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 2008-02-18 Ben Pfaff <blp@gnu.org>
11 Patch #6426. Reviewed by John Darrington.
13 * format.c (min_width): Renamed fmt_min_width and made public.
14 Updated all references.
15 (max_width): Renamed fmt_max_width and made public. Updated all
17 (max_decimals): Renamed fmt_max_decimals and made public. Updated
19 (var_create): Use the new functions for default variable
21 (var_default_formats): New function.
22 (var_default_measure): New function.
23 (var_default_alignment): New function.
25 * format.h (macro FMT_MAX_NUMERIC_WIDTH): New macro.
27 2008-02-09 Ben Pfaff <blp@gnu.org>
29 Add a couple of extensions to GET DATA TYPE=TXT. Patch #6412.
30 Thanks to John Darrington for review.
32 * data-in.c (data_in): Add new argument to designate the last
33 column of the data field being parsed, for use in error messages.
36 * data-parser (struct data_parser): New member `quote_escape'.
37 (data_parser_create): Initialize quote_escape.
38 (data_parser_set_quotes): New function.
39 (cut_field): Support escaped quotes.
40 (parse_delimited_span): Ditto.
41 (parse_delimited_no_span): Ditto.
43 * get-data.c (parse_get_txt): Support ESCAPE extension subcommand
44 in enhanced mode. Only support multiple quote characters in
47 2008-02-06 John Darrington <john@darrington.wattle.id.au>
49 psql-reader.c psql-reader.h: Read more than one tuple at
50 once. Fix bug reading a query which returns no data. Fix bug
51 when transformation followed a reader.
52 Ask the server for the number of records in the query, for the
55 2008-02-05 John Darrington <john@darrington.wattle.id.au>
57 psql-reader.c: So yesterday they release postgresql 8.3.0
58 which has money represented with 64 bits. They must get
61 2008-02-02 John Darrington <john@darrington.wattle.id.au>
63 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
64 for reviewing this code.
66 2008-02-02 Ben Pfaff <blp@gnu.org>
70 * sys-file-reader.c (read_variable_record): Allow missing values
71 to be specified on long string variables, but warn about them
72 because PSPP does not yet support them.
73 (read_extension_record): Ignore extension records 20 and 21, which
74 PSPP does not yet support.
75 (read_header): Fix error message when floating-point format cannot
78 2008-02-01 Ben Pfaff <blp@gnu.org>
80 Patch #6386. Thanks to John Darrington for review and for the
81 updates to gnumeric-reader.c.
83 * dictionary.c (make_hinted_name): New function.
84 (make_numeric_name): New function.
85 (dict_make_unique_var_name): New function.
87 * gnumeric-reader.c (devise_name): Removed.
88 (munge_name): Removed.
89 (gnumeric_open_reader): Use new function
90 dict_make_unique_var_name.
92 * short-names.c (set_var_short_name_suffix): Use new function
95 2008-01-19 John Darrington <john@darrington.wattle.id.au>
97 * settings.c settings.h: Moved static variables into a
98 single struct. Renamed functions to have a settings_ prefix.
100 2008-01-14 John Darrington <john@darrington.wattle.id.au>
102 * data-out.c (output_AHEX): Corrected number of bytes in
103 call to output_hex. Closes bug #22011
105 2008-01-02 John Darrington <john@darrington.wattle.id.au>
107 * variable.c variable.h: Replaced var_get_value_name with
108 var_append_value_name which doesn't use any static data.
109 Thanks to Ben for review.
111 2007-12-07 Ben Pfaff <blp@gnu.org>
115 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
116 made from an empty casereader produce a casegrouper with no
117 groups, instead of one with one group that has no cases.
119 * casereader.c (casereader_is_empty): New function.
121 2007-12-06 Ben Pfaff <blp@gnu.org>
125 * sys-file-reader.c (read_display_parameters): Handle variable
126 display parameters record with only 2 data items per variable.
127 Reported by Guido Gay <gay@irer.it>.
129 2007-12-04 Ben Pfaff <blp@gnu.org>
131 * identifier.c (lex_id_match_n): New function.
132 (lex_id_match): Reimplement in terms of lex_id_match_n.
134 2007-11-24 Ben Pfaff <blp@gnu.org>
136 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
139 2007-11-24 Ben Pfaff <blp@gnu.org>
141 Fix warning reported when reading back system files that include
142 very long string variables. Thanks to Guido Gay <gay@irer.it> for
145 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
146 was written where `j' was meant.
148 2007-11-18 Ben Pfaff <blp@gnu.org>
150 Properly write variables that include a range of missing values to
151 system files. Thanks to Guido Gay <gay@irer.it> for reporting
154 * sys-file-writer.c (write_variable): Correctly calculate
155 n_missing_values field when writing variables that include a range
158 2007-11-10 Ben Pfaff <blp@gnu.org>
160 Cleanups and bug fixes devised while writing up documentation.
163 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
165 * dict-class.c: New file.
166 (dict_class_from_id): Move here.
167 (dict_class_to_name): Move here.
169 * dict-class.h: New file.
170 (enum dict_class): Move here. Change from consecutive integers to
171 consecutive bits, to make testing for any of multiple values
172 easier. Add new DC_ALL constant.
174 * dictionary.c (struct dictionary): Change `case_limit' from
175 size_t to casenumber.
176 (dict_get_vars): Make final argument an enum dict_class.
177 (dict_get_vars_mutable): Ditto.
178 (dict_get_case_limit): Change return value to casenumber.
179 (dict_set_case_limit): Change final argument to a casenumber.
180 (dict_unset_split_var): Add assertion. Rephrase slightly.
181 (dict_set_label): Use xstrndup to simplify.
183 * format.c (fmt_step_width): AHEX format also needs 2-byte
185 (fmt_set_style): Simplify assertions.
187 * missing-values.c (mv_add_num_range): Rename mv_add_range.
188 Simplify implementation.
189 (mv_has_value): Simplify implementation.
190 (mv_pop_value): Remove the first value, not the last, to avoid
191 having GET followed by SAVE reverse the order of missing values.
192 (mv_peek_value): Rename mv_get_value. Simplify assertion.
193 (mv_has_range): Simplify implementation.
194 (mv_peek_range): Rename mv_get_range.
195 (can_resize_string): Removed.
196 (mv_is_resizable): Use value_is_resizable.
197 (mv_resize): Use value_resize.
199 * short-names.h (SHORT_NAME_LEN): Move here.
201 * val-type.h: New file, for definitions related to type and width
202 of abstract values. Before, these definitions were mixed among
203 those related to "union value" and those related to variables.
204 (macro SYSMIS): Move here.
205 (macro LOWEST): Move here.
206 (macro HIGHEST): Move here.
207 (macro MAX_STRING); Move here.
208 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
209 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
210 All references updated.
211 (val_type_is_valid): New function. Replaces var_type_is_valid.
212 All references updated.
213 (val_type_from_width): New function. Replaces
214 var_type_from_width. All references updated.
216 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
218 (val_labs_can_set_width): Use value_is_resizable.
219 (val_labs_add): Simply return false if the value labels set is too
220 wide, instead of having undefined behavior.
221 (val_labs_replace): Ditto.
222 (val_labs_replace): Ditto.
223 (val_labs_first): Set iterator to null if iteration is complete.
224 (val_labs_first_sorted): Ditto.
225 (val_labs_done): Become a no-op if the iterator is null.
227 * value.c (value_is_resizable): New function.
228 (value_resize): New function.
230 * variable.c (var_get_dict_class): New function.
232 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
235 2007-11-08 Ben Pfaff <blp@gnu.org>
237 * data-in.c: Make formatted data parsing locale-independent.
238 (parse_number): Use c_strtod instead of strtod, to avoid
239 locale-specific behavior.
242 2007-11-06 Ben Pfaff <blp@gnu.org>
244 Patch #6256: add support for binary, 360 file formats. Reviewed
247 * data-in.c (struct data_in): Add `encoding' member.
248 (data_in): Add `encoding' parameter, and re-encode the data passed
249 in where appropriate. Update all callers to pass it in.
250 (parse_A): Implement EBCDIC recoding wart described in manual.
251 (parse_AHEX): Implement EBCDIC recoding.
253 * data-out.c (data_out_legacy): New function.
254 (data_out): Make into a wrapper around data_out_legacy.
256 * file-handle-def.c (struct file_handle): New member `encoding'.
257 (fh_create_file): Set encoding.
258 (fh_default_properties): Set default encoding.
259 (fh_get_legacy_encoding): New function.
261 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
262 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
263 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
264 (struct fh_properties): New member `encoding'.
266 2007-11-05 Ben Pfaff <blp@gnu.org>
268 Patch #6258. Reviewed by John Darrington.
270 * file-handle-def.c (fh_lock): Add comment that TYPE should be
271 marked with N_() in the caller. Added these markings to each
272 caller too. Should make i18n easier.
273 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
275 2007-11-03 Ben Pfaff <blp@gnu.org>
277 Allow output files to overwrite input files (bug #21280).
279 * file-handle-def.c: Separate locking of files for input and for
280 output, to allow both to take place at once. Also, distinguish a
281 file handle from the identity of the underlying file, because the
282 identity of a file changes over time and the file handle can't
283 represent two different identities.
284 (struct file_handle): Remove `next', `open_cnt', `deleted',
285 `type', `open_mode', `aux', `identity' members. Change `id' from
286 char array to char *. Add `ref_cnt' member.
287 (file_handle_from_ll) New function.
288 (file_handles) Removed.
289 (named_handles) New variable.
290 (fh_init) Initialize named_handles.
291 (fh_done) Remove name from all named_handles.
292 (free_handle) Updated for modified struct file_handle.
293 (unname_handle) New function.
294 (fh_ref) New function.
295 (fh_from_file_name) Removed.
296 (fh_unref) New function.
297 (fh_unname) New function.
298 (fh_from_id) Rewritten.
299 (create_handle) Updated for modified struct file_handle.
300 (fh_inline_file) Increment the handle's ref count.
301 (fh_create_file) Updated for modified struct file_handle.
302 (fh_create_scratch) Ditto.
307 (fh_is_open) Removed.
308 (fh_get_id) Updated for modified struct file_handle.
309 (fh_get_default_handle) Increment the handle's ref count.
310 (fh_set_default_handle) Handle ref counts.
311 (struct fh_lock) New structure.
312 (locks) New static var.
313 (fh_lock) New function.
314 (fh_unlock) New function.
315 (fh_lock_get_aux) New function.
316 (fh_lock_set_aux) New function.
317 (fh_is_locked) New function.
318 (make_key) New function.
319 (free_key) New function.
320 (compare_fh_locks) New function.
321 (hash_fh_lock) New function.
323 * file-handle-def.h (enum fh_access) New enum.
325 * file-name.c: Made file_identity the same in all supported
327 (struct file_identity): New `name' member.
328 (fn_get_identity): For a file that doesn't exist, get the
329 dev/inode of its directory plus its name. If even the directory
330 doesn't exist, just use its name. Merge the Windows
331 implementation into the Unix one.
332 (fn_compare_file_identities): Rewritten. Merge the Windows
333 implementation into the Unix one.
334 (fn_hash_identity): New function.
336 * make-file.c (struct replace_file): New structure.
337 (all_files): New static var.
338 (replace_file_start): New function.
339 (replace_file_commit): New function.
340 (replace_file_abort): New function.
341 (free_replace_file): New function.
342 (unlink_replace_files): New function.
344 * por-file-reader.c (struct pfm_reader): Add `lock' member.
345 (close_reader): Unlock file.
346 (pfm_open_reader): Lock file.
348 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
350 (pfm_open_writer): Lock file and prepare for its replacement.
351 (close_writer): Unlock file.
353 * scratch-handle.h (struct scratch_handle): Add unique_id so that
354 different generations of a scratch file can be distinguished.
356 * scratch-reader.c (scratch_reader_open): Verify that the file is
359 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
360 remove scratch_handle member.
361 (scratch_writer_open): Lock handle. Prepare to replace handle
362 data, instead of doing it immediately.
363 (scratch_writer_casewriter_destroy): Replace handle data and
366 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
367 (sfm_open_reader): Lock file.
368 (close_reader): Unlock file.
370 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
372 (sfm_open_writer): Lock file and prepare for its replacement.
373 (close_writer): Unlock file.
375 2007-11-02 Ben Pfaff <blp@gnu.org>
377 * data-out.c (output_number): Use isfinite (from C99) instead of
378 gsl_isfinite, since we now have universal support for it in
380 (output_infinite): Ditto.
381 [!HAVE_ROUND] (round): Remove definition, since we now have a
382 replacement in gnulib.
384 2007-10-27 John Darrington <john@darrington.wattle.id.au>
386 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
387 wasn't initialised correctly. Closes bugs #21330 and 21397
390 2007-10-08 Ben Pfaff <blp@gnu.org>
392 Fix bug #21280. Thanks to John Darrington for review.
394 * file-name.c (create_stream): New function.
396 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
397 file handle before creating the file, to ensure that we don't
398 truncate a file that we're reading. Make code easier to read by
401 * sys-file-write.c (sfm_open_writer): Ditto.
403 2007-10-01 Ben Pfaff <blp@gnu.org>
405 Fix bug #21192. Thanks to John Darrington for review.
407 * casereader.c (casereader_read): Decrement case_cnt before
408 calling the casereader's "read" member function, so that we
409 interact properly with lazy_casereader.
411 * datasheet.c: Add regression test for above bug fix.
412 (clone_datasheet): New function.
413 (lazy_callback): New function.
414 (check_datasheet_casereader): New function.
415 (check_datasheet): Check datasheet contents are reported correctly
416 through an ordinary casereader and a lazy casereader.
417 (clone_model): Use clone_datasheet.
419 2007-09-24 Ben Pfaff <blp@gnu.org>
421 Patch #6210. Reviewed by John Darrington.
422 * settings.c: Drop "long view", which is not needed any longer.
423 (static var long_view): Removed.
424 (force_long_view): Removed.
425 (get_viewwidth): Removed.
426 (init_viewport): Removed long_view reference.
428 2007-09-19 John Darrington <john@darrington.wattle.id.au>
430 * settings.c settings.h: Changed viewport's length and width to be
431 owned by the user interface which uses the data library. This allows
432 better abstraction, and makes dynamically adjustable dimensions easier.
434 2007-09-18 Ben Pfaff <blp@gnu.org>
436 * procedure.c (proc_extract_active_file_data): New function.
438 * lazy-casereader.h: New file.
440 * lazy-casereader.c: New file.
442 * casereader.c (casereader_dynamic_cast): New function.
444 2007-09-14 Ben Pfaff <blp@gnu.org>
446 * dictionary.c (dict_clone): Copy case indexes from cloned
447 dictionary. Fixes bug #21061. Reviewed by John Darrington.
449 2007-09-13 John Darrington <john@darrington.wattle.id.au>
451 * value.c value.h (create_value): New function.
453 2007-09-12 Ben Pfaff <blp@gnu.org>
455 Make it clear that translator casereader and casewriter translate
456 functions are supposed to destroy their input case.
458 * casereader-translator.c (struct casereader_translator): Change
459 input case parameter of translate member function from const
460 struct ccase * to struct ccase *.
461 (casereader_create_translator): Ditto, for translate parameter.
463 * casewriter-translator.c (struct casewriter_translator): Ditto.
464 (casewriter_create_translator): Ditto.
466 2007-08-27 John Darrington <john@darrington.wattle.id.au>
468 * sys-file-reader.c (read_display_parameters): Force display width
469 to 8, if the sys file says 0 (like SPSS does).
471 2007-08-12 Ben Pfaff <blp@gnu.org>
473 * dictionary.c (dict_dump): New function.
475 2007-08-12 Ben Pfaff <blp@gnu.org>
477 Drop dict_compactor in favor of using the new struct case_map.
479 * dictionary.c (struct copy_map): Removed.
480 (struct dict_compactor): Removed.
481 (dict_make_compactor): Removed.
482 (dict_compactor_compact): Removed.
483 (dict_compactor_destroy): Removed.
485 * procedure.c (struct data_set): Change `compactor' member to be a
487 (proc_open): Use case_map_to_compact_dict instead of
489 (proc_casereader_read): Use case_map_execute instead of
490 dict_compactor_compact.
491 (proc_commit): Use case_map_destroy instead of
492 dict_compactor_destroy.
494 * scratch-writer.c (struct scratch_writer): Change `compactor'
495 member to be a struct case_map *.
496 (scratch_writer_open): Use case_map_to_compact_dict instead of
498 (scratch_writer_casewriter_write): Use case_map_execute instead of
499 dict_compactor_compact.
501 2007-08-12 Ben Pfaff <blp@gnu.org>
503 * automake.mk: Add case-map.c, case-map.h.
505 * case-map.c: New file.
507 * case-map.h: New file.
509 2007-08-12 Ben Pfaff <blp@gnu.org>
511 * dictionary.c (dict_compact_values): Don't delete scratch
512 variables as well as compacting case indexes. Update all callers.
513 (dict_get_compacted_value_cnt): Rename dict_count_values and
514 change interface. Update all callers.
515 (dict_get_compacted_value_cnt): Remove.
516 (dict_compacting_would_shrink): Remove.
517 (dict_compacting_would_change): Remove.
518 (dict_make_compactor): Add new parameter. Update all callers.
520 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
521 instead of dict_count_values, changing an O(N) operation into
524 2007-08-12 Ben Pfaff <blp@gnu.org>
526 * casereader.c (casereader_read): Don't require cases read by a
527 casereader to be exactly the expected size: as long as they're big
530 2007-08-12 Ben Pfaff <blp@gnu.org>
532 Make casewriters keep track of the number of `union value's in
533 each case. This is useful for two reasons: casewriter_write can
534 then check that the case being written is large enough, and later
535 recipients of the casewriter can determine the size of the case.
537 * casewriter-translator.c (casewriter_create_translator): Add
540 * casewriter.c (struct casewriter): Add value_cnt member.
541 (casewriter_write): Check that the case passed in is big enough.
542 (casewriter_get_value_cnt): New function.
543 (casewriter_create): Add value_cnt parameter.
545 2007-08-09 Ben Pfaff <blp@gnu.org>
547 Fix bug reported by Jason Stover.
548 * settings.c: Move get_termcap_viewport and in particular the
549 #include for <curses.h> to the end of the file. curses.h
550 redefines bool on some systems (e.g. OpenBSD), which causes
551 disagreement between uses of bool before its inclusion and
553 Tested by Jason Stover.
555 2007-07-29 Ben Pfaff <blp@gnu.org>
557 Provisional fix for bug #18692 and bug #20161. Reviewed by John
560 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
561 argument (never "rb" or "wb") because SUSv3 says that only those
562 modes are defined, and glibc in fact rejects other modes.
564 Open portable files with fn_open so that they can be read from
565 pipes. Fix missing fh_close call to go along with fh_open.
566 Report an error if the file close reports an error.
567 * por-file-reader.c (close_reader): New function.
568 (por_file_casereader_destroy): Use close_reader.
569 (pfm_open_reader): Open file with fn_open.
571 2007-07-28 Ben Pfaff <blp@gnu.org>
573 Make PSPP able to read all the portable files I could find on the
574 web. Thanks to John Darrington for review. Bug #17620.
575 * por-file-reader.c (struct pfm_reader): New member `line_length'.
576 (error): Print file offset in hexadecimal.
577 (warning): New function.
578 (advance): Treat lines less than 80 bytes long as padded to 80
580 (pfm_open_reader): Call read_documents if we find an "E" record.
581 (convert_format): Convert invalid formats to the default format
582 instead of aborting reading the file.
583 (read_variables): Rename duplicate variable names instead of
584 aborting reading the file.
585 (read_value_label): Allow string variables of different widths to
586 be assigned value labels in the same record. Replace duplicate
587 value labels instead of aborting.
588 (read_documents): New function.
590 * por-file-writer.c (pfm_open_writer): Call write_documents if the
591 dictionary has documents.
592 (write_documents): New function.
594 2007-07-25 Ben Pfaff <blp@gnu.org>
596 Fix bugs related to bug #17213.
598 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
599 since the former is what config.h has. Include the needed ncurses
601 (static var echo) Rename to `do_echo' because the original name is
602 the same as an ncurses identifier.
603 (get_termcap_viewport) Use error instead of msg.
605 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
606 (fn_close): Don't close stdin, stdout, stderr.
608 2007-07-26 John Darrington <john@darrington.wattle.id.au>
610 * procedure.c procedure.h: Added callbacks which get invoked whenever
611 a dataset's transformation chain changes.
613 2007-07-24 Ben Pfaff <blp@gnu.org>
616 * sys-file-writer.c (write_variable_display_parameters): Use new
617 var_default_display_width function to choose display width of
618 segments after the first one in a given variable.
619 * variable.c (var_create): Use var_default_display_width to pick
620 new variable's display width.
621 (var_default_display_width): New function.
622 Reviewed by John Darrington.
624 2007-07-24 Ben Pfaff <blp@gnu.org>
627 * por-file-writer.c (write_variables): Write weight variable.
628 Reviewed by John Darrington.
630 2007-07-23 Ben Pfaff <blp@gnu.org>
632 Improvements to system file reader and writer.
634 First, move all detailed knowledge of very long strings into
635 sys-file-private.[ch], so that this nasty stuff can be isolated.
637 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
638 (EFFECTIVE_VLS_CHUNK): New macro.
639 (min_int): New function.
640 (max_int): New function.
641 (sfm_width_to_bytes): Rewrite.
642 (sfm_width_to_octs): New function.
643 (sfm_segment_alloc_width): New function.
644 (sfm_segment_alloc_bytes): New function.
645 (sfm_segment_used_bytes): New function.
646 (sfm_segment_offset): New function.
647 (sfm_segment_effective_offset): New function.
648 (sfm_dictionary_to_sfm_vars): New function.
650 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
651 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
652 (struct sfm_var): New structure.
654 Next, improvements to the system file reader.
656 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
657 type casenumber. Added `version_major', `version_minor',
660 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
661 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
662 Change `case_cnt' to type casenumber. Removed `has_vls'.
663 (struct sfm_var): Removed.
664 (sfm_open_reader): Don't warn on wrong case size if the file was
665 written by SPSS 13, which tends to get it wrong. Use
666 sfm_dictionary_to_sfm_vars.
667 (read_header): Always output system file info.
668 (read_variable_record): Simplify code for reading missing values.
669 (read_machine_int32_info): Save version numbers from system file
670 into info struct passed as new argument.
671 (read_long_string_map): Restructured to use new sys-file-private
673 (read_value_labels): Use size_overflow_p.
674 (sys_file_casereader_read): Get rid of distinction between fast
675 and slow paths. Use information provided by sys-file-primate's
676 struct sfm_var to simplify code.
677 (skip_whole_strings): New function.
678 (read_int32): Renamed read_int. Changed return value to int.
680 (read_flt64): Renamed read_float. Changed return value to
681 double. Updated all callers.
682 (int32_to_native): Removed. Changed callers to use
684 (flt64_to_double): Removed. Changed callers to use float_convert.
686 Finally, get rid of int32, flt64 terminology and types in system
687 file writer. The former wasn't very useful since a POSIX "int"
688 can hold the whole range of int32 and we generally didn't have a
689 need for it to be exactly-32-bits, just at-least-32-bits. The
690 latter was inconvenient because we had to assume that it could be
691 different from double and thereby convert special values SYSMIS,
692 HIGHEST, LOWEST to and from it in multiple places. Instead, now
693 we just use "int" and "double" in most places, and do conversions,
694 if necessary, very close to where we do I/O. This change meant
695 that the writer code couldn't represent records in the file as C
696 structs any longer, but that's no great loss. The code actually
697 seems to be more readable without them.
699 Simplify the compression buffering code: only buffer as much as
700 necessary, which is no more than eight 8-byte units at any given
703 * sys-file-writer.c (typedef flt64): Removed.
704 (macro second_lowest_flt64): Removed.
705 (struct sysfile_header): Removed.
706 (struct sysfile_variable): Removed.
707 (struct sfm_writer): Removed `needs_translation', `has_vls',
708 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
709 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
710 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
711 compression buffering by `opcodes', `opcode_cnt', `data',
712 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
713 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
714 data writer functions instead of structures.
715 (calc_oct_idx): New function.
716 (write_header): Use simple data writer functions instead of
718 (write_format_spec): Renamed write_format. New argument.
719 (write_variable_continuation_records): New function.
720 (write_variable): Use simple data writer functions instead of
721 structures. Use write_variable_continuation_records. Write
722 entire very long string instead of requiring caller to understand
724 (write_value_labels): Use simple data writer functions instead of
726 (write_documents): Ditto.
727 (write_variable_display_parameters): Use sys-file-private
728 functions to simplify. Use simple data writer functions instead
730 (write_vls_length_table): Use simple data writer functions instead
732 (write_longvar_table): Ditto.
733 (write_rec_7_34): Break into new functions
734 write_integer_info_record, write_float_info_record. Use simple
735 data writer functions instead of structures.
736 (buf_write): Removed.
737 (append_string_max): Removed.
738 (ensure_buf_space): Removed.
739 (sys_file_casewriter_write): Get rid of the distinction between
740 fast and slow paths, which didn't seem to be too useful. Use new
741 functions write_case_uncompressed, write_case_compressed.
742 (put_instruction): Removed.
743 (put_element): Removed.
744 (write_compressed_data): Removed.
745 (close_writer): Use flush_compressed. Only write case count to
746 system file if it will fit in the field.
747 (write_case_compressed): New function.
748 (write_case_uncompressed): New function.
749 (flush_compressed): New function.
750 (put_cmp_opcode): New function.
751 (put_cmp_number): New function.
752 (write_int): New function.
753 (convert_double_to_output_format): New function.
754 (write_float): New function.
755 (write_value): New function.
756 (write_string): New function.
757 (write_bytes): New function.
758 (write_zeros): New function.
759 (write_spaces): New function.
761 Reviewed by John Darrington.
763 2007-07-22 Ben Pfaff <blp@gnu.org>
765 Don't try to write very long strings to portable files. The
766 format does not support it.
768 * por-file-writer.c (MAX_POR_WIDTH): New macro.
769 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
770 (write_format): Add arg to take width to resize format to.
771 (write_value): Limit width of value written to MAX_POR_WIDTH.
772 (write_variables): Limit width of variable and its output formats
774 Reviewed by John Darrington.
776 2007-07-22 Ben Pfaff <blp@gnu.org>
778 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
779 local variable instead of literal 5.
780 Reviewed by John Darrington.
782 2007-07-22 Ben Pfaff <blp@gnu.org>
784 Fix problems with uniqueness of short names in system files with
785 very long string variables. Now a variable may have multiple
788 * automake.mk (src_data_libdata_a_SOURCES): Add new files
789 short-names.c, short-names.h.
791 * dictionary.c (dict_clone): Clone all the short names.
792 (compare_strings): Move into short-names.c.
793 (hash_strings): Ditto.
794 (set_var_short_name_suffix): Ditto.
795 (dict_assign_short_names): Ditto, rename short_names_assign,
796 change to assign all short names.
798 * por-file-writer.c (write_variables): Use short_names_assign
799 instead of dict_assign_short_names.
801 * short-names.c: New file.
803 * short-names.h: New file.
805 * sys-file-private.c (sfm_width_to_segments): New function.
807 * sys-file-reader.c (read_long_var_name_map): Save and restore all
808 the short names, not just the first one.
810 * sys-file-writer.c (cont_var_name): Removed.
811 (sfm_open_writer): Use short_names_assign instead of
812 dict_assign_short_names. Use unique short names assigned by
813 short_names_assign instead of those generated by cont_var_name.
815 * variable.c (struct variable): Remove `short_name' member,
816 replace by `short_names' and `short_name_cnt'.
817 (var_create) Initialize new members.
818 (var_get_short_name_cnt): New function.
819 (var_get_short_name): Now takes an index argument. Changed most
821 (var_set_short_name): Ditto.
822 (var_clear_short_name): Renamed var_clear_short_names, changed to
823 clear all short names.
825 Reviewed by John Darrington.
827 2007-07-22 Ben Pfaff <blp@gnu.org>
829 * variable.c (var_set_width): Use new var_set_width function.
831 * missing-values.c (mv_n_values): Drop assertion, which was not
834 * format.c (fmt_default_for_width): New function.
835 (fmt_resize): New function.
837 Reviewed by John Darrington.
839 2007-07-18 John Darrington <john@darrington.wattle.id.au>
841 * datasheet.c (datasheet_delete_columns): Added assertion to check
842 we're not deleting outside the range of the sheet.
845 * dictionary.c dictionary.h variable.c: Added the ability for string
846 variables to be resized.
848 * vardict.h: Added some prototypes (moved from dictionary.h) as
849 these should only be called by variable.c
852 2007-07-14 John Darrington <john@darrington.wattle.id.au>
854 * sfm-reader.c: Respect case_cnt field in file header.
856 2007-07-01 John Darrington <john@darrington.wattle.id.au>
858 * transformation.c transformation.h (trns_chain_execute): Changed the
859 signature (Patch #6057)
861 2007-06-10 Ben Pfaff <blp@gnu.org>
863 * casereader-filter.c (casereader_filter_destroy): Make sure to
864 write all the remaining excluded cases to the casewriter, if any.
866 * caseinit.c (init_list_destroy): Rewrite.
867 (init_list_clear): Ditto.
869 * casegrouper.c (casegrouper_get_next_group): Always set *reader
870 to null when returning false.
872 2007-06-06 Ben Pfaff <blp@gnu.org>
874 Actually implement the new procedure code and adapt all of its
875 clients to match. Also adapt all of the other case sources and
876 sinks in the tree and their clients to use the
877 casereader/casewriter infrastructure.
879 * automake.mk: Add and remove files.
881 * any-reader.c: Change into a casereader.
882 * por-file-reader.c: Ditto.
883 * scratch-reader.c: Ditto.
884 * sys-file-reader.c: Ditto.
886 * any-writer.c: Change into a casewriter.
887 * por-file-writer.c: Ditto.
888 * scratch-writer.c: Ditto.
889 * sys-file-writer.c: Ditto.
891 * procedure.c: Change to use casereader, casewriter, caseinit, and
892 other new infrastructure.
894 * scratch-handle.c: Adapt to new infrastructure.
896 * case-sink.c: Removed, now dead code.
897 * case-sink.h: Ditto.
898 * case-source.c: Ditto.
899 * case-source.h: Ditto.
900 * casefile-factory.c: Ditto.
901 * casefile-private.h: Ditto.
904 * casefilter.c: Ditto.
905 * casefilter.h: Ditto.
908 * fastfile-factory.c: Ditto.
909 * fastfile-factory.h: Ditto.
910 * storage-stream.c: Ditto.
911 * storage-stream.h: Ditto.
913 2007-06-06 Ben Pfaff <blp@gnu.org>
917 * automake.mk: Add new files.
919 * datasheet.c: New file.
921 * datasheet.h: New file.
923 2007-06-06 Ben Pfaff <blp@gnu.org>
925 Until now, the procedure code has provided a case to the
926 case_source, which has filled in the data values that come from
927 the active file. "Left" data values that don't come from the
928 active file naturally stay the same from case to case, because the
929 procedure code keeps using that same case.
931 One of the compromises that comes with the new procedure code is
932 that the active file allocates and provides its own case, which
933 the procedure code then has to resize to provide room for any
934 other variables that should go in the case and then fill in the
935 values of "left" variables. Then, when we're done with that case,
936 we have to save the values of "left" variables to copy into the
937 next case read from the active file.
939 The caseinit code helps with this.
941 * automake.mk: Add new files.
943 * caseinit.c: New file.
945 * caseinit.h: New file.
947 2007-06-06 Ben Pfaff <blp@gnu.org>
949 * value.h (value_cnt_from_width): New function.
951 * variable.c (var_get_value_cnt): Use new function.
953 2007-06-06 Ben Pfaff <blp@gnu.org>
955 Add casegrouper, to allow cases read from a given casereader to be
956 broken into groups, each of which has its own casereader.
957 Generally cases are grouped based on having equal values for some
960 * automake.mk: Add new files.
962 * casegrouper.c: New file.
964 * casegrouper.h: New file.
966 2007-06-06 Ben Pfaff <blp@gnu.org>
968 Add interface to lexicographical ordering of cases.
970 * automake.mk: Add new files.
972 * case-ordering.c: New file.
974 * case-ordering.h: New file.
976 2007-06-06 Ben Pfaff <blp@gnu.org>
978 Add casereaders and casewriters, the basis of the new data processing
979 implementation. A casereader is a uniform interface to reading cases
980 from a data source; a casewriter is a uniform interface to writing
981 cases to a data sink.
983 * automake.mk: Add new files.
985 * casereader-filter.c: New file.
987 * casereader-provider.h: New file.
989 * casereader-translator.c: New file.
991 * casereader.c: New file.
993 * casereader.h: New file.
995 * casewriter-provider.h: New file.
997 * casewriter-translator.c: New file.
999 * casewriter.c: New file.
1001 * casewriter.h: New file.
1003 2007-06-06 Ben Pfaff <blp@gnu.org>
1005 "casewindow" data structure that extends the deque (from libpspp)
1006 of cases with the ability to dump cases to disk if we get too many
1009 * automake.mk: Add new files.
1011 * casewindow.c: New file.
1013 * casewindow.h: New file.
1015 2007-06-06 Ben Pfaff <blp@gnu.org>
1017 sparse_cases data structure that augments a sparse_array of cases
1018 with the ability to dump cases to disk if we get too many cases in
1021 * automake.mk: Add new files.
1023 * sparse-cases.c: New file.
1025 * sparse-cases.h: New file.
1027 2007-06-06 Ben Pfaff <blp@gnu.org>
1029 Adds a low-level on-disk case array data structure.
1031 * automake.mk: Add new files.
1033 * case-tmpfile.c: New file.
1035 * case-tmpfile.h: New file.
1037 2007-06-06 Ben Pfaff <blp@gnu.org>
1039 In a couple of places we calculate the maximum number of cases to
1040 keep in memory based on the user-defined workspace. Enable
1041 centralizing the calculation through a new function.
1043 * settings.c (get_workspace_cases): New function.
1045 2007-06-06 Ben Pfaff <blp@gnu.org>
1047 The casenumber type is defined in transformations.h, but case.h is
1048 a more sensible place. Move it.
1050 * case.h (CASENUMBER_MAX): New macro.
1051 (typedef casenumber): Move here, from transformations.h.
1053 2007-06-03 Ben Pfaff <blp@gnu.org>
1055 Slightly generalize case_to_values and case_from_values functions.
1057 * case.c (case_to_values): Rename case_copy_out, change interface.
1058 (case_from_values): Rename case_copy_in, change interface.
1060 * fastfile.c (fastfilereader_get_next_case): Update caller.
1061 (write_case_to_disk): Ditto.
1063 2007-06-02 Ben Pfaff <blp@gnu.org>
1065 Clean up after a forgotten part of patch #5829.
1067 * casedeque.h: Remove unused file.
1069 * automake.mk: Remove casedeque.h from sources.
1071 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1073 * category.c: Removed redundant #include
1075 2007-05-06 Ben Pfaff <blp@gnu.org>
1077 Abstract the documents within a dictionary a little better.
1078 Thanks to John Darrington for suggestion, initial version, and
1079 review. Patch #5917.
1081 * dictionary.c (struct dictionary): Change `documents' member from
1082 char * to struct string.
1083 (dict_clear): Destroy struct string.
1084 (dict_get_documents): Convert struct string to char *.
1085 (dict_set_documents): Set struct string. Pad to 80-character
1087 (dict_clear_documents): New function.
1088 (dict_add_document_line): New function.
1089 (dict_get_document_line_cnt): New function.
1090 (dict_get_document_line): New function.
1092 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1094 * sys-file-reader.c (read_documents): Use new document functions.
1096 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1098 * sys-file-reader.c: When reading a system file which has no
1099 long name table, automatically create one where the long names
1100 are the lower case versions of the short names.
1102 2007-04-22 Ben Pfaff <blp@gnu.org>
1104 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1105 dict_clear will free most data related to the dictionary.
1106 dict_clear does a decent job, except that dict_set_split_vars on
1107 some systems won't actually free the dict's "split" member.
1108 Instead, it'll allocate a 1-byte region. Fix this.
1110 * value.c (value_copy): New function.
1111 (value_set_missing): Ditto.
1113 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1115 * Deleted existing category.h and moved cat-routines.h into
1116 category.h Encapsulated struct cat_vals better.
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 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1126 * sys-file-reader.c: Some versions of Other Software seem to
1127 produce system files with string variables' measure set to
1128 zero. We'll assume these are supposed to be nominal variables.
1130 2007-03-30 Ben Pfaff <blp@gnu.org>
1132 * procedure.c: Adapt to new deque data structure.
1134 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1135 Ben Pfaff <blp@gnu.org>
1137 * file-name.c: Mingw compatibility fixes.
1138 (fn_search_path): Use ISSLASH instead of comparing against '/'
1140 (fn_dir_name): Use dir_name from gnulib.
1141 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1142 (fn_get_identity): Use GetFullPathName instead of canonicalize
1143 from gnulib, because the latter does not fully support
1144 Windows-style path names. Use this implementation based on the
1145 detected presence of Windows instead of the absence of Unix, since
1146 the new implementation is Windows-specific.
1147 (fn_compare_file_identities): In Windows implementation, compare
1148 names case-insensitively.
1150 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1152 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1153 <stdlib.h> to have the same effect.
1155 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1157 * por-file-reader.c: Add missing _() around messages.
1159 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1161 * make-file.c: Include "mkstemp.h", without which linking on
1164 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1166 Reduce platform dependence.
1168 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1169 Everywhere we using this, we really should have just depended on
1170 the shell to expand tildes.
1171 (fn_search_path): Simplify, given that we don't do tilde expansion
1173 (fn_normalize): Removed. Caller changed to use the canonicalize
1175 (fn_get_cwd): Removed. Only user was fn_normalize.
1176 (fn_is_absolute): Really only test for absolute names.
1177 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1179 (fn_readlink): Removed, as it was only used fn_normalize.
1180 (fn_exists): Assume the stat function is available; gnulib does.
1181 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1183 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1185 * dictionary.c (dict_create_vector_assert): New function.
1187 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1189 * file-name.c (fn_normalize): Correct name of function
1190 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1191 for pointing this out.
1193 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1195 We are using a single member in struct file_handle, the "name"
1196 field, for more than one purpose. When it begins with '"', it's a
1197 file name; otherwise, it's a token that can be used to identify
1198 it. When that assertion fires, it's because we searched for the
1199 name case-sensitively as a file name (so that there was no match),
1200 and then we try to insert it case-insensitively as a token, which
1201 fails because duplicates aren't allowed.
1203 Solution: break the two purposes into two separate fields. This
1204 fixes the problem and likely makes the code easier to read too.
1206 Fixes bug #18922. Thanks to John Darrington for bug report and
1209 * file-handle-def.c (struct file_handle): New `id' member.
1210 (fh_from_name): Rename fh_from_id. Update all callers.
1211 (create_handle): New `id' parameter. Update all callers.
1212 (fh_create_file): Ditto.
1213 (fh_get_id): New function.
1215 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1217 * case.c (case_is_null): Change return type to bool.
1219 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1223 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1224 reference counting to share data. This makes it easy for
1225 valgrind, etc. to find accesses to cases that have been destroyed
1226 but have been kept around by another user's ref-count. This often
1227 happens when the data set is small enough to find in memory; if a
1228 bigger data set that would overflow to disk were used, then data
1229 corruption would occur.
1231 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1235 * case.c (case_unshare): Make it check internally whether the
1236 ref_cnt is greater than 1, so that the callers don't have to.
1237 Update callers not to check.
1239 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1241 Before, I was thinking that I might want to get rid of reference
1242 counting at some point. Now, I'm pretty sure that it's here to
1243 stay. Thus, because we have to store the value_cnt anyway for
1244 reference-counted cases, we might as well expose it to users.
1246 * case.c (case_get_value_cnt): New function.
1247 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1248 resize case if its size actually changed.
1250 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1251 instead of peeking inside struct case directly.
1252 (casefile_append): Ditto.
1254 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1256 Get rid of the inlines for the case functions, which made the
1257 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1258 the inlines didn't speed up "make check" at all, which is not a
1259 perfect benchmark but seems indicative.)
1261 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1262 definitions. Remove some assertions on nonnull pointers that were
1263 redundant with a pointer dereference soon after in the function.
1265 (struct case_data): Move definition here from case.h.
1269 (case_data_wr): Ditto.
1271 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1273 * automake.mk: Add casedeque.h to sources.
1275 * casedeque.h: New file.
1277 * procedure.c: (struct dataset) Change lag_count, lag_head,
1278 lag_queue member into single struct casedeque member. Update all
1279 users to use the casedeque instead.
1282 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1284 * procedure.c: Simplify lagged cases interface. Updated all
1285 clients--well, the only client--to use the simplified interface.
1286 (dataset_n_lag) Removed.
1287 (dataset_set_n_lag) Removed.
1288 (dataset_need_lag) New function.
1290 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1292 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1293 whenever appropriate, but only when the dataset/dictionary is in a
1296 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1298 * dictionary.c dictionary.h : Added callbacks for change of filter and
1299 split variables. Refactored some code to ensure that callbacks get
1300 invoked when appropriate.
1302 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1303 replace_dict callback occurs when permanent_dict replaces the current
1306 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1308 * dictionary.c dictionary.h : Added callback for when the weight
1309 variable of a dictionary changes.
1311 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1313 * dictionary.c dictionary.h : Added replace_source and replace_dict
1314 callbacks, and functions to deal with them.
1316 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1318 Simplify missing value handling.
1320 * missing-values.h (enum mv_class): New type.
1321 (enum mv_type): Moved definition into missing-values.c and renamed
1322 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1323 enums. Updated all uses.
1324 (struct missing_values): Changed type of `type' from `enum
1325 mv_type' to `int' because the definition is no longer exposed.
1327 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1328 parameter. Update all callers.
1329 (mv_is_num_missing): Ditto.
1330 (mv_is_str_missing): Ditto.
1331 (mv_is_value_user_missing): Removed. Changed callers to use
1332 mv_is_value_missing.
1333 (mv_is_num_user_missing): Removed. Changed callers to use
1335 (mv_is_str_user_missing): Removed. Changed callers to use
1337 (mv_is_value_system_missing): Removed. Changed callers to use
1338 mv_is_value_missing.
1339 (mv_set_type): Removed. Changed callers to use mv_clear.
1340 (mv_clear): New function.
1342 * variable.c (var_is_value_missing): Add new enum mv_class
1343 parameter. Update all callers.
1344 (var_is_num_missing): Ditto.
1345 (var_is_str_missing): Ditto.
1346 (var_is_value_user_missing): Removed. Changed callers to use
1347 var_is_value_missing.
1348 (var_is_num_user_missing): Removed. Changed callers to use
1350 (var_is_str_user_missing): Removed. Changed callers to use
1352 (var_is_value_system_missing): Removed. Changed callers to use
1353 var_is_value_missing.
1355 * casefilter.c (struct casefilter): Use enum mv_class in place of
1357 (casefilter_variable_missing): Adapt to new member.
1358 (casefilter_create): Change signature to take enum mv_class,
1361 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1363 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1365 * case-sink.c case-sink.h procedure.c procedure.h
1366 storage-stream.c: Now uses the factory.
1368 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1370 Make it possible to pull cases from the active file with a
1371 function call, instead of requiring indirection through a callback
1374 * case-source.h (struct case_source_class): Change ->read function
1375 to return a single case, instead of calling a callback function
1376 for each case. Change ->destroy function to return an error
1379 * case-source.c (free_case_source): Pass along the value returned
1380 by the case_source ->destroy function.
1382 * procedure.c (struct write_case_data): Removed.
1383 (struct dataset): Added some members to track procedure state.
1384 (procedure): Optimize the trivial case at this level.
1385 (internal_procedure): Re-implement in terms of proc_open,
1386 proc_read, proc_close.
1387 (proc_open) New function.
1388 (proc_read) New function.
1389 (proc_close) New function.
1390 (write_case) Moved into proc_read.
1391 (close_active_file) Moved closing of data source into proc_close.
1393 * storage-source.c: Rewrote to conform with modified
1394 case_source_class interface.
1396 * transformations.c (trns_chain_execute): Added argument to allow
1397 starting execution from an arbitrary transformation. Updated
1400 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1402 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1404 * sys-file-reader.c (read_display_parameters): Don't assume that
1405 MEASURE_* and ALIGN_* have the same values found in system files.
1407 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1409 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1410 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1411 the value of n_MEASURES, which was off by 1 (at least from my
1414 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1416 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1417 callbacks which are invoked when the dictionary or its
1418 variables are changed.
1420 * missing-values.c missing-values.h value-labels.c: Tidied up
1421 consistency checks, and made some of them return false
1422 instead of assert-failing.
1424 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1426 * calendar.c (calendar_days_in_month): New function.
1428 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1430 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1431 the value label's width.
1433 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1435 * sfm-private.h: Move contents into sys-file-writer.c, which is
1436 the only remaining user. Removed Borland C++-specific directives.
1438 * sys-file-reader.c: Clean up and rewrite entire file. The
1439 rewritten version is simpler and better abstracted, and should be
1440 easier to maintain and extend. It avoids using structures to read
1441 file data, which is prone to padding variations among compilers.
1442 It should also handle non-IEEE 754 system files, although I
1443 haven't been able to find any. It has been tested against many
1444 .sav files obtained from the Web and found to produce the same
1445 results as the earlier version of the code, or in some cases
1446 improved results. It is more tolerant of format variations found
1449 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1450 member, putting an enum integer_format in its place. New member
1451 `float_format'. Changed `compressed' member to type bool.
1453 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1455 * dictionary.c (dict_delete_consecutive_vars): New function.
1457 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1459 * file-name.c (fn_search_path): Remove prefix arg that was unused
1460 by any caller. Updated all callers.
1462 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1464 * format.c (fmt_dollar_template): Use user's decimal point
1465 character. Add assertion.
1467 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1469 * format.c (fmt_dollar_template): New function, based on
1470 dollar_format_template from var-type-dialog.c.
1472 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1474 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1477 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1479 Finish converting struct variable to an opaque type. In this
1480 phase, we add remaining setter and getter functions, convert the
1481 remaining PSPP code to use them, and do a bunch of cleanup. The
1482 resulting changes are pervasive but mostly trivial, and only the
1483 notable changes are logged.
1485 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1488 * case.c (case_data): Renamed case_data_idx.
1489 (case_num): Renamed case_num_idx.
1490 (case_str): Renamed case_str_idx.
1491 (case_data_rw): Renamed case_data_rw_idx.
1493 * case.h (case_data): New function with old name and an interface
1494 that takes a variable instead of an index, which is easier to
1495 use. Updated all callers to use the new interface, or to use the
1496 new *_idx function (see above).
1499 (case_data_rw): Ditto.
1501 * category.c (cat_stored_values_destroy): Changed interface to
1502 take a struct cat_vals * instead of a struct variable *.
1504 * dictionary.c (dict_clone): Use new vector_clone function.
1505 (dict_clear) Use new var_destroy function.
1506 (add_var) New function.
1507 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1508 (dict_create_var_assert) Rewrite in terms of add_var.
1509 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1510 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1511 (dict_lookup_var) Use new var_create, var_destroy functions.
1512 (dict_contains_var) Rewrite in terms of new vardict functionality.
1513 (set_var_dict_index) New function.
1514 (set_var_case_index) New function.
1515 (reindex_vars) New function.
1516 (dict_delete_var) Rewrite in terms of new vardict functionality.
1517 (dict_reorder_var) Ditto.
1518 (dict_reorder_vars) Ditto.
1519 (rename_var) New function.
1520 (dict_rename_var) Use rename_var.
1521 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1522 (dict_get_compacted_idx_to_fv) Rename
1523 dict_get_compacted_dict_index_to_case_index, update callers.
1524 (dict_create_vector) Use new vector_create function.
1525 (dict_clear_vectors) Use new vector_destroy function.
1526 (set_var_short_name_suffix) Move here from variable.c, renamed
1527 from var_set_short_name_suffix, make static, update caller.
1529 * sys-file-private.c: New file.
1530 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1531 width_to_bytes, update callers.
1533 * sys-file-private.h: New file. Later it will supplant
1534 sfm-private.h; for now it supplements it.
1535 (macro MIN_VERY_LONG_STRING) New macro.
1536 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1538 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1539 MAX_LONG_STRING was used before.
1541 * sys-file-writer.c: Ditto.
1543 * value-labels.c: Change the paradigm here to be that a null
1544 pointer is OK for a struct val_labs * in most cases; it just
1545 represents an empty set of value labels.
1546 (val_labs_copy) A copy of a null set is a null set.
1547 (val_labs_count) A null set has 0 labels.
1548 (val_labs_replace) Change return type to void. Rewrite for
1550 (val_labs_find) A null set does not contain the value.
1551 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1552 transposed argument order, updated all callers.
1554 * value.c: New file.
1555 (value_dup) Moved here from variable.c.
1556 (compare_values) Ditto.
1559 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1560 (macro MAX_LONG_STRING) Removed, because it was only interesting
1561 for system files, not for general code.
1562 (macro MAX_VERY_LONG_STRING) Ditto.
1563 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1564 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1566 * vardict.h: New file, for an interface between variables and
1569 * variable.c: A lot of functions were moved around, for better
1571 (struct variable) Move definition here, from variable.h.
1572 (var_type_adj) Removed--makes i18n hard.
1573 (var_type_noun) Ditto.
1574 (var_create) New function.
1575 (var_clone) New function.
1576 (var_destroy) New function.
1577 (var_set_name) Assert that variable is not in a dictionary.
1578 (compare_var_names) Rename compare_vars_by_name and fix a couple
1579 of callers who thought the args were strings.
1580 (hash_var_name) Rename hash_var_by_name.
1581 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1582 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1583 (var_is_very_long_string) Removed, because it was only interesting
1584 to system file code.
1585 (var_set_missing_values) Allow the argument to be the wrong width,
1586 as long as we can resize it. Simplify callers who were doing the
1587 resizing themselves.
1588 (var_get_value_labels) New function.
1589 (var_has_value_labels) New function.
1590 (var_set_value_labels) New function.
1591 (alloc_value_labels) New function.
1592 (var_add_value_label) New function.
1593 (var_replace_value_label) New function.
1594 (var_clear_value_labels) New function.
1595 (var_lookup_value_label) New function.
1596 (var_get_value_name) Moved here from variable.c, renamed from
1597 var_get_value_name, transposed argument order, updated all
1599 (var_to_string) Moved here, from variable-label.c.
1600 (var_set_leave) New function.
1601 (var_get_leave) New function.
1602 (var_must_leave) New function.
1603 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1604 set_var_short_name_suffix.
1605 (var_get_dict_index) New function.
1606 (var_get_case_index) New function.
1607 (var_get_obs_vals) New function.
1608 (var_set_obs_vals) New function.
1609 (var_has_obs_vals) New function.
1610 (var_get_vardict) New function.
1611 (var_set_vardict) New function.
1612 (var_has_vardict) New function.
1613 (var_clear_vardict) New function.
1614 (value_dup) Moved to value.c.
1615 (compare_values) Ditto.
1618 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1619 (enum ALPHA) Rename VAR_STRING, update all users.
1621 * vector.c: New file.
1622 (struct vector) Moved here, from variable.h.
1623 (check_widths) New function.
1624 (vector_create) New function.
1625 (vector_clone) New function.
1626 (vector_destroy) New function.
1627 (vector_get_name) New function.
1628 (vector_get_var) New function.
1629 (vector_get_var_cnt) New function.
1630 (compare_vector_ptrs_by_name) New function.
1632 * vector.h: New file.
1634 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1636 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1637 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1640 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1642 * value-labels.c (destroy_atoms): New function.
1643 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1645 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1647 Thanks to Jason Stover for pointing out this problem.
1649 * data-out.c (output_number): Use gsl_finite from GSL, which is
1650 portable, instead of isfinite, which is not.
1653 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1655 * variable.c variable.h (value_dup): New function.
1657 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1659 Start converting struct variable to an opaque type. In this
1660 phase, we add a bunch of setter and getter functions and convert
1661 most of the PSPP code to use them. The resulting changes are
1662 pervasive but mostly trivial, and only the notable changes are
1665 * format.c (fmt_equal): New function.
1667 * variable.c (var_type_is_valid): New function.
1668 (measure_is_valid) Moved here, from format.c.
1669 (alignment_is_valid) Moved here, from format.c.
1670 (var_get_name) New function.
1671 (var_set_name) New function.
1672 (width_to_type) New function.
1673 (var_get_type) New function.
1674 (var_get_width) New function.
1675 (var_set_width) New function.
1676 (var_is_numeric) New function.
1677 (var_is_alpha) New function.
1678 (var_is_short_string) New function.
1679 (var_is_long_string) New function.
1680 (var_is_very_long_string) New function.
1681 (var_get_missing_values) New function.
1682 (var_set_missing_values) New function.
1683 (var_clear_missing_values) New function.
1684 (var_has_missing_values) New function.
1685 (var_is_value_missing) New function.
1686 (var_is_num_missing) New function.
1687 (var_is_str_missing) New function.
1688 (var_is_value_user_missing) New function.
1689 (var_is_num_user_missing) New function.
1690 (var_is_str_user_missing) New function.
1691 (var_is_value_system_missing) New function.
1692 (var_get_print_format) New function.
1693 (var_set_print_format) New function.
1694 (var_get_write_format) New function.
1695 (var_set_write_format) New function.
1696 (var_set_both_formats) New function.
1697 (var_get_label) New function.
1698 (var_set_label) New function.
1699 (var_clear_label) New function.
1700 (var_has_label) New function.
1701 (var_get_measure) New function.
1702 (var_set_measure) New function.
1703 (var_get_display_width) New function.
1704 (var_set_display_width) New function.
1705 (var_get_alignment) New function.
1706 (var_set_alignment) New function.
1707 (var_get_value_cnt) New function.
1708 (var_get_leave) New function.
1709 (var_get_short_name) New function.
1711 * variable.h: (struct variable) Removed "type" and "nv" members;
1712 they are now computed from "width" where needed.
1714 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1716 * missing-values.c (mv_resize): Don't write beyond end of the
1717 allocated buffer when resizing a long string.
1719 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1721 Clean up identifier code: don't require identifier enumerations to
1722 be in a particular order; make better use of string library;
1723 expose less of the internals.
1725 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1726 change interface. Updated all callers.
1727 (lex_id_match) Change interface to use struct substring, update
1729 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1730 (global array keywords[]) Make static, change form. Update all
1731 users to use lex_id_name instead.
1732 (lex_is_keyword) New function.
1733 (lex_id_to_token) Change interface to use struct substring, update
1735 (lex_id_name) New function.
1737 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1738 lex_is_keyword instead.
1739 (T_LAST_KEYWORD) Removed.
1740 (T_N_KEYWORDS) Removed.
1742 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1744 * format.c: (fmt_date_template) Distinguish characters for which a
1745 space is output and any date delimiter is allowed on input, from
1746 those for which a space is output and only a space is allowed on
1747 input. The former is represented by X, the latter by a space.
1748 Also, drop distinction between h and H, changing the former to the
1751 * data-in.c: Completely rewrite internals to conform to SPSS input
1752 formats as closely as possible.
1753 (data_in) Changed external interface by replacing the structure
1754 that was used as a single argument by a set of arguments. Updated
1756 (data_in_finite_line) Removed. Converted all callers to use plain
1758 (data_in_get_integer_format) New function.
1759 (data_in_set_integer_format) New function.
1760 (data_in_get_float_format) New function.
1761 (data_in_set_float_format) New function.
1763 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1764 (struct data_in) Removed.
1766 * data-out.c: (output_date) Drop each component from the input as
1767 it is output, to allow us to drop the distinction between h (a
1768 count of hours) and H (the hour of day) template characters.
1769 Also, handle new X template character.
1770 (output_scientific) Follow more rational rule on when to drop
1771 fraction introduced between SPSS 13 and 15. Updated test case to
1774 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1776 Fix buffer overflow reported by John Darrington.
1778 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1779 realize that DIGITS is a count of nibbles, not of bytes.
1781 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1783 * calendar.c (calendar_offset_to_gregorian) Also return the
1784 year-of-day. Change callers to new interface.
1786 * data-out.c: Completely rewrite internals to conform to SPSS
1787 output formats as completely as possible.
1788 (data_out) Change interface to put input parameters before output
1789 parameters, for consistency with the style I now prefer. Update
1791 (data_out_get_integer_format) New public function.
1792 (data_out_set_integer_format) New public function.
1793 (data_out_get_float_format) New public function.
1794 (data_out_set_float_format) New public function.
1796 * data-out.h: New file. Move prototype for data_out here, from
1799 * format.c: (fmt_step_width) Use equality comparison instead of
1800 bitwise and, for clarity.
1801 (fmt_is_string) Ditto.
1802 (fmt_input_to_output) Fix categories that are translated to F
1805 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1807 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1808 casefile-private.h: Added casefilter to assist commands with missing
1811 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1813 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1815 * settings.c: (route_errors_to_terminal) New variable.
1816 (route_errors_to_listing) New variable.
1817 (get_error_routing_to_terminal) New function.
1818 (set_error_routing_to_terminal) New function.
1819 (get_error_routing_to_listing) New function.
1820 (set_error_routing_to_listing) New function.
1822 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1824 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1826 * format.c: Completely rewrite, to achieve better abstraction.
1827 Rewrite all references to formats in other files.
1829 * format.def: Rewrite and reorganize.
1831 * settings.c: Move everything related to custom currency formats
1832 into format.[ch], changing them in form, so as to group related
1833 code and definitions better. Changed all references to use the
1835 (static var decimal) Removed.
1836 (static var grouping) Removed.
1837 (static var cc) Removed.
1838 (get_decimal) Removed.
1839 (set_decimal) Removed.
1840 (get_grouping) Removed.
1841 (set_grouping) Removed.
1845 * settings.h: (macro CC_CNT) Removed.
1846 (macro CC_WIDTH) Removed.
1847 (struct custom_currency) Removed.
1849 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1851 * data-in.c (data_in): Use switch statement instead of table, to
1852 avoid dependence on the order of the FMT_* enums.
1854 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1856 * data-out.c: (num_to_string) Removed, because it was dead code.
1858 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1860 * data-in.c (parse_trailer): Fix error message.
1862 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1864 * format.c (fmt_is_binary): New function.
1866 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1868 * procedure.c procedure.h: Encapsulated the static data into a single
1871 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1873 * casefile.c (casereader_read_xfer): Always initialize the case,
1874 even on an error condition.
1876 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1878 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1879 entire function useless.
1881 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1883 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1884 casereader method casereader_clone.
1886 * procedure.c pransformations.h: Introduced new type casenum_t
1888 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1890 * variable.c: (width_to_bytes) Rephrase code for clarify.
1892 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1894 * format.c: (fmt_type_from_string) New function.
1895 (fmt_to_string) Include decimals in output if the format has
1896 decimals, even if the format type does not. This way, we can
1897 accurately reproduce incorrect formats in user output.
1898 (check_common_specifier) Make the check for a bad format type an
1899 assertion, so we get bug reports if they show up. Fix message.
1900 Check for decimal places with a format type that doesn't allow
1902 (check_input_specifier) Remove check for FMT_X, which has been
1904 (check_output_specifier) Ditto.
1906 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1908 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1909 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1910 (enum fmt_parse_flags) Removed.
1912 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1914 * casefile.c casefile.h: Converted to an abstract base class.
1915 * casefile-private.h fastfile.c fastfile.h: New files.
1916 * automake.mk procedure.c scratch-writer.c storage-stream.c
1918 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1920 * procedure.c (internal_procedure): Create sink_case with only as
1921 many values as the compacted dictionary.
1923 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1925 Remove "debugging" code that caused plenty of false positives and
1928 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1930 * case.c: Remove all references to `this' member.
1932 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1934 Fix link error noted by Jason Stover.
1936 * storage-stream.c: Include <assert.h>.
1938 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1940 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1941 ALL) and additional underlying system file issues.
1943 Thanks to John Darrington for review.
1945 First problem: var_hash points to variables not owned by the
1946 sys-file-reader, which the caller may free or modify. Use an
1947 array of sfm_vars instead, as done earlier (e.g. CVS version
1950 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1951 members and remove all code that references it. Add vars, var_cnt
1952 members. Remove fix_specials member, which was unused.
1953 (struct sfm_var) Remove name member, which was unused.
1954 (sfm_close_reader) Free vars member instead of var_hash.
1955 (compare_var_shortnames) Removed.
1956 (hash_var_shortname) Removed.
1957 (sfm_open_reader) Fill out vars array.
1958 (compare_var_index) Removed.
1959 (sfm_read_case) Use vars instead of var_hash.
1961 Second problem: we're confused about when we actually have very
1962 long strings, causing us to choose incorrectly between slow path
1963 and fast path in sfm_read_case.
1965 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1966 have very long strings, not when we have long variable names,
1967 which is an unrelated feature.
1969 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1971 * variable.h: Move var_set and variable parsing declarations to
1972 new header, src/language/lexer/variable-parser.h. Modified lots
1973 of files to include the new header.
1975 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1977 * value-labels.c (value_to_string): When there's no value label,
1978 format the variable according to its print format, instead of
1979 always effectively using A or F format.
1981 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1983 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1986 * format.c: Removed tortological assertion.
1988 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1990 Reform string library.
1992 * file-name.c (fn_interp_vars): Change interface to take a
1993 substring as input. Updated all users.
1995 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1997 * format.c (measure_is_valid): Really return false when m >=
2000 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
2002 Implement random access to casefiles, for use in GUI.
2004 * casefile.c: (struct casereader) Add `random', `file_ofs',
2005 `buffer_ofs' members.
2006 (casefile_get_random_reader) New function.
2007 (read_open_file) Break part into new function
2008 seek_and_fill_buffer().
2009 (fill_buffer) Update buffer_ofs, file_ofs.
2010 (casereader_seek) New function.
2012 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2014 * settings.c: Added call to i18n{done, init}.
2016 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
2018 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
2021 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
2023 * casefile.c: Convert many uses of `int' to `bool'.
2025 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
2027 * transformations.c (trns_chain_destroy): Destroy chain's trns
2028 member, to fix memory leak.
2030 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
2032 * storage-stream.c (storage_source_decapsulate): Destroy case
2033 source to fix memory leak.
2035 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
2037 * scratch-reader.c (scratch_reader_read_case): Copy into existing
2038 case passed as argument instead of initializing the argument as a
2039 case. Fixes memory leak that showed up in
2040 tests/command/aggregate.sh with scratch files.
2042 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
2044 * procedure.c (proc_done): Destroy default_dict, to fix memory
2047 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2049 Simplify procedure_with_splits().
2051 * procedure.c (struct split_aux_data): Removed case_count member.
2052 (procedure_with_splits) Don't initialize case_count.
2053 (split_procedure_case_func) Check whether prev_case is null
2054 instead of case_count.
2055 (split_procedure_end_func) Ditto.
2057 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2059 * case.c (case_move): Do nothing if dst and src are the same
2061 (case_try_create) Merge two similar cases.
2062 (case_copy) Unshare only if data must be actually copied.
2064 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2066 * data-in.c data-out.c dictionary.c sys-file-reader.c
2067 sys-file-writer.c variable.c variable.h: Reworked very long string
2068 support for better encapsulation.
2070 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2072 * value-labels.c (val_labs_can_set_width): New function.
2073 (val_labs_set_width) Clear labels if increasing width to long
2075 (val_labs_destroy) Remove unneeded test for null.
2077 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2079 * value-labels.h: Remove unneeded dependency on variable.h.
2081 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2083 Get rid of `char *c' member in union value, for cleanliness.
2085 * value.h: (union value) Remove `c' member.
2087 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2089 Make dictionary compacting functions a little more general.
2091 * sys-file-writer.c (sfm_open_writer): Use
2092 dict_compacting_would_change().
2093 (does_dict_need_translation) Removed.
2095 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2097 Make dictionary compacting functions a little more general.
2099 * dictionary.c (dict_needs_compaction): Rename
2100 dict_compacting_would_shrink(). Update all callers.
2101 (dict_compacting_would_change) New function.
2103 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2105 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2106 inverted return value (!).
2108 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2110 Continue reforming procedure execution.
2112 * procedure.c: Search and replace "vfm" by "proc". Notably:
2113 (static var vfm_source) Rename proc_source. Update all
2115 (static var vfm_sink) Rename proc_sink. Update all references.
2117 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2119 Continue reforming procedure execution. In this phase, remove
2120 PROCESS IF, which was deprecated anyway and can be easily
2121 simulated with TEMPORARY followed by SELECT IF.
2123 * procedure.c: (open_active_file) Don't call
2124 add_process_if_trns().
2125 (discard_variables) Get rid of redundant call to
2126 proc_cancel_all_transformations().
2127 (add_process_if_trns) Removed.
2128 (process_if_trns_proc) Removed.
2129 (process_if_trns_free) Removed.
2131 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2133 Continue reforming procedure execution. In this phase, add
2134 `const' to the case passed to procedure()'s callback.
2136 Updated all users of procedure() as well.
2138 * procedure.c: (struct write_case_data) Add "const" to ccase
2139 parameter for case_func member.
2140 (procedure) Add "const" to ccase parameter for proc_func
2142 (multipass_case_func) Make ccase parameter const.
2143 (internal_procedure) Add "const" to ccase parameter for case_func
2145 (split_procedure_case_func) Make ccase parameter const.
2146 (multipass_split_case_func) Make ccase parameter const.
2148 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2150 Continue reforming procedure execution. In this phase, get rid of
2151 the output code for SPLIT FILE groups in procedure.c, which really
2152 shouldn't be doing any output. Move it into the individual
2153 procedures instead. This also adds some flexibility.
2155 Updated many users of procedure_with_splits() and
2156 multipass_procedure_with_splits() to call
2157 output_split_file_values() and to deal with increased use of
2160 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2161 parameter to begin_func member.
2162 (procedure_with_splits) Add "const struct ccase *" parameter to
2163 begin_func parameter. Make ccase parameter const in proc_func
2165 (split_procedure_case_func) Don't dump split file group. Pass
2167 (dump_splits) Moved to language/dictionary/split-file.c as
2168 output_split_file_values().
2169 (struct multipass_split_aux_data) Add "const struct ccase *"
2170 parameter to split_func member.
2171 (multipass_procedure_with_splits) Add "const struct ccase *"
2172 parameter to split_func parameter.
2173 (multipass_split_case_func) Save new SPLIT FILE case before
2175 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2177 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2179 Continue reforming procedure execution. Change
2180 internal_procedure() so that it calls open_active_file() and
2181 close_active_file(), which isolates most of the actual procedure
2184 * procedure.c: (struct write_case_data) Rename `proc_func' member
2185 to `case_func' and update all references.
2186 (procedure) Rewrite as one-line wrapper around
2187 internal_procedure().
2188 (struct multipass_aux_data) New.
2189 (multipass_callback) Renamed multipass_case_func(). Use struct
2190 multipass_aux_data as auxiliary data.
2191 (multipass_end_func) New function.
2192 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2193 that uses multipass_case_func, multipass_end_func.
2194 (internal_procedure) Add `end_func' argument. Move optimization
2195 of trivial case in here. Move call to open_active_file() and
2196 close_active_file() in here. Now assert that vfm_source is
2198 (procedure_with_splits_callback) Rename
2199 split_procedure_case_func().
2200 (split_procedure_end_func) New function.
2201 (multipass_split_callback) Rename multipass_split_case_func.
2202 (multipass_split_end_func) New function.
2203 (discard_variables) No need to test for nonnull vfm_source.
2205 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2207 Continue reforming procedure execution. Get rid of unused member.
2209 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2211 (write_case) Don't increment cases_analyzed.
2213 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2215 Continue reforming procedure execution. In this phase, move
2216 procedure.c and procedure.h from src to src/data. Update
2217 makefiles and #includes accordingly.
2219 * procedure.c: Moved here from src/.
2221 * procedure.h: Moved here from src/.
2223 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2225 Continue reforming procedure execution. In this phase, get rid of
2226 many global variables, consolidating procedure execution in
2227 procedure.c. Encapsulate transformations in new "struct
2228 trns_chain". Also, change implementation of N OF CASES, FILTER,
2229 and PROCESS IF from special cases to transformations.
2231 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2234 * dictionary.c: (global variable default_dict) Move to
2237 * variable.h: (TRNS_*) Move to transformations.h.
2238 (struct transformation) Move to transformations.c.
2240 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2242 * sys-file-reader.c: Fixed invalid read problems.
2244 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2246 * storage-stream.c: Add missing function comments.
2248 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2250 Continue reforming procedure execution. In this phase, add some
2251 new, needed functionality to storage-stream.
2253 * storage-stream.c: (storage_source_decapsulate) New function.
2255 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2257 * variable.c (width_to_bytes): Declarations must precede
2258 statements for C90 compliance.
2260 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2262 * data-out.c, data-in.c, variable.c, variable.h: New functions
2263 copy_mangle and copy_demangle for reading/writing cases; emulates the
2264 way SPSS deals with strings > 255 bytes.
2266 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2267 subtype 14 needed for strings longer than 255 bytes.
2269 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2270 instead of literal values. Also fixed some constness issues.
2272 * format.h: Constness
2274 * sfm-private.h: Renamed the case_size identifier, since I discovered
2275 that SPSS's respect for this variable is very nominal.
2277 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2279 Change case limit type from int to size_t.
2281 * dictionary.c: (struct dictionary) Change type of case_limit
2283 (dict_get_case_limit) Change return type.
2284 (dict_set_case_limit) Change parameter type.
2286 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2288 * variable.h: (struct variable) Rename `reinit' member as `leave'
2289 and invert sense. Fix up all references.
2291 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2293 Continue reforming procedure execution. In this phase, break
2294 procedure.c into multiple files.
2296 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2298 * case-sink.c: New file.
2300 * case-sink.h: New file.
2302 * case-source.c: New file.
2304 * case-source.h: New file.
2306 * storage-stream.c: New file.
2308 * storage-stream.h: New file.
2310 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2312 * variable.h: (struct variable) Remove `init' member and all
2313 references to it from other files. It was initialized in several
2314 places, but nothing really ever used it for anything worthwhile.
2315 Thanks to Jason Stover for pointing out how confusing this
2318 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2320 Continue reforming error message support. In this phase, get rid
2321 of message "titles" and put the message text in `struct error'.
2322 Now `struct error' encapsulates a message more properly.
2324 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2325 Format message ourselves.
2327 * data-in.c: (vdls_error) Ditto.
2329 * por-file-reader.c: (error) Ditto.
2331 * sys-file-reader.c: (corrupt_msg) Ditto.
2333 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2335 GNU standards require "file name" instead of "filename" in
2336 documentation. It's nice for our code to follow the convention
2339 * casefile.c: (struct casefile) Rename `filename' member to
2340 `file_name'. Updated all references.
2342 * file-name.c: [!unix] (struct file_identity) Rename
2343 normalized_filename member to normalized_file_name. Updated all
2346 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2348 We don't really support anything but Unix-like environments well,
2349 so we might as well de-obfuscate by writing directory and path
2350 separators explicitly.
2352 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2354 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2355 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2357 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2360 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2362 GNU standards require "file name" instead of "filename" in
2363 documentation. It's nice for our code to follow the convention
2366 * filename.c: Rename to file-name.c.
2368 * filename.h: Rename to file-name.h. Update all inclusions.
2369 Update header guards.
2371 * automake.mk: Update file names.
2373 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2375 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2377 (fn_basename) Removed (dead code).
2378 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2379 (fn_special_p) Renamed fn_is_special(), all references updated.
2380 (fn_exists_p) Renamed fn_exists(), all references updated.
2382 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2384 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2385 Also, now it only tilde-expands file names, not paths.
2386 (fn_search_path) Tilde-expand one directory at a time.
2388 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2390 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2391 get rid of non-Unixlike version of the code, which has probably
2393 (fn_prepend_dir) Removed (dead code).
2395 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2396 (macro PATH_SEPARATOR_STRING) New.
2397 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2399 Continue reforming error message support. In this phase, we get
2400 rid of VM() and the other msg() support for "verbosity", replacing
2401 it by a new function verbose_msg().
2403 * filename.c: (fn_search_path) Use verbose_msg() instead of
2406 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2408 * sfm-private.h: Get rid of #defines after #error, which makes no
2411 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2413 Get rid of our own int32 type in favor of the standard int32_t
2416 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2419 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2421 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2423 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2425 Remove ill-considered file routines that are no longer used.
2427 * filename.c: (fn_open_ext) Removed.
2428 (fn_close_ext) Removed.
2430 * filename.h: (struct file_ext) Removed.
2432 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2434 * variable.c (var_is_valid_name): Move declarations before code
2437 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2439 * filename.ch (fn_interp_vars): Fixed small buglet.
2441 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2443 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2444 so as to modify the string inline. Thus makeing it easier to
2445 destroy the results when no longer needed.
2447 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2449 * category.c (cat_stored_values_destroy): Fixed memory leak.
2451 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2453 Add some missing frees. Thanks to John Darrington for reporting
2456 * any-writer.c (any_writer_close): Free writer.
2458 * any-reader.c (any_reader_close): Free reader.
2460 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2462 * por-file-reader.c: (error) Mark as NO_RETURN.
2464 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2466 * settings.c: Changed default value of scompress to true.
2468 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2470 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2472 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2475 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2476 here from pref.h.orig.
2478 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2480 * sys-file-reader.c: Fixed bug reading compressed files.
2482 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2484 * Numerous renames. See src/ChangeLog for details.
2486 * Moved files from src directory