1 2008-03-06 Ben Pfaff <blp@gnu.org>
3 * dictionary.c (var_name_is_insertable): New function.
4 (make_hinted_name): Don't accept variable names that match PSPP
5 keywords. Thanks to Jason Stover for reporting the problem.
7 2008-03-06 Ben Pfaff <blp@gnu.org>
9 * format-guesser.c (syntax): Require month names to be spelled out
10 as English words, so that single characters that happen to be
11 Roman numerals don't get detected as months. Thanks to John
12 Darrington for reporting this bug.
14 2008-03-04 Ben Pfaff <blp@gnu.org>
16 Patch #6441. Reviewed by John Darrington.
18 * format.c (fmt_fix): New function.
19 (fmt_fix_input): New function.
20 (fmt_fix_output): New function.
22 * format.def: Correct minimum width for DATETIME format. It was
23 7, should have been 17.
25 * automake.mk: Add new files.
27 * format-guesser.c: New file.
29 * format-guesser.h: New file.
31 2008-02-18 Ben Pfaff <blp@gnu.org>
33 Patch #6426. Reviewed by John Darrington.
35 * format.c (min_width): Renamed fmt_min_width and made public.
36 Updated all references.
37 (max_width): Renamed fmt_max_width and made public. Updated all
39 (max_decimals): Renamed fmt_max_decimals and made public. Updated
41 (var_create): Use the new functions for default variable
43 (var_default_formats): New function.
44 (var_default_measure): New function.
45 (var_default_alignment): New function.
47 * format.h (macro FMT_MAX_NUMERIC_WIDTH): New macro.
49 2008-02-09 Ben Pfaff <blp@gnu.org>
51 Add a couple of extensions to GET DATA TYPE=TXT. Patch #6412.
52 Thanks to John Darrington for review.
54 * data-in.c (data_in): Add new argument to designate the last
55 column of the data field being parsed, for use in error messages.
58 * data-parser (struct data_parser): New member `quote_escape'.
59 (data_parser_create): Initialize quote_escape.
60 (data_parser_set_quotes): New function.
61 (cut_field): Support escaped quotes.
62 (parse_delimited_span): Ditto.
63 (parse_delimited_no_span): Ditto.
65 * get-data.c (parse_get_txt): Support ESCAPE extension subcommand
66 in enhanced mode. Only support multiple quote characters in
69 2008-02-06 John Darrington <john@darrington.wattle.id.au>
71 psql-reader.c psql-reader.h: Read more than one tuple at
72 once. Fix bug reading a query which returns no data. Fix bug
73 when transformation followed a reader.
74 Ask the server for the number of records in the query, for the
77 2008-02-05 John Darrington <john@darrington.wattle.id.au>
79 psql-reader.c: So yesterday they release postgresql 8.3.0
80 which has money represented with 64 bits. They must get
83 2008-02-02 John Darrington <john@darrington.wattle.id.au>
85 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
86 for reviewing this code.
88 2008-02-02 Ben Pfaff <blp@gnu.org>
92 * sys-file-reader.c (read_variable_record): Allow missing values
93 to be specified on long string variables, but warn about them
94 because PSPP does not yet support them.
95 (read_extension_record): Ignore extension records 20 and 21, which
96 PSPP does not yet support.
97 (read_header): Fix error message when floating-point format cannot
100 2008-02-01 Ben Pfaff <blp@gnu.org>
102 Patch #6386. Thanks to John Darrington for review and for the
103 updates to gnumeric-reader.c.
105 * dictionary.c (make_hinted_name): New function.
106 (make_numeric_name): New function.
107 (dict_make_unique_var_name): New function.
109 * gnumeric-reader.c (devise_name): Removed.
110 (munge_name): Removed.
111 (gnumeric_open_reader): Use new function
112 dict_make_unique_var_name.
114 * short-names.c (set_var_short_name_suffix): Use new function
117 2008-01-19 John Darrington <john@darrington.wattle.id.au>
119 * settings.c settings.h: Moved static variables into a
120 single struct. Renamed functions to have a settings_ prefix.
122 2008-01-14 John Darrington <john@darrington.wattle.id.au>
124 * data-out.c (output_AHEX): Corrected number of bytes in
125 call to output_hex. Closes bug #22011
127 2008-01-02 John Darrington <john@darrington.wattle.id.au>
129 * variable.c variable.h: Replaced var_get_value_name with
130 var_append_value_name which doesn't use any static data.
131 Thanks to Ben for review.
133 2007-12-07 Ben Pfaff <blp@gnu.org>
137 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
138 made from an empty casereader produce a casegrouper with no
139 groups, instead of one with one group that has no cases.
141 * casereader.c (casereader_is_empty): New function.
143 2007-12-06 Ben Pfaff <blp@gnu.org>
147 * sys-file-reader.c (read_display_parameters): Handle variable
148 display parameters record with only 2 data items per variable.
149 Reported by Guido Gay <gay@irer.it>.
151 2007-12-04 Ben Pfaff <blp@gnu.org>
153 * identifier.c (lex_id_match_n): New function.
154 (lex_id_match): Reimplement in terms of lex_id_match_n.
156 2007-11-24 Ben Pfaff <blp@gnu.org>
158 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
161 2007-11-24 Ben Pfaff <blp@gnu.org>
163 Fix warning reported when reading back system files that include
164 very long string variables. Thanks to Guido Gay <gay@irer.it> for
167 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
168 was written where `j' was meant.
170 2007-11-18 Ben Pfaff <blp@gnu.org>
172 Properly write variables that include a range of missing values to
173 system files. Thanks to Guido Gay <gay@irer.it> for reporting
176 * sys-file-writer.c (write_variable): Correctly calculate
177 n_missing_values field when writing variables that include a range
180 2007-11-10 Ben Pfaff <blp@gnu.org>
182 Cleanups and bug fixes devised while writing up documentation.
185 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
187 * dict-class.c: New file.
188 (dict_class_from_id): Move here.
189 (dict_class_to_name): Move here.
191 * dict-class.h: New file.
192 (enum dict_class): Move here. Change from consecutive integers to
193 consecutive bits, to make testing for any of multiple values
194 easier. Add new DC_ALL constant.
196 * dictionary.c (struct dictionary): Change `case_limit' from
197 size_t to casenumber.
198 (dict_get_vars): Make final argument an enum dict_class.
199 (dict_get_vars_mutable): Ditto.
200 (dict_get_case_limit): Change return value to casenumber.
201 (dict_set_case_limit): Change final argument to a casenumber.
202 (dict_unset_split_var): Add assertion. Rephrase slightly.
203 (dict_set_label): Use xstrndup to simplify.
205 * format.c (fmt_step_width): AHEX format also needs 2-byte
207 (fmt_set_style): Simplify assertions.
209 * missing-values.c (mv_add_num_range): Rename mv_add_range.
210 Simplify implementation.
211 (mv_has_value): Simplify implementation.
212 (mv_pop_value): Remove the first value, not the last, to avoid
213 having GET followed by SAVE reverse the order of missing values.
214 (mv_peek_value): Rename mv_get_value. Simplify assertion.
215 (mv_has_range): Simplify implementation.
216 (mv_peek_range): Rename mv_get_range.
217 (can_resize_string): Removed.
218 (mv_is_resizable): Use value_is_resizable.
219 (mv_resize): Use value_resize.
221 * short-names.h (SHORT_NAME_LEN): Move here.
223 * val-type.h: New file, for definitions related to type and width
224 of abstract values. Before, these definitions were mixed among
225 those related to "union value" and those related to variables.
226 (macro SYSMIS): Move here.
227 (macro LOWEST): Move here.
228 (macro HIGHEST): Move here.
229 (macro MAX_STRING); Move here.
230 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
231 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
232 All references updated.
233 (val_type_is_valid): New function. Replaces var_type_is_valid.
234 All references updated.
235 (val_type_from_width): New function. Replaces
236 var_type_from_width. All references updated.
238 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
240 (val_labs_can_set_width): Use value_is_resizable.
241 (val_labs_add): Simply return false if the value labels set is too
242 wide, instead of having undefined behavior.
243 (val_labs_replace): Ditto.
244 (val_labs_replace): Ditto.
245 (val_labs_first): Set iterator to null if iteration is complete.
246 (val_labs_first_sorted): Ditto.
247 (val_labs_done): Become a no-op if the iterator is null.
249 * value.c (value_is_resizable): New function.
250 (value_resize): New function.
252 * variable.c (var_get_dict_class): New function.
254 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
257 2007-11-08 Ben Pfaff <blp@gnu.org>
259 * data-in.c: Make formatted data parsing locale-independent.
260 (parse_number): Use c_strtod instead of strtod, to avoid
261 locale-specific behavior.
264 2007-11-06 Ben Pfaff <blp@gnu.org>
266 Patch #6256: add support for binary, 360 file formats. Reviewed
269 * data-in.c (struct data_in): Add `encoding' member.
270 (data_in): Add `encoding' parameter, and re-encode the data passed
271 in where appropriate. Update all callers to pass it in.
272 (parse_A): Implement EBCDIC recoding wart described in manual.
273 (parse_AHEX): Implement EBCDIC recoding.
275 * data-out.c (data_out_legacy): New function.
276 (data_out): Make into a wrapper around data_out_legacy.
278 * file-handle-def.c (struct file_handle): New member `encoding'.
279 (fh_create_file): Set encoding.
280 (fh_default_properties): Set default encoding.
281 (fh_get_legacy_encoding): New function.
283 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
284 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
285 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
286 (struct fh_properties): New member `encoding'.
288 2007-11-05 Ben Pfaff <blp@gnu.org>
290 Patch #6258. Reviewed by John Darrington.
292 * file-handle-def.c (fh_lock): Add comment that TYPE should be
293 marked with N_() in the caller. Added these markings to each
294 caller too. Should make i18n easier.
295 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
297 2007-11-03 Ben Pfaff <blp@gnu.org>
299 Allow output files to overwrite input files (bug #21280).
301 * file-handle-def.c: Separate locking of files for input and for
302 output, to allow both to take place at once. Also, distinguish a
303 file handle from the identity of the underlying file, because the
304 identity of a file changes over time and the file handle can't
305 represent two different identities.
306 (struct file_handle): Remove `next', `open_cnt', `deleted',
307 `type', `open_mode', `aux', `identity' members. Change `id' from
308 char array to char *. Add `ref_cnt' member.
309 (file_handle_from_ll) New function.
310 (file_handles) Removed.
311 (named_handles) New variable.
312 (fh_init) Initialize named_handles.
313 (fh_done) Remove name from all named_handles.
314 (free_handle) Updated for modified struct file_handle.
315 (unname_handle) New function.
316 (fh_ref) New function.
317 (fh_from_file_name) Removed.
318 (fh_unref) New function.
319 (fh_unname) New function.
320 (fh_from_id) Rewritten.
321 (create_handle) Updated for modified struct file_handle.
322 (fh_inline_file) Increment the handle's ref count.
323 (fh_create_file) Updated for modified struct file_handle.
324 (fh_create_scratch) Ditto.
329 (fh_is_open) Removed.
330 (fh_get_id) Updated for modified struct file_handle.
331 (fh_get_default_handle) Increment the handle's ref count.
332 (fh_set_default_handle) Handle ref counts.
333 (struct fh_lock) New structure.
334 (locks) New static var.
335 (fh_lock) New function.
336 (fh_unlock) New function.
337 (fh_lock_get_aux) New function.
338 (fh_lock_set_aux) New function.
339 (fh_is_locked) New function.
340 (make_key) New function.
341 (free_key) New function.
342 (compare_fh_locks) New function.
343 (hash_fh_lock) New function.
345 * file-handle-def.h (enum fh_access) New enum.
347 * file-name.c: Made file_identity the same in all supported
349 (struct file_identity): New `name' member.
350 (fn_get_identity): For a file that doesn't exist, get the
351 dev/inode of its directory plus its name. If even the directory
352 doesn't exist, just use its name. Merge the Windows
353 implementation into the Unix one.
354 (fn_compare_file_identities): Rewritten. Merge the Windows
355 implementation into the Unix one.
356 (fn_hash_identity): New function.
358 * make-file.c (struct replace_file): New structure.
359 (all_files): New static var.
360 (replace_file_start): New function.
361 (replace_file_commit): New function.
362 (replace_file_abort): New function.
363 (free_replace_file): New function.
364 (unlink_replace_files): New function.
366 * por-file-reader.c (struct pfm_reader): Add `lock' member.
367 (close_reader): Unlock file.
368 (pfm_open_reader): Lock file.
370 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
372 (pfm_open_writer): Lock file and prepare for its replacement.
373 (close_writer): Unlock file.
375 * scratch-handle.h (struct scratch_handle): Add unique_id so that
376 different generations of a scratch file can be distinguished.
378 * scratch-reader.c (scratch_reader_open): Verify that the file is
381 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
382 remove scratch_handle member.
383 (scratch_writer_open): Lock handle. Prepare to replace handle
384 data, instead of doing it immediately.
385 (scratch_writer_casewriter_destroy): Replace handle data and
388 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
389 (sfm_open_reader): Lock file.
390 (close_reader): Unlock file.
392 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
394 (sfm_open_writer): Lock file and prepare for its replacement.
395 (close_writer): Unlock file.
397 2007-11-02 Ben Pfaff <blp@gnu.org>
399 * data-out.c (output_number): Use isfinite (from C99) instead of
400 gsl_isfinite, since we now have universal support for it in
402 (output_infinite): Ditto.
403 [!HAVE_ROUND] (round): Remove definition, since we now have a
404 replacement in gnulib.
406 2007-10-27 John Darrington <john@darrington.wattle.id.au>
408 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
409 wasn't initialised correctly. Closes bugs #21330 and 21397
412 2007-10-08 Ben Pfaff <blp@gnu.org>
414 Fix bug #21280. Thanks to John Darrington for review.
416 * file-name.c (create_stream): New function.
418 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
419 file handle before creating the file, to ensure that we don't
420 truncate a file that we're reading. Make code easier to read by
423 * sys-file-write.c (sfm_open_writer): Ditto.
425 2007-10-01 Ben Pfaff <blp@gnu.org>
427 Fix bug #21192. Thanks to John Darrington for review.
429 * casereader.c (casereader_read): Decrement case_cnt before
430 calling the casereader's "read" member function, so that we
431 interact properly with lazy_casereader.
433 * datasheet.c: Add regression test for above bug fix.
434 (clone_datasheet): New function.
435 (lazy_callback): New function.
436 (check_datasheet_casereader): New function.
437 (check_datasheet): Check datasheet contents are reported correctly
438 through an ordinary casereader and a lazy casereader.
439 (clone_model): Use clone_datasheet.
441 2007-09-24 Ben Pfaff <blp@gnu.org>
443 Patch #6210. Reviewed by John Darrington.
444 * settings.c: Drop "long view", which is not needed any longer.
445 (static var long_view): Removed.
446 (force_long_view): Removed.
447 (get_viewwidth): Removed.
448 (init_viewport): Removed long_view reference.
450 2007-09-19 John Darrington <john@darrington.wattle.id.au>
452 * settings.c settings.h: Changed viewport's length and width to be
453 owned by the user interface which uses the data library. This allows
454 better abstraction, and makes dynamically adjustable dimensions easier.
456 2007-09-18 Ben Pfaff <blp@gnu.org>
458 * procedure.c (proc_extract_active_file_data): New function.
460 * lazy-casereader.h: New file.
462 * lazy-casereader.c: New file.
464 * casereader.c (casereader_dynamic_cast): New function.
466 2007-09-14 Ben Pfaff <blp@gnu.org>
468 * dictionary.c (dict_clone): Copy case indexes from cloned
469 dictionary. Fixes bug #21061. Reviewed by John Darrington.
471 2007-09-13 John Darrington <john@darrington.wattle.id.au>
473 * value.c value.h (create_value): New function.
475 2007-09-12 Ben Pfaff <blp@gnu.org>
477 Make it clear that translator casereader and casewriter translate
478 functions are supposed to destroy their input case.
480 * casereader-translator.c (struct casereader_translator): Change
481 input case parameter of translate member function from const
482 struct ccase * to struct ccase *.
483 (casereader_create_translator): Ditto, for translate parameter.
485 * casewriter-translator.c (struct casewriter_translator): Ditto.
486 (casewriter_create_translator): Ditto.
488 2007-08-27 John Darrington <john@darrington.wattle.id.au>
490 * sys-file-reader.c (read_display_parameters): Force display width
491 to 8, if the sys file says 0 (like SPSS does).
493 2007-08-12 Ben Pfaff <blp@gnu.org>
495 * dictionary.c (dict_dump): New function.
497 2007-08-12 Ben Pfaff <blp@gnu.org>
499 Drop dict_compactor in favor of using the new struct case_map.
501 * dictionary.c (struct copy_map): Removed.
502 (struct dict_compactor): Removed.
503 (dict_make_compactor): Removed.
504 (dict_compactor_compact): Removed.
505 (dict_compactor_destroy): Removed.
507 * procedure.c (struct data_set): Change `compactor' member to be a
509 (proc_open): Use case_map_to_compact_dict instead of
511 (proc_casereader_read): Use case_map_execute instead of
512 dict_compactor_compact.
513 (proc_commit): Use case_map_destroy instead of
514 dict_compactor_destroy.
516 * scratch-writer.c (struct scratch_writer): Change `compactor'
517 member to be a struct case_map *.
518 (scratch_writer_open): Use case_map_to_compact_dict instead of
520 (scratch_writer_casewriter_write): Use case_map_execute instead of
521 dict_compactor_compact.
523 2007-08-12 Ben Pfaff <blp@gnu.org>
525 * automake.mk: Add case-map.c, case-map.h.
527 * case-map.c: New file.
529 * case-map.h: New file.
531 2007-08-12 Ben Pfaff <blp@gnu.org>
533 * dictionary.c (dict_compact_values): Don't delete scratch
534 variables as well as compacting case indexes. Update all callers.
535 (dict_get_compacted_value_cnt): Rename dict_count_values and
536 change interface. Update all callers.
537 (dict_get_compacted_value_cnt): Remove.
538 (dict_compacting_would_shrink): Remove.
539 (dict_compacting_would_change): Remove.
540 (dict_make_compactor): Add new parameter. Update all callers.
542 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
543 instead of dict_count_values, changing an O(N) operation into
546 2007-08-12 Ben Pfaff <blp@gnu.org>
548 * casereader.c (casereader_read): Don't require cases read by a
549 casereader to be exactly the expected size: as long as they're big
552 2007-08-12 Ben Pfaff <blp@gnu.org>
554 Make casewriters keep track of the number of `union value's in
555 each case. This is useful for two reasons: casewriter_write can
556 then check that the case being written is large enough, and later
557 recipients of the casewriter can determine the size of the case.
559 * casewriter-translator.c (casewriter_create_translator): Add
562 * casewriter.c (struct casewriter): Add value_cnt member.
563 (casewriter_write): Check that the case passed in is big enough.
564 (casewriter_get_value_cnt): New function.
565 (casewriter_create): Add value_cnt parameter.
567 2007-08-09 Ben Pfaff <blp@gnu.org>
569 Fix bug reported by Jason Stover.
570 * settings.c: Move get_termcap_viewport and in particular the
571 #include for <curses.h> to the end of the file. curses.h
572 redefines bool on some systems (e.g. OpenBSD), which causes
573 disagreement between uses of bool before its inclusion and
575 Tested by Jason Stover.
577 2007-07-29 Ben Pfaff <blp@gnu.org>
579 Provisional fix for bug #18692 and bug #20161. Reviewed by John
582 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
583 argument (never "rb" or "wb") because SUSv3 says that only those
584 modes are defined, and glibc in fact rejects other modes.
586 Open portable files with fn_open so that they can be read from
587 pipes. Fix missing fh_close call to go along with fh_open.
588 Report an error if the file close reports an error.
589 * por-file-reader.c (close_reader): New function.
590 (por_file_casereader_destroy): Use close_reader.
591 (pfm_open_reader): Open file with fn_open.
593 2007-07-28 Ben Pfaff <blp@gnu.org>
595 Make PSPP able to read all the portable files I could find on the
596 web. Thanks to John Darrington for review. Bug #17620.
597 * por-file-reader.c (struct pfm_reader): New member `line_length'.
598 (error): Print file offset in hexadecimal.
599 (warning): New function.
600 (advance): Treat lines less than 80 bytes long as padded to 80
602 (pfm_open_reader): Call read_documents if we find an "E" record.
603 (convert_format): Convert invalid formats to the default format
604 instead of aborting reading the file.
605 (read_variables): Rename duplicate variable names instead of
606 aborting reading the file.
607 (read_value_label): Allow string variables of different widths to
608 be assigned value labels in the same record. Replace duplicate
609 value labels instead of aborting.
610 (read_documents): New function.
612 * por-file-writer.c (pfm_open_writer): Call write_documents if the
613 dictionary has documents.
614 (write_documents): New function.
616 2007-07-25 Ben Pfaff <blp@gnu.org>
618 Fix bugs related to bug #17213.
620 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
621 since the former is what config.h has. Include the needed ncurses
623 (static var echo) Rename to `do_echo' because the original name is
624 the same as an ncurses identifier.
625 (get_termcap_viewport) Use error instead of msg.
627 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
628 (fn_close): Don't close stdin, stdout, stderr.
630 2007-07-26 John Darrington <john@darrington.wattle.id.au>
632 * procedure.c procedure.h: Added callbacks which get invoked whenever
633 a dataset's transformation chain changes.
635 2007-07-24 Ben Pfaff <blp@gnu.org>
638 * sys-file-writer.c (write_variable_display_parameters): Use new
639 var_default_display_width function to choose display width of
640 segments after the first one in a given variable.
641 * variable.c (var_create): Use var_default_display_width to pick
642 new variable's display width.
643 (var_default_display_width): New function.
644 Reviewed by John Darrington.
646 2007-07-24 Ben Pfaff <blp@gnu.org>
649 * por-file-writer.c (write_variables): Write weight variable.
650 Reviewed by John Darrington.
652 2007-07-23 Ben Pfaff <blp@gnu.org>
654 Improvements to system file reader and writer.
656 First, move all detailed knowledge of very long strings into
657 sys-file-private.[ch], so that this nasty stuff can be isolated.
659 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
660 (EFFECTIVE_VLS_CHUNK): New macro.
661 (min_int): New function.
662 (max_int): New function.
663 (sfm_width_to_bytes): Rewrite.
664 (sfm_width_to_octs): New function.
665 (sfm_segment_alloc_width): New function.
666 (sfm_segment_alloc_bytes): New function.
667 (sfm_segment_used_bytes): New function.
668 (sfm_segment_offset): New function.
669 (sfm_segment_effective_offset): New function.
670 (sfm_dictionary_to_sfm_vars): New function.
672 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
673 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
674 (struct sfm_var): New structure.
676 Next, improvements to the system file reader.
678 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
679 type casenumber. Added `version_major', `version_minor',
682 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
683 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
684 Change `case_cnt' to type casenumber. Removed `has_vls'.
685 (struct sfm_var): Removed.
686 (sfm_open_reader): Don't warn on wrong case size if the file was
687 written by SPSS 13, which tends to get it wrong. Use
688 sfm_dictionary_to_sfm_vars.
689 (read_header): Always output system file info.
690 (read_variable_record): Simplify code for reading missing values.
691 (read_machine_int32_info): Save version numbers from system file
692 into info struct passed as new argument.
693 (read_long_string_map): Restructured to use new sys-file-private
695 (read_value_labels): Use size_overflow_p.
696 (sys_file_casereader_read): Get rid of distinction between fast
697 and slow paths. Use information provided by sys-file-primate's
698 struct sfm_var to simplify code.
699 (skip_whole_strings): New function.
700 (read_int32): Renamed read_int. Changed return value to int.
702 (read_flt64): Renamed read_float. Changed return value to
703 double. Updated all callers.
704 (int32_to_native): Removed. Changed callers to use
706 (flt64_to_double): Removed. Changed callers to use float_convert.
708 Finally, get rid of int32, flt64 terminology and types in system
709 file writer. The former wasn't very useful since a POSIX "int"
710 can hold the whole range of int32 and we generally didn't have a
711 need for it to be exactly-32-bits, just at-least-32-bits. The
712 latter was inconvenient because we had to assume that it could be
713 different from double and thereby convert special values SYSMIS,
714 HIGHEST, LOWEST to and from it in multiple places. Instead, now
715 we just use "int" and "double" in most places, and do conversions,
716 if necessary, very close to where we do I/O. This change meant
717 that the writer code couldn't represent records in the file as C
718 structs any longer, but that's no great loss. The code actually
719 seems to be more readable without them.
721 Simplify the compression buffering code: only buffer as much as
722 necessary, which is no more than eight 8-byte units at any given
725 * sys-file-writer.c (typedef flt64): Removed.
726 (macro second_lowest_flt64): Removed.
727 (struct sysfile_header): Removed.
728 (struct sysfile_variable): Removed.
729 (struct sfm_writer): Removed `needs_translation', `has_vls',
730 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
731 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
732 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
733 compression buffering by `opcodes', `opcode_cnt', `data',
734 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
735 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
736 data writer functions instead of structures.
737 (calc_oct_idx): New function.
738 (write_header): Use simple data writer functions instead of
740 (write_format_spec): Renamed write_format. New argument.
741 (write_variable_continuation_records): New function.
742 (write_variable): Use simple data writer functions instead of
743 structures. Use write_variable_continuation_records. Write
744 entire very long string instead of requiring caller to understand
746 (write_value_labels): Use simple data writer functions instead of
748 (write_documents): Ditto.
749 (write_variable_display_parameters): Use sys-file-private
750 functions to simplify. Use simple data writer functions instead
752 (write_vls_length_table): Use simple data writer functions instead
754 (write_longvar_table): Ditto.
755 (write_rec_7_34): Break into new functions
756 write_integer_info_record, write_float_info_record. Use simple
757 data writer functions instead of structures.
758 (buf_write): Removed.
759 (append_string_max): Removed.
760 (ensure_buf_space): Removed.
761 (sys_file_casewriter_write): Get rid of the distinction between
762 fast and slow paths, which didn't seem to be too useful. Use new
763 functions write_case_uncompressed, write_case_compressed.
764 (put_instruction): Removed.
765 (put_element): Removed.
766 (write_compressed_data): Removed.
767 (close_writer): Use flush_compressed. Only write case count to
768 system file if it will fit in the field.
769 (write_case_compressed): New function.
770 (write_case_uncompressed): New function.
771 (flush_compressed): New function.
772 (put_cmp_opcode): New function.
773 (put_cmp_number): New function.
774 (write_int): New function.
775 (convert_double_to_output_format): New function.
776 (write_float): New function.
777 (write_value): New function.
778 (write_string): New function.
779 (write_bytes): New function.
780 (write_zeros): New function.
781 (write_spaces): New function.
783 Reviewed by John Darrington.
785 2007-07-22 Ben Pfaff <blp@gnu.org>
787 Don't try to write very long strings to portable files. The
788 format does not support it.
790 * por-file-writer.c (MAX_POR_WIDTH): New macro.
791 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
792 (write_format): Add arg to take width to resize format to.
793 (write_value): Limit width of value written to MAX_POR_WIDTH.
794 (write_variables): Limit width of variable and its output formats
796 Reviewed by John Darrington.
798 2007-07-22 Ben Pfaff <blp@gnu.org>
800 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
801 local variable instead of literal 5.
802 Reviewed by John Darrington.
804 2007-07-22 Ben Pfaff <blp@gnu.org>
806 Fix problems with uniqueness of short names in system files with
807 very long string variables. Now a variable may have multiple
810 * automake.mk (src_data_libdata_a_SOURCES): Add new files
811 short-names.c, short-names.h.
813 * dictionary.c (dict_clone): Clone all the short names.
814 (compare_strings): Move into short-names.c.
815 (hash_strings): Ditto.
816 (set_var_short_name_suffix): Ditto.
817 (dict_assign_short_names): Ditto, rename short_names_assign,
818 change to assign all short names.
820 * por-file-writer.c (write_variables): Use short_names_assign
821 instead of dict_assign_short_names.
823 * short-names.c: New file.
825 * short-names.h: New file.
827 * sys-file-private.c (sfm_width_to_segments): New function.
829 * sys-file-reader.c (read_long_var_name_map): Save and restore all
830 the short names, not just the first one.
832 * sys-file-writer.c (cont_var_name): Removed.
833 (sfm_open_writer): Use short_names_assign instead of
834 dict_assign_short_names. Use unique short names assigned by
835 short_names_assign instead of those generated by cont_var_name.
837 * variable.c (struct variable): Remove `short_name' member,
838 replace by `short_names' and `short_name_cnt'.
839 (var_create) Initialize new members.
840 (var_get_short_name_cnt): New function.
841 (var_get_short_name): Now takes an index argument. Changed most
843 (var_set_short_name): Ditto.
844 (var_clear_short_name): Renamed var_clear_short_names, changed to
845 clear all short names.
847 Reviewed by John Darrington.
849 2007-07-22 Ben Pfaff <blp@gnu.org>
851 * variable.c (var_set_width): Use new var_set_width function.
853 * missing-values.c (mv_n_values): Drop assertion, which was not
856 * format.c (fmt_default_for_width): New function.
857 (fmt_resize): New function.
859 Reviewed by John Darrington.
861 2007-07-18 John Darrington <john@darrington.wattle.id.au>
863 * datasheet.c (datasheet_delete_columns): Added assertion to check
864 we're not deleting outside the range of the sheet.
867 * dictionary.c dictionary.h variable.c: Added the ability for string
868 variables to be resized.
870 * vardict.h: Added some prototypes (moved from dictionary.h) as
871 these should only be called by variable.c
874 2007-07-14 John Darrington <john@darrington.wattle.id.au>
876 * sfm-reader.c: Respect case_cnt field in file header.
878 2007-07-01 John Darrington <john@darrington.wattle.id.au>
880 * transformation.c transformation.h (trns_chain_execute): Changed the
881 signature (Patch #6057)
883 2007-06-10 Ben Pfaff <blp@gnu.org>
885 * casereader-filter.c (casereader_filter_destroy): Make sure to
886 write all the remaining excluded cases to the casewriter, if any.
888 * caseinit.c (init_list_destroy): Rewrite.
889 (init_list_clear): Ditto.
891 * casegrouper.c (casegrouper_get_next_group): Always set *reader
892 to null when returning false.
894 2007-06-06 Ben Pfaff <blp@gnu.org>
896 Actually implement the new procedure code and adapt all of its
897 clients to match. Also adapt all of the other case sources and
898 sinks in the tree and their clients to use the
899 casereader/casewriter infrastructure.
901 * automake.mk: Add and remove files.
903 * any-reader.c: Change into a casereader.
904 * por-file-reader.c: Ditto.
905 * scratch-reader.c: Ditto.
906 * sys-file-reader.c: Ditto.
908 * any-writer.c: Change into a casewriter.
909 * por-file-writer.c: Ditto.
910 * scratch-writer.c: Ditto.
911 * sys-file-writer.c: Ditto.
913 * procedure.c: Change to use casereader, casewriter, caseinit, and
914 other new infrastructure.
916 * scratch-handle.c: Adapt to new infrastructure.
918 * case-sink.c: Removed, now dead code.
919 * case-sink.h: Ditto.
920 * case-source.c: Ditto.
921 * case-source.h: Ditto.
922 * casefile-factory.c: Ditto.
923 * casefile-private.h: Ditto.
926 * casefilter.c: Ditto.
927 * casefilter.h: Ditto.
930 * fastfile-factory.c: Ditto.
931 * fastfile-factory.h: Ditto.
932 * storage-stream.c: Ditto.
933 * storage-stream.h: Ditto.
935 2007-06-06 Ben Pfaff <blp@gnu.org>
939 * automake.mk: Add new files.
941 * datasheet.c: New file.
943 * datasheet.h: New file.
945 2007-06-06 Ben Pfaff <blp@gnu.org>
947 Until now, the procedure code has provided a case to the
948 case_source, which has filled in the data values that come from
949 the active file. "Left" data values that don't come from the
950 active file naturally stay the same from case to case, because the
951 procedure code keeps using that same case.
953 One of the compromises that comes with the new procedure code is
954 that the active file allocates and provides its own case, which
955 the procedure code then has to resize to provide room for any
956 other variables that should go in the case and then fill in the
957 values of "left" variables. Then, when we're done with that case,
958 we have to save the values of "left" variables to copy into the
959 next case read from the active file.
961 The caseinit code helps with this.
963 * automake.mk: Add new files.
965 * caseinit.c: New file.
967 * caseinit.h: New file.
969 2007-06-06 Ben Pfaff <blp@gnu.org>
971 * value.h (value_cnt_from_width): New function.
973 * variable.c (var_get_value_cnt): Use new function.
975 2007-06-06 Ben Pfaff <blp@gnu.org>
977 Add casegrouper, to allow cases read from a given casereader to be
978 broken into groups, each of which has its own casereader.
979 Generally cases are grouped based on having equal values for some
982 * automake.mk: Add new files.
984 * casegrouper.c: New file.
986 * casegrouper.h: New file.
988 2007-06-06 Ben Pfaff <blp@gnu.org>
990 Add interface to lexicographical ordering of cases.
992 * automake.mk: Add new files.
994 * case-ordering.c: New file.
996 * case-ordering.h: New file.
998 2007-06-06 Ben Pfaff <blp@gnu.org>
1000 Add casereaders and casewriters, the basis of the new data processing
1001 implementation. A casereader is a uniform interface to reading cases
1002 from a data source; a casewriter is a uniform interface to writing
1003 cases to a data sink.
1005 * automake.mk: Add new files.
1007 * casereader-filter.c: New file.
1009 * casereader-provider.h: New file.
1011 * casereader-translator.c: New file.
1013 * casereader.c: New file.
1015 * casereader.h: New file.
1017 * casewriter-provider.h: New file.
1019 * casewriter-translator.c: New file.
1021 * casewriter.c: New file.
1023 * casewriter.h: New file.
1025 2007-06-06 Ben Pfaff <blp@gnu.org>
1027 "casewindow" data structure that extends the deque (from libpspp)
1028 of cases with the ability to dump cases to disk if we get too many
1031 * automake.mk: Add new files.
1033 * casewindow.c: New file.
1035 * casewindow.h: New file.
1037 2007-06-06 Ben Pfaff <blp@gnu.org>
1039 sparse_cases data structure that augments a sparse_array of cases
1040 with the ability to dump cases to disk if we get too many cases in
1043 * automake.mk: Add new files.
1045 * sparse-cases.c: New file.
1047 * sparse-cases.h: New file.
1049 2007-06-06 Ben Pfaff <blp@gnu.org>
1051 Adds a low-level on-disk case array data structure.
1053 * automake.mk: Add new files.
1055 * case-tmpfile.c: New file.
1057 * case-tmpfile.h: New file.
1059 2007-06-06 Ben Pfaff <blp@gnu.org>
1061 In a couple of places we calculate the maximum number of cases to
1062 keep in memory based on the user-defined workspace. Enable
1063 centralizing the calculation through a new function.
1065 * settings.c (get_workspace_cases): New function.
1067 2007-06-06 Ben Pfaff <blp@gnu.org>
1069 The casenumber type is defined in transformations.h, but case.h is
1070 a more sensible place. Move it.
1072 * case.h (CASENUMBER_MAX): New macro.
1073 (typedef casenumber): Move here, from transformations.h.
1075 2007-06-03 Ben Pfaff <blp@gnu.org>
1077 Slightly generalize case_to_values and case_from_values functions.
1079 * case.c (case_to_values): Rename case_copy_out, change interface.
1080 (case_from_values): Rename case_copy_in, change interface.
1082 * fastfile.c (fastfilereader_get_next_case): Update caller.
1083 (write_case_to_disk): Ditto.
1085 2007-06-02 Ben Pfaff <blp@gnu.org>
1087 Clean up after a forgotten part of patch #5829.
1089 * casedeque.h: Remove unused file.
1091 * automake.mk: Remove casedeque.h from sources.
1093 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1095 * category.c: Removed redundant #include
1097 2007-05-06 Ben Pfaff <blp@gnu.org>
1099 Abstract the documents within a dictionary a little better.
1100 Thanks to John Darrington for suggestion, initial version, and
1101 review. Patch #5917.
1103 * dictionary.c (struct dictionary): Change `documents' member from
1104 char * to struct string.
1105 (dict_clear): Destroy struct string.
1106 (dict_get_documents): Convert struct string to char *.
1107 (dict_set_documents): Set struct string. Pad to 80-character
1109 (dict_clear_documents): New function.
1110 (dict_add_document_line): New function.
1111 (dict_get_document_line_cnt): New function.
1112 (dict_get_document_line): New function.
1114 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1116 * sys-file-reader.c (read_documents): Use new document functions.
1118 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1120 * sys-file-reader.c: When reading a system file which has no
1121 long name table, automatically create one where the long names
1122 are the lower case versions of the short names.
1124 2007-04-22 Ben Pfaff <blp@gnu.org>
1126 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1127 dict_clear will free most data related to the dictionary.
1128 dict_clear does a decent job, except that dict_set_split_vars on
1129 some systems won't actually free the dict's "split" member.
1130 Instead, it'll allocate a 1-byte region. Fix this.
1132 * value.c (value_copy): New function.
1133 (value_set_missing): Ditto.
1135 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1137 * Deleted existing category.h and moved cat-routines.h into
1138 category.h Encapsulated struct cat_vals better.
1140 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1142 * sys-file-reader.c: When reading a system file which has no
1143 long name table, automatically create one where the long names
1144 are the lower case versions of the short names.
1146 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1148 * sys-file-reader.c: Some versions of Other Software seem to
1149 produce system files with string variables' measure set to
1150 zero. We'll assume these are supposed to be nominal variables.
1152 2007-03-30 Ben Pfaff <blp@gnu.org>
1154 * procedure.c: Adapt to new deque data structure.
1156 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1157 Ben Pfaff <blp@gnu.org>
1159 * file-name.c: Mingw compatibility fixes.
1160 (fn_search_path): Use ISSLASH instead of comparing against '/'
1162 (fn_dir_name): Use dir_name from gnulib.
1163 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1164 (fn_get_identity): Use GetFullPathName instead of canonicalize
1165 from gnulib, because the latter does not fully support
1166 Windows-style path names. Use this implementation based on the
1167 detected presence of Windows instead of the absence of Unix, since
1168 the new implementation is Windows-specific.
1169 (fn_compare_file_identities): In Windows implementation, compare
1170 names case-insensitively.
1172 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1174 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1175 <stdlib.h> to have the same effect.
1177 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1179 * por-file-reader.c: Add missing _() around messages.
1181 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1183 * make-file.c: Include "mkstemp.h", without which linking on
1186 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1188 Reduce platform dependence.
1190 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1191 Everywhere we using this, we really should have just depended on
1192 the shell to expand tildes.
1193 (fn_search_path): Simplify, given that we don't do tilde expansion
1195 (fn_normalize): Removed. Caller changed to use the canonicalize
1197 (fn_get_cwd): Removed. Only user was fn_normalize.
1198 (fn_is_absolute): Really only test for absolute names.
1199 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1201 (fn_readlink): Removed, as it was only used fn_normalize.
1202 (fn_exists): Assume the stat function is available; gnulib does.
1203 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1205 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1207 * dictionary.c (dict_create_vector_assert): New function.
1209 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1211 * file-name.c (fn_normalize): Correct name of function
1212 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1213 for pointing this out.
1215 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1217 We are using a single member in struct file_handle, the "name"
1218 field, for more than one purpose. When it begins with '"', it's a
1219 file name; otherwise, it's a token that can be used to identify
1220 it. When that assertion fires, it's because we searched for the
1221 name case-sensitively as a file name (so that there was no match),
1222 and then we try to insert it case-insensitively as a token, which
1223 fails because duplicates aren't allowed.
1225 Solution: break the two purposes into two separate fields. This
1226 fixes the problem and likely makes the code easier to read too.
1228 Fixes bug #18922. Thanks to John Darrington for bug report and
1231 * file-handle-def.c (struct file_handle): New `id' member.
1232 (fh_from_name): Rename fh_from_id. Update all callers.
1233 (create_handle): New `id' parameter. Update all callers.
1234 (fh_create_file): Ditto.
1235 (fh_get_id): New function.
1237 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1239 * case.c (case_is_null): Change return type to bool.
1241 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1245 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1246 reference counting to share data. This makes it easy for
1247 valgrind, etc. to find accesses to cases that have been destroyed
1248 but have been kept around by another user's ref-count. This often
1249 happens when the data set is small enough to find in memory; if a
1250 bigger data set that would overflow to disk were used, then data
1251 corruption would occur.
1253 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1257 * case.c (case_unshare): Make it check internally whether the
1258 ref_cnt is greater than 1, so that the callers don't have to.
1259 Update callers not to check.
1261 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1263 Before, I was thinking that I might want to get rid of reference
1264 counting at some point. Now, I'm pretty sure that it's here to
1265 stay. Thus, because we have to store the value_cnt anyway for
1266 reference-counted cases, we might as well expose it to users.
1268 * case.c (case_get_value_cnt): New function.
1269 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1270 resize case if its size actually changed.
1272 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1273 instead of peeking inside struct case directly.
1274 (casefile_append): Ditto.
1276 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1278 Get rid of the inlines for the case functions, which made the
1279 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1280 the inlines didn't speed up "make check" at all, which is not a
1281 perfect benchmark but seems indicative.)
1283 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1284 definitions. Remove some assertions on nonnull pointers that were
1285 redundant with a pointer dereference soon after in the function.
1287 (struct case_data): Move definition here from case.h.
1291 (case_data_wr): Ditto.
1293 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1295 * automake.mk: Add casedeque.h to sources.
1297 * casedeque.h: New file.
1299 * procedure.c: (struct dataset) Change lag_count, lag_head,
1300 lag_queue member into single struct casedeque member. Update all
1301 users to use the casedeque instead.
1304 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1306 * procedure.c: Simplify lagged cases interface. Updated all
1307 clients--well, the only client--to use the simplified interface.
1308 (dataset_n_lag) Removed.
1309 (dataset_set_n_lag) Removed.
1310 (dataset_need_lag) New function.
1312 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1314 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1315 whenever appropriate, but only when the dataset/dictionary is in a
1318 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1320 * dictionary.c dictionary.h : Added callbacks for change of filter and
1321 split variables. Refactored some code to ensure that callbacks get
1322 invoked when appropriate.
1324 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1325 replace_dict callback occurs when permanent_dict replaces the current
1328 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1330 * dictionary.c dictionary.h : Added callback for when the weight
1331 variable of a dictionary changes.
1333 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1335 * dictionary.c dictionary.h : Added replace_source and replace_dict
1336 callbacks, and functions to deal with them.
1338 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1340 Simplify missing value handling.
1342 * missing-values.h (enum mv_class): New type.
1343 (enum mv_type): Moved definition into missing-values.c and renamed
1344 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1345 enums. Updated all uses.
1346 (struct missing_values): Changed type of `type' from `enum
1347 mv_type' to `int' because the definition is no longer exposed.
1349 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1350 parameter. Update all callers.
1351 (mv_is_num_missing): Ditto.
1352 (mv_is_str_missing): Ditto.
1353 (mv_is_value_user_missing): Removed. Changed callers to use
1354 mv_is_value_missing.
1355 (mv_is_num_user_missing): Removed. Changed callers to use
1357 (mv_is_str_user_missing): Removed. Changed callers to use
1359 (mv_is_value_system_missing): Removed. Changed callers to use
1360 mv_is_value_missing.
1361 (mv_set_type): Removed. Changed callers to use mv_clear.
1362 (mv_clear): New function.
1364 * variable.c (var_is_value_missing): Add new enum mv_class
1365 parameter. Update all callers.
1366 (var_is_num_missing): Ditto.
1367 (var_is_str_missing): Ditto.
1368 (var_is_value_user_missing): Removed. Changed callers to use
1369 var_is_value_missing.
1370 (var_is_num_user_missing): Removed. Changed callers to use
1372 (var_is_str_user_missing): Removed. Changed callers to use
1374 (var_is_value_system_missing): Removed. Changed callers to use
1375 var_is_value_missing.
1377 * casefilter.c (struct casefilter): Use enum mv_class in place of
1379 (casefilter_variable_missing): Adapt to new member.
1380 (casefilter_create): Change signature to take enum mv_class,
1383 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1385 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1387 * case-sink.c case-sink.h procedure.c procedure.h
1388 storage-stream.c: Now uses the factory.
1390 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1392 Make it possible to pull cases from the active file with a
1393 function call, instead of requiring indirection through a callback
1396 * case-source.h (struct case_source_class): Change ->read function
1397 to return a single case, instead of calling a callback function
1398 for each case. Change ->destroy function to return an error
1401 * case-source.c (free_case_source): Pass along the value returned
1402 by the case_source ->destroy function.
1404 * procedure.c (struct write_case_data): Removed.
1405 (struct dataset): Added some members to track procedure state.
1406 (procedure): Optimize the trivial case at this level.
1407 (internal_procedure): Re-implement in terms of proc_open,
1408 proc_read, proc_close.
1409 (proc_open) New function.
1410 (proc_read) New function.
1411 (proc_close) New function.
1412 (write_case) Moved into proc_read.
1413 (close_active_file) Moved closing of data source into proc_close.
1415 * storage-source.c: Rewrote to conform with modified
1416 case_source_class interface.
1418 * transformations.c (trns_chain_execute): Added argument to allow
1419 starting execution from an arbitrary transformation. Updated
1422 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1424 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1426 * sys-file-reader.c (read_display_parameters): Don't assume that
1427 MEASURE_* and ALIGN_* have the same values found in system files.
1429 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1431 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1432 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1433 the value of n_MEASURES, which was off by 1 (at least from my
1436 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1438 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1439 callbacks which are invoked when the dictionary or its
1440 variables are changed.
1442 * missing-values.c missing-values.h value-labels.c: Tidied up
1443 consistency checks, and made some of them return false
1444 instead of assert-failing.
1446 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1448 * calendar.c (calendar_days_in_month): New function.
1450 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1452 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1453 the value label's width.
1455 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1457 * sfm-private.h: Move contents into sys-file-writer.c, which is
1458 the only remaining user. Removed Borland C++-specific directives.
1460 * sys-file-reader.c: Clean up and rewrite entire file. The
1461 rewritten version is simpler and better abstracted, and should be
1462 easier to maintain and extend. It avoids using structures to read
1463 file data, which is prone to padding variations among compilers.
1464 It should also handle non-IEEE 754 system files, although I
1465 haven't been able to find any. It has been tested against many
1466 .sav files obtained from the Web and found to produce the same
1467 results as the earlier version of the code, or in some cases
1468 improved results. It is more tolerant of format variations found
1471 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1472 member, putting an enum integer_format in its place. New member
1473 `float_format'. Changed `compressed' member to type bool.
1475 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1477 * dictionary.c (dict_delete_consecutive_vars): New function.
1479 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1481 * file-name.c (fn_search_path): Remove prefix arg that was unused
1482 by any caller. Updated all callers.
1484 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1486 * format.c (fmt_dollar_template): Use user's decimal point
1487 character. Add assertion.
1489 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1491 * format.c (fmt_dollar_template): New function, based on
1492 dollar_format_template from var-type-dialog.c.
1494 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1496 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1499 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1501 Finish converting struct variable to an opaque type. In this
1502 phase, we add remaining setter and getter functions, convert the
1503 remaining PSPP code to use them, and do a bunch of cleanup. The
1504 resulting changes are pervasive but mostly trivial, and only the
1505 notable changes are logged.
1507 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1510 * case.c (case_data): Renamed case_data_idx.
1511 (case_num): Renamed case_num_idx.
1512 (case_str): Renamed case_str_idx.
1513 (case_data_rw): Renamed case_data_rw_idx.
1515 * case.h (case_data): New function with old name and an interface
1516 that takes a variable instead of an index, which is easier to
1517 use. Updated all callers to use the new interface, or to use the
1518 new *_idx function (see above).
1521 (case_data_rw): Ditto.
1523 * category.c (cat_stored_values_destroy): Changed interface to
1524 take a struct cat_vals * instead of a struct variable *.
1526 * dictionary.c (dict_clone): Use new vector_clone function.
1527 (dict_clear) Use new var_destroy function.
1528 (add_var) New function.
1529 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1530 (dict_create_var_assert) Rewrite in terms of add_var.
1531 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1532 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1533 (dict_lookup_var) Use new var_create, var_destroy functions.
1534 (dict_contains_var) Rewrite in terms of new vardict functionality.
1535 (set_var_dict_index) New function.
1536 (set_var_case_index) New function.
1537 (reindex_vars) New function.
1538 (dict_delete_var) Rewrite in terms of new vardict functionality.
1539 (dict_reorder_var) Ditto.
1540 (dict_reorder_vars) Ditto.
1541 (rename_var) New function.
1542 (dict_rename_var) Use rename_var.
1543 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1544 (dict_get_compacted_idx_to_fv) Rename
1545 dict_get_compacted_dict_index_to_case_index, update callers.
1546 (dict_create_vector) Use new vector_create function.
1547 (dict_clear_vectors) Use new vector_destroy function.
1548 (set_var_short_name_suffix) Move here from variable.c, renamed
1549 from var_set_short_name_suffix, make static, update caller.
1551 * sys-file-private.c: New file.
1552 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1553 width_to_bytes, update callers.
1555 * sys-file-private.h: New file. Later it will supplant
1556 sfm-private.h; for now it supplements it.
1557 (macro MIN_VERY_LONG_STRING) New macro.
1558 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1560 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1561 MAX_LONG_STRING was used before.
1563 * sys-file-writer.c: Ditto.
1565 * value-labels.c: Change the paradigm here to be that a null
1566 pointer is OK for a struct val_labs * in most cases; it just
1567 represents an empty set of value labels.
1568 (val_labs_copy) A copy of a null set is a null set.
1569 (val_labs_count) A null set has 0 labels.
1570 (val_labs_replace) Change return type to void. Rewrite for
1572 (val_labs_find) A null set does not contain the value.
1573 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1574 transposed argument order, updated all callers.
1576 * value.c: New file.
1577 (value_dup) Moved here from variable.c.
1578 (compare_values) Ditto.
1581 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1582 (macro MAX_LONG_STRING) Removed, because it was only interesting
1583 for system files, not for general code.
1584 (macro MAX_VERY_LONG_STRING) Ditto.
1585 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1586 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1588 * vardict.h: New file, for an interface between variables and
1591 * variable.c: A lot of functions were moved around, for better
1593 (struct variable) Move definition here, from variable.h.
1594 (var_type_adj) Removed--makes i18n hard.
1595 (var_type_noun) Ditto.
1596 (var_create) New function.
1597 (var_clone) New function.
1598 (var_destroy) New function.
1599 (var_set_name) Assert that variable is not in a dictionary.
1600 (compare_var_names) Rename compare_vars_by_name and fix a couple
1601 of callers who thought the args were strings.
1602 (hash_var_name) Rename hash_var_by_name.
1603 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1604 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1605 (var_is_very_long_string) Removed, because it was only interesting
1606 to system file code.
1607 (var_set_missing_values) Allow the argument to be the wrong width,
1608 as long as we can resize it. Simplify callers who were doing the
1609 resizing themselves.
1610 (var_get_value_labels) New function.
1611 (var_has_value_labels) New function.
1612 (var_set_value_labels) New function.
1613 (alloc_value_labels) New function.
1614 (var_add_value_label) New function.
1615 (var_replace_value_label) New function.
1616 (var_clear_value_labels) New function.
1617 (var_lookup_value_label) New function.
1618 (var_get_value_name) Moved here from variable.c, renamed from
1619 var_get_value_name, transposed argument order, updated all
1621 (var_to_string) Moved here, from variable-label.c.
1622 (var_set_leave) New function.
1623 (var_get_leave) New function.
1624 (var_must_leave) New function.
1625 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1626 set_var_short_name_suffix.
1627 (var_get_dict_index) New function.
1628 (var_get_case_index) New function.
1629 (var_get_obs_vals) New function.
1630 (var_set_obs_vals) New function.
1631 (var_has_obs_vals) New function.
1632 (var_get_vardict) New function.
1633 (var_set_vardict) New function.
1634 (var_has_vardict) New function.
1635 (var_clear_vardict) New function.
1636 (value_dup) Moved to value.c.
1637 (compare_values) Ditto.
1640 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1641 (enum ALPHA) Rename VAR_STRING, update all users.
1643 * vector.c: New file.
1644 (struct vector) Moved here, from variable.h.
1645 (check_widths) New function.
1646 (vector_create) New function.
1647 (vector_clone) New function.
1648 (vector_destroy) New function.
1649 (vector_get_name) New function.
1650 (vector_get_var) New function.
1651 (vector_get_var_cnt) New function.
1652 (compare_vector_ptrs_by_name) New function.
1654 * vector.h: New file.
1656 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1658 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1659 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1662 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1664 * value-labels.c (destroy_atoms): New function.
1665 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1667 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1669 Thanks to Jason Stover for pointing out this problem.
1671 * data-out.c (output_number): Use gsl_finite from GSL, which is
1672 portable, instead of isfinite, which is not.
1675 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1677 * variable.c variable.h (value_dup): New function.
1679 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1681 Start converting struct variable to an opaque type. In this
1682 phase, we add a bunch of setter and getter functions and convert
1683 most of the PSPP code to use them. The resulting changes are
1684 pervasive but mostly trivial, and only the notable changes are
1687 * format.c (fmt_equal): New function.
1689 * variable.c (var_type_is_valid): New function.
1690 (measure_is_valid) Moved here, from format.c.
1691 (alignment_is_valid) Moved here, from format.c.
1692 (var_get_name) New function.
1693 (var_set_name) New function.
1694 (width_to_type) New function.
1695 (var_get_type) New function.
1696 (var_get_width) New function.
1697 (var_set_width) New function.
1698 (var_is_numeric) New function.
1699 (var_is_alpha) New function.
1700 (var_is_short_string) New function.
1701 (var_is_long_string) New function.
1702 (var_is_very_long_string) New function.
1703 (var_get_missing_values) New function.
1704 (var_set_missing_values) New function.
1705 (var_clear_missing_values) New function.
1706 (var_has_missing_values) New function.
1707 (var_is_value_missing) New function.
1708 (var_is_num_missing) New function.
1709 (var_is_str_missing) New function.
1710 (var_is_value_user_missing) New function.
1711 (var_is_num_user_missing) New function.
1712 (var_is_str_user_missing) New function.
1713 (var_is_value_system_missing) New function.
1714 (var_get_print_format) New function.
1715 (var_set_print_format) New function.
1716 (var_get_write_format) New function.
1717 (var_set_write_format) New function.
1718 (var_set_both_formats) New function.
1719 (var_get_label) New function.
1720 (var_set_label) New function.
1721 (var_clear_label) New function.
1722 (var_has_label) New function.
1723 (var_get_measure) New function.
1724 (var_set_measure) New function.
1725 (var_get_display_width) New function.
1726 (var_set_display_width) New function.
1727 (var_get_alignment) New function.
1728 (var_set_alignment) New function.
1729 (var_get_value_cnt) New function.
1730 (var_get_leave) New function.
1731 (var_get_short_name) New function.
1733 * variable.h: (struct variable) Removed "type" and "nv" members;
1734 they are now computed from "width" where needed.
1736 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1738 * missing-values.c (mv_resize): Don't write beyond end of the
1739 allocated buffer when resizing a long string.
1741 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1743 Clean up identifier code: don't require identifier enumerations to
1744 be in a particular order; make better use of string library;
1745 expose less of the internals.
1747 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1748 change interface. Updated all callers.
1749 (lex_id_match) Change interface to use struct substring, update
1751 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1752 (global array keywords[]) Make static, change form. Update all
1753 users to use lex_id_name instead.
1754 (lex_is_keyword) New function.
1755 (lex_id_to_token) Change interface to use struct substring, update
1757 (lex_id_name) New function.
1759 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1760 lex_is_keyword instead.
1761 (T_LAST_KEYWORD) Removed.
1762 (T_N_KEYWORDS) Removed.
1764 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1766 * format.c: (fmt_date_template) Distinguish characters for which a
1767 space is output and any date delimiter is allowed on input, from
1768 those for which a space is output and only a space is allowed on
1769 input. The former is represented by X, the latter by a space.
1770 Also, drop distinction between h and H, changing the former to the
1773 * data-in.c: Completely rewrite internals to conform to SPSS input
1774 formats as closely as possible.
1775 (data_in) Changed external interface by replacing the structure
1776 that was used as a single argument by a set of arguments. Updated
1778 (data_in_finite_line) Removed. Converted all callers to use plain
1780 (data_in_get_integer_format) New function.
1781 (data_in_set_integer_format) New function.
1782 (data_in_get_float_format) New function.
1783 (data_in_set_float_format) New function.
1785 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1786 (struct data_in) Removed.
1788 * data-out.c: (output_date) Drop each component from the input as
1789 it is output, to allow us to drop the distinction between h (a
1790 count of hours) and H (the hour of day) template characters.
1791 Also, handle new X template character.
1792 (output_scientific) Follow more rational rule on when to drop
1793 fraction introduced between SPSS 13 and 15. Updated test case to
1796 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1798 Fix buffer overflow reported by John Darrington.
1800 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1801 realize that DIGITS is a count of nibbles, not of bytes.
1803 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1805 * calendar.c (calendar_offset_to_gregorian) Also return the
1806 year-of-day. Change callers to new interface.
1808 * data-out.c: Completely rewrite internals to conform to SPSS
1809 output formats as completely as possible.
1810 (data_out) Change interface to put input parameters before output
1811 parameters, for consistency with the style I now prefer. Update
1813 (data_out_get_integer_format) New public function.
1814 (data_out_set_integer_format) New public function.
1815 (data_out_get_float_format) New public function.
1816 (data_out_set_float_format) New public function.
1818 * data-out.h: New file. Move prototype for data_out here, from
1821 * format.c: (fmt_step_width) Use equality comparison instead of
1822 bitwise and, for clarity.
1823 (fmt_is_string) Ditto.
1824 (fmt_input_to_output) Fix categories that are translated to F
1827 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1829 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1830 casefile-private.h: Added casefilter to assist commands with missing
1833 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1835 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1837 * settings.c: (route_errors_to_terminal) New variable.
1838 (route_errors_to_listing) New variable.
1839 (get_error_routing_to_terminal) New function.
1840 (set_error_routing_to_terminal) New function.
1841 (get_error_routing_to_listing) New function.
1842 (set_error_routing_to_listing) New function.
1844 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1846 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1848 * format.c: Completely rewrite, to achieve better abstraction.
1849 Rewrite all references to formats in other files.
1851 * format.def: Rewrite and reorganize.
1853 * settings.c: Move everything related to custom currency formats
1854 into format.[ch], changing them in form, so as to group related
1855 code and definitions better. Changed all references to use the
1857 (static var decimal) Removed.
1858 (static var grouping) Removed.
1859 (static var cc) Removed.
1860 (get_decimal) Removed.
1861 (set_decimal) Removed.
1862 (get_grouping) Removed.
1863 (set_grouping) Removed.
1867 * settings.h: (macro CC_CNT) Removed.
1868 (macro CC_WIDTH) Removed.
1869 (struct custom_currency) Removed.
1871 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1873 * data-in.c (data_in): Use switch statement instead of table, to
1874 avoid dependence on the order of the FMT_* enums.
1876 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1878 * data-out.c: (num_to_string) Removed, because it was dead code.
1880 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1882 * data-in.c (parse_trailer): Fix error message.
1884 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1886 * format.c (fmt_is_binary): New function.
1888 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1890 * procedure.c procedure.h: Encapsulated the static data into a single
1893 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1895 * casefile.c (casereader_read_xfer): Always initialize the case,
1896 even on an error condition.
1898 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1900 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1901 entire function useless.
1903 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1905 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1906 casereader method casereader_clone.
1908 * procedure.c pransformations.h: Introduced new type casenum_t
1910 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1912 * variable.c: (width_to_bytes) Rephrase code for clarify.
1914 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1916 * format.c: (fmt_type_from_string) New function.
1917 (fmt_to_string) Include decimals in output if the format has
1918 decimals, even if the format type does not. This way, we can
1919 accurately reproduce incorrect formats in user output.
1920 (check_common_specifier) Make the check for a bad format type an
1921 assertion, so we get bug reports if they show up. Fix message.
1922 Check for decimal places with a format type that doesn't allow
1924 (check_input_specifier) Remove check for FMT_X, which has been
1926 (check_output_specifier) Ditto.
1928 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1930 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1931 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1932 (enum fmt_parse_flags) Removed.
1934 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1936 * casefile.c casefile.h: Converted to an abstract base class.
1937 * casefile-private.h fastfile.c fastfile.h: New files.
1938 * automake.mk procedure.c scratch-writer.c storage-stream.c
1940 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1942 * procedure.c (internal_procedure): Create sink_case with only as
1943 many values as the compacted dictionary.
1945 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1947 Remove "debugging" code that caused plenty of false positives and
1950 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1952 * case.c: Remove all references to `this' member.
1954 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1956 Fix link error noted by Jason Stover.
1958 * storage-stream.c: Include <assert.h>.
1960 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1962 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1963 ALL) and additional underlying system file issues.
1965 Thanks to John Darrington for review.
1967 First problem: var_hash points to variables not owned by the
1968 sys-file-reader, which the caller may free or modify. Use an
1969 array of sfm_vars instead, as done earlier (e.g. CVS version
1972 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1973 members and remove all code that references it. Add vars, var_cnt
1974 members. Remove fix_specials member, which was unused.
1975 (struct sfm_var) Remove name member, which was unused.
1976 (sfm_close_reader) Free vars member instead of var_hash.
1977 (compare_var_shortnames) Removed.
1978 (hash_var_shortname) Removed.
1979 (sfm_open_reader) Fill out vars array.
1980 (compare_var_index) Removed.
1981 (sfm_read_case) Use vars instead of var_hash.
1983 Second problem: we're confused about when we actually have very
1984 long strings, causing us to choose incorrectly between slow path
1985 and fast path in sfm_read_case.
1987 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1988 have very long strings, not when we have long variable names,
1989 which is an unrelated feature.
1991 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1993 * variable.h: Move var_set and variable parsing declarations to
1994 new header, src/language/lexer/variable-parser.h. Modified lots
1995 of files to include the new header.
1997 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1999 * value-labels.c (value_to_string): When there's no value label,
2000 format the variable according to its print format, instead of
2001 always effectively using A or F format.
2003 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
2005 * casefile.c (casefile_get_random_reader): Nasty hack to get around
2008 * format.c: Removed tortological assertion.
2010 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
2012 Reform string library.
2014 * file-name.c (fn_interp_vars): Change interface to take a
2015 substring as input. Updated all users.
2017 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
2019 * format.c (measure_is_valid): Really return false when m >=
2022 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
2024 Implement random access to casefiles, for use in GUI.
2026 * casefile.c: (struct casereader) Add `random', `file_ofs',
2027 `buffer_ofs' members.
2028 (casefile_get_random_reader) New function.
2029 (read_open_file) Break part into new function
2030 seek_and_fill_buffer().
2031 (fill_buffer) Update buffer_ofs, file_ofs.
2032 (casereader_seek) New function.
2034 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2036 * settings.c: Added call to i18n{done, init}.
2038 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
2040 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
2043 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
2045 * casefile.c: Convert many uses of `int' to `bool'.
2047 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
2049 * transformations.c (trns_chain_destroy): Destroy chain's trns
2050 member, to fix memory leak.
2052 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
2054 * storage-stream.c (storage_source_decapsulate): Destroy case
2055 source to fix memory leak.
2057 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
2059 * scratch-reader.c (scratch_reader_read_case): Copy into existing
2060 case passed as argument instead of initializing the argument as a
2061 case. Fixes memory leak that showed up in
2062 tests/command/aggregate.sh with scratch files.
2064 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
2066 * procedure.c (proc_done): Destroy default_dict, to fix memory
2069 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2071 Simplify procedure_with_splits().
2073 * procedure.c (struct split_aux_data): Removed case_count member.
2074 (procedure_with_splits) Don't initialize case_count.
2075 (split_procedure_case_func) Check whether prev_case is null
2076 instead of case_count.
2077 (split_procedure_end_func) Ditto.
2079 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2081 * case.c (case_move): Do nothing if dst and src are the same
2083 (case_try_create) Merge two similar cases.
2084 (case_copy) Unshare only if data must be actually copied.
2086 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2088 * data-in.c data-out.c dictionary.c sys-file-reader.c
2089 sys-file-writer.c variable.c variable.h: Reworked very long string
2090 support for better encapsulation.
2092 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2094 * value-labels.c (val_labs_can_set_width): New function.
2095 (val_labs_set_width) Clear labels if increasing width to long
2097 (val_labs_destroy) Remove unneeded test for null.
2099 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2101 * value-labels.h: Remove unneeded dependency on variable.h.
2103 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2105 Get rid of `char *c' member in union value, for cleanliness.
2107 * value.h: (union value) Remove `c' member.
2109 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2111 Make dictionary compacting functions a little more general.
2113 * sys-file-writer.c (sfm_open_writer): Use
2114 dict_compacting_would_change().
2115 (does_dict_need_translation) Removed.
2117 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2119 Make dictionary compacting functions a little more general.
2121 * dictionary.c (dict_needs_compaction): Rename
2122 dict_compacting_would_shrink(). Update all callers.
2123 (dict_compacting_would_change) New function.
2125 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2127 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2128 inverted return value (!).
2130 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2132 Continue reforming procedure execution.
2134 * procedure.c: Search and replace "vfm" by "proc". Notably:
2135 (static var vfm_source) Rename proc_source. Update all
2137 (static var vfm_sink) Rename proc_sink. Update all references.
2139 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2141 Continue reforming procedure execution. In this phase, remove
2142 PROCESS IF, which was deprecated anyway and can be easily
2143 simulated with TEMPORARY followed by SELECT IF.
2145 * procedure.c: (open_active_file) Don't call
2146 add_process_if_trns().
2147 (discard_variables) Get rid of redundant call to
2148 proc_cancel_all_transformations().
2149 (add_process_if_trns) Removed.
2150 (process_if_trns_proc) Removed.
2151 (process_if_trns_free) Removed.
2153 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2155 Continue reforming procedure execution. In this phase, add
2156 `const' to the case passed to procedure()'s callback.
2158 Updated all users of procedure() as well.
2160 * procedure.c: (struct write_case_data) Add "const" to ccase
2161 parameter for case_func member.
2162 (procedure) Add "const" to ccase parameter for proc_func
2164 (multipass_case_func) Make ccase parameter const.
2165 (internal_procedure) Add "const" to ccase parameter for case_func
2167 (split_procedure_case_func) Make ccase parameter const.
2168 (multipass_split_case_func) Make ccase parameter const.
2170 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2172 Continue reforming procedure execution. In this phase, get rid of
2173 the output code for SPLIT FILE groups in procedure.c, which really
2174 shouldn't be doing any output. Move it into the individual
2175 procedures instead. This also adds some flexibility.
2177 Updated many users of procedure_with_splits() and
2178 multipass_procedure_with_splits() to call
2179 output_split_file_values() and to deal with increased use of
2182 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2183 parameter to begin_func member.
2184 (procedure_with_splits) Add "const struct ccase *" parameter to
2185 begin_func parameter. Make ccase parameter const in proc_func
2187 (split_procedure_case_func) Don't dump split file group. Pass
2189 (dump_splits) Moved to language/dictionary/split-file.c as
2190 output_split_file_values().
2191 (struct multipass_split_aux_data) Add "const struct ccase *"
2192 parameter to split_func member.
2193 (multipass_procedure_with_splits) Add "const struct ccase *"
2194 parameter to split_func parameter.
2195 (multipass_split_case_func) Save new SPLIT FILE case before
2197 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2199 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2201 Continue reforming procedure execution. Change
2202 internal_procedure() so that it calls open_active_file() and
2203 close_active_file(), which isolates most of the actual procedure
2206 * procedure.c: (struct write_case_data) Rename `proc_func' member
2207 to `case_func' and update all references.
2208 (procedure) Rewrite as one-line wrapper around
2209 internal_procedure().
2210 (struct multipass_aux_data) New.
2211 (multipass_callback) Renamed multipass_case_func(). Use struct
2212 multipass_aux_data as auxiliary data.
2213 (multipass_end_func) New function.
2214 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2215 that uses multipass_case_func, multipass_end_func.
2216 (internal_procedure) Add `end_func' argument. Move optimization
2217 of trivial case in here. Move call to open_active_file() and
2218 close_active_file() in here. Now assert that vfm_source is
2220 (procedure_with_splits_callback) Rename
2221 split_procedure_case_func().
2222 (split_procedure_end_func) New function.
2223 (multipass_split_callback) Rename multipass_split_case_func.
2224 (multipass_split_end_func) New function.
2225 (discard_variables) No need to test for nonnull vfm_source.
2227 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2229 Continue reforming procedure execution. Get rid of unused member.
2231 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2233 (write_case) Don't increment cases_analyzed.
2235 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2237 Continue reforming procedure execution. In this phase, move
2238 procedure.c and procedure.h from src to src/data. Update
2239 makefiles and #includes accordingly.
2241 * procedure.c: Moved here from src/.
2243 * procedure.h: Moved here from src/.
2245 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2247 Continue reforming procedure execution. In this phase, get rid of
2248 many global variables, consolidating procedure execution in
2249 procedure.c. Encapsulate transformations in new "struct
2250 trns_chain". Also, change implementation of N OF CASES, FILTER,
2251 and PROCESS IF from special cases to transformations.
2253 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2256 * dictionary.c: (global variable default_dict) Move to
2259 * variable.h: (TRNS_*) Move to transformations.h.
2260 (struct transformation) Move to transformations.c.
2262 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2264 * sys-file-reader.c: Fixed invalid read problems.
2266 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2268 * storage-stream.c: Add missing function comments.
2270 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2272 Continue reforming procedure execution. In this phase, add some
2273 new, needed functionality to storage-stream.
2275 * storage-stream.c: (storage_source_decapsulate) New function.
2277 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2279 * variable.c (width_to_bytes): Declarations must precede
2280 statements for C90 compliance.
2282 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2284 * data-out.c, data-in.c, variable.c, variable.h: New functions
2285 copy_mangle and copy_demangle for reading/writing cases; emulates the
2286 way SPSS deals with strings > 255 bytes.
2288 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2289 subtype 14 needed for strings longer than 255 bytes.
2291 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2292 instead of literal values. Also fixed some constness issues.
2294 * format.h: Constness
2296 * sfm-private.h: Renamed the case_size identifier, since I discovered
2297 that SPSS's respect for this variable is very nominal.
2299 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2301 Change case limit type from int to size_t.
2303 * dictionary.c: (struct dictionary) Change type of case_limit
2305 (dict_get_case_limit) Change return type.
2306 (dict_set_case_limit) Change parameter type.
2308 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2310 * variable.h: (struct variable) Rename `reinit' member as `leave'
2311 and invert sense. Fix up all references.
2313 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2315 Continue reforming procedure execution. In this phase, break
2316 procedure.c into multiple files.
2318 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2320 * case-sink.c: New file.
2322 * case-sink.h: New file.
2324 * case-source.c: New file.
2326 * case-source.h: New file.
2328 * storage-stream.c: New file.
2330 * storage-stream.h: New file.
2332 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2334 * variable.h: (struct variable) Remove `init' member and all
2335 references to it from other files. It was initialized in several
2336 places, but nothing really ever used it for anything worthwhile.
2337 Thanks to Jason Stover for pointing out how confusing this
2340 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2342 Continue reforming error message support. In this phase, get rid
2343 of message "titles" and put the message text in `struct error'.
2344 Now `struct error' encapsulates a message more properly.
2346 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2347 Format message ourselves.
2349 * data-in.c: (vdls_error) Ditto.
2351 * por-file-reader.c: (error) Ditto.
2353 * sys-file-reader.c: (corrupt_msg) Ditto.
2355 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2357 GNU standards require "file name" instead of "filename" in
2358 documentation. It's nice for our code to follow the convention
2361 * casefile.c: (struct casefile) Rename `filename' member to
2362 `file_name'. Updated all references.
2364 * file-name.c: [!unix] (struct file_identity) Rename
2365 normalized_filename member to normalized_file_name. Updated all
2368 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2370 We don't really support anything but Unix-like environments well,
2371 so we might as well de-obfuscate by writing directory and path
2372 separators explicitly.
2374 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2376 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2377 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2379 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2382 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2384 GNU standards require "file name" instead of "filename" in
2385 documentation. It's nice for our code to follow the convention
2388 * filename.c: Rename to file-name.c.
2390 * filename.h: Rename to file-name.h. Update all inclusions.
2391 Update header guards.
2393 * automake.mk: Update file names.
2395 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2397 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2399 (fn_basename) Removed (dead code).
2400 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2401 (fn_special_p) Renamed fn_is_special(), all references updated.
2402 (fn_exists_p) Renamed fn_exists(), all references updated.
2404 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2406 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2407 Also, now it only tilde-expands file names, not paths.
2408 (fn_search_path) Tilde-expand one directory at a time.
2410 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2412 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2413 get rid of non-Unixlike version of the code, which has probably
2415 (fn_prepend_dir) Removed (dead code).
2417 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2418 (macro PATH_SEPARATOR_STRING) New.
2419 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2421 Continue reforming error message support. In this phase, we get
2422 rid of VM() and the other msg() support for "verbosity", replacing
2423 it by a new function verbose_msg().
2425 * filename.c: (fn_search_path) Use verbose_msg() instead of
2428 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2430 * sfm-private.h: Get rid of #defines after #error, which makes no
2433 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2435 Get rid of our own int32 type in favor of the standard int32_t
2438 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2441 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2443 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2445 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2447 Remove ill-considered file routines that are no longer used.
2449 * filename.c: (fn_open_ext) Removed.
2450 (fn_close_ext) Removed.
2452 * filename.h: (struct file_ext) Removed.
2454 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2456 * variable.c (var_is_valid_name): Move declarations before code
2459 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2461 * filename.ch (fn_interp_vars): Fixed small buglet.
2463 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2465 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2466 so as to modify the string inline. Thus makeing it easier to
2467 destroy the results when no longer needed.
2469 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2471 * category.c (cat_stored_values_destroy): Fixed memory leak.
2473 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2475 Add some missing frees. Thanks to John Darrington for reporting
2478 * any-writer.c (any_writer_close): Free writer.
2480 * any-reader.c (any_reader_close): Free reader.
2482 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2484 * por-file-reader.c: (error) Mark as NO_RETURN.
2486 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2488 * settings.c: Changed default value of scompress to true.
2490 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2492 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2494 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2497 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2498 here from pref.h.orig.
2500 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2502 * sys-file-reader.c: Fixed bug reading compressed files.
2504 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2506 * Numerous renames. See src/ChangeLog for details.
2508 * Moved files from src directory
projects / pspp-builds.git / blob