
1494 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1495 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1496 (dict_lookup_var) Use new var_create, var_destroy functions.
1497 (dict_contains_var) Rewrite in terms of new vardict functionality.
1498 (set_var_dict_index) New function.
1499 (set_var_case_index) New function.
1500 (reindex_vars) New function.
1501 (dict_delete_var) Rewrite in terms of new vardict functionality.
1502 (dict_reorder_var) Ditto.
1503 (dict_reorder_vars) Ditto.
1504 (rename_var) New function.
1505 (dict_rename_var) Use rename_var.
1506 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1507 (dict_get_compacted_idx_to_fv) Rename
1508 dict_get_compacted_dict_index_to_case_index, update callers.
1509 (dict_create_vector) Use new vector_create function.
1510 (dict_clear_vectors) Use new vector_destroy function.
1511 (set_var_short_name_suffix) Move here from variable.c, renamed
1512 from var_set_short_name_suffix, make static, update caller.
1514 * sys-file-private.c: New file.
1515 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1516 width_to_bytes, update callers.
1518 * sys-file-private.h: New file. Later it will supplant
1519 sfm-private.h; for now it supplements it.
1520 (macro MIN_VERY_LONG_STRING) New macro.
1521 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1523 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1524 MAX_LONG_STRING was used before.
1526 * sys-file-writer.c: Ditto.
1528 * value-labels.c: Change the paradigm here to be that a null
1529 pointer is OK for a struct val_labs * in most cases; it just
1530 represents an empty set of value labels.
1531 (val_labs_copy) A copy of a null set is a null set.
1532 (val_labs_count) A null set has 0 labels.
1533 (val_labs_replace) Change return type to void. Rewrite for
1535 (val_labs_find) A null set does not contain the value.
1536 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1537 transposed argument order, updated all callers.
1539 * value.c: New file.
1540 (value_dup) Moved here from variable.c.
1541 (compare_values) Ditto.
1544 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1545 (macro MAX_LONG_STRING) Removed, because it was only interesting
1546 for system files, not for general code.
1547 (macro MAX_VERY_LONG_STRING) Ditto.
1548 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1549 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1551 * vardict.h: New file, for an interface between variables and
1554 * variable.c: A lot of functions were moved around, for better
1556 (struct variable) Move definition here, from variable.h.
1557 (var_type_adj) Removed--makes i18n hard.
1558 (var_type_noun) Ditto.
1559 (var_create) New function.
1560 (var_clone) New function.
1561 (var_destroy) New function.
1562 (var_set_name) Assert that variable is not in a dictionary.
1563 (compare_var_names) Rename compare_vars_by_name and fix a couple
1564 of callers who thought the args were strings.
1565 (hash_var_name) Rename hash_var_by_name.
1566 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1567 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1568 (var_is_very_long_string) Removed, because it was only interesting
1569 to system file code.
1570 (var_set_missing_values) Allow the argument to be the wrong width,
1571 as long as we can resize it. Simplify callers who were doing the
1572 resizing themselves.
1573 (var_get_value_labels) New function.
1574 (var_has_value_labels) New function.
1575 (var_set_value_labels) New function.
1576 (alloc_value_labels) New function.
1577 (var_add_value_label) New function.
1578 (var_replace_value_label) New function.
1579 (var_clear_value_labels) New function.
1580 (var_lookup_value_label) New function.
1581 (var_get_value_name) Moved here from variable.c, renamed from
1582 var_get_value_name, transposed argument order, updated all
1584 (var_to_string) Moved here, from variable-label.c.
1585 (var_set_leave) New function.
1586 (var_get_leave) New function.
1587 (var_must_leave) New function.
1588 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1589 set_var_short_name_suffix.
1590 (var_get_dict_index) New function.
1591 (var_get_case_index) New function.
1592 (var_get_obs_vals) New function.
1593 (var_set_obs_vals) New function.
1594 (var_has_obs_vals) New function.
1595 (var_get_vardict) New function.
1596 (var_set_vardict) New function.
1597 (var_has_vardict) New function.
1598 (var_clear_vardict) New function.
1599 (value_dup) Moved to value.c.
1600 (compare_values) Ditto.
1603 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1604 (enum ALPHA) Rename VAR_STRING, update all users.
1606 * vector.c: New file.
1607 (struct vector) Moved here, from variable.h.
1608 (check_widths) New function.
1609 (vector_create) New function.
1610 (vector_clone) New function.
1611 (vector_destroy) New function.
1612 (vector_get_name) New function.
1613 (vector_get_var) New function.
1614 (vector_get_var_cnt) New function.
1615 (compare_vector_ptrs_by_name) New function.
1617 * vector.h: New file.
1619 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1621 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1622 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1625 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1627 * value-labels.c (destroy_atoms): New function.
1628 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1630 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1632 Thanks to Jason Stover for pointing out this problem.
1634 * data-out.c (output_number): Use gsl_finite from GSL, which is
1635 portable, instead of isfinite, which is not.
1638 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1640 * variable.c variable.h (value_dup): New function.
1642 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1644 Start converting struct variable to an opaque type. In this
1645 phase, we add a bunch of setter and getter functions and convert
1646 most of the PSPP code to use them. The resulting changes are
1647 pervasive but mostly trivial, and only the notable changes are
1650 * format.c (fmt_equal): New function.
1652 * variable.c (var_type_is_valid): New function.
1653 (measure_is_valid) Moved here, from format.c.
1654 (alignment_is_valid) Moved here, from format.c.
1655 (var_get_name) New function.
1656 (var_set_name) New function.
1657 (width_to_type) New function.
1658 (var_get_type) New function.
1659 (var_get_width) New function.
1660 (var_set_width) New function.
1661 (var_is_numeric) New function.
1662 (var_is_alpha) New function.
1663 (var_is_short_string) New function.
1664 (var_is_long_string) New function.
1665 (var_is_very_long_string) New function.
1666 (var_get_missing_values) New function.
1667 (var_set_missing_values) New function.
1668 (var_clear_missing_values) New function.
1669 (var_has_missing_values) New function.
1670 (var_is_value_missing) New function.
1671 (var_is_num_missing) New function.
1672 (var_is_str_missing) New function.
1673 (var_is_value_user_missing) New function.
1674 (var_is_num_user_missing) New function.
1675 (var_is_str_user_missing) New function.
1676 (var_is_value_system_missing) New function.
1677 (var_get_print_format) New function.
1678 (var_set_print_format) New function.
1679 (var_get_write_format) New function.
1680 (var_set_write_format) New function.
1681 (var_set_both_formats) New function.
1682 (var_get_label) New function.
1683 (var_set_label) New function.
1684 (var_clear_label) New function.
1685 (var_has_label) New function.
1686 (var_get_measure) New function.
1687 (var_set_measure) New function.
1688 (var_get_display_width) New function.
1689 (var_set_display_width) New function.
1690 (var_get_alignment) New function.
1691 (var_set_alignment) New function.
1692 (var_get_value_cnt) New function.
1693 (var_get_leave) New function.
1694 (var_get_short_name) New function.
1696 * variable.h: (struct variable) Removed "type" and "nv" members;
1697 they are now computed from "width" where needed.
1699 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1701 * missing-values.c (mv_resize): Don't write beyond end of the
1702 allocated buffer when resizing a long string.
1704 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1706 Clean up identifier code: don't require identifier enumerations to
1707 be in a particular order; make better use of string library;
1708 expose less of the internals.
1710 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1711 change interface. Updated all callers.
1712 (lex_id_match) Change interface to use struct substring, update
1714 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1715 (global array keywords[]) Make static, change form. Update all
1716 users to use lex_id_name instead.
1717 (lex_is_keyword) New function.
1718 (lex_id_to_token) Change interface to use struct substring, update
1720 (lex_id_name) New function.
1722 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1723 lex_is_keyword instead.
1724 (T_LAST_KEYWORD) Removed.
1725 (T_N_KEYWORDS) Removed.
1727 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1729 * format.c: (fmt_date_template) Distinguish characters for which a
1730 space is output and any date delimiter is allowed on input, from
1731 those for which a space is output and only a space is allowed on
1732 input. The former is represented by X, the latter by a space.
1733 Also, drop distinction between h and H, changing the former to the
1736 * data-in.c: Completely rewrite internals to conform to SPSS input
1737 formats as closely as possible.
1738 (data_in) Changed external interface by replacing the structure
1739 that was used as a single argument by a set of arguments. Updated
1741 (data_in_finite_line) Removed. Converted all callers to use plain
1743 (data_in_get_integer_format) New function.
1744 (data_in_set_integer_format) New function.
1745 (data_in_get_float_format) New function.
1746 (data_in_set_float_format) New function.
1748 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1749 (struct data_in) Removed.
1751 * data-out.c: (output_date) Drop each component from the input as
1752 it is output, to allow us to drop the distinction between h (a
1753 count of hours) and H (the hour of day) template characters.
1754 Also, handle new X template character.
1755 (output_scientific) Follow more rational rule on when to drop
1756 fraction introduced between SPSS 13 and 15. Updated test case to
1759 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1761 Fix buffer overflow reported by John Darrington.
1763 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1764 realize that DIGITS is a count of nibbles, not of bytes.
1766 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1768 * calendar.c (calendar_offset_to_gregorian) Also return the
1769 year-of-day. Change callers to new interface.
1771 * data-out.c: Completely rewrite internals to conform to SPSS
1772 output formats as completely as possible.
1773 (data_out) Change interface to put input parameters before output
1774 parameters, for consistency with the style I now prefer. Update
1776 (data_out_get_integer_format) New public function.
1777 (data_out_set_integer_format) New public function.
1778 (data_out_get_float_format) New public function.
1779 (data_out_set_float_format) New public function.
1781 * data-out.h: New file. Move prototype for data_out here, from
1784 * format.c: (fmt_step_width) Use equality comparison instead of
1785 bitwise and, for clarity.
1786 (fmt_is_string) Ditto.
1787 (fmt_input_to_output) Fix categories that are translated to F
1790 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1792 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1793 casefile-private.h: Added casefilter to assist commands with missing
1796 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1798 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1800 * settings.c: (route_errors_to_terminal) New variable.
1801 (route_errors_to_listing) New variable.
1802 (get_error_routing_to_terminal) New function.
1803 (set_error_routing_to_terminal) New function.
1804 (get_error_routing_to_listing) New function.
1805 (set_error_routing_to_listing) New function.
1807 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1809 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1811 * format.c: Completely rewrite, to achieve better abstraction.
1812 Rewrite all references to formats in other files.
1814 * format.def: Rewrite and reorganize.
1816 * settings.c: Move everything related to custom currency formats
1817 into format.[ch], changing them in form, so as to group related
1818 code and definitions better. Changed all references to use the
1820 (static var decimal) Removed.
1821 (static var grouping) Removed.
1822 (static var cc) Removed.
1823 (get_decimal) Removed.
1824 (set_decimal) Removed.
1825 (get_grouping) Removed.
1826 (set_grouping) Removed.
1830 * settings.h: (macro CC_CNT) Removed.
1831 (macro CC_WIDTH) Removed.
1832 (struct custom_currency) Removed.
1834 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1836 * data-in.c (data_in): Use switch statement instead of table, to
1837 avoid dependence on the order of the FMT_* enums.
1839 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1841 * data-out.c: (num_to_string) Removed, because it was dead code.
1843 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1845 * data-in.c (parse_trailer): Fix error message.
1847 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1849 * format.c (fmt_is_binary): New function.
1851 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1853 * procedure.c procedure.h: Encapsulated the static data into a single
1856 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1858 * casefile.c (casereader_read_xfer): Always initialize the case,
1859 even on an error condition.
1861 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1863 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1864 entire function useless.
1866 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1868 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1869 casereader method casereader_clone.
1871 * procedure.c pransformations.h: Introduced new type casenum_t
1873 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1875 * variable.c: (width_to_bytes) Rephrase code for clarify.
1877 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1879 * format.c: (fmt_type_from_string) New function.
1880 (fmt_to_string) Include decimals in output if the format has
1881 decimals, even if the format type does not. This way, we can
1882 accurately reproduce incorrect formats in user output.
1883 (check_common_specifier) Make the check for a bad format type an
1884 assertion, so we get bug reports if they show up. Fix message.
1885 Check for decimal places with a format type that doesn't allow
1887 (check_input_specifier) Remove check for FMT_X, which has been
1889 (check_output_specifier) Ditto.
1891 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1893 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1894 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1895 (enum fmt_parse_flags) Removed.
1897 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1899 * casefile.c casefile.h: Converted to an abstract base class.
1900 * casefile-private.h fastfile.c fastfile.h: New files.
1901 * automake.mk procedure.c scratch-writer.c storage-stream.c
1903 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1905 * procedure.c (internal_procedure): Create sink_case with only as
1906 many values as the compacted dictionary.
1908 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1910 Remove "debugging" code that caused plenty of false positives and
1913 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1915 * case.c: Remove all references to `this' member.
1917 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1919 Fix link error noted by Jason Stover.
1921 * storage-stream.c: Include <assert.h>.
1923 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1925 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1926 ALL) and additional underlying system file issues.
1928 Thanks to John Darrington for review.
1930 First problem: var_hash points to variables not owned by the
1931 sys-file-reader, which the caller may free or modify. Use an
1932 array of sfm_vars instead, as done earlier (e.g. CVS version
1935 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1936 members and remove all code that references it. Add vars, var_cnt
1937 members. Remove fix_specials member, which was unused.
1938 (struct sfm_var) Remove name member, which was unused.
1939 (sfm_close_reader) Free vars member instead of var_hash.
1940 (compare_var_shortnames) Removed.
1941 (hash_var_shortname) Removed.
1942 (sfm_open_reader) Fill out vars array.
1943 (compare_var_index) Removed.
1944 (sfm_read_case) Use vars instead of var_hash.
1946 Second problem: we're confused about when we actually have very
1947 long strings, causing us to choose incorrectly between slow path
1948 and fast path in sfm_read_case.
1950 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1951 have very long strings, not when we have long variable names,
1952 which is an unrelated feature.
1954 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1956 * variable.h: Move var_set and variable parsing declarations to
1957 new header, src/language/lexer/variable-parser.h. Modified lots
1958 of files to include the new header.
1960 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1962 * value-labels.c (value_to_string): When there's no value label,
1963 format the variable according to its print format, instead of
1964 always effectively using A or F format.
1966 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1968 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1971 * format.c: Removed tortological assertion.
1973 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1975 Reform string library.
1977 * file-name.c (fn_interp_vars): Change interface to take a
1978 substring as input. Updated all users.
1980 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1982 * format.c (measure_is_valid): Really return false when m >=
1985 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1987 Implement random access to casefiles, for use in GUI.
1989 * casefile.c: (struct casereader) Add `random', `file_ofs',
1990 `buffer_ofs' members.
1991 (casefile_get_random_reader) New function.
1992 (read_open_file) Break part into new function
1993 seek_and_fill_buffer().
1994 (fill_buffer) Update buffer_ofs, file_ofs.
1995 (casereader_seek) New function.
1997 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1999 * settings.c: Added call to i18n{done, init}.
2001 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
2003 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
2006 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
2008 * casefile.c: Convert many uses of `int' to `bool'.
2010 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
2012 * transformations.c (trns_chain_destroy): Destroy chain's trns
2013 member, to fix memory leak.
2015 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
2017 * storage-stream.c (storage_source_decapsulate): Destroy case
2018 source to fix memory leak.
2020 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
2022 * scratch-reader.c (scratch_reader_read_case): Copy into existing
2023 case passed as argument instead of initializing the argument as a
2024 case. Fixes memory leak that showed up in
2025 tests/command/aggregate.sh with scratch files.
2027 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
2029 * procedure.c (proc_done): Destroy default_dict, to fix memory
2032 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2034 Simplify procedure_with_splits().
2036 * procedure.c (struct split_aux_data): Removed case_count member.
2037 (procedure_with_splits) Don't initialize case_count.
2038 (split_procedure_case_func) Check whether prev_case is null
2039 instead of case_count.
2040 (split_procedure_end_func) Ditto.
2042 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2044 * case.c (case_move): Do nothing if dst and src are the same
2046 (case_try_create) Merge two similar cases.
2047 (case_copy) Unshare only if data must be actually copied.
2049 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2051 * data-in.c data-out.c dictionary.c sys-file-reader.c
2052 sys-file-writer.c variable.c variable.h: Reworked very long string
2053 support for better encapsulation.
2055 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2057 * value-labels.c (val_labs_can_set_width): New function.
2058 (val_labs_set_width) Clear labels if increasing width to long
2060 (val_labs_destroy) Remove unneeded test for null.
2062 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2064 * value-labels.h: Remove unneeded dependency on variable.h.
2066 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2068 Get rid of `char *c' member in union value, for cleanliness.
2070 * value.h: (union value) Remove `c' member.
2072 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2074 Make dictionary compacting functions a little more general.
2076 * sys-file-writer.c (sfm_open_writer): Use
2077 dict_compacting_would_change().
2078 (does_dict_need_translation) Removed.
2080 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2082 Make dictionary compacting functions a little more general.
2084 * dictionary.c (dict_needs_compaction): Rename
2085 dict_compacting_would_shrink(). Update all callers.
2086 (dict_compacting_would_change) New function.
2088 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2090 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2091 inverted return value (!).
2093 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2095 Continue reforming procedure execution.
2097 * procedure.c: Search and replace "vfm" by "proc". Notably:
2098 (static var vfm_source) Rename proc_source. Update all
2100 (static var vfm_sink) Rename proc_sink. Update all references.
2102 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2104 Continue reforming procedure execution. In this phase, remove
2105 PROCESS IF, which was deprecated anyway and can be easily
2106 simulated with TEMPORARY followed by SELECT IF.
2108 * procedure.c: (open_active_file) Don't call
2109 add_process_if_trns().
2110 (discard_variables) Get rid of redundant call to
2111 proc_cancel_all_transformations().
2112 (add_process_if_trns) Removed.
2113 (process_if_trns_proc) Removed.
2114 (process_if_trns_free) Removed.
2116 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2118 Continue reforming procedure execution. In this phase, add
2119 `const' to the case passed to procedure()'s callback.
2121 Updated all users of procedure() as well.
2123 * procedure.c: (struct write_case_data) Add "const" to ccase
2124 parameter for case_func member.
2125 (procedure) Add "const" to ccase parameter for proc_func
2127 (multipass_case_func) Make ccase parameter const.
2128 (internal_procedure) Add "const" to ccase parameter for case_func
2130 (split_procedure_case_func) Make ccase parameter const.
2131 (multipass_split_case_func) Make ccase parameter const.
2133 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2135 Continue reforming procedure execution. In this phase, get rid of
2136 the output code for SPLIT FILE groups in procedure.c, which really
2137 shouldn't be doing any output. Move it into the individual
2138 procedures instead. This also adds some flexibility.
2140 Updated many users of procedure_with_splits() and
2141 multipass_procedure_with_splits() to call
2142 output_split_file_values() and to deal with increased use of
2145 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2146 parameter to begin_func member.
2147 (procedure_with_splits) Add "const struct ccase *" parameter to
2148 begin_func parameter. Make ccase parameter const in proc_func
2150 (split_procedure_case_func) Don't dump split file group. Pass
2152 (dump_splits) Moved to language/dictionary/split-file.c as
2153 output_split_file_values().
2154 (struct multipass_split_aux_data) Add "const struct ccase *"
2155 parameter to split_func member.
2156 (multipass_procedure_with_splits) Add "const struct ccase *"
2157 parameter to split_func parameter.
2158 (multipass_split_case_func) Save new SPLIT FILE case before
2160 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2162 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2164 Continue reforming procedure execution. Change
2165 internal_procedure() so that it calls open_active_file() and
2166 close_active_file(), which isolates most of the actual procedure
2169 * procedure.c: (struct write_case_data) Rename `proc_func' member
2170 to `case_func' and update all references.
2171 (procedure) Rewrite as one-line wrapper around
2172 internal_procedure().
2173 (struct multipass_aux_data) New.
2174 (multipass_callback) Renamed multipass_case_func(). Use struct
2175 multipass_aux_data as auxiliary data.
2176 (multipass_end_func) New function.
2177 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2178 that uses multipass_case_func, multipass_end_func.
2179 (internal_procedure) Add `end_func' argument. Move optimization
2180 of trivial case in here. Move call to open_active_file() and
2181 close_active_file() in here. Now assert that vfm_source is
2183 (procedure_with_splits_callback) Rename
2184 split_procedure_case_func().
2185 (split_procedure_end_func) New function.
2186 (multipass_split_callback) Rename multipass_split_case_func.
2187 (multipass_split_end_func) New function.
2188 (discard_variables) No need to test for nonnull vfm_source.
2190 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2192 Continue reforming procedure execution. Get rid of unused member.
2194 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2196 (write_case) Don't increment cases_analyzed.
2198 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2200 Continue reforming procedure execution. In this phase, move
2201 procedure.c and procedure.h from src to src/data. Update
2202 makefiles and #includes accordingly.
2204 * procedure.c: Moved here from src/.
2206 * procedure.h: Moved here from src/.
2208 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2210 Continue reforming procedure execution. In this phase, get rid of
2211 many global variables, consolidating procedure execution in
2212 procedure.c. Encapsulate transformations in new "struct
2213 trns_chain". Also, change implementation of N OF CASES, FILTER,
2214 and PROCESS IF from special cases to transformations.
2216 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2219 * dictionary.c: (global variable default_dict) Move to
2222 * variable.h: (TRNS_*) Move to transformations.h.
2223 (struct transformation) Move to transformations.c.
2225 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2227 * sys-file-reader.c: Fixed invalid read problems.
2229 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2231 * storage-stream.c: Add missing function comments.
2233 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2235 Continue reforming procedure execution. In this phase, add some
2236 new, needed functionality to storage-stream.
2238 * storage-stream.c: (storage_source_decapsulate) New function.
2240 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2242 * variable.c (width_to_bytes): Declarations must precede
2243 statements for C90 compliance.
2245 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2247 * data-out.c, data-in.c, variable.c, variable.h: New functions
2248 copy_mangle and copy_demangle for reading/writing cases; emulates the
2249 way SPSS deals with strings > 255 bytes.
2251 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2252 subtype 14 needed for strings longer than 255 bytes.
2254 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2255 instead of literal values. Also fixed some constness issues.
2257 * format.h: Constness
2259 * sfm-private.h: Renamed the case_size identifier, since I discovered
2260 that SPSS's respect for this variable is very nominal.
2262 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2264 Change case limit type from int to size_t.
2266 * dictionary.c: (struct dictionary) Change type of case_limit
2268 (dict_get_case_limit) Change return type.
2269 (dict_set_case_limit) Change parameter type.
2271 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2273 * variable.h: (struct variable) Rename `reinit' member as `leave'
2274 and invert sense. Fix up all references.
2276 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2278 Continue reforming procedure execution. In this phase, break
2279 procedure.c into multiple files.
2281 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2283 * case-sink.c: New file.
2285 * case-sink.h: New file.
2287 * case-source.c: New file.
2289 * case-source.h: New file.
2291 * storage-stream.c: New file.
2293 * storage-stream.h: New file.
2295 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2297 * variable.h: (struct variable) Remove `init' member and all
2298 references to it from other files. It was initialized in several
2299 places, but nothing really ever used it for anything worthwhile.
2300 Thanks to Jason Stover for pointing out how confusing this
2303 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2305 Continue reforming error message support. In this phase, get rid
2306 of message "titles" and put the message text in `struct error'.
2307 Now `struct error' encapsulates a message more properly.
2309 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2310 Format message ourselves.
2312 * data-in.c: (vdls_error) Ditto.
2314 * por-file-reader.c: (error) Ditto.
2316 * sys-file-reader.c: (corrupt_msg) Ditto.
2318 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2320 GNU standards require "file name" instead of "filename" in
2321 documentation. It's nice for our code to follow the convention
2324 * casefile.c: (struct casefile) Rename `filename' member to
2325 `file_name'. Updated all references.
2327 * file-name.c: [!unix] (struct file_identity) Rename
2328 normalized_filename member to normalized_file_name. Updated all
2331 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2333 We don't really support anything but Unix-like environments well,
2334 so we might as well de-obfuscate by writing directory and path
2335 separators explicitly.
2337 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2339 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2340 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2342 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2345 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2347 GNU standards require "file name" instead of "filename" in
2348 documentation. It's nice for our code to follow the convention
2351 * filename.c: Rename to file-name.c.
2353 * filename.h: Rename to file-name.h. Update all inclusions.
2354 Update header guards.
2356 * automake.mk: Update file names.
2358 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2360 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2362 (fn_basename) Removed (dead code).
2363 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2364 (fn_special_p) Renamed fn_is_special(), all references updated.
2365 (fn_exists_p) Renamed fn_exists(), all references updated.
2367 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2369 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2370 Also, now it only tilde-expands file names, not paths.
2371 (fn_search_path) Tilde-expand one directory at a time.
2373 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2375 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2376 get rid of non-Unixlike version of the code, which has probably
2378 (fn_prepend_dir) Removed (dead code).
2380 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2381 (macro PATH_SEPARATOR_STRING) New.
2382 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2384 Continue reforming error message support. In this phase, we get
2385 rid of VM() and the other msg() support for "verbosity", replacing
2386 it by a new function verbose_msg().
2388 * filename.c: (fn_search_path) Use verbose_msg() instead of
2391 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2393 * sfm-private.h: Get rid of #defines after #error, which makes no
2396 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2398 Get rid of our own int32 type in favor of the standard int32_t
2401 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2404 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2406 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2408 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2410 Remove ill-considered file routines that are no longer used.
2412 * filename.c: (fn_open_ext) Removed.
2413 (fn_close_ext) Removed.
2415 * filename.h: (struct file_ext) Removed.
2417 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2419 * variable.c (var_is_valid_name): Move declarations before code
2422 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2424 * filename.ch (fn_interp_vars): Fixed small buglet.
2426 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2428 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2429 so as to modify the string inline. Thus makeing it easier to
2430 destroy the results when no longer needed.
2432 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2434 * category.c (cat_stored_values_destroy): Fixed memory leak.
2436 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2438 Add some missing frees. Thanks to John Darrington for reporting
2441 * any-writer.c (any_writer_close): Free writer.
2443 * any-reader.c (any_reader_close): Free reader.
2445 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2447 * por-file-reader.c: (error) Mark as NO_RETURN.
2449 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2451 * settings.c: Changed default value of scompress to true.
2453 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2455 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2457 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2460 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2461 here from pref.h.orig.
2463 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2465 * sys-file-reader.c: Fixed bug reading compressed files.
2467 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2469 * Numerous renames. See src/ChangeLog for details.
2471 * Moved files from src directory
projects / pspp-builds.git / blob