1 2008-03-04 Ben Pfaff <blp@gnu.org>
3 Patch #6441. Reviewed by John Darrington.
5 * format.c (fmt_fix): New function.
6 (fmt_fix_input): New function.
7 (fmt_fix_output): New function.
9 * format.def: Correct minimum width for DATETIME format. It was
10 7, should have been 17.
12 2008-02-18 Ben Pfaff <blp@gnu.org>
14 Patch #6426. Reviewed by John Darrington.
16 * format.c (min_width): Renamed fmt_min_width and made public.
17 Updated all references.
18 (max_width): Renamed fmt_max_width and made public. Updated all
20 (max_decimals): Renamed fmt_max_decimals and made public. Updated
22 (var_create): Use the new functions for default variable
24 (var_default_formats): New function.
25 (var_default_measure): New function.
26 (var_default_alignment): New function.
28 * format.h (macro FMT_MAX_NUMERIC_WIDTH): New macro.
30 2008-02-09 Ben Pfaff <blp@gnu.org>
32 Add a couple of extensions to GET DATA TYPE=TXT. Patch #6412.
33 Thanks to John Darrington for review.
35 * data-in.c (data_in): Add new argument to designate the last
36 column of the data field being parsed, for use in error messages.
39 * data-parser (struct data_parser): New member `quote_escape'.
40 (data_parser_create): Initialize quote_escape.
41 (data_parser_set_quotes): New function.
42 (cut_field): Support escaped quotes.
43 (parse_delimited_span): Ditto.
44 (parse_delimited_no_span): Ditto.
46 * get-data.c (parse_get_txt): Support ESCAPE extension subcommand
47 in enhanced mode. Only support multiple quote characters in
50 2008-02-06 John Darrington <john@darrington.wattle.id.au>
52 psql-reader.c psql-reader.h: Read more than one tuple at
53 once. Fix bug reading a query which returns no data. Fix bug
54 when transformation followed a reader.
55 Ask the server for the number of records in the query, for the
58 2008-02-05 John Darrington <john@darrington.wattle.id.au>
60 psql-reader.c: So yesterday they release postgresql 8.3.0
61 which has money represented with 64 bits. They must get
64 2008-02-02 John Darrington <john@darrington.wattle.id.au>
66 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
67 for reviewing this code.
69 2008-02-02 Ben Pfaff <blp@gnu.org>
73 * sys-file-reader.c (read_variable_record): Allow missing values
74 to be specified on long string variables, but warn about them
75 because PSPP does not yet support them.
76 (read_extension_record): Ignore extension records 20 and 21, which
77 PSPP does not yet support.
78 (read_header): Fix error message when floating-point format cannot
81 2008-02-01 Ben Pfaff <blp@gnu.org>
83 Patch #6386. Thanks to John Darrington for review and for the
84 updates to gnumeric-reader.c.
86 * dictionary.c (make_hinted_name): New function.
87 (make_numeric_name): New function.
88 (dict_make_unique_var_name): New function.
90 * gnumeric-reader.c (devise_name): Removed.
91 (munge_name): Removed.
92 (gnumeric_open_reader): Use new function
93 dict_make_unique_var_name.
95 * short-names.c (set_var_short_name_suffix): Use new function
98 2008-01-19 John Darrington <john@darrington.wattle.id.au>
100 * settings.c settings.h: Moved static variables into a
101 single struct. Renamed functions to have a settings_ prefix.
103 2008-01-14 John Darrington <john@darrington.wattle.id.au>
105 * data-out.c (output_AHEX): Corrected number of bytes in
106 call to output_hex. Closes bug #22011
108 2008-01-02 John Darrington <john@darrington.wattle.id.au>
110 * variable.c variable.h: Replaced var_get_value_name with
111 var_append_value_name which doesn't use any static data.
112 Thanks to Ben for review.
114 2007-12-07 Ben Pfaff <blp@gnu.org>
118 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
119 made from an empty casereader produce a casegrouper with no
120 groups, instead of one with one group that has no cases.
122 * casereader.c (casereader_is_empty): New function.
124 2007-12-06 Ben Pfaff <blp@gnu.org>
128 * sys-file-reader.c (read_display_parameters): Handle variable
129 display parameters record with only 2 data items per variable.
130 Reported by Guido Gay <gay@irer.it>.
132 2007-12-04 Ben Pfaff <blp@gnu.org>
134 * identifier.c (lex_id_match_n): New function.
135 (lex_id_match): Reimplement in terms of lex_id_match_n.
137 2007-11-24 Ben Pfaff <blp@gnu.org>
139 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
142 2007-11-24 Ben Pfaff <blp@gnu.org>
144 Fix warning reported when reading back system files that include
145 very long string variables. Thanks to Guido Gay <gay@irer.it> for
148 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
149 was written where `j' was meant.
151 2007-11-18 Ben Pfaff <blp@gnu.org>
153 Properly write variables that include a range of missing values to
154 system files. Thanks to Guido Gay <gay@irer.it> for reporting
157 * sys-file-writer.c (write_variable): Correctly calculate
158 n_missing_values field when writing variables that include a range
161 2007-11-10 Ben Pfaff <blp@gnu.org>
163 Cleanups and bug fixes devised while writing up documentation.
166 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
168 * dict-class.c: New file.
169 (dict_class_from_id): Move here.
170 (dict_class_to_name): Move here.
172 * dict-class.h: New file.
173 (enum dict_class): Move here. Change from consecutive integers to
174 consecutive bits, to make testing for any of multiple values
175 easier. Add new DC_ALL constant.
177 * dictionary.c (struct dictionary): Change `case_limit' from
178 size_t to casenumber.
179 (dict_get_vars): Make final argument an enum dict_class.
180 (dict_get_vars_mutable): Ditto.
181 (dict_get_case_limit): Change return value to casenumber.
182 (dict_set_case_limit): Change final argument to a casenumber.
183 (dict_unset_split_var): Add assertion. Rephrase slightly.
184 (dict_set_label): Use xstrndup to simplify.
186 * format.c (fmt_step_width): AHEX format also needs 2-byte
188 (fmt_set_style): Simplify assertions.
190 * missing-values.c (mv_add_num_range): Rename mv_add_range.
191 Simplify implementation.
192 (mv_has_value): Simplify implementation.
193 (mv_pop_value): Remove the first value, not the last, to avoid
194 having GET followed by SAVE reverse the order of missing values.
195 (mv_peek_value): Rename mv_get_value. Simplify assertion.
196 (mv_has_range): Simplify implementation.
197 (mv_peek_range): Rename mv_get_range.
198 (can_resize_string): Removed.
199 (mv_is_resizable): Use value_is_resizable.
200 (mv_resize): Use value_resize.
202 * short-names.h (SHORT_NAME_LEN): Move here.
204 * val-type.h: New file, for definitions related to type and width
205 of abstract values. Before, these definitions were mixed among
206 those related to "union value" and those related to variables.
207 (macro SYSMIS): Move here.
208 (macro LOWEST): Move here.
209 (macro HIGHEST): Move here.
210 (macro MAX_STRING); Move here.
211 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
212 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
213 All references updated.
214 (val_type_is_valid): New function. Replaces var_type_is_valid.
215 All references updated.
216 (val_type_from_width): New function. Replaces
217 var_type_from_width. All references updated.
219 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
221 (val_labs_can_set_width): Use value_is_resizable.
222 (val_labs_add): Simply return false if the value labels set is too
223 wide, instead of having undefined behavior.
224 (val_labs_replace): Ditto.
225 (val_labs_replace): Ditto.
226 (val_labs_first): Set iterator to null if iteration is complete.
227 (val_labs_first_sorted): Ditto.
228 (val_labs_done): Become a no-op if the iterator is null.
230 * value.c (value_is_resizable): New function.
231 (value_resize): New function.
233 * variable.c (var_get_dict_class): New function.
235 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
238 2007-11-08 Ben Pfaff <blp@gnu.org>
240 * data-in.c: Make formatted data parsing locale-independent.
241 (parse_number): Use c_strtod instead of strtod, to avoid
242 locale-specific behavior.
245 2007-11-06 Ben Pfaff <blp@gnu.org>
247 Patch #6256: add support for binary, 360 file formats. Reviewed
250 * data-in.c (struct data_in): Add `encoding' member.
251 (data_in): Add `encoding' parameter, and re-encode the data passed
252 in where appropriate. Update all callers to pass it in.
253 (parse_A): Implement EBCDIC recoding wart described in manual.
254 (parse_AHEX): Implement EBCDIC recoding.
256 * data-out.c (data_out_legacy): New function.
257 (data_out): Make into a wrapper around data_out_legacy.
259 * file-handle-def.c (struct file_handle): New member `encoding'.
260 (fh_create_file): Set encoding.
261 (fh_default_properties): Set default encoding.
262 (fh_get_legacy_encoding): New function.
264 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
265 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
266 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
267 (struct fh_properties): New member `encoding'.
269 2007-11-05 Ben Pfaff <blp@gnu.org>
271 Patch #6258. Reviewed by John Darrington.
273 * file-handle-def.c (fh_lock): Add comment that TYPE should be
274 marked with N_() in the caller. Added these markings to each
275 caller too. Should make i18n easier.
276 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
278 2007-11-03 Ben Pfaff <blp@gnu.org>
280 Allow output files to overwrite input files (bug #21280).
282 * file-handle-def.c: Separate locking of files for input and for
283 output, to allow both to take place at once. Also, distinguish a
284 file handle from the identity of the underlying file, because the
285 identity of a file changes over time and the file handle can't
286 represent two different identities.
287 (struct file_handle): Remove `next', `open_cnt', `deleted',
288 `type', `open_mode', `aux', `identity' members. Change `id' from
289 char array to char *. Add `ref_cnt' member.
290 (file_handle_from_ll) New function.
291 (file_handles) Removed.
292 (named_handles) New variable.
293 (fh_init) Initialize named_handles.
294 (fh_done) Remove name from all named_handles.
295 (free_handle) Updated for modified struct file_handle.
296 (unname_handle) New function.
297 (fh_ref) New function.
298 (fh_from_file_name) Removed.
299 (fh_unref) New function.
300 (fh_unname) New function.
301 (fh_from_id) Rewritten.
302 (create_handle) Updated for modified struct file_handle.
303 (fh_inline_file) Increment the handle's ref count.
304 (fh_create_file) Updated for modified struct file_handle.
305 (fh_create_scratch) Ditto.
310 (fh_is_open) Removed.
311 (fh_get_id) Updated for modified struct file_handle.
312 (fh_get_default_handle) Increment the handle's ref count.
313 (fh_set_default_handle) Handle ref counts.
314 (struct fh_lock) New structure.
315 (locks) New static var.
316 (fh_lock) New function.
317 (fh_unlock) New function.
318 (fh_lock_get_aux) New function.
319 (fh_lock_set_aux) New function.
320 (fh_is_locked) New function.
321 (make_key) New function.
322 (free_key) New function.
323 (compare_fh_locks) New function.
324 (hash_fh_lock) New function.
326 * file-handle-def.h (enum fh_access) New enum.
328 * file-name.c: Made file_identity the same in all supported
330 (struct file_identity): New `name' member.
331 (fn_get_identity): For a file that doesn't exist, get the
332 dev/inode of its directory plus its name. If even the directory
333 doesn't exist, just use its name. Merge the Windows
334 implementation into the Unix one.
335 (fn_compare_file_identities): Rewritten. Merge the Windows
336 implementation into the Unix one.
337 (fn_hash_identity): New function.
339 * make-file.c (struct replace_file): New structure.
340 (all_files): New static var.
341 (replace_file_start): New function.
342 (replace_file_commit): New function.
343 (replace_file_abort): New function.
344 (free_replace_file): New function.
345 (unlink_replace_files): New function.
347 * por-file-reader.c (struct pfm_reader): Add `lock' member.
348 (close_reader): Unlock file.
349 (pfm_open_reader): Lock file.
351 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
353 (pfm_open_writer): Lock file and prepare for its replacement.
354 (close_writer): Unlock file.
356 * scratch-handle.h (struct scratch_handle): Add unique_id so that
357 different generations of a scratch file can be distinguished.
359 * scratch-reader.c (scratch_reader_open): Verify that the file is
362 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
363 remove scratch_handle member.
364 (scratch_writer_open): Lock handle. Prepare to replace handle
365 data, instead of doing it immediately.
366 (scratch_writer_casewriter_destroy): Replace handle data and
369 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
370 (sfm_open_reader): Lock file.
371 (close_reader): Unlock file.
373 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
375 (sfm_open_writer): Lock file and prepare for its replacement.
376 (close_writer): Unlock file.
378 2007-11-02 Ben Pfaff <blp@gnu.org>
380 * data-out.c (output_number): Use isfinite (from C99) instead of
381 gsl_isfinite, since we now have universal support for it in
383 (output_infinite): Ditto.
384 [!HAVE_ROUND] (round): Remove definition, since we now have a
385 replacement in gnulib.
387 2007-10-27 John Darrington <john@darrington.wattle.id.au>
389 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
390 wasn't initialised correctly. Closes bugs #21330 and 21397
393 2007-10-08 Ben Pfaff <blp@gnu.org>
395 Fix bug #21280. Thanks to John Darrington for review.
397 * file-name.c (create_stream): New function.
399 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
400 file handle before creating the file, to ensure that we don't
401 truncate a file that we're reading. Make code easier to read by
404 * sys-file-write.c (sfm_open_writer): Ditto.
406 2007-10-01 Ben Pfaff <blp@gnu.org>
408 Fix bug #21192. Thanks to John Darrington for review.
410 * casereader.c (casereader_read): Decrement case_cnt before
411 calling the casereader's "read" member function, so that we
412 interact properly with lazy_casereader.
414 * datasheet.c: Add regression test for above bug fix.
415 (clone_datasheet): New function.
416 (lazy_callback): New function.
417 (check_datasheet_casereader): New function.
418 (check_datasheet): Check datasheet contents are reported correctly
419 through an ordinary casereader and a lazy casereader.
420 (clone_model): Use clone_datasheet.
422 2007-09-24 Ben Pfaff <blp@gnu.org>
424 Patch #6210. Reviewed by John Darrington.
425 * settings.c: Drop "long view", which is not needed any longer.
426 (static var long_view): Removed.
427 (force_long_view): Removed.
428 (get_viewwidth): Removed.
429 (init_viewport): Removed long_view reference.
431 2007-09-19 John Darrington <john@darrington.wattle.id.au>
433 * settings.c settings.h: Changed viewport's length and width to be
434 owned by the user interface which uses the data library. This allows
435 better abstraction, and makes dynamically adjustable dimensions easier.
437 2007-09-18 Ben Pfaff <blp@gnu.org>
439 * procedure.c (proc_extract_active_file_data): New function.
441 * lazy-casereader.h: New file.
443 * lazy-casereader.c: New file.
445 * casereader.c (casereader_dynamic_cast): New function.
447 2007-09-14 Ben Pfaff <blp@gnu.org>
449 * dictionary.c (dict_clone): Copy case indexes from cloned
450 dictionary. Fixes bug #21061. Reviewed by John Darrington.
452 2007-09-13 John Darrington <john@darrington.wattle.id.au>
454 * value.c value.h (create_value): New function.
456 2007-09-12 Ben Pfaff <blp@gnu.org>
458 Make it clear that translator casereader and casewriter translate
459 functions are supposed to destroy their input case.
461 * casereader-translator.c (struct casereader_translator): Change
462 input case parameter of translate member function from const
463 struct ccase * to struct ccase *.
464 (casereader_create_translator): Ditto, for translate parameter.
466 * casewriter-translator.c (struct casewriter_translator): Ditto.
467 (casewriter_create_translator): Ditto.
469 2007-08-27 John Darrington <john@darrington.wattle.id.au>
471 * sys-file-reader.c (read_display_parameters): Force display width
472 to 8, if the sys file says 0 (like SPSS does).
474 2007-08-12 Ben Pfaff <blp@gnu.org>
476 * dictionary.c (dict_dump): New function.
478 2007-08-12 Ben Pfaff <blp@gnu.org>
480 Drop dict_compactor in favor of using the new struct case_map.
482 * dictionary.c (struct copy_map): Removed.
483 (struct dict_compactor): Removed.
484 (dict_make_compactor): Removed.
485 (dict_compactor_compact): Removed.
486 (dict_compactor_destroy): Removed.
488 * procedure.c (struct data_set): Change `compactor' member to be a
490 (proc_open): Use case_map_to_compact_dict instead of
492 (proc_casereader_read): Use case_map_execute instead of
493 dict_compactor_compact.
494 (proc_commit): Use case_map_destroy instead of
495 dict_compactor_destroy.
497 * scratch-writer.c (struct scratch_writer): Change `compactor'
498 member to be a struct case_map *.
499 (scratch_writer_open): Use case_map_to_compact_dict instead of
501 (scratch_writer_casewriter_write): Use case_map_execute instead of
502 dict_compactor_compact.
504 2007-08-12 Ben Pfaff <blp@gnu.org>
506 * automake.mk: Add case-map.c, case-map.h.
508 * case-map.c: New file.
510 * case-map.h: New file.
512 2007-08-12 Ben Pfaff <blp@gnu.org>
514 * dictionary.c (dict_compact_values): Don't delete scratch
515 variables as well as compacting case indexes. Update all callers.
516 (dict_get_compacted_value_cnt): Rename dict_count_values and
517 change interface. Update all callers.
518 (dict_get_compacted_value_cnt): Remove.
519 (dict_compacting_would_shrink): Remove.
520 (dict_compacting_would_change): Remove.
521 (dict_make_compactor): Add new parameter. Update all callers.
523 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
524 instead of dict_count_values, changing an O(N) operation into
527 2007-08-12 Ben Pfaff <blp@gnu.org>
529 * casereader.c (casereader_read): Don't require cases read by a
530 casereader to be exactly the expected size: as long as they're big
533 2007-08-12 Ben Pfaff <blp@gnu.org>
535 Make casewriters keep track of the number of `union value's in
536 each case. This is useful for two reasons: casewriter_write can
537 then check that the case being written is large enough, and later
538 recipients of the casewriter can determine the size of the case.
540 * casewriter-translator.c (casewriter_create_translator): Add
543 * casewriter.c (struct casewriter): Add value_cnt member.
544 (casewriter_write): Check that the case passed in is big enough.
545 (casewriter_get_value_cnt): New function.
546 (casewriter_create): Add value_cnt parameter.
548 2007-08-09 Ben Pfaff <blp@gnu.org>
550 Fix bug reported by Jason Stover.
551 * settings.c: Move get_termcap_viewport and in particular the
552 #include for <curses.h> to the end of the file. curses.h
553 redefines bool on some systems (e.g. OpenBSD), which causes
554 disagreement between uses of bool before its inclusion and
556 Tested by Jason Stover.
558 2007-07-29 Ben Pfaff <blp@gnu.org>
560 Provisional fix for bug #18692 and bug #20161. Reviewed by John
563 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
564 argument (never "rb" or "wb") because SUSv3 says that only those
565 modes are defined, and glibc in fact rejects other modes.
567 Open portable files with fn_open so that they can be read from
568 pipes. Fix missing fh_close call to go along with fh_open.
569 Report an error if the file close reports an error.
570 * por-file-reader.c (close_reader): New function.
571 (por_file_casereader_destroy): Use close_reader.
572 (pfm_open_reader): Open file with fn_open.
574 2007-07-28 Ben Pfaff <blp@gnu.org>
576 Make PSPP able to read all the portable files I could find on the
577 web. Thanks to John Darrington for review. Bug #17620.
578 * por-file-reader.c (struct pfm_reader): New member `line_length'.
579 (error): Print file offset in hexadecimal.
580 (warning): New function.
581 (advance): Treat lines less than 80 bytes long as padded to 80
583 (pfm_open_reader): Call read_documents if we find an "E" record.
584 (convert_format): Convert invalid formats to the default format
585 instead of aborting reading the file.
586 (read_variables): Rename duplicate variable names instead of
587 aborting reading the file.
588 (read_value_label): Allow string variables of different widths to
589 be assigned value labels in the same record. Replace duplicate
590 value labels instead of aborting.
591 (read_documents): New function.
593 * por-file-writer.c (pfm_open_writer): Call write_documents if the
594 dictionary has documents.
595 (write_documents): New function.
597 2007-07-25 Ben Pfaff <blp@gnu.org>
599 Fix bugs related to bug #17213.
601 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
602 since the former is what config.h has. Include the needed ncurses
604 (static var echo) Rename to `do_echo' because the original name is
605 the same as an ncurses identifier.
606 (get_termcap_viewport) Use error instead of msg.
608 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
609 (fn_close): Don't close stdin, stdout, stderr.
611 2007-07-26 John Darrington <john@darrington.wattle.id.au>
613 * procedure.c procedure.h: Added callbacks which get invoked whenever
614 a dataset's transformation chain changes.
616 2007-07-24 Ben Pfaff <blp@gnu.org>
619 * sys-file-writer.c (write_variable_display_parameters): Use new
620 var_default_display_width function to choose display width of
621 segments after the first one in a given variable.
622 * variable.c (var_create): Use var_default_display_width to pick
623 new variable's display width.
624 (var_default_display_width): New function.
625 Reviewed by John Darrington.
627 2007-07-24 Ben Pfaff <blp@gnu.org>
630 * por-file-writer.c (write_variables): Write weight variable.
631 Reviewed by John Darrington.
633 2007-07-23 Ben Pfaff <blp@gnu.org>
635 Improvements to system file reader and writer.
637 First, move all detailed knowledge of very long strings into
638 sys-file-private.[ch], so that this nasty stuff can be isolated.
640 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
641 (EFFECTIVE_VLS_CHUNK): New macro.
642 (min_int): New function.
643 (max_int): New function.
644 (sfm_width_to_bytes): Rewrite.
645 (sfm_width_to_octs): New function.
646 (sfm_segment_alloc_width): New function.
647 (sfm_segment_alloc_bytes): New function.
648 (sfm_segment_used_bytes): New function.
649 (sfm_segment_offset): New function.
650 (sfm_segment_effective_offset): New function.
651 (sfm_dictionary_to_sfm_vars): New function.
653 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
654 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
655 (struct sfm_var): New structure.
657 Next, improvements to the system file reader.
659 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
660 type casenumber. Added `version_major', `version_minor',
663 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
664 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
665 Change `case_cnt' to type casenumber. Removed `has_vls'.
666 (struct sfm_var): Removed.
667 (sfm_open_reader): Don't warn on wrong case size if the file was
668 written by SPSS 13, which tends to get it wrong. Use
669 sfm_dictionary_to_sfm_vars.
670 (read_header): Always output system file info.
671 (read_variable_record): Simplify code for reading missing values.
672 (read_machine_int32_info): Save version numbers from system file
673 into info struct passed as new argument.
674 (read_long_string_map): Restructured to use new sys-file-private
676 (read_value_labels): Use size_overflow_p.
677 (sys_file_casereader_read): Get rid of distinction between fast
678 and slow paths. Use information provided by sys-file-primate's
679 struct sfm_var to simplify code.
680 (skip_whole_strings): New function.
681 (read_int32): Renamed read_int. Changed return value to int.
683 (read_flt64): Renamed read_float. Changed return value to
684 double. Updated all callers.
685 (int32_to_native): Removed. Changed callers to use
687 (flt64_to_double): Removed. Changed callers to use float_convert.
689 Finally, get rid of int32, flt64 terminology and types in system
690 file writer. The former wasn't very useful since a POSIX "int"
691 can hold the whole range of int32 and we generally didn't have a
692 need for it to be exactly-32-bits, just at-least-32-bits. The
693 latter was inconvenient because we had to assume that it could be
694 different from double and thereby convert special values SYSMIS,
695 HIGHEST, LOWEST to and from it in multiple places. Instead, now
696 we just use "int" and "double" in most places, and do conversions,
697 if necessary, very close to where we do I/O. This change meant
698 that the writer code couldn't represent records in the file as C
699 structs any longer, but that's no great loss. The code actually
700 seems to be more readable without them.
702 Simplify the compression buffering code: only buffer as much as
703 necessary, which is no more than eight 8-byte units at any given
706 * sys-file-writer.c (typedef flt64): Removed.
707 (macro second_lowest_flt64): Removed.
708 (struct sysfile_header): Removed.
709 (struct sysfile_variable): Removed.
710 (struct sfm_writer): Removed `needs_translation', `has_vls',
711 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
712 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
713 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
714 compression buffering by `opcodes', `opcode_cnt', `data',
715 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
716 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
717 data writer functions instead of structures.
718 (calc_oct_idx): New function.
719 (write_header): Use simple data writer functions instead of
721 (write_format_spec): Renamed write_format. New argument.
722 (write_variable_continuation_records): New function.
723 (write_variable): Use simple data writer functions instead of
724 structures. Use write_variable_continuation_records. Write
725 entire very long string instead of requiring caller to understand
727 (write_value_labels): Use simple data writer functions instead of
729 (write_documents): Ditto.
730 (write_variable_display_parameters): Use sys-file-private
731 functions to simplify. Use simple data writer functions instead
733 (write_vls_length_table): Use simple data writer functions instead
735 (write_longvar_table): Ditto.
736 (write_rec_7_34): Break into new functions
737 write_integer_info_record, write_float_info_record. Use simple
738 data writer functions instead of structures.
739 (buf_write): Removed.
740 (append_string_max): Removed.
741 (ensure_buf_space): Removed.
742 (sys_file_casewriter_write): Get rid of the distinction between
743 fast and slow paths, which didn't seem to be too useful. Use new
744 functions write_case_uncompressed, write_case_compressed.
745 (put_instruction): Removed.
746 (put_element): Removed.
747 (write_compressed_data): Removed.
748 (close_writer): Use flush_compressed. Only write case count to
749 system file if it will fit in the field.
750 (write_case_compressed): New function.
751 (write_case_uncompressed): New function.
752 (flush_compressed): New function.
753 (put_cmp_opcode): New function.
754 (put_cmp_number): New function.
755 (write_int): New function.
756 (convert_double_to_output_format): New function.
757 (write_float): New function.
758 (write_value): New function.
759 (write_string): New function.
760 (write_bytes): New function.
761 (write_zeros): New function.
762 (write_spaces): New function.
764 Reviewed by John Darrington.
766 2007-07-22 Ben Pfaff <blp@gnu.org>
768 Don't try to write very long strings to portable files. The
769 format does not support it.
771 * por-file-writer.c (MAX_POR_WIDTH): New macro.
772 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
773 (write_format): Add arg to take width to resize format to.
774 (write_value): Limit width of value written to MAX_POR_WIDTH.
775 (write_variables): Limit width of variable and its output formats
777 Reviewed by John Darrington.
779 2007-07-22 Ben Pfaff <blp@gnu.org>
781 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
782 local variable instead of literal 5.
783 Reviewed by John Darrington.
785 2007-07-22 Ben Pfaff <blp@gnu.org>
787 Fix problems with uniqueness of short names in system files with
788 very long string variables. Now a variable may have multiple
791 * automake.mk (src_data_libdata_a_SOURCES): Add new files
792 short-names.c, short-names.h.
794 * dictionary.c (dict_clone): Clone all the short names.
795 (compare_strings): Move into short-names.c.
796 (hash_strings): Ditto.
797 (set_var_short_name_suffix): Ditto.
798 (dict_assign_short_names): Ditto, rename short_names_assign,
799 change to assign all short names.
801 * por-file-writer.c (write_variables): Use short_names_assign
802 instead of dict_assign_short_names.
804 * short-names.c: New file.
806 * short-names.h: New file.
808 * sys-file-private.c (sfm_width_to_segments): New function.
810 * sys-file-reader.c (read_long_var_name_map): Save and restore all
811 the short names, not just the first one.
813 * sys-file-writer.c (cont_var_name): Removed.
814 (sfm_open_writer): Use short_names_assign instead of
815 dict_assign_short_names. Use unique short names assigned by
816 short_names_assign instead of those generated by cont_var_name.
818 * variable.c (struct variable): Remove `short_name' member,
819 replace by `short_names' and `short_name_cnt'.
820 (var_create) Initialize new members.
821 (var_get_short_name_cnt): New function.
822 (var_get_short_name): Now takes an index argument. Changed most
824 (var_set_short_name): Ditto.
825 (var_clear_short_name): Renamed var_clear_short_names, changed to
826 clear all short names.
828 Reviewed by John Darrington.
830 2007-07-22 Ben Pfaff <blp@gnu.org>
832 * variable.c (var_set_width): Use new var_set_width function.
834 * missing-values.c (mv_n_values): Drop assertion, which was not
837 * format.c (fmt_default_for_width): New function.
838 (fmt_resize): New function.
840 Reviewed by John Darrington.
842 2007-07-18 John Darrington <john@darrington.wattle.id.au>
844 * datasheet.c (datasheet_delete_columns): Added assertion to check
845 we're not deleting outside the range of the sheet.
848 * dictionary.c dictionary.h variable.c: Added the ability for string
849 variables to be resized.
851 * vardict.h: Added some prototypes (moved from dictionary.h) as
852 these should only be called by variable.c
855 2007-07-14 John Darrington <john@darrington.wattle.id.au>
857 * sfm-reader.c: Respect case_cnt field in file header.
859 2007-07-01 John Darrington <john@darrington.wattle.id.au>
861 * transformation.c transformation.h (trns_chain_execute): Changed the
862 signature (Patch #6057)
864 2007-06-10 Ben Pfaff <blp@gnu.org>
866 * casereader-filter.c (casereader_filter_destroy): Make sure to
867 write all the remaining excluded cases to the casewriter, if any.
869 * caseinit.c (init_list_destroy): Rewrite.
870 (init_list_clear): Ditto.
872 * casegrouper.c (casegrouper_get_next_group): Always set *reader
873 to null when returning false.
875 2007-06-06 Ben Pfaff <blp@gnu.org>
877 Actually implement the new procedure code and adapt all of its
878 clients to match. Also adapt all of the other case sources and
879 sinks in the tree and their clients to use the
880 casereader/casewriter infrastructure.
882 * automake.mk: Add and remove files.
884 * any-reader.c: Change into a casereader.
885 * por-file-reader.c: Ditto.
886 * scratch-reader.c: Ditto.
887 * sys-file-reader.c: Ditto.
889 * any-writer.c: Change into a casewriter.
890 * por-file-writer.c: Ditto.
891 * scratch-writer.c: Ditto.
892 * sys-file-writer.c: Ditto.
894 * procedure.c: Change to use casereader, casewriter, caseinit, and
895 other new infrastructure.
897 * scratch-handle.c: Adapt to new infrastructure.
899 * case-sink.c: Removed, now dead code.
900 * case-sink.h: Ditto.
901 * case-source.c: Ditto.
902 * case-source.h: Ditto.
903 * casefile-factory.c: Ditto.
904 * casefile-private.h: Ditto.
907 * casefilter.c: Ditto.
908 * casefilter.h: Ditto.
911 * fastfile-factory.c: Ditto.
912 * fastfile-factory.h: Ditto.
913 * storage-stream.c: Ditto.
914 * storage-stream.h: Ditto.
916 2007-06-06 Ben Pfaff <blp@gnu.org>
920 * automake.mk: Add new files.
922 * datasheet.c: New file.
924 * datasheet.h: New file.
926 2007-06-06 Ben Pfaff <blp@gnu.org>
928 Until now, the procedure code has provided a case to the
929 case_source, which has filled in the data values that come from
930 the active file. "Left" data values that don't come from the
931 active file naturally stay the same from case to case, because the
932 procedure code keeps using that same case.
934 One of the compromises that comes with the new procedure code is
935 that the active file allocates and provides its own case, which
936 the procedure code then has to resize to provide room for any
937 other variables that should go in the case and then fill in the
938 values of "left" variables. Then, when we're done with that case,
939 we have to save the values of "left" variables to copy into the
940 next case read from the active file.
942 The caseinit code helps with this.
944 * automake.mk: Add new files.
946 * caseinit.c: New file.
948 * caseinit.h: New file.
950 2007-06-06 Ben Pfaff <blp@gnu.org>
952 * value.h (value_cnt_from_width): New function.
954 * variable.c (var_get_value_cnt): Use new function.
956 2007-06-06 Ben Pfaff <blp@gnu.org>
958 Add casegrouper, to allow cases read from a given casereader to be
959 broken into groups, each of which has its own casereader.
960 Generally cases are grouped based on having equal values for some
963 * automake.mk: Add new files.
965 * casegrouper.c: New file.
967 * casegrouper.h: New file.
969 2007-06-06 Ben Pfaff <blp@gnu.org>
971 Add interface to lexicographical ordering of cases.
973 * automake.mk: Add new files.
975 * case-ordering.c: New file.
977 * case-ordering.h: New file.
979 2007-06-06 Ben Pfaff <blp@gnu.org>
981 Add casereaders and casewriters, the basis of the new data processing
982 implementation. A casereader is a uniform interface to reading cases
983 from a data source; a casewriter is a uniform interface to writing
984 cases to a data sink.
986 * automake.mk: Add new files.
988 * casereader-filter.c: New file.
990 * casereader-provider.h: New file.
992 * casereader-translator.c: New file.
994 * casereader.c: New file.
996 * casereader.h: New file.
998 * casewriter-provider.h: New file.
1000 * casewriter-translator.c: New file.
1002 * casewriter.c: New file.
1004 * casewriter.h: New file.
1006 2007-06-06 Ben Pfaff <blp@gnu.org>
1008 "casewindow" data structure that extends the deque (from libpspp)
1009 of cases with the ability to dump cases to disk if we get too many
1012 * automake.mk: Add new files.
1014 * casewindow.c: New file.
1016 * casewindow.h: New file.
1018 2007-06-06 Ben Pfaff <blp@gnu.org>
1020 sparse_cases data structure that augments a sparse_array of cases
1021 with the ability to dump cases to disk if we get too many cases in
1024 * automake.mk: Add new files.
1026 * sparse-cases.c: New file.
1028 * sparse-cases.h: New file.
1030 2007-06-06 Ben Pfaff <blp@gnu.org>
1032 Adds a low-level on-disk case array data structure.
1034 * automake.mk: Add new files.
1036 * case-tmpfile.c: New file.
1038 * case-tmpfile.h: New file.
1040 2007-06-06 Ben Pfaff <blp@gnu.org>
1042 In a couple of places we calculate the maximum number of cases to
1043 keep in memory based on the user-defined workspace. Enable
1044 centralizing the calculation through a new function.
1046 * settings.c (get_workspace_cases): New function.
1048 2007-06-06 Ben Pfaff <blp@gnu.org>
1050 The casenumber type is defined in transformations.h, but case.h is
1051 a more sensible place. Move it.
1053 * case.h (CASENUMBER_MAX): New macro.
1054 (typedef casenumber): Move here, from transformations.h.
1056 2007-06-03 Ben Pfaff <blp@gnu.org>
1058 Slightly generalize case_to_values and case_from_values functions.
1060 * case.c (case_to_values): Rename case_copy_out, change interface.
1061 (case_from_values): Rename case_copy_in, change interface.
1063 * fastfile.c (fastfilereader_get_next_case): Update caller.
1064 (write_case_to_disk): Ditto.
1066 2007-06-02 Ben Pfaff <blp@gnu.org>
1068 Clean up after a forgotten part of patch #5829.
1070 * casedeque.h: Remove unused file.
1072 * automake.mk: Remove casedeque.h from sources.
1074 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1076 * category.c: Removed redundant #include
1078 2007-05-06 Ben Pfaff <blp@gnu.org>
1080 Abstract the documents within a dictionary a little better.
1081 Thanks to John Darrington for suggestion, initial version, and
1082 review. Patch #5917.
1084 * dictionary.c (struct dictionary): Change `documents' member from
1085 char * to struct string.
1086 (dict_clear): Destroy struct string.
1087 (dict_get_documents): Convert struct string to char *.
1088 (dict_set_documents): Set struct string. Pad to 80-character
1090 (dict_clear_documents): New function.
1091 (dict_add_document_line): New function.
1092 (dict_get_document_line_cnt): New function.
1093 (dict_get_document_line): New function.
1095 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1097 * sys-file-reader.c (read_documents): Use new document functions.
1099 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1101 * sys-file-reader.c: When reading a system file which has no
1102 long name table, automatically create one where the long names
1103 are the lower case versions of the short names.
1105 2007-04-22 Ben Pfaff <blp@gnu.org>
1107 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1108 dict_clear will free most data related to the dictionary.
1109 dict_clear does a decent job, except that dict_set_split_vars on
1110 some systems won't actually free the dict's "split" member.
1111 Instead, it'll allocate a 1-byte region. Fix this.
1113 * value.c (value_copy): New function.
1114 (value_set_missing): Ditto.
1116 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1118 * Deleted existing category.h and moved cat-routines.h into
1119 category.h Encapsulated struct cat_vals better.
1121 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1123 * sys-file-reader.c: When reading a system file which has no
1124 long name table, automatically create one where the long names
1125 are the lower case versions of the short names.
1127 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1129 * sys-file-reader.c: Some versions of Other Software seem to
1130 produce system files with string variables' measure set to
1131 zero. We'll assume these are supposed to be nominal variables.
1133 2007-03-30 Ben Pfaff <blp@gnu.org>
1135 * procedure.c: Adapt to new deque data structure.
1137 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1138 Ben Pfaff <blp@gnu.org>
1140 * file-name.c: Mingw compatibility fixes.
1141 (fn_search_path): Use ISSLASH instead of comparing against '/'
1143 (fn_dir_name): Use dir_name from gnulib.
1144 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1145 (fn_get_identity): Use GetFullPathName instead of canonicalize
1146 from gnulib, because the latter does not fully support
1147 Windows-style path names. Use this implementation based on the
1148 detected presence of Windows instead of the absence of Unix, since
1149 the new implementation is Windows-specific.
1150 (fn_compare_file_identities): In Windows implementation, compare
1151 names case-insensitively.
1153 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1155 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1156 <stdlib.h> to have the same effect.
1158 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1160 * por-file-reader.c: Add missing _() around messages.
1162 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1164 * make-file.c: Include "mkstemp.h", without which linking on
1167 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1169 Reduce platform dependence.
1171 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1172 Everywhere we using this, we really should have just depended on
1173 the shell to expand tildes.
1174 (fn_search_path): Simplify, given that we don't do tilde expansion
1176 (fn_normalize): Removed. Caller changed to use the canonicalize
1178 (fn_get_cwd): Removed. Only user was fn_normalize.
1179 (fn_is_absolute): Really only test for absolute names.
1180 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1182 (fn_readlink): Removed, as it was only used fn_normalize.
1183 (fn_exists): Assume the stat function is available; gnulib does.
1184 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1186 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1188 * dictionary.c (dict_create_vector_assert): New function.
1190 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1192 * file-name.c (fn_normalize): Correct name of function
1193 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1194 for pointing this out.
1196 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1198 We are using a single member in struct file_handle, the "name"
1199 field, for more than one purpose. When it begins with '"', it's a
1200 file name; otherwise, it's a token that can be used to identify
1201 it. When that assertion fires, it's because we searched for the
1202 name case-sensitively as a file name (so that there was no match),
1203 and then we try to insert it case-insensitively as a token, which
1204 fails because duplicates aren't allowed.
1206 Solution: break the two purposes into two separate fields. This
1207 fixes the problem and likely makes the code easier to read too.
1209 Fixes bug #18922. Thanks to John Darrington for bug report and
1212 * file-handle-def.c (struct file_handle): New `id' member.
1213 (fh_from_name): Rename fh_from_id. Update all callers.
1214 (create_handle): New `id' parameter. Update all callers.
1215 (fh_create_file): Ditto.
1216 (fh_get_id): New function.
1218 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1220 * case.c (case_is_null): Change return type to bool.
1222 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1226 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1227 reference counting to share data. This makes it easy for
1228 valgrind, etc. to find accesses to cases that have been destroyed
1229 but have been kept around by another user's ref-count. This often
1230 happens when the data set is small enough to find in memory; if a
1231 bigger data set that would overflow to disk were used, then data
1232 corruption would occur.
1234 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1238 * case.c (case_unshare): Make it check internally whether the
1239 ref_cnt is greater than 1, so that the callers don't have to.
1240 Update callers not to check.
1242 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1244 Before, I was thinking that I might want to get rid of reference
1245 counting at some point. Now, I'm pretty sure that it's here to
1246 stay. Thus, because we have to store the value_cnt anyway for
1247 reference-counted cases, we might as well expose it to users.
1249 * case.c (case_get_value_cnt): New function.
1250 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1251 resize case if its size actually changed.
1253 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1254 instead of peeking inside struct case directly.
1255 (casefile_append): Ditto.
1257 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1259 Get rid of the inlines for the case functions, which made the
1260 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1261 the inlines didn't speed up "make check" at all, which is not a
1262 perfect benchmark but seems indicative.)
1264 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1265 definitions. Remove some assertions on nonnull pointers that were
1266 redundant with a pointer dereference soon after in the function.
1268 (struct case_data): Move definition here from case.h.
1272 (case_data_wr): Ditto.
1274 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1276 * automake.mk: Add casedeque.h to sources.
1278 * casedeque.h: New file.
1280 * procedure.c: (struct dataset) Change lag_count, lag_head,
1281 lag_queue member into single struct casedeque member. Update all
1282 users to use the casedeque instead.
1285 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1287 * procedure.c: Simplify lagged cases interface. Updated all
1288 clients--well, the only client--to use the simplified interface.
1289 (dataset_n_lag) Removed.
1290 (dataset_set_n_lag) Removed.
1291 (dataset_need_lag) New function.
1293 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1295 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1296 whenever appropriate, but only when the dataset/dictionary is in a
1299 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1301 * dictionary.c dictionary.h : Added callbacks for change of filter and
1302 split variables. Refactored some code to ensure that callbacks get
1303 invoked when appropriate.
1305 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1306 replace_dict callback occurs when permanent_dict replaces the current
1309 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1311 * dictionary.c dictionary.h : Added callback for when the weight
1312 variable of a dictionary changes.
1314 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1316 * dictionary.c dictionary.h : Added replace_source and replace_dict
1317 callbacks, and functions to deal with them.
1319 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1321 Simplify missing value handling.
1323 * missing-values.h (enum mv_class): New type.
1324 (enum mv_type): Moved definition into missing-values.c and renamed
1325 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1326 enums. Updated all uses.
1327 (struct missing_values): Changed type of `type' from `enum
1328 mv_type' to `int' because the definition is no longer exposed.
1330 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1331 parameter. Update all callers.
1332 (mv_is_num_missing): Ditto.
1333 (mv_is_str_missing): Ditto.
1334 (mv_is_value_user_missing): Removed. Changed callers to use
1335 mv_is_value_missing.
1336 (mv_is_num_user_missing): Removed. Changed callers to use
1338 (mv_is_str_user_missing): Removed. Changed callers to use
1340 (mv_is_value_system_missing): Removed. Changed callers to use
1341 mv_is_value_missing.
1342 (mv_set_type): Removed. Changed callers to use mv_clear.
1343 (mv_clear): New function.
1345 * variable.c (var_is_value_missing): Add new enum mv_class
1346 parameter. Update all callers.
1347 (var_is_num_missing): Ditto.
1348 (var_is_str_missing): Ditto.
1349 (var_is_value_user_missing): Removed. Changed callers to use
1350 var_is_value_missing.
1351 (var_is_num_user_missing): Removed. Changed callers to use
1353 (var_is_str_user_missing): Removed. Changed callers to use
1355 (var_is_value_system_missing): Removed. Changed callers to use
1356 var_is_value_missing.
1358 * casefilter.c (struct casefilter): Use enum mv_class in place of
1360 (casefilter_variable_missing): Adapt to new member.
1361 (casefilter_create): Change signature to take enum mv_class,
1364 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1366 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1368 * case-sink.c case-sink.h procedure.c procedure.h
1369 storage-stream.c: Now uses the factory.
1371 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1373 Make it possible to pull cases from the active file with a
1374 function call, instead of requiring indirection through a callback
1377 * case-source.h (struct case_source_class): Change ->read function
1378 to return a single case, instead of calling a callback function
1379 for each case. Change ->destroy function to return an error
1382 * case-source.c (free_case_source): Pass along the value returned
1383 by the case_source ->destroy function.
1385 * procedure.c (struct write_case_data): Removed.
1386 (struct dataset): Added some members to track procedure state.
1387 (procedure): Optimize the trivial case at this level.
1388 (internal_procedure): Re-implement in terms of proc_open,
1389 proc_read, proc_close.
1390 (proc_open) New function.
1391 (proc_read) New function.
1392 (proc_close) New function.
1393 (write_case) Moved into proc_read.
1394 (close_active_file) Moved closing of data source into proc_close.
1396 * storage-source.c: Rewrote to conform with modified
1397 case_source_class interface.
1399 * transformations.c (trns_chain_execute): Added argument to allow
1400 starting execution from an arbitrary transformation. Updated
1403 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1405 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1407 * sys-file-reader.c (read_display_parameters): Don't assume that
1408 MEASURE_* and ALIGN_* have the same values found in system files.
1410 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1412 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1413 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1414 the value of n_MEASURES, which was off by 1 (at least from my
1417 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1419 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1420 callbacks which are invoked when the dictionary or its
1421 variables are changed.
1423 * missing-values.c missing-values.h value-labels.c: Tidied up
1424 consistency checks, and made some of them return false
1425 instead of assert-failing.
1427 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1429 * calendar.c (calendar_days_in_month): New function.
1431 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1433 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1434 the value label's width.
1436 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1438 * sfm-private.h: Move contents into sys-file-writer.c, which is
1439 the only remaining user. Removed Borland C++-specific directives.
1441 * sys-file-reader.c: Clean up and rewrite entire file. The
1442 rewritten version is simpler and better abstracted, and should be
1443 easier to maintain and extend. It avoids using structures to read
1444 file data, which is prone to padding variations among compilers.
1445 It should also handle non-IEEE 754 system files, although I
1446 haven't been able to find any. It has been tested against many
1447 .sav files obtained from the Web and found to produce the same
1448 results as the earlier version of the code, or in some cases
1449 improved results. It is more tolerant of format variations found
1452 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1453 member, putting an enum integer_format in its place. New member
1454 `float_format'. Changed `compressed' member to type bool.
1456 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1458 * dictionary.c (dict_delete_consecutive_vars): New function.
1460 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1462 * file-name.c (fn_search_path): Remove prefix arg that was unused
1463 by any caller. Updated all callers.
1465 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1467 * format.c (fmt_dollar_template): Use user's decimal point
1468 character. Add assertion.
1470 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1472 * format.c (fmt_dollar_template): New function, based on
1473 dollar_format_template from var-type-dialog.c.
1475 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1477 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1480 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1482 Finish converting struct variable to an opaque type. In this
1483 phase, we add remaining setter and getter functions, convert the
1484 remaining PSPP code to use them, and do a bunch of cleanup. The
1485 resulting changes are pervasive but mostly trivial, and only the
1486 notable changes are logged.
1488 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1491 * case.c (case_data): Renamed case_data_idx.
1492 (case_num): Renamed case_num_idx.
1493 (case_str): Renamed case_str_idx.
1494 (case_data_rw): Renamed case_data_rw_idx.
1496 * case.h (case_data): New function with old name and an interface
1497 that takes a variable instead of an index, which is easier to
1498 use. Updated all callers to use the new interface, or to use the
1499 new *_idx function (see above).
1502 (case_data_rw): Ditto.
1504 * category.c (cat_stored_values_destroy): Changed interface to
1505 take a struct cat_vals * instead of a struct variable *.
1507 * dictionary.c (dict_clone): Use new vector_clone function.
1508 (dict_clear) Use new var_destroy function.
1509 (add_var) New function.
1510 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1511 (dict_create_var_assert) Rewrite in terms of add_var.
1512 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1513 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1514 (dict_lookup_var) Use new var_create, var_destroy functions.
1515 (dict_contains_var) Rewrite in terms of new vardict functionality.
1516 (set_var_dict_index) New function.
1517 (set_var_case_index) New function.
1518 (reindex_vars) New function.
1519 (dict_delete_var) Rewrite in terms of new vardict functionality.
1520 (dict_reorder_var) Ditto.
1521 (dict_reorder_vars) Ditto.
1522 (rename_var) New function.
1523 (dict_rename_var) Use rename_var.
1524 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1525 (dict_get_compacted_idx_to_fv) Rename
1526 dict_get_compacted_dict_index_to_case_index, update callers.
1527 (dict_create_vector) Use new vector_create function.
1528 (dict_clear_vectors) Use new vector_destroy function.
1529 (set_var_short_name_suffix) Move here from variable.c, renamed
1530 from var_set_short_name_suffix, make static, update caller.
1532 * sys-file-private.c: New file.
1533 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1534 width_to_bytes, update callers.
1536 * sys-file-private.h: New file. Later it will supplant
1537 sfm-private.h; for now it supplements it.
1538 (macro MIN_VERY_LONG_STRING) New macro.
1539 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1541 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1542 MAX_LONG_STRING was used before.
1544 * sys-file-writer.c: Ditto.
1546 * value-labels.c: Change the paradigm here to be that a null
1547 pointer is OK for a struct val_labs * in most cases; it just
1548 represents an empty set of value labels.
1549 (val_labs_copy) A copy of a null set is a null set.
1550 (val_labs_count) A null set has 0 labels.
1551 (val_labs_replace) Change return type to void. Rewrite for
1553 (val_labs_find) A null set does not contain the value.
1554 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1555 transposed argument order, updated all callers.
1557 * value.c: New file.
1558 (value_dup) Moved here from variable.c.
1559 (compare_values) Ditto.
1562 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1563 (macro MAX_LONG_STRING) Removed, because it was only interesting
1564 for system files, not for general code.
1565 (macro MAX_VERY_LONG_STRING) Ditto.
1566 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1567 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1569 * vardict.h: New file, for an interface between variables and
1572 * variable.c: A lot of functions were moved around, for better
1574 (struct variable) Move definition here, from variable.h.
1575 (var_type_adj) Removed--makes i18n hard.
1576 (var_type_noun) Ditto.
1577 (var_create) New function.
1578 (var_clone) New function.
1579 (var_destroy) New function.
1580 (var_set_name) Assert that variable is not in a dictionary.
1581 (compare_var_names) Rename compare_vars_by_name and fix a couple
1582 of callers who thought the args were strings.
1583 (hash_var_name) Rename hash_var_by_name.
1584 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1585 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1586 (var_is_very_long_string) Removed, because it was only interesting
1587 to system file code.
1588 (var_set_missing_values) Allow the argument to be the wrong width,
1589 as long as we can resize it. Simplify callers who were doing the
1590 resizing themselves.
1591 (var_get_value_labels) New function.
1592 (var_has_value_labels) New function.
1593 (var_set_value_labels) New function.
1594 (alloc_value_labels) New function.
1595 (var_add_value_label) New function.
1596 (var_replace_value_label) New function.
1597 (var_clear_value_labels) New function.
1598 (var_lookup_value_label) New function.
1599 (var_get_value_name) Moved here from variable.c, renamed from
1600 var_get_value_name, transposed argument order, updated all
1602 (var_to_string) Moved here, from variable-label.c.
1603 (var_set_leave) New function.
1604 (var_get_leave) New function.
1605 (var_must_leave) New function.
1606 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1607 set_var_short_name_suffix.
1608 (var_get_dict_index) New function.
1609 (var_get_case_index) New function.
1610 (var_get_obs_vals) New function.
1611 (var_set_obs_vals) New function.
1612 (var_has_obs_vals) New function.
1613 (var_get_vardict) New function.
1614 (var_set_vardict) New function.
1615 (var_has_vardict) New function.
1616 (var_clear_vardict) New function.
1617 (value_dup) Moved to value.c.
1618 (compare_values) Ditto.
1621 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1622 (enum ALPHA) Rename VAR_STRING, update all users.
1624 * vector.c: New file.
1625 (struct vector) Moved here, from variable.h.
1626 (check_widths) New function.
1627 (vector_create) New function.
1628 (vector_clone) New function.
1629 (vector_destroy) New function.
1630 (vector_get_name) New function.
1631 (vector_get_var) New function.
1632 (vector_get_var_cnt) New function.
1633 (compare_vector_ptrs_by_name) New function.
1635 * vector.h: New file.
1637 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1639 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1640 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1643 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1645 * value-labels.c (destroy_atoms): New function.
1646 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1648 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1650 Thanks to Jason Stover for pointing out this problem.
1652 * data-out.c (output_number): Use gsl_finite from GSL, which is
1653 portable, instead of isfinite, which is not.
1656 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1658 * variable.c variable.h (value_dup): New function.
1660 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1662 Start converting struct variable to an opaque type. In this
1663 phase, we add a bunch of setter and getter functions and convert
1664 most of the PSPP code to use them. The resulting changes are
1665 pervasive but mostly trivial, and only the notable changes are
1668 * format.c (fmt_equal): New function.
1670 * variable.c (var_type_is_valid): New function.
1671 (measure_is_valid) Moved here, from format.c.
1672 (alignment_is_valid) Moved here, from format.c.
1673 (var_get_name) New function.
1674 (var_set_name) New function.
1675 (width_to_type) New function.
1676 (var_get_type) New function.
1677 (var_get_width) New function.
1678 (var_set_width) New function.
1679 (var_is_numeric) New function.
1680 (var_is_alpha) New function.
1681 (var_is_short_string) New function.
1682 (var_is_long_string) New function.
1683 (var_is_very_long_string) New function.
1684 (var_get_missing_values) New function.
1685 (var_set_missing_values) New function.
1686 (var_clear_missing_values) New function.
1687 (var_has_missing_values) New function.
1688 (var_is_value_missing) New function.
1689 (var_is_num_missing) New function.
1690 (var_is_str_missing) New function.
1691 (var_is_value_user_missing) New function.
1692 (var_is_num_user_missing) New function.
1693 (var_is_str_user_missing) New function.
1694 (var_is_value_system_missing) New function.
1695 (var_get_print_format) New function.
1696 (var_set_print_format) New function.
1697 (var_get_write_format) New function.
1698 (var_set_write_format) New function.
1699 (var_set_both_formats) New function.
1700 (var_get_label) New function.
1701 (var_set_label) New function.
1702 (var_clear_label) New function.
1703 (var_has_label) New function.
1704 (var_get_measure) New function.
1705 (var_set_measure) New function.
1706 (var_get_display_width) New function.
1707 (var_set_display_width) New function.
1708 (var_get_alignment) New function.
1709 (var_set_alignment) New function.
1710 (var_get_value_cnt) New function.
1711 (var_get_leave) New function.
1712 (var_get_short_name) New function.
1714 * variable.h: (struct variable) Removed "type" and "nv" members;
1715 they are now computed from "width" where needed.
1717 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1719 * missing-values.c (mv_resize): Don't write beyond end of the
1720 allocated buffer when resizing a long string.
1722 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1724 Clean up identifier code: don't require identifier enumerations to
1725 be in a particular order; make better use of string library;
1726 expose less of the internals.
1728 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1729 change interface. Updated all callers.
1730 (lex_id_match) Change interface to use struct substring, update
1732 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1733 (global array keywords[]) Make static, change form. Update all
1734 users to use lex_id_name instead.
1735 (lex_is_keyword) New function.
1736 (lex_id_to_token) Change interface to use struct substring, update
1738 (lex_id_name) New function.
1740 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1741 lex_is_keyword instead.
1742 (T_LAST_KEYWORD) Removed.
1743 (T_N_KEYWORDS) Removed.
1745 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1747 * format.c: (fmt_date_template) Distinguish characters for which a
1748 space is output and any date delimiter is allowed on input, from
1749 those for which a space is output and only a space is allowed on
1750 input. The former is represented by X, the latter by a space.
1751 Also, drop distinction between h and H, changing the former to the
1754 * data-in.c: Completely rewrite internals to conform to SPSS input
1755 formats as closely as possible.
1756 (data_in) Changed external interface by replacing the structure
1757 that was used as a single argument by a set of arguments. Updated
1759 (data_in_finite_line) Removed. Converted all callers to use plain
1761 (data_in_get_integer_format) New function.
1762 (data_in_set_integer_format) New function.
1763 (data_in_get_float_format) New function.
1764 (data_in_set_float_format) New function.
1766 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1767 (struct data_in) Removed.
1769 * data-out.c: (output_date) Drop each component from the input as
1770 it is output, to allow us to drop the distinction between h (a
1771 count of hours) and H (the hour of day) template characters.
1772 Also, handle new X template character.
1773 (output_scientific) Follow more rational rule on when to drop
1774 fraction introduced between SPSS 13 and 15. Updated test case to
1777 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1779 Fix buffer overflow reported by John Darrington.
1781 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1782 realize that DIGITS is a count of nibbles, not of bytes.
1784 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1786 * calendar.c (calendar_offset_to_gregorian) Also return the
1787 year-of-day. Change callers to new interface.
1789 * data-out.c: Completely rewrite internals to conform to SPSS
1790 output formats as completely as possible.
1791 (data_out) Change interface to put input parameters before output
1792 parameters, for consistency with the style I now prefer. Update
1794 (data_out_get_integer_format) New public function.
1795 (data_out_set_integer_format) New public function.
1796 (data_out_get_float_format) New public function.
1797 (data_out_set_float_format) New public function.
1799 * data-out.h: New file. Move prototype for data_out here, from
1802 * format.c: (fmt_step_width) Use equality comparison instead of
1803 bitwise and, for clarity.
1804 (fmt_is_string) Ditto.
1805 (fmt_input_to_output) Fix categories that are translated to F
1808 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1810 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1811 casefile-private.h: Added casefilter to assist commands with missing
1814 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1816 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1818 * settings.c: (route_errors_to_terminal) New variable.
1819 (route_errors_to_listing) New variable.
1820 (get_error_routing_to_terminal) New function.
1821 (set_error_routing_to_terminal) New function.
1822 (get_error_routing_to_listing) New function.
1823 (set_error_routing_to_listing) New function.
1825 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1827 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1829 * format.c: Completely rewrite, to achieve better abstraction.
1830 Rewrite all references to formats in other files.
1832 * format.def: Rewrite and reorganize.
1834 * settings.c: Move everything related to custom currency formats
1835 into format.[ch], changing them in form, so as to group related
1836 code and definitions better. Changed all references to use the
1838 (static var decimal) Removed.
1839 (static var grouping) Removed.
1840 (static var cc) Removed.
1841 (get_decimal) Removed.
1842 (set_decimal) Removed.
1843 (get_grouping) Removed.
1844 (set_grouping) Removed.
1848 * settings.h: (macro CC_CNT) Removed.
1849 (macro CC_WIDTH) Removed.
1850 (struct custom_currency) Removed.
1852 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1854 * data-in.c (data_in): Use switch statement instead of table, to
1855 avoid dependence on the order of the FMT_* enums.
1857 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1859 * data-out.c: (num_to_string) Removed, because it was dead code.
1861 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1863 * data-in.c (parse_trailer): Fix error message.
1865 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1867 * format.c (fmt_is_binary): New function.
1869 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1871 * procedure.c procedure.h: Encapsulated the static data into a single
1874 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1876 * casefile.c (casereader_read_xfer): Always initialize the case,
1877 even on an error condition.
1879 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1881 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1882 entire function useless.
1884 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1886 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1887 casereader method casereader_clone.
1889 * procedure.c pransformations.h: Introduced new type casenum_t
1891 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1893 * variable.c: (width_to_bytes) Rephrase code for clarify.
1895 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1897 * format.c: (fmt_type_from_string) New function.
1898 (fmt_to_string) Include decimals in output if the format has
1899 decimals, even if the format type does not. This way, we can
1900 accurately reproduce incorrect formats in user output.
1901 (check_common_specifier) Make the check for a bad format type an
1902 assertion, so we get bug reports if they show up. Fix message.
1903 Check for decimal places with a format type that doesn't allow
1905 (check_input_specifier) Remove check for FMT_X, which has been
1907 (check_output_specifier) Ditto.
1909 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1911 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1912 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1913 (enum fmt_parse_flags) Removed.
1915 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1917 * casefile.c casefile.h: Converted to an abstract base class.
1918 * casefile-private.h fastfile.c fastfile.h: New files.
1919 * automake.mk procedure.c scratch-writer.c storage-stream.c
1921 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1923 * procedure.c (internal_procedure): Create sink_case with only as
1924 many values as the compacted dictionary.
1926 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1928 Remove "debugging" code that caused plenty of false positives and
1931 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1933 * case.c: Remove all references to `this' member.
1935 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1937 Fix link error noted by Jason Stover.
1939 * storage-stream.c: Include <assert.h>.
1941 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1943 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1944 ALL) and additional underlying system file issues.
1946 Thanks to John Darrington for review.
1948 First problem: var_hash points to variables not owned by the
1949 sys-file-reader, which the caller may free or modify. Use an
1950 array of sfm_vars instead, as done earlier (e.g. CVS version
1953 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1954 members and remove all code that references it. Add vars, var_cnt
1955 members. Remove fix_specials member, which was unused.
1956 (struct sfm_var) Remove name member, which was unused.
1957 (sfm_close_reader) Free vars member instead of var_hash.
1958 (compare_var_shortnames) Removed.
1959 (hash_var_shortname) Removed.
1960 (sfm_open_reader) Fill out vars array.
1961 (compare_var_index) Removed.
1962 (sfm_read_case) Use vars instead of var_hash.
1964 Second problem: we're confused about when we actually have very
1965 long strings, causing us to choose incorrectly between slow path
1966 and fast path in sfm_read_case.
1968 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1969 have very long strings, not when we have long variable names,
1970 which is an unrelated feature.
1972 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1974 * variable.h: Move var_set and variable parsing declarations to
1975 new header, src/language/lexer/variable-parser.h. Modified lots
1976 of files to include the new header.
1978 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1980 * value-labels.c (value_to_string): When there's no value label,
1981 format the variable according to its print format, instead of
1982 always effectively using A or F format.
1984 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1986 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1989 * format.c: Removed tortological assertion.
1991 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1993 Reform string library.
1995 * file-name.c (fn_interp_vars): Change interface to take a
1996 substring as input. Updated all users.
1998 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
2000 * format.c (measure_is_valid): Really return false when m >=
2003 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
2005 Implement random access to casefiles, for use in GUI.
2007 * casefile.c: (struct casereader) Add `random', `file_ofs',
2008 `buffer_ofs' members.
2009 (casefile_get_random_reader) New function.
2010 (read_open_file) Break part into new function
2011 seek_and_fill_buffer().
2012 (fill_buffer) Update buffer_ofs, file_ofs.
2013 (casereader_seek) New function.
2015 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2017 * settings.c: Added call to i18n{done, init}.
2019 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
2021 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
2024 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
2026 * casefile.c: Convert many uses of `int' to `bool'.
2028 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
2030 * transformations.c (trns_chain_destroy): Destroy chain's trns
2031 member, to fix memory leak.
2033 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
2035 * storage-stream.c (storage_source_decapsulate): Destroy case
2036 source to fix memory leak.
2038 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
2040 * scratch-reader.c (scratch_reader_read_case): Copy into existing
2041 case passed as argument instead of initializing the argument as a
2042 case. Fixes memory leak that showed up in
2043 tests/command/aggregate.sh with scratch files.
2045 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
2047 * procedure.c (proc_done): Destroy default_dict, to fix memory
2050 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2052 Simplify procedure_with_splits().
2054 * procedure.c (struct split_aux_data): Removed case_count member.
2055 (procedure_with_splits) Don't initialize case_count.
2056 (split_procedure_case_func) Check whether prev_case is null
2057 instead of case_count.
2058 (split_procedure_end_func) Ditto.
2060 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2062 * case.c (case_move): Do nothing if dst and src are the same
2064 (case_try_create) Merge two similar cases.
2065 (case_copy) Unshare only if data must be actually copied.
2067 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2069 * data-in.c data-out.c dictionary.c sys-file-reader.c
2070 sys-file-writer.c variable.c variable.h: Reworked very long string
2071 support for better encapsulation.
2073 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2075 * value-labels.c (val_labs_can_set_width): New function.
2076 (val_labs_set_width) Clear labels if increasing width to long
2078 (val_labs_destroy) Remove unneeded test for null.
2080 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2082 * value-labels.h: Remove unneeded dependency on variable.h.
2084 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2086 Get rid of `char *c' member in union value, for cleanliness.
2088 * value.h: (union value) Remove `c' member.
2090 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2092 Make dictionary compacting functions a little more general.
2094 * sys-file-writer.c (sfm_open_writer): Use
2095 dict_compacting_would_change().
2096 (does_dict_need_translation) Removed.
2098 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2100 Make dictionary compacting functions a little more general.
2102 * dictionary.c (dict_needs_compaction): Rename
2103 dict_compacting_would_shrink(). Update all callers.
2104 (dict_compacting_would_change) New function.
2106 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2108 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2109 inverted return value (!).
2111 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2113 Continue reforming procedure execution.
2115 * procedure.c: Search and replace "vfm" by "proc". Notably:
2116 (static var vfm_source) Rename proc_source. Update all
2118 (static var vfm_sink) Rename proc_sink. Update all references.
2120 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2122 Continue reforming procedure execution. In this phase, remove
2123 PROCESS IF, which was deprecated anyway and can be easily
2124 simulated with TEMPORARY followed by SELECT IF.
2126 * procedure.c: (open_active_file) Don't call
2127 add_process_if_trns().
2128 (discard_variables) Get rid of redundant call to
2129 proc_cancel_all_transformations().
2130 (add_process_if_trns) Removed.
2131 (process_if_trns_proc) Removed.
2132 (process_if_trns_free) Removed.
2134 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2136 Continue reforming procedure execution. In this phase, add
2137 `const' to the case passed to procedure()'s callback.
2139 Updated all users of procedure() as well.
2141 * procedure.c: (struct write_case_data) Add "const" to ccase
2142 parameter for case_func member.
2143 (procedure) Add "const" to ccase parameter for proc_func
2145 (multipass_case_func) Make ccase parameter const.
2146 (internal_procedure) Add "const" to ccase parameter for case_func
2148 (split_procedure_case_func) Make ccase parameter const.
2149 (multipass_split_case_func) Make ccase parameter const.
2151 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2153 Continue reforming procedure execution. In this phase, get rid of
2154 the output code for SPLIT FILE groups in procedure.c, which really
2155 shouldn't be doing any output. Move it into the individual
2156 procedures instead. This also adds some flexibility.
2158 Updated many users of procedure_with_splits() and
2159 multipass_procedure_with_splits() to call
2160 output_split_file_values() and to deal with increased use of
2163 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2164 parameter to begin_func member.
2165 (procedure_with_splits) Add "const struct ccase *" parameter to
2166 begin_func parameter. Make ccase parameter const in proc_func
2168 (split_procedure_case_func) Don't dump split file group. Pass
2170 (dump_splits) Moved to language/dictionary/split-file.c as
2171 output_split_file_values().
2172 (struct multipass_split_aux_data) Add "const struct ccase *"
2173 parameter to split_func member.
2174 (multipass_procedure_with_splits) Add "const struct ccase *"
2175 parameter to split_func parameter.
2176 (multipass_split_case_func) Save new SPLIT FILE case before
2178 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2180 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2182 Continue reforming procedure execution. Change
2183 internal_procedure() so that it calls open_active_file() and
2184 close_active_file(), which isolates most of the actual procedure
2187 * procedure.c: (struct write_case_data) Rename `proc_func' member
2188 to `case_func' and update all references.
2189 (procedure) Rewrite as one-line wrapper around
2190 internal_procedure().
2191 (struct multipass_aux_data) New.
2192 (multipass_callback) Renamed multipass_case_func(). Use struct
2193 multipass_aux_data as auxiliary data.
2194 (multipass_end_func) New function.
2195 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2196 that uses multipass_case_func, multipass_end_func.
2197 (internal_procedure) Add `end_func' argument. Move optimization
2198 of trivial case in here. Move call to open_active_file() and
2199 close_active_file() in here. Now assert that vfm_source is
2201 (procedure_with_splits_callback) Rename
2202 split_procedure_case_func().
2203 (split_procedure_end_func) New function.
2204 (multipass_split_callback) Rename multipass_split_case_func.
2205 (multipass_split_end_func) New function.
2206 (discard_variables) No need to test for nonnull vfm_source.
2208 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2210 Continue reforming procedure execution. Get rid of unused member.
2212 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2214 (write_case) Don't increment cases_analyzed.
2216 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2218 Continue reforming procedure execution. In this phase, move
2219 procedure.c and procedure.h from src to src/data. Update
2220 makefiles and #includes accordingly.
2222 * procedure.c: Moved here from src/.
2224 * procedure.h: Moved here from src/.
2226 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2228 Continue reforming procedure execution. In this phase, get rid of
2229 many global variables, consolidating procedure execution in
2230 procedure.c. Encapsulate transformations in new "struct
2231 trns_chain". Also, change implementation of N OF CASES, FILTER,
2232 and PROCESS IF from special cases to transformations.
2234 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2237 * dictionary.c: (global variable default_dict) Move to
2240 * variable.h: (TRNS_*) Move to transformations.h.
2241 (struct transformation) Move to transformations.c.
2243 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2245 * sys-file-reader.c: Fixed invalid read problems.
2247 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2249 * storage-stream.c: Add missing function comments.
2251 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2253 Continue reforming procedure execution. In this phase, add some
2254 new, needed functionality to storage-stream.
2256 * storage-stream.c: (storage_source_decapsulate) New function.
2258 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2260 * variable.c (width_to_bytes): Declarations must precede
2261 statements for C90 compliance.
2263 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2265 * data-out.c, data-in.c, variable.c, variable.h: New functions
2266 copy_mangle and copy_demangle for reading/writing cases; emulates the
2267 way SPSS deals with strings > 255 bytes.
2269 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2270 subtype 14 needed for strings longer than 255 bytes.
2272 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2273 instead of literal values. Also fixed some constness issues.
2275 * format.h: Constness
2277 * sfm-private.h: Renamed the case_size identifier, since I discovered
2278 that SPSS's respect for this variable is very nominal.
2280 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2282 Change case limit type from int to size_t.
2284 * dictionary.c: (struct dictionary) Change type of case_limit
2286 (dict_get_case_limit) Change return type.
2287 (dict_set_case_limit) Change parameter type.
2289 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2291 * variable.h: (struct variable) Rename `reinit' member as `leave'
2292 and invert sense. Fix up all references.
2294 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2296 Continue reforming procedure execution. In this phase, break
2297 procedure.c into multiple files.
2299 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2301 * case-sink.c: New file.
2303 * case-sink.h: New file.
2305 * case-source.c: New file.
2307 * case-source.h: New file.
2309 * storage-stream.c: New file.
2311 * storage-stream.h: New file.
2313 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2315 * variable.h: (struct variable) Remove `init' member and all
2316 references to it from other files. It was initialized in several
2317 places, but nothing really ever used it for anything worthwhile.
2318 Thanks to Jason Stover for pointing out how confusing this
2321 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2323 Continue reforming error message support. In this phase, get rid
2324 of message "titles" and put the message text in `struct error'.
2325 Now `struct error' encapsulates a message more properly.
2327 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2328 Format message ourselves.
2330 * data-in.c: (vdls_error) Ditto.
2332 * por-file-reader.c: (error) Ditto.
2334 * sys-file-reader.c: (corrupt_msg) Ditto.
2336 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2338 GNU standards require "file name" instead of "filename" in
2339 documentation. It's nice for our code to follow the convention
2342 * casefile.c: (struct casefile) Rename `filename' member to
2343 `file_name'. Updated all references.
2345 * file-name.c: [!unix] (struct file_identity) Rename
2346 normalized_filename member to normalized_file_name. Updated all
2349 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2351 We don't really support anything but Unix-like environments well,
2352 so we might as well de-obfuscate by writing directory and path
2353 separators explicitly.
2355 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2357 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2358 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2360 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2363 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2365 GNU standards require "file name" instead of "filename" in
2366 documentation. It's nice for our code to follow the convention
2369 * filename.c: Rename to file-name.c.
2371 * filename.h: Rename to file-name.h. Update all inclusions.
2372 Update header guards.
2374 * automake.mk: Update file names.
2376 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2378 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2380 (fn_basename) Removed (dead code).
2381 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2382 (fn_special_p) Renamed fn_is_special(), all references updated.
2383 (fn_exists_p) Renamed fn_exists(), all references updated.
2385 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2387 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2388 Also, now it only tilde-expands file names, not paths.
2389 (fn_search_path) Tilde-expand one directory at a time.
2391 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2393 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2394 get rid of non-Unixlike version of the code, which has probably
2396 (fn_prepend_dir) Removed (dead code).
2398 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2399 (macro PATH_SEPARATOR_STRING) New.
2400 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2402 Continue reforming error message support. In this phase, we get
2403 rid of VM() and the other msg() support for "verbosity", replacing
2404 it by a new function verbose_msg().
2406 * filename.c: (fn_search_path) Use verbose_msg() instead of
2409 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2411 * sfm-private.h: Get rid of #defines after #error, which makes no
2414 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2416 Get rid of our own int32 type in favor of the standard int32_t
2419 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2422 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2424 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2426 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2428 Remove ill-considered file routines that are no longer used.
2430 * filename.c: (fn_open_ext) Removed.
2431 (fn_close_ext) Removed.
2433 * filename.h: (struct file_ext) Removed.
2435 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2437 * variable.c (var_is_valid_name): Move declarations before code
2440 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2442 * filename.ch (fn_interp_vars): Fixed small buglet.
2444 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2446 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2447 so as to modify the string inline. Thus makeing it easier to
2448 destroy the results when no longer needed.
2450 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2452 * category.c (cat_stored_values_destroy): Fixed memory leak.
2454 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2456 Add some missing frees. Thanks to John Darrington for reporting
2459 * any-writer.c (any_writer_close): Free writer.
2461 * any-reader.c (any_reader_close): Free reader.
2463 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2465 * por-file-reader.c: (error) Mark as NO_RETURN.
2467 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2469 * settings.c: Changed default value of scompress to true.
2471 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2473 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2475 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2478 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2479 here from pref.h.orig.
2481 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2483 * sys-file-reader.c: Fixed bug reading compressed files.
2485 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2487 * Numerous renames. See src/ChangeLog for details.
2489 * Moved files from src directory