1 2008-01-02 John Darrington <john@darrington.wattle.id.au>
3 * variable.c variable.h: Replaced var_get_value_name with
4 var_append_value_name which doesn't use any static data.
5 Thanks to Ben for review.
7 2007-12-07 Ben Pfaff <blp@gnu.org>
11 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
12 made from an empty casereader produce a casegrouper with no
13 groups, instead of one with one group that has no cases.
15 * casereader.c (casereader_is_empty): New function.
17 2007-12-06 Ben Pfaff <blp@gnu.org>
21 * sys-file-reader.c (read_display_parameters): Handle variable
22 display parameters record with only 2 data items per variable.
23 Reported by Guido Gay <gay@irer.it>.
25 2007-12-04 Ben Pfaff <blp@gnu.org>
27 * identifier.c (lex_id_match_n): New function.
28 (lex_id_match): Reimplement in terms of lex_id_match_n.
30 2007-11-24 Ben Pfaff <blp@gnu.org>
32 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
35 2007-11-24 Ben Pfaff <blp@gnu.org>
37 Fix warning reported when reading back system files that include
38 very long string variables. Thanks to Guido Gay <gay@irer.it> for
41 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
42 was written where `j' was meant.
44 2007-11-18 Ben Pfaff <blp@gnu.org>
46 Properly write variables that include a range of missing values to
47 system files. Thanks to Guido Gay <gay@irer.it> for reporting
50 * sys-file-writer.c (write_variable): Correctly calculate
51 n_missing_values field when writing variables that include a range
54 2007-11-10 Ben Pfaff <blp@gnu.org>
56 Cleanups and bug fixes devised while writing up documentation.
59 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
61 * dict-class.c: New file.
62 (dict_class_from_id): Move here.
63 (dict_class_to_name): Move here.
65 * dict-class.h: New file.
66 (enum dict_class): Move here. Change from consecutive integers to
67 consecutive bits, to make testing for any of multiple values
68 easier. Add new DC_ALL constant.
70 * dictionary.c (struct dictionary): Change `case_limit' from
72 (dict_get_vars): Make final argument an enum dict_class.
73 (dict_get_vars_mutable): Ditto.
74 (dict_get_case_limit): Change return value to casenumber.
75 (dict_set_case_limit): Change final argument to a casenumber.
76 (dict_unset_split_var): Add assertion. Rephrase slightly.
77 (dict_set_label): Use xstrndup to simplify.
79 * format.c (fmt_step_width): AHEX format also needs 2-byte
81 (fmt_set_style): Simplify assertions.
83 * missing-values.c (mv_add_num_range): Rename mv_add_range.
84 Simplify implementation.
85 (mv_has_value): Simplify implementation.
86 (mv_pop_value): Remove the first value, not the last, to avoid
87 having GET followed by SAVE reverse the order of missing values.
88 (mv_peek_value): Rename mv_get_value. Simplify assertion.
89 (mv_has_range): Simplify implementation.
90 (mv_peek_range): Rename mv_get_range.
91 (can_resize_string): Removed.
92 (mv_is_resizable): Use value_is_resizable.
93 (mv_resize): Use value_resize.
95 * short-names.h (SHORT_NAME_LEN): Move here.
97 * val-type.h: New file, for definitions related to type and width
98 of abstract values. Before, these definitions were mixed among
99 those related to "union value" and those related to variables.
100 (macro SYSMIS): Move here.
101 (macro LOWEST): Move here.
102 (macro HIGHEST): Move here.
103 (macro MAX_STRING); Move here.
104 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
105 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
106 All references updated.
107 (val_type_is_valid): New function. Replaces var_type_is_valid.
108 All references updated.
109 (val_type_from_width): New function. Replaces
110 var_type_from_width. All references updated.
112 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
114 (val_labs_can_set_width): Use value_is_resizable.
115 (val_labs_add): Simply return false if the value labels set is too
116 wide, instead of having undefined behavior.
117 (val_labs_replace): Ditto.
118 (val_labs_replace): Ditto.
119 (val_labs_first): Set iterator to null if iteration is complete.
120 (val_labs_first_sorted): Ditto.
121 (val_labs_done): Become a no-op if the iterator is null.
123 * value.c (value_is_resizable): New function.
124 (value_resize): New function.
126 * variable.c (var_get_dict_class): New function.
128 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
131 2007-11-08 Ben Pfaff <blp@gnu.org>
133 * data-in.c: Make formatted data parsing locale-independent.
134 (parse_number): Use c_strtod instead of strtod, to avoid
135 locale-specific behavior.
138 2007-11-06 Ben Pfaff <blp@gnu.org>
140 Patch #6256: add support for binary, 360 file formats. Reviewed
143 * data-in.c (struct data_in): Add `encoding' member.
144 (data_in): Add `encoding' parameter, and re-encode the data passed
145 in where appropriate. Update all callers to pass it in.
146 (parse_A): Implement EBCDIC recoding wart described in manual.
147 (parse_AHEX): Implement EBCDIC recoding.
149 * data-out.c (data_out_legacy): New function.
150 (data_out): Make into a wrapper around data_out_legacy.
152 * file-handle-def.c (struct file_handle): New member `encoding'.
153 (fh_create_file): Set encoding.
154 (fh_default_properties): Set default encoding.
155 (fh_get_legacy_encoding): New function.
157 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
158 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
159 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
160 (struct fh_properties): New member `encoding'.
162 2007-11-05 Ben Pfaff <blp@gnu.org>
164 Patch #6258. Reviewed by John Darrington.
166 * file-handle-def.c (fh_lock): Add comment that TYPE should be
167 marked with N_() in the caller. Added these markings to each
168 caller too. Should make i18n easier.
169 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
171 2007-11-03 Ben Pfaff <blp@gnu.org>
173 Allow output files to overwrite input files (bug #21280).
175 * file-handle-def.c: Separate locking of files for input and for
176 output, to allow both to take place at once. Also, distinguish a
177 file handle from the identity of the underlying file, because the
178 identity of a file changes over time and the file handle can't
179 represent two different identities.
180 (struct file_handle): Remove `next', `open_cnt', `deleted',
181 `type', `open_mode', `aux', `identity' members. Change `id' from
182 char array to char *. Add `ref_cnt' member.
183 (file_handle_from_ll) New function.
184 (file_handles) Removed.
185 (named_handles) New variable.
186 (fh_init) Initialize named_handles.
187 (fh_done) Remove name from all named_handles.
188 (free_handle) Updated for modified struct file_handle.
189 (unname_handle) New function.
190 (fh_ref) New function.
191 (fh_from_file_name) Removed.
192 (fh_unref) New function.
193 (fh_unname) New function.
194 (fh_from_id) Rewritten.
195 (create_handle) Updated for modified struct file_handle.
196 (fh_inline_file) Increment the handle's ref count.
197 (fh_create_file) Updated for modified struct file_handle.
198 (fh_create_scratch) Ditto.
203 (fh_is_open) Removed.
204 (fh_get_id) Updated for modified struct file_handle.
205 (fh_get_default_handle) Increment the handle's ref count.
206 (fh_set_default_handle) Handle ref counts.
207 (struct fh_lock) New structure.
208 (locks) New static var.
209 (fh_lock) New function.
210 (fh_unlock) New function.
211 (fh_lock_get_aux) New function.
212 (fh_lock_set_aux) New function.
213 (fh_is_locked) New function.
214 (make_key) New function.
215 (free_key) New function.
216 (compare_fh_locks) New function.
217 (hash_fh_lock) New function.
219 * file-handle-def.h (enum fh_access) New enum.
221 * file-name.c: Made file_identity the same in all supported
223 (struct file_identity): New `name' member.
224 (fn_get_identity): For a file that doesn't exist, get the
225 dev/inode of its directory plus its name. If even the directory
226 doesn't exist, just use its name. Merge the Windows
227 implementation into the Unix one.
228 (fn_compare_file_identities): Rewritten. Merge the Windows
229 implementation into the Unix one.
230 (fn_hash_identity): New function.
232 * make-file.c (struct replace_file): New structure.
233 (all_files): New static var.
234 (replace_file_start): New function.
235 (replace_file_commit): New function.
236 (replace_file_abort): New function.
237 (free_replace_file): New function.
238 (unlink_replace_files): New function.
240 * por-file-reader.c (struct pfm_reader): Add `lock' member.
241 (close_reader): Unlock file.
242 (pfm_open_reader): Lock file.
244 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
246 (pfm_open_writer): Lock file and prepare for its replacement.
247 (close_writer): Unlock file.
249 * scratch-handle.h (struct scratch_handle): Add unique_id so that
250 different generations of a scratch file can be distinguished.
252 * scratch-reader.c (scratch_reader_open): Verify that the file is
255 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
256 remove scratch_handle member.
257 (scratch_writer_open): Lock handle. Prepare to replace handle
258 data, instead of doing it immediately.
259 (scratch_writer_casewriter_destroy): Replace handle data and
262 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
263 (sfm_open_reader): Lock file.
264 (close_reader): Unlock file.
266 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
268 (sfm_open_writer): Lock file and prepare for its replacement.
269 (close_writer): Unlock file.
271 2007-11-02 Ben Pfaff <blp@gnu.org>
273 * data-out.c (output_number): Use isfinite (from C99) instead of
274 gsl_isfinite, since we now have universal support for it in
276 (output_infinite): Ditto.
277 [!HAVE_ROUND] (round): Remove definition, since we now have a
278 replacement in gnulib.
280 2007-10-27 John Darrington <john@darrington.wattle.id.au>
282 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
283 wasn't initialised correctly. Closes bugs #21330 and 21397
286 2007-10-08 Ben Pfaff <blp@gnu.org>
288 Fix bug #21280. Thanks to John Darrington for review.
290 * file-name.c (create_stream): New function.
292 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
293 file handle before creating the file, to ensure that we don't
294 truncate a file that we're reading. Make code easier to read by
297 * sys-file-write.c (sfm_open_writer): Ditto.
299 2007-10-01 Ben Pfaff <blp@gnu.org>
301 Fix bug #21192. Thanks to John Darrington for review.
303 * casereader.c (casereader_read): Decrement case_cnt before
304 calling the casereader's "read" member function, so that we
305 interact properly with lazy_casereader.
307 * datasheet.c: Add regression test for above bug fix.
308 (clone_datasheet): New function.
309 (lazy_callback): New function.
310 (check_datasheet_casereader): New function.
311 (check_datasheet): Check datasheet contents are reported correctly
312 through an ordinary casereader and a lazy casereader.
313 (clone_model): Use clone_datasheet.
315 2007-09-24 Ben Pfaff <blp@gnu.org>
317 Patch #6210. Reviewed by John Darrington.
318 * settings.c: Drop "long view", which is not needed any longer.
319 (static var long_view): Removed.
320 (force_long_view): Removed.
321 (get_viewwidth): Removed.
322 (init_viewport): Removed long_view reference.
324 2007-09-19 John Darrington <john@darrington.wattle.id.au>
326 * settings.c settings.h: Changed viewport's length and width to be
327 owned by the user interface which uses the data library. This allows
328 better abstraction, and makes dynamically adjustable dimensions easier.
330 2007-09-18 Ben Pfaff <blp@gnu.org>
332 * procedure.c (proc_extract_active_file_data): New function.
334 * lazy-casereader.h: New file.
336 * lazy-casereader.c: New file.
338 * casereader.c (casereader_dynamic_cast): New function.
340 2007-09-14 Ben Pfaff <blp@gnu.org>
342 * dictionary.c (dict_clone): Copy case indexes from cloned
343 dictionary. Fixes bug #21061. Reviewed by John Darrington.
345 2007-09-13 John Darrington <john@darrington.wattle.id.au>
347 * value.c value.h (create_value): New function.
349 2007-09-12 Ben Pfaff <blp@gnu.org>
351 Make it clear that translator casereader and casewriter translate
352 functions are supposed to destroy their input case.
354 * casereader-translator.c (struct casereader_translator): Change
355 input case parameter of translate member function from const
356 struct ccase * to struct ccase *.
357 (casereader_create_translator): Ditto, for translate parameter.
359 * casewriter-translator.c (struct casewriter_translator): Ditto.
360 (casewriter_create_translator): Ditto.
362 2007-08-27 John Darrington <john@darrington.wattle.id.au>
364 * sys-file-reader.c (read_display_parameters): Force display width
365 to 8, if the sys file says 0 (like SPSS does).
367 2007-08-12 Ben Pfaff <blp@gnu.org>
369 * dictionary.c (dict_dump): New function.
371 2007-08-12 Ben Pfaff <blp@gnu.org>
373 Drop dict_compactor in favor of using the new struct case_map.
375 * dictionary.c (struct copy_map): Removed.
376 (struct dict_compactor): Removed.
377 (dict_make_compactor): Removed.
378 (dict_compactor_compact): Removed.
379 (dict_compactor_destroy): Removed.
381 * procedure.c (struct data_set): Change `compactor' member to be a
383 (proc_open): Use case_map_to_compact_dict instead of
385 (proc_casereader_read): Use case_map_execute instead of
386 dict_compactor_compact.
387 (proc_commit): Use case_map_destroy instead of
388 dict_compactor_destroy.
390 * scratch-writer.c (struct scratch_writer): Change `compactor'
391 member to be a struct case_map *.
392 (scratch_writer_open): Use case_map_to_compact_dict instead of
394 (scratch_writer_casewriter_write): Use case_map_execute instead of
395 dict_compactor_compact.
397 2007-08-12 Ben Pfaff <blp@gnu.org>
399 * automake.mk: Add case-map.c, case-map.h.
401 * case-map.c: New file.
403 * case-map.h: New file.
405 2007-08-12 Ben Pfaff <blp@gnu.org>
407 * dictionary.c (dict_compact_values): Don't delete scratch
408 variables as well as compacting case indexes. Update all callers.
409 (dict_get_compacted_value_cnt): Rename dict_count_values and
410 change interface. Update all callers.
411 (dict_get_compacted_value_cnt): Remove.
412 (dict_compacting_would_shrink): Remove.
413 (dict_compacting_would_change): Remove.
414 (dict_make_compactor): Add new parameter. Update all callers.
416 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
417 instead of dict_count_values, changing an O(N) operation into
420 2007-08-12 Ben Pfaff <blp@gnu.org>
422 * casereader.c (casereader_read): Don't require cases read by a
423 casereader to be exactly the expected size: as long as they're big
426 2007-08-12 Ben Pfaff <blp@gnu.org>
428 Make casewriters keep track of the number of `union value's in
429 each case. This is useful for two reasons: casewriter_write can
430 then check that the case being written is large enough, and later
431 recipients of the casewriter can determine the size of the case.
433 * casewriter-translator.c (casewriter_create_translator): Add
436 * casewriter.c (struct casewriter): Add value_cnt member.
437 (casewriter_write): Check that the case passed in is big enough.
438 (casewriter_get_value_cnt): New function.
439 (casewriter_create): Add value_cnt parameter.
441 2007-08-09 Ben Pfaff <blp@gnu.org>
443 Fix bug reported by Jason Stover.
444 * settings.c: Move get_termcap_viewport and in particular the
445 #include for <curses.h> to the end of the file. curses.h
446 redefines bool on some systems (e.g. OpenBSD), which causes
447 disagreement between uses of bool before its inclusion and
449 Tested by Jason Stover.
451 2007-07-29 Ben Pfaff <blp@gnu.org>
453 Provisional fix for bug #18692 and bug #20161. Reviewed by John
456 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
457 argument (never "rb" or "wb") because SUSv3 says that only those
458 modes are defined, and glibc in fact rejects other modes.
460 Open portable files with fn_open so that they can be read from
461 pipes. Fix missing fh_close call to go along with fh_open.
462 Report an error if the file close reports an error.
463 * por-file-reader.c (close_reader): New function.
464 (por_file_casereader_destroy): Use close_reader.
465 (pfm_open_reader): Open file with fn_open.
467 2007-07-28 Ben Pfaff <blp@gnu.org>
469 Make PSPP able to read all the portable files I could find on the
470 web. Thanks to John Darrington for review. Bug #17620.
471 * por-file-reader.c (struct pfm_reader): New member `line_length'.
472 (error): Print file offset in hexadecimal.
473 (warning): New function.
474 (advance): Treat lines less than 80 bytes long as padded to 80
476 (pfm_open_reader): Call read_documents if we find an "E" record.
477 (convert_format): Convert invalid formats to the default format
478 instead of aborting reading the file.
479 (read_variables): Rename duplicate variable names instead of
480 aborting reading the file.
481 (read_value_label): Allow string variables of different widths to
482 be assigned value labels in the same record. Replace duplicate
483 value labels instead of aborting.
484 (read_documents): New function.
486 * por-file-writer.c (pfm_open_writer): Call write_documents if the
487 dictionary has documents.
488 (write_documents): New function.
490 2007-07-25 Ben Pfaff <blp@gnu.org>
492 Fix bugs related to bug #17213.
494 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
495 since the former is what config.h has. Include the needed ncurses
497 (static var echo) Rename to `do_echo' because the original name is
498 the same as an ncurses identifier.
499 (get_termcap_viewport) Use error instead of msg.
501 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
502 (fn_close): Don't close stdin, stdout, stderr.
504 2007-07-26 John Darrington <john@darrington.wattle.id.au>
506 * procedure.c procedure.h: Added callbacks which get invoked whenever
507 a dataset's transformation chain changes.
509 2007-07-24 Ben Pfaff <blp@gnu.org>
512 * sys-file-writer.c (write_variable_display_parameters): Use new
513 var_default_display_width function to choose display width of
514 segments after the first one in a given variable.
515 * variable.c (var_create): Use var_default_display_width to pick
516 new variable's display width.
517 (var_default_display_width): New function.
518 Reviewed by John Darrington.
520 2007-07-24 Ben Pfaff <blp@gnu.org>
523 * por-file-writer.c (write_variables): Write weight variable.
524 Reviewed by John Darrington.
526 2007-07-23 Ben Pfaff <blp@gnu.org>
528 Improvements to system file reader and writer.
530 First, move all detailed knowledge of very long strings into
531 sys-file-private.[ch], so that this nasty stuff can be isolated.
533 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
534 (EFFECTIVE_VLS_CHUNK): New macro.
535 (min_int): New function.
536 (max_int): New function.
537 (sfm_width_to_bytes): Rewrite.
538 (sfm_width_to_octs): New function.
539 (sfm_segment_alloc_width): New function.
540 (sfm_segment_alloc_bytes): New function.
541 (sfm_segment_used_bytes): New function.
542 (sfm_segment_offset): New function.
543 (sfm_segment_effective_offset): New function.
544 (sfm_dictionary_to_sfm_vars): New function.
546 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
547 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
548 (struct sfm_var): New structure.
550 Next, improvements to the system file reader.
552 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
553 type casenumber. Added `version_major', `version_minor',
556 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
557 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
558 Change `case_cnt' to type casenumber. Removed `has_vls'.
559 (struct sfm_var): Removed.
560 (sfm_open_reader): Don't warn on wrong case size if the file was
561 written by SPSS 13, which tends to get it wrong. Use
562 sfm_dictionary_to_sfm_vars.
563 (read_header): Always output system file info.
564 (read_variable_record): Simplify code for reading missing values.
565 (read_machine_int32_info): Save version numbers from system file
566 into info struct passed as new argument.
567 (read_long_string_map): Restructured to use new sys-file-private
569 (read_value_labels): Use size_overflow_p.
570 (sys_file_casereader_read): Get rid of distinction between fast
571 and slow paths. Use information provided by sys-file-primate's
572 struct sfm_var to simplify code.
573 (skip_whole_strings): New function.
574 (read_int32): Renamed read_int. Changed return value to int.
576 (read_flt64): Renamed read_float. Changed return value to
577 double. Updated all callers.
578 (int32_to_native): Removed. Changed callers to use
580 (flt64_to_double): Removed. Changed callers to use float_convert.
582 Finally, get rid of int32, flt64 terminology and types in system
583 file writer. The former wasn't very useful since a POSIX "int"
584 can hold the whole range of int32 and we generally didn't have a
585 need for it to be exactly-32-bits, just at-least-32-bits. The
586 latter was inconvenient because we had to assume that it could be
587 different from double and thereby convert special values SYSMIS,
588 HIGHEST, LOWEST to and from it in multiple places. Instead, now
589 we just use "int" and "double" in most places, and do conversions,
590 if necessary, very close to where we do I/O. This change meant
591 that the writer code couldn't represent records in the file as C
592 structs any longer, but that's no great loss. The code actually
593 seems to be more readable without them.
595 Simplify the compression buffering code: only buffer as much as
596 necessary, which is no more than eight 8-byte units at any given
599 * sys-file-writer.c (typedef flt64): Removed.
600 (macro second_lowest_flt64): Removed.
601 (struct sysfile_header): Removed.
602 (struct sysfile_variable): Removed.
603 (struct sfm_writer): Removed `needs_translation', `has_vls',
604 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
605 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
606 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
607 compression buffering by `opcodes', `opcode_cnt', `data',
608 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
609 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
610 data writer functions instead of structures.
611 (calc_oct_idx): New function.
612 (write_header): Use simple data writer functions instead of
614 (write_format_spec): Renamed write_format. New argument.
615 (write_variable_continuation_records): New function.
616 (write_variable): Use simple data writer functions instead of
617 structures. Use write_variable_continuation_records. Write
618 entire very long string instead of requiring caller to understand
620 (write_value_labels): Use simple data writer functions instead of
622 (write_documents): Ditto.
623 (write_variable_display_parameters): Use sys-file-private
624 functions to simplify. Use simple data writer functions instead
626 (write_vls_length_table): Use simple data writer functions instead
628 (write_longvar_table): Ditto.
629 (write_rec_7_34): Break into new functions
630 write_integer_info_record, write_float_info_record. Use simple
631 data writer functions instead of structures.
632 (buf_write): Removed.
633 (append_string_max): Removed.
634 (ensure_buf_space): Removed.
635 (sys_file_casewriter_write): Get rid of the distinction between
636 fast and slow paths, which didn't seem to be too useful. Use new
637 functions write_case_uncompressed, write_case_compressed.
638 (put_instruction): Removed.
639 (put_element): Removed.
640 (write_compressed_data): Removed.
641 (close_writer): Use flush_compressed. Only write case count to
642 system file if it will fit in the field.
643 (write_case_compressed): New function.
644 (write_case_uncompressed): New function.
645 (flush_compressed): New function.
646 (put_cmp_opcode): New function.
647 (put_cmp_number): New function.
648 (write_int): New function.
649 (convert_double_to_output_format): New function.
650 (write_float): New function.
651 (write_value): New function.
652 (write_string): New function.
653 (write_bytes): New function.
654 (write_zeros): New function.
655 (write_spaces): New function.
657 Reviewed by John Darrington.
659 2007-07-22 Ben Pfaff <blp@gnu.org>
661 Don't try to write very long strings to portable files. The
662 format does not support it.
664 * por-file-writer.c (MAX_POR_WIDTH): New macro.
665 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
666 (write_format): Add arg to take width to resize format to.
667 (write_value): Limit width of value written to MAX_POR_WIDTH.
668 (write_variables): Limit width of variable and its output formats
670 Reviewed by John Darrington.
672 2007-07-22 Ben Pfaff <blp@gnu.org>
674 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
675 local variable instead of literal 5.
676 Reviewed by John Darrington.
678 2007-07-22 Ben Pfaff <blp@gnu.org>
680 Fix problems with uniqueness of short names in system files with
681 very long string variables. Now a variable may have multiple
684 * automake.mk (src_data_libdata_a_SOURCES): Add new files
685 short-names.c, short-names.h.
687 * dictionary.c (dict_clone): Clone all the short names.
688 (compare_strings): Move into short-names.c.
689 (hash_strings): Ditto.
690 (set_var_short_name_suffix): Ditto.
691 (dict_assign_short_names): Ditto, rename short_names_assign,
692 change to assign all short names.
694 * por-file-writer.c (write_variables): Use short_names_assign
695 instead of dict_assign_short_names.
697 * short-names.c: New file.
699 * short-names.h: New file.
701 * sys-file-private.c (sfm_width_to_segments): New function.
703 * sys-file-reader.c (read_long_var_name_map): Save and restore all
704 the short names, not just the first one.
706 * sys-file-writer.c (cont_var_name): Removed.
707 (sfm_open_writer): Use short_names_assign instead of
708 dict_assign_short_names. Use unique short names assigned by
709 short_names_assign instead of those generated by cont_var_name.
711 * variable.c (struct variable): Remove `short_name' member,
712 replace by `short_names' and `short_name_cnt'.
713 (var_create) Initialize new members.
714 (var_get_short_name_cnt): New function.
715 (var_get_short_name): Now takes an index argument. Changed most
717 (var_set_short_name): Ditto.
718 (var_clear_short_name): Renamed var_clear_short_names, changed to
719 clear all short names.
721 Reviewed by John Darrington.
723 2007-07-22 Ben Pfaff <blp@gnu.org>
725 * variable.c (var_set_width): Use new var_set_width function.
727 * missing-values.c (mv_n_values): Drop assertion, which was not
730 * format.c (fmt_default_for_width): New function.
731 (fmt_resize): New function.
733 Reviewed by John Darrington.
735 2007-07-18 John Darrington <john@darrington.wattle.id.au>
737 * datasheet.c (datasheet_delete_columns): Added assertion to check
738 we're not deleting outside the range of the sheet.
741 * dictionary.c dictionary.h variable.c: Added the ability for string
742 variables to be resized.
744 * vardict.h: Added some prototypes (moved from dictionary.h) as
745 these should only be called by variable.c
748 2007-07-14 John Darrington <john@darrington.wattle.id.au>
750 * sfm-reader.c: Respect case_cnt field in file header.
752 2007-07-01 John Darrington <john@darrington.wattle.id.au>
754 * transformation.c transformation.h (trns_chain_execute): Changed the
755 signature (Patch #6057)
757 2007-06-10 Ben Pfaff <blp@gnu.org>
759 * casereader-filter.c (casereader_filter_destroy): Make sure to
760 write all the remaining excluded cases to the casewriter, if any.
762 * caseinit.c (init_list_destroy): Rewrite.
763 (init_list_clear): Ditto.
765 * casegrouper.c (casegrouper_get_next_group): Always set *reader
766 to null when returning false.
768 2007-06-06 Ben Pfaff <blp@gnu.org>
770 Actually implement the new procedure code and adapt all of its
771 clients to match. Also adapt all of the other case sources and
772 sinks in the tree and their clients to use the
773 casereader/casewriter infrastructure.
775 * automake.mk: Add and remove files.
777 * any-reader.c: Change into a casereader.
778 * por-file-reader.c: Ditto.
779 * scratch-reader.c: Ditto.
780 * sys-file-reader.c: Ditto.
782 * any-writer.c: Change into a casewriter.
783 * por-file-writer.c: Ditto.
784 * scratch-writer.c: Ditto.
785 * sys-file-writer.c: Ditto.
787 * procedure.c: Change to use casereader, casewriter, caseinit, and
788 other new infrastructure.
790 * scratch-handle.c: Adapt to new infrastructure.
792 * case-sink.c: Removed, now dead code.
793 * case-sink.h: Ditto.
794 * case-source.c: Ditto.
795 * case-source.h: Ditto.
796 * casefile-factory.c: Ditto.
797 * casefile-private.h: Ditto.
800 * casefilter.c: Ditto.
801 * casefilter.h: Ditto.
804 * fastfile-factory.c: Ditto.
805 * fastfile-factory.h: Ditto.
806 * storage-stream.c: Ditto.
807 * storage-stream.h: Ditto.
809 2007-06-06 Ben Pfaff <blp@gnu.org>
813 * automake.mk: Add new files.
815 * datasheet.c: New file.
817 * datasheet.h: New file.
819 2007-06-06 Ben Pfaff <blp@gnu.org>
821 Until now, the procedure code has provided a case to the
822 case_source, which has filled in the data values that come from
823 the active file. "Left" data values that don't come from the
824 active file naturally stay the same from case to case, because the
825 procedure code keeps using that same case.
827 One of the compromises that comes with the new procedure code is
828 that the active file allocates and provides its own case, which
829 the procedure code then has to resize to provide room for any
830 other variables that should go in the case and then fill in the
831 values of "left" variables. Then, when we're done with that case,
832 we have to save the values of "left" variables to copy into the
833 next case read from the active file.
835 The caseinit code helps with this.
837 * automake.mk: Add new files.
839 * caseinit.c: New file.
841 * caseinit.h: New file.
843 2007-06-06 Ben Pfaff <blp@gnu.org>
845 * value.h (value_cnt_from_width): New function.
847 * variable.c (var_get_value_cnt): Use new function.
849 2007-06-06 Ben Pfaff <blp@gnu.org>
851 Add casegrouper, to allow cases read from a given casereader to be
852 broken into groups, each of which has its own casereader.
853 Generally cases are grouped based on having equal values for some
856 * automake.mk: Add new files.
858 * casegrouper.c: New file.
860 * casegrouper.h: New file.
862 2007-06-06 Ben Pfaff <blp@gnu.org>
864 Add interface to lexicographical ordering of cases.
866 * automake.mk: Add new files.
868 * case-ordering.c: New file.
870 * case-ordering.h: New file.
872 2007-06-06 Ben Pfaff <blp@gnu.org>
874 Add casereaders and casewriters, the basis of the new data processing
875 implementation. A casereader is a uniform interface to reading cases
876 from a data source; a casewriter is a uniform interface to writing
877 cases to a data sink.
879 * automake.mk: Add new files.
881 * casereader-filter.c: New file.
883 * casereader-provider.h: New file.
885 * casereader-translator.c: New file.
887 * casereader.c: New file.
889 * casereader.h: New file.
891 * casewriter-provider.h: New file.
893 * casewriter-translator.c: New file.
895 * casewriter.c: New file.
897 * casewriter.h: New file.
899 2007-06-06 Ben Pfaff <blp@gnu.org>
901 "casewindow" data structure that extends the deque (from libpspp)
902 of cases with the ability to dump cases to disk if we get too many
905 * automake.mk: Add new files.
907 * casewindow.c: New file.
909 * casewindow.h: New file.
911 2007-06-06 Ben Pfaff <blp@gnu.org>
913 sparse_cases data structure that augments a sparse_array of cases
914 with the ability to dump cases to disk if we get too many cases in
917 * automake.mk: Add new files.
919 * sparse-cases.c: New file.
921 * sparse-cases.h: New file.
923 2007-06-06 Ben Pfaff <blp@gnu.org>
925 Adds a low-level on-disk case array data structure.
927 * automake.mk: Add new files.
929 * case-tmpfile.c: New file.
931 * case-tmpfile.h: New file.
933 2007-06-06 Ben Pfaff <blp@gnu.org>
935 In a couple of places we calculate the maximum number of cases to
936 keep in memory based on the user-defined workspace. Enable
937 centralizing the calculation through a new function.
939 * settings.c (get_workspace_cases): New function.
941 2007-06-06 Ben Pfaff <blp@gnu.org>
943 The casenumber type is defined in transformations.h, but case.h is
944 a more sensible place. Move it.
946 * case.h (CASENUMBER_MAX): New macro.
947 (typedef casenumber): Move here, from transformations.h.
949 2007-06-03 Ben Pfaff <blp@gnu.org>
951 Slightly generalize case_to_values and case_from_values functions.
953 * case.c (case_to_values): Rename case_copy_out, change interface.
954 (case_from_values): Rename case_copy_in, change interface.
956 * fastfile.c (fastfilereader_get_next_case): Update caller.
957 (write_case_to_disk): Ditto.
959 2007-06-02 Ben Pfaff <blp@gnu.org>
961 Clean up after a forgotten part of patch #5829.
963 * casedeque.h: Remove unused file.
965 * automake.mk: Remove casedeque.h from sources.
967 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
969 * category.c: Removed redundant #include
971 2007-05-06 Ben Pfaff <blp@gnu.org>
973 Abstract the documents within a dictionary a little better.
974 Thanks to John Darrington for suggestion, initial version, and
977 * dictionary.c (struct dictionary): Change `documents' member from
978 char * to struct string.
979 (dict_clear): Destroy struct string.
980 (dict_get_documents): Convert struct string to char *.
981 (dict_set_documents): Set struct string. Pad to 80-character
983 (dict_clear_documents): New function.
984 (dict_add_document_line): New function.
985 (dict_get_document_line_cnt): New function.
986 (dict_get_document_line): New function.
988 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
990 * sys-file-reader.c (read_documents): Use new document functions.
992 2007-04-19 John Darrington <john@darrington.wattle.id.au>
994 * sys-file-reader.c: When reading a system file which has no
995 long name table, automatically create one where the long names
996 are the lower case versions of the short names.
998 2007-04-22 Ben Pfaff <blp@gnu.org>
1000 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1001 dict_clear will free most data related to the dictionary.
1002 dict_clear does a decent job, except that dict_set_split_vars on
1003 some systems won't actually free the dict's "split" member.
1004 Instead, it'll allocate a 1-byte region. Fix this.
1006 * value.c (value_copy): New function.
1007 (value_set_missing): Ditto.
1009 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1011 * Deleted existing category.h and moved cat-routines.h into
1012 category.h Encapsulated struct cat_vals better.
1014 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1016 * sys-file-reader.c: When reading a system file which has no
1017 long name table, automatically create one where the long names
1018 are the lower case versions of the short names.
1020 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1022 * sys-file-reader.c: Some versions of Other Software seem to
1023 produce system files with string variables' measure set to
1024 zero. We'll assume these are supposed to be nominal variables.
1026 2007-03-30 Ben Pfaff <blp@gnu.org>
1028 * procedure.c: Adapt to new deque data structure.
1030 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1031 Ben Pfaff <blp@gnu.org>
1033 * file-name.c: Mingw compatibility fixes.
1034 (fn_search_path): Use ISSLASH instead of comparing against '/'
1036 (fn_dir_name): Use dir_name from gnulib.
1037 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1038 (fn_get_identity): Use GetFullPathName instead of canonicalize
1039 from gnulib, because the latter does not fully support
1040 Windows-style path names. Use this implementation based on the
1041 detected presence of Windows instead of the absence of Unix, since
1042 the new implementation is Windows-specific.
1043 (fn_compare_file_identities): In Windows implementation, compare
1044 names case-insensitively.
1046 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1048 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1049 <stdlib.h> to have the same effect.
1051 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1053 * por-file-reader.c: Add missing _() around messages.
1055 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1057 * make-file.c: Include "mkstemp.h", without which linking on
1060 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1062 Reduce platform dependence.
1064 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1065 Everywhere we using this, we really should have just depended on
1066 the shell to expand tildes.
1067 (fn_search_path): Simplify, given that we don't do tilde expansion
1069 (fn_normalize): Removed. Caller changed to use the canonicalize
1071 (fn_get_cwd): Removed. Only user was fn_normalize.
1072 (fn_is_absolute): Really only test for absolute names.
1073 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1075 (fn_readlink): Removed, as it was only used fn_normalize.
1076 (fn_exists): Assume the stat function is available; gnulib does.
1077 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1079 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1081 * dictionary.c (dict_create_vector_assert): New function.
1083 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1085 * file-name.c (fn_normalize): Correct name of function
1086 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1087 for pointing this out.
1089 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1091 We are using a single member in struct file_handle, the "name"
1092 field, for more than one purpose. When it begins with '"', it's a
1093 file name; otherwise, it's a token that can be used to identify
1094 it. When that assertion fires, it's because we searched for the
1095 name case-sensitively as a file name (so that there was no match),
1096 and then we try to insert it case-insensitively as a token, which
1097 fails because duplicates aren't allowed.
1099 Solution: break the two purposes into two separate fields. This
1100 fixes the problem and likely makes the code easier to read too.
1102 Fixes bug #18922. Thanks to John Darrington for bug report and
1105 * file-handle-def.c (struct file_handle): New `id' member.
1106 (fh_from_name): Rename fh_from_id. Update all callers.
1107 (create_handle): New `id' parameter. Update all callers.
1108 (fh_create_file): Ditto.
1109 (fh_get_id): New function.
1111 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1113 * case.c (case_is_null): Change return type to bool.
1115 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1119 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1120 reference counting to share data. This makes it easy for
1121 valgrind, etc. to find accesses to cases that have been destroyed
1122 but have been kept around by another user's ref-count. This often
1123 happens when the data set is small enough to find in memory; if a
1124 bigger data set that would overflow to disk were used, then data
1125 corruption would occur.
1127 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1131 * case.c (case_unshare): Make it check internally whether the
1132 ref_cnt is greater than 1, so that the callers don't have to.
1133 Update callers not to check.
1135 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1137 Before, I was thinking that I might want to get rid of reference
1138 counting at some point. Now, I'm pretty sure that it's here to
1139 stay. Thus, because we have to store the value_cnt anyway for
1140 reference-counted cases, we might as well expose it to users.
1142 * case.c (case_get_value_cnt): New function.
1143 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1144 resize case if its size actually changed.
1146 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1147 instead of peeking inside struct case directly.
1148 (casefile_append): Ditto.
1150 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1152 Get rid of the inlines for the case functions, which made the
1153 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1154 the inlines didn't speed up "make check" at all, which is not a
1155 perfect benchmark but seems indicative.)
1157 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1158 definitions. Remove some assertions on nonnull pointers that were
1159 redundant with a pointer dereference soon after in the function.
1161 (struct case_data): Move definition here from case.h.
1165 (case_data_wr): Ditto.
1167 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1169 * automake.mk: Add casedeque.h to sources.
1171 * casedeque.h: New file.
1173 * procedure.c: (struct dataset) Change lag_count, lag_head,
1174 lag_queue member into single struct casedeque member. Update all
1175 users to use the casedeque instead.
1178 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1180 * procedure.c: Simplify lagged cases interface. Updated all
1181 clients--well, the only client--to use the simplified interface.
1182 (dataset_n_lag) Removed.
1183 (dataset_set_n_lag) Removed.
1184 (dataset_need_lag) New function.
1186 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1188 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1189 whenever appropriate, but only when the dataset/dictionary is in a
1192 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1194 * dictionary.c dictionary.h : Added callbacks for change of filter and
1195 split variables. Refactored some code to ensure that callbacks get
1196 invoked when appropriate.
1198 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1199 replace_dict callback occurs when permanent_dict replaces the current
1202 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1204 * dictionary.c dictionary.h : Added callback for when the weight
1205 variable of a dictionary changes.
1207 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1209 * dictionary.c dictionary.h : Added replace_source and replace_dict
1210 callbacks, and functions to deal with them.
1212 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1214 Simplify missing value handling.
1216 * missing-values.h (enum mv_class): New type.
1217 (enum mv_type): Moved definition into missing-values.c and renamed
1218 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1219 enums. Updated all uses.
1220 (struct missing_values): Changed type of `type' from `enum
1221 mv_type' to `int' because the definition is no longer exposed.
1223 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1224 parameter. Update all callers.
1225 (mv_is_num_missing): Ditto.
1226 (mv_is_str_missing): Ditto.
1227 (mv_is_value_user_missing): Removed. Changed callers to use
1228 mv_is_value_missing.
1229 (mv_is_num_user_missing): Removed. Changed callers to use
1231 (mv_is_str_user_missing): Removed. Changed callers to use
1233 (mv_is_value_system_missing): Removed. Changed callers to use
1234 mv_is_value_missing.
1235 (mv_set_type): Removed. Changed callers to use mv_clear.
1236 (mv_clear): New function.
1238 * variable.c (var_is_value_missing): Add new enum mv_class
1239 parameter. Update all callers.
1240 (var_is_num_missing): Ditto.
1241 (var_is_str_missing): Ditto.
1242 (var_is_value_user_missing): Removed. Changed callers to use
1243 var_is_value_missing.
1244 (var_is_num_user_missing): Removed. Changed callers to use
1246 (var_is_str_user_missing): Removed. Changed callers to use
1248 (var_is_value_system_missing): Removed. Changed callers to use
1249 var_is_value_missing.
1251 * casefilter.c (struct casefilter): Use enum mv_class in place of
1253 (casefilter_variable_missing): Adapt to new member.
1254 (casefilter_create): Change signature to take enum mv_class,
1257 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1259 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1261 * case-sink.c case-sink.h procedure.c procedure.h
1262 storage-stream.c: Now uses the factory.
1264 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1266 Make it possible to pull cases from the active file with a
1267 function call, instead of requiring indirection through a callback
1270 * case-source.h (struct case_source_class): Change ->read function
1271 to return a single case, instead of calling a callback function
1272 for each case. Change ->destroy function to return an error
1275 * case-source.c (free_case_source): Pass along the value returned
1276 by the case_source ->destroy function.
1278 * procedure.c (struct write_case_data): Removed.
1279 (struct dataset): Added some members to track procedure state.
1280 (procedure): Optimize the trivial case at this level.
1281 (internal_procedure): Re-implement in terms of proc_open,
1282 proc_read, proc_close.
1283 (proc_open) New function.
1284 (proc_read) New function.
1285 (proc_close) New function.
1286 (write_case) Moved into proc_read.
1287 (close_active_file) Moved closing of data source into proc_close.
1289 * storage-source.c: Rewrote to conform with modified
1290 case_source_class interface.
1292 * transformations.c (trns_chain_execute): Added argument to allow
1293 starting execution from an arbitrary transformation. Updated
1296 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1298 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1300 * sys-file-reader.c (read_display_parameters): Don't assume that
1301 MEASURE_* and ALIGN_* have the same values found in system files.
1303 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1305 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1306 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1307 the value of n_MEASURES, which was off by 1 (at least from my
1310 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1312 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1313 callbacks which are invoked when the dictionary or its
1314 variables are changed.
1316 * missing-values.c missing-values.h value-labels.c: Tidied up
1317 consistency checks, and made some of them return false
1318 instead of assert-failing.
1320 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1322 * calendar.c (calendar_days_in_month): New function.
1324 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1326 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1327 the value label's width.
1329 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1331 * sfm-private.h: Move contents into sys-file-writer.c, which is
1332 the only remaining user. Removed Borland C++-specific directives.
1334 * sys-file-reader.c: Clean up and rewrite entire file. The
1335 rewritten version is simpler and better abstracted, and should be
1336 easier to maintain and extend. It avoids using structures to read
1337 file data, which is prone to padding variations among compilers.
1338 It should also handle non-IEEE 754 system files, although I
1339 haven't been able to find any. It has been tested against many
1340 .sav files obtained from the Web and found to produce the same
1341 results as the earlier version of the code, or in some cases
1342 improved results. It is more tolerant of format variations found
1345 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1346 member, putting an enum integer_format in its place. New member
1347 `float_format'. Changed `compressed' member to type bool.
1349 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1351 * dictionary.c (dict_delete_consecutive_vars): New function.
1353 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1355 * file-name.c (fn_search_path): Remove prefix arg that was unused
1356 by any caller. Updated all callers.
1358 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1360 * format.c (fmt_dollar_template): Use user's decimal point
1361 character. Add assertion.
1363 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1365 * format.c (fmt_dollar_template): New function, based on
1366 dollar_format_template from var-type-dialog.c.
1368 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1370 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1373 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1375 Finish converting struct variable to an opaque type. In this
1376 phase, we add remaining setter and getter functions, convert the
1377 remaining PSPP code to use them, and do a bunch of cleanup. The
1378 resulting changes are pervasive but mostly trivial, and only the
1379 notable changes are logged.
1381 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1384 * case.c (case_data): Renamed case_data_idx.
1385 (case_num): Renamed case_num_idx.
1386 (case_str): Renamed case_str_idx.
1387 (case_data_rw): Renamed case_data_rw_idx.
1389 * case.h (case_data): New function with old name and an interface
1390 that takes a variable instead of an index, which is easier to
1391 use. Updated all callers to use the new interface, or to use the
1392 new *_idx function (see above).
1395 (case_data_rw): Ditto.
1397 * category.c (cat_stored_values_destroy): Changed interface to
1398 take a struct cat_vals * instead of a struct variable *.
1400 * dictionary.c (dict_clone): Use new vector_clone function.
1401 (dict_clear) Use new var_destroy function.
1402 (add_var) New function.
1403 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1404 (dict_create_var_assert) Rewrite in terms of add_var.
1405 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1406 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1407 (dict_lookup_var) Use new var_create, var_destroy functions.
1408 (dict_contains_var) Rewrite in terms of new vardict functionality.
1409 (set_var_dict_index) New function.
1410 (set_var_case_index) New function.
1411 (reindex_vars) New function.
1412 (dict_delete_var) Rewrite in terms of new vardict functionality.
1413 (dict_reorder_var) Ditto.
1414 (dict_reorder_vars) Ditto.
1415 (rename_var) New function.
1416 (dict_rename_var) Use rename_var.
1417 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1418 (dict_get_compacted_idx_to_fv) Rename
1419 dict_get_compacted_dict_index_to_case_index, update callers.
1420 (dict_create_vector) Use new vector_create function.
1421 (dict_clear_vectors) Use new vector_destroy function.
1422 (set_var_short_name_suffix) Move here from variable.c, renamed
1423 from var_set_short_name_suffix, make static, update caller.
1425 * sys-file-private.c: New file.
1426 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1427 width_to_bytes, update callers.
1429 * sys-file-private.h: New file. Later it will supplant
1430 sfm-private.h; for now it supplements it.
1431 (macro MIN_VERY_LONG_STRING) New macro.
1432 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1434 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1435 MAX_LONG_STRING was used before.
1437 * sys-file-writer.c: Ditto.
1439 * value-labels.c: Change the paradigm here to be that a null
1440 pointer is OK for a struct val_labs * in most cases; it just
1441 represents an empty set of value labels.
1442 (val_labs_copy) A copy of a null set is a null set.
1443 (val_labs_count) A null set has 0 labels.
1444 (val_labs_replace) Change return type to void. Rewrite for
1446 (val_labs_find) A null set does not contain the value.
1447 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1448 transposed argument order, updated all callers.
1450 * value.c: New file.
1451 (value_dup) Moved here from variable.c.
1452 (compare_values) Ditto.
1455 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1456 (macro MAX_LONG_STRING) Removed, because it was only interesting
1457 for system files, not for general code.
1458 (macro MAX_VERY_LONG_STRING) Ditto.
1459 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1460 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1462 * vardict.h: New file, for an interface between variables and
1465 * variable.c: A lot of functions were moved around, for better
1467 (struct variable) Move definition here, from variable.h.
1468 (var_type_adj) Removed--makes i18n hard.
1469 (var_type_noun) Ditto.
1470 (var_create) New function.
1471 (var_clone) New function.
1472 (var_destroy) New function.
1473 (var_set_name) Assert that variable is not in a dictionary.
1474 (compare_var_names) Rename compare_vars_by_name and fix a couple
1475 of callers who thought the args were strings.
1476 (hash_var_name) Rename hash_var_by_name.
1477 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1478 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1479 (var_is_very_long_string) Removed, because it was only interesting
1480 to system file code.
1481 (var_set_missing_values) Allow the argument to be the wrong width,
1482 as long as we can resize it. Simplify callers who were doing the
1483 resizing themselves.
1484 (var_get_value_labels) New function.
1485 (var_has_value_labels) New function.
1486 (var_set_value_labels) New function.
1487 (alloc_value_labels) New function.
1488 (var_add_value_label) New function.
1489 (var_replace_value_label) New function.
1490 (var_clear_value_labels) New function.
1491 (var_lookup_value_label) New function.
1492 (var_get_value_name) Moved here from variable.c, renamed from
1493 var_get_value_name, transposed argument order, updated all
1495 (var_to_string) Moved here, from variable-label.c.
1496 (var_set_leave) New function.
1497 (var_get_leave) New function.
1498 (var_must_leave) New function.
1499 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1500 set_var_short_name_suffix.
1501 (var_get_dict_index) New function.
1502 (var_get_case_index) New function.
1503 (var_get_obs_vals) New function.
1504 (var_set_obs_vals) New function.
1505 (var_has_obs_vals) New function.
1506 (var_get_vardict) New function.
1507 (var_set_vardict) New function.
1508 (var_has_vardict) New function.
1509 (var_clear_vardict) New function.
1510 (value_dup) Moved to value.c.
1511 (compare_values) Ditto.
1514 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1515 (enum ALPHA) Rename VAR_STRING, update all users.
1517 * vector.c: New file.
1518 (struct vector) Moved here, from variable.h.
1519 (check_widths) New function.
1520 (vector_create) New function.
1521 (vector_clone) New function.
1522 (vector_destroy) New function.
1523 (vector_get_name) New function.
1524 (vector_get_var) New function.
1525 (vector_get_var_cnt) New function.
1526 (compare_vector_ptrs_by_name) New function.
1528 * vector.h: New file.
1530 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1532 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1533 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1536 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1538 * value-labels.c (destroy_atoms): New function.
1539 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1541 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1543 Thanks to Jason Stover for pointing out this problem.
1545 * data-out.c (output_number): Use gsl_finite from GSL, which is
1546 portable, instead of isfinite, which is not.
1549 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1551 * variable.c variable.h (value_dup): New function.
1553 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1555 Start converting struct variable to an opaque type. In this
1556 phase, we add a bunch of setter and getter functions and convert
1557 most of the PSPP code to use them. The resulting changes are
1558 pervasive but mostly trivial, and only the notable changes are
1561 * format.c (fmt_equal): New function.
1563 * variable.c (var_type_is_valid): New function.
1564 (measure_is_valid) Moved here, from format.c.
1565 (alignment_is_valid) Moved here, from format.c.
1566 (var_get_name) New function.
1567 (var_set_name) New function.
1568 (width_to_type) New function.
1569 (var_get_type) New function.
1570 (var_get_width) New function.
1571 (var_set_width) New function.
1572 (var_is_numeric) New function.
1573 (var_is_alpha) New function.
1574 (var_is_short_string) New function.
1575 (var_is_long_string) New function.
1576 (var_is_very_long_string) New function.
1577 (var_get_missing_values) New function.
1578 (var_set_missing_values) New function.
1579 (var_clear_missing_values) New function.
1580 (var_has_missing_values) New function.
1581 (var_is_value_missing) New function.
1582 (var_is_num_missing) New function.
1583 (var_is_str_missing) New function.
1584 (var_is_value_user_missing) New function.
1585 (var_is_num_user_missing) New function.
1586 (var_is_str_user_missing) New function.
1587 (var_is_value_system_missing) New function.
1588 (var_get_print_format) New function.
1589 (var_set_print_format) New function.
1590 (var_get_write_format) New function.
1591 (var_set_write_format) New function.
1592 (var_set_both_formats) New function.
1593 (var_get_label) New function.
1594 (var_set_label) New function.
1595 (var_clear_label) New function.
1596 (var_has_label) New function.
1597 (var_get_measure) New function.
1598 (var_set_measure) New function.
1599 (var_get_display_width) New function.
1600 (var_set_display_width) New function.
1601 (var_get_alignment) New function.
1602 (var_set_alignment) New function.
1603 (var_get_value_cnt) New function.
1604 (var_get_leave) New function.
1605 (var_get_short_name) New function.
1607 * variable.h: (struct variable) Removed "type" and "nv" members;
1608 they are now computed from "width" where needed.
1610 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1612 * missing-values.c (mv_resize): Don't write beyond end of the
1613 allocated buffer when resizing a long string.
1615 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1617 Clean up identifier code: don't require identifier enumerations to
1618 be in a particular order; make better use of string library;
1619 expose less of the internals.
1621 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1622 change interface. Updated all callers.
1623 (lex_id_match) Change interface to use struct substring, update
1625 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1626 (global array keywords[]) Make static, change form. Update all
1627 users to use lex_id_name instead.
1628 (lex_is_keyword) New function.
1629 (lex_id_to_token) Change interface to use struct substring, update
1631 (lex_id_name) New function.
1633 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1634 lex_is_keyword instead.
1635 (T_LAST_KEYWORD) Removed.
1636 (T_N_KEYWORDS) Removed.
1638 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1640 * format.c: (fmt_date_template) Distinguish characters for which a
1641 space is output and any date delimiter is allowed on input, from
1642 those for which a space is output and only a space is allowed on
1643 input. The former is represented by X, the latter by a space.
1644 Also, drop distinction between h and H, changing the former to the
1647 * data-in.c: Completely rewrite internals to conform to SPSS input
1648 formats as closely as possible.
1649 (data_in) Changed external interface by replacing the structure
1650 that was used as a single argument by a set of arguments. Updated
1652 (data_in_finite_line) Removed. Converted all callers to use plain
1654 (data_in_get_integer_format) New function.
1655 (data_in_set_integer_format) New function.
1656 (data_in_get_float_format) New function.
1657 (data_in_set_float_format) New function.
1659 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1660 (struct data_in) Removed.
1662 * data-out.c: (output_date) Drop each component from the input as
1663 it is output, to allow us to drop the distinction between h (a
1664 count of hours) and H (the hour of day) template characters.
1665 Also, handle new X template character.
1666 (output_scientific) Follow more rational rule on when to drop
1667 fraction introduced between SPSS 13 and 15. Updated test case to
1670 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1672 Fix buffer overflow reported by John Darrington.
1674 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1675 realize that DIGITS is a count of nibbles, not of bytes.
1677 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1679 * calendar.c (calendar_offset_to_gregorian) Also return the
1680 year-of-day. Change callers to new interface.
1682 * data-out.c: Completely rewrite internals to conform to SPSS
1683 output formats as completely as possible.
1684 (data_out) Change interface to put input parameters before output
1685 parameters, for consistency with the style I now prefer. Update
1687 (data_out_get_integer_format) New public function.
1688 (data_out_set_integer_format) New public function.
1689 (data_out_get_float_format) New public function.
1690 (data_out_set_float_format) New public function.
1692 * data-out.h: New file. Move prototype for data_out here, from
1695 * format.c: (fmt_step_width) Use equality comparison instead of
1696 bitwise and, for clarity.
1697 (fmt_is_string) Ditto.
1698 (fmt_input_to_output) Fix categories that are translated to F
1701 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1703 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1704 casefile-private.h: Added casefilter to assist commands with missing
1707 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1709 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1711 * settings.c: (route_errors_to_terminal) New variable.
1712 (route_errors_to_listing) New variable.
1713 (get_error_routing_to_terminal) New function.
1714 (set_error_routing_to_terminal) New function.
1715 (get_error_routing_to_listing) New function.
1716 (set_error_routing_to_listing) New function.
1718 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1720 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1722 * format.c: Completely rewrite, to achieve better abstraction.
1723 Rewrite all references to formats in other files.
1725 * format.def: Rewrite and reorganize.
1727 * settings.c: Move everything related to custom currency formats
1728 into format.[ch], changing them in form, so as to group related
1729 code and definitions better. Changed all references to use the
1731 (static var decimal) Removed.
1732 (static var grouping) Removed.
1733 (static var cc) Removed.
1734 (get_decimal) Removed.
1735 (set_decimal) Removed.
1736 (get_grouping) Removed.
1737 (set_grouping) Removed.
1741 * settings.h: (macro CC_CNT) Removed.
1742 (macro CC_WIDTH) Removed.
1743 (struct custom_currency) Removed.
1745 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1747 * data-in.c (data_in): Use switch statement instead of table, to
1748 avoid dependence on the order of the FMT_* enums.
1750 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1752 * data-out.c: (num_to_string) Removed, because it was dead code.
1754 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1756 * data-in.c (parse_trailer): Fix error message.
1758 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1760 * format.c (fmt_is_binary): New function.
1762 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1764 * procedure.c procedure.h: Encapsulated the static data into a single
1767 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1769 * casefile.c (casereader_read_xfer): Always initialize the case,
1770 even on an error condition.
1772 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1774 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1775 entire function useless.
1777 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1779 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1780 casereader method casereader_clone.
1782 * procedure.c pransformations.h: Introduced new type casenum_t
1784 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1786 * variable.c: (width_to_bytes) Rephrase code for clarify.
1788 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1790 * format.c: (fmt_type_from_string) New function.
1791 (fmt_to_string) Include decimals in output if the format has
1792 decimals, even if the format type does not. This way, we can
1793 accurately reproduce incorrect formats in user output.
1794 (check_common_specifier) Make the check for a bad format type an
1795 assertion, so we get bug reports if they show up. Fix message.
1796 Check for decimal places with a format type that doesn't allow
1798 (check_input_specifier) Remove check for FMT_X, which has been
1800 (check_output_specifier) Ditto.
1802 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1804 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1805 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1806 (enum fmt_parse_flags) Removed.
1808 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1810 * casefile.c casefile.h: Converted to an abstract base class.
1811 * casefile-private.h fastfile.c fastfile.h: New files.
1812 * automake.mk procedure.c scratch-writer.c storage-stream.c
1814 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1816 * procedure.c (internal_procedure): Create sink_case with only as
1817 many values as the compacted dictionary.
1819 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1821 Remove "debugging" code that caused plenty of false positives and
1824 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1826 * case.c: Remove all references to `this' member.
1828 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1830 Fix link error noted by Jason Stover.
1832 * storage-stream.c: Include <assert.h>.
1834 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1836 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1837 ALL) and additional underlying system file issues.
1839 Thanks to John Darrington for review.
1841 First problem: var_hash points to variables not owned by the
1842 sys-file-reader, which the caller may free or modify. Use an
1843 array of sfm_vars instead, as done earlier (e.g. CVS version
1846 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1847 members and remove all code that references it. Add vars, var_cnt
1848 members. Remove fix_specials member, which was unused.
1849 (struct sfm_var) Remove name member, which was unused.
1850 (sfm_close_reader) Free vars member instead of var_hash.
1851 (compare_var_shortnames) Removed.
1852 (hash_var_shortname) Removed.
1853 (sfm_open_reader) Fill out vars array.
1854 (compare_var_index) Removed.
1855 (sfm_read_case) Use vars instead of var_hash.
1857 Second problem: we're confused about when we actually have very
1858 long strings, causing us to choose incorrectly between slow path
1859 and fast path in sfm_read_case.
1861 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1862 have very long strings, not when we have long variable names,
1863 which is an unrelated feature.
1865 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1867 * variable.h: Move var_set and variable parsing declarations to
1868 new header, src/language/lexer/variable-parser.h. Modified lots
1869 of files to include the new header.
1871 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1873 * value-labels.c (value_to_string): When there's no value label,
1874 format the variable according to its print format, instead of
1875 always effectively using A or F format.
1877 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1879 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1882 * format.c: Removed tortological assertion.
1884 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1886 Reform string library.
1888 * file-name.c (fn_interp_vars): Change interface to take a
1889 substring as input. Updated all users.
1891 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1893 * format.c (measure_is_valid): Really return false when m >=
1896 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1898 Implement random access to casefiles, for use in GUI.
1900 * casefile.c: (struct casereader) Add `random', `file_ofs',
1901 `buffer_ofs' members.
1902 (casefile_get_random_reader) New function.
1903 (read_open_file) Break part into new function
1904 seek_and_fill_buffer().
1905 (fill_buffer) Update buffer_ofs, file_ofs.
1906 (casereader_seek) New function.
1908 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1910 * settings.c: Added call to i18n{done, init}.
1912 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1914 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1917 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1919 * casefile.c: Convert many uses of `int' to `bool'.
1921 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1923 * transformations.c (trns_chain_destroy): Destroy chain's trns
1924 member, to fix memory leak.
1926 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1928 * storage-stream.c (storage_source_decapsulate): Destroy case
1929 source to fix memory leak.
1931 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1933 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1934 case passed as argument instead of initializing the argument as a
1935 case. Fixes memory leak that showed up in
1936 tests/command/aggregate.sh with scratch files.
1938 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1940 * procedure.c (proc_done): Destroy default_dict, to fix memory
1943 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1945 Simplify procedure_with_splits().
1947 * procedure.c (struct split_aux_data): Removed case_count member.
1948 (procedure_with_splits) Don't initialize case_count.
1949 (split_procedure_case_func) Check whether prev_case is null
1950 instead of case_count.
1951 (split_procedure_end_func) Ditto.
1953 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1955 * case.c (case_move): Do nothing if dst and src are the same
1957 (case_try_create) Merge two similar cases.
1958 (case_copy) Unshare only if data must be actually copied.
1960 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1962 * data-in.c data-out.c dictionary.c sys-file-reader.c
1963 sys-file-writer.c variable.c variable.h: Reworked very long string
1964 support for better encapsulation.
1966 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1968 * value-labels.c (val_labs_can_set_width): New function.
1969 (val_labs_set_width) Clear labels if increasing width to long
1971 (val_labs_destroy) Remove unneeded test for null.
1973 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1975 * value-labels.h: Remove unneeded dependency on variable.h.
1977 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1979 Get rid of `char *c' member in union value, for cleanliness.
1981 * value.h: (union value) Remove `c' member.
1983 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1985 Make dictionary compacting functions a little more general.
1987 * sys-file-writer.c (sfm_open_writer): Use
1988 dict_compacting_would_change().
1989 (does_dict_need_translation) Removed.
1991 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1993 Make dictionary compacting functions a little more general.
1995 * dictionary.c (dict_needs_compaction): Rename
1996 dict_compacting_would_shrink(). Update all callers.
1997 (dict_compacting_would_change) New function.
1999 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2001 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2002 inverted return value (!).
2004 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2006 Continue reforming procedure execution.
2008 * procedure.c: Search and replace "vfm" by "proc". Notably:
2009 (static var vfm_source) Rename proc_source. Update all
2011 (static var vfm_sink) Rename proc_sink. Update all references.
2013 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2015 Continue reforming procedure execution. In this phase, remove
2016 PROCESS IF, which was deprecated anyway and can be easily
2017 simulated with TEMPORARY followed by SELECT IF.
2019 * procedure.c: (open_active_file) Don't call
2020 add_process_if_trns().
2021 (discard_variables) Get rid of redundant call to
2022 proc_cancel_all_transformations().
2023 (add_process_if_trns) Removed.
2024 (process_if_trns_proc) Removed.
2025 (process_if_trns_free) Removed.
2027 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2029 Continue reforming procedure execution. In this phase, add
2030 `const' to the case passed to procedure()'s callback.
2032 Updated all users of procedure() as well.
2034 * procedure.c: (struct write_case_data) Add "const" to ccase
2035 parameter for case_func member.
2036 (procedure) Add "const" to ccase parameter for proc_func
2038 (multipass_case_func) Make ccase parameter const.
2039 (internal_procedure) Add "const" to ccase parameter for case_func
2041 (split_procedure_case_func) Make ccase parameter const.
2042 (multipass_split_case_func) Make ccase parameter const.
2044 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2046 Continue reforming procedure execution. In this phase, get rid of
2047 the output code for SPLIT FILE groups in procedure.c, which really
2048 shouldn't be doing any output. Move it into the individual
2049 procedures instead. This also adds some flexibility.
2051 Updated many users of procedure_with_splits() and
2052 multipass_procedure_with_splits() to call
2053 output_split_file_values() and to deal with increased use of
2056 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2057 parameter to begin_func member.
2058 (procedure_with_splits) Add "const struct ccase *" parameter to
2059 begin_func parameter. Make ccase parameter const in proc_func
2061 (split_procedure_case_func) Don't dump split file group. Pass
2063 (dump_splits) Moved to language/dictionary/split-file.c as
2064 output_split_file_values().
2065 (struct multipass_split_aux_data) Add "const struct ccase *"
2066 parameter to split_func member.
2067 (multipass_procedure_with_splits) Add "const struct ccase *"
2068 parameter to split_func parameter.
2069 (multipass_split_case_func) Save new SPLIT FILE case before
2071 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2073 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2075 Continue reforming procedure execution. Change
2076 internal_procedure() so that it calls open_active_file() and
2077 close_active_file(), which isolates most of the actual procedure
2080 * procedure.c: (struct write_case_data) Rename `proc_func' member
2081 to `case_func' and update all references.
2082 (procedure) Rewrite as one-line wrapper around
2083 internal_procedure().
2084 (struct multipass_aux_data) New.
2085 (multipass_callback) Renamed multipass_case_func(). Use struct
2086 multipass_aux_data as auxiliary data.
2087 (multipass_end_func) New function.
2088 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2089 that uses multipass_case_func, multipass_end_func.
2090 (internal_procedure) Add `end_func' argument. Move optimization
2091 of trivial case in here. Move call to open_active_file() and
2092 close_active_file() in here. Now assert that vfm_source is
2094 (procedure_with_splits_callback) Rename
2095 split_procedure_case_func().
2096 (split_procedure_end_func) New function.
2097 (multipass_split_callback) Rename multipass_split_case_func.
2098 (multipass_split_end_func) New function.
2099 (discard_variables) No need to test for nonnull vfm_source.
2101 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2103 Continue reforming procedure execution. Get rid of unused member.
2105 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2107 (write_case) Don't increment cases_analyzed.
2109 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2111 Continue reforming procedure execution. In this phase, move
2112 procedure.c and procedure.h from src to src/data. Update
2113 makefiles and #includes accordingly.
2115 * procedure.c: Moved here from src/.
2117 * procedure.h: Moved here from src/.
2119 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2121 Continue reforming procedure execution. In this phase, get rid of
2122 many global variables, consolidating procedure execution in
2123 procedure.c. Encapsulate transformations in new "struct
2124 trns_chain". Also, change implementation of N OF CASES, FILTER,
2125 and PROCESS IF from special cases to transformations.
2127 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2130 * dictionary.c: (global variable default_dict) Move to
2133 * variable.h: (TRNS_*) Move to transformations.h.
2134 (struct transformation) Move to transformations.c.
2136 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2138 * sys-file-reader.c: Fixed invalid read problems.
2140 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2142 * storage-stream.c: Add missing function comments.
2144 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2146 Continue reforming procedure execution. In this phase, add some
2147 new, needed functionality to storage-stream.
2149 * storage-stream.c: (storage_source_decapsulate) New function.
2151 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2153 * variable.c (width_to_bytes): Declarations must precede
2154 statements for C90 compliance.
2156 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2158 * data-out.c, data-in.c, variable.c, variable.h: New functions
2159 copy_mangle and copy_demangle for reading/writing cases; emulates the
2160 way SPSS deals with strings > 255 bytes.
2162 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2163 subtype 14 needed for strings longer than 255 bytes.
2165 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2166 instead of literal values. Also fixed some constness issues.
2168 * format.h: Constness
2170 * sfm-private.h: Renamed the case_size identifier, since I discovered
2171 that SPSS's respect for this variable is very nominal.
2173 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2175 Change case limit type from int to size_t.
2177 * dictionary.c: (struct dictionary) Change type of case_limit
2179 (dict_get_case_limit) Change return type.
2180 (dict_set_case_limit) Change parameter type.
2182 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2184 * variable.h: (struct variable) Rename `reinit' member as `leave'
2185 and invert sense. Fix up all references.
2187 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2189 Continue reforming procedure execution. In this phase, break
2190 procedure.c into multiple files.
2192 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2194 * case-sink.c: New file.
2196 * case-sink.h: New file.
2198 * case-source.c: New file.
2200 * case-source.h: New file.
2202 * storage-stream.c: New file.
2204 * storage-stream.h: New file.
2206 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2208 * variable.h: (struct variable) Remove `init' member and all
2209 references to it from other files. It was initialized in several
2210 places, but nothing really ever used it for anything worthwhile.
2211 Thanks to Jason Stover for pointing out how confusing this
2214 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2216 Continue reforming error message support. In this phase, get rid
2217 of message "titles" and put the message text in `struct error'.
2218 Now `struct error' encapsulates a message more properly.
2220 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2221 Format message ourselves.
2223 * data-in.c: (vdls_error) Ditto.
2225 * por-file-reader.c: (error) Ditto.
2227 * sys-file-reader.c: (corrupt_msg) Ditto.
2229 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2231 GNU standards require "file name" instead of "filename" in
2232 documentation. It's nice for our code to follow the convention
2235 * casefile.c: (struct casefile) Rename `filename' member to
2236 `file_name'. Updated all references.
2238 * file-name.c: [!unix] (struct file_identity) Rename
2239 normalized_filename member to normalized_file_name. Updated all
2242 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2244 We don't really support anything but Unix-like environments well,
2245 so we might as well de-obfuscate by writing directory and path
2246 separators explicitly.
2248 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2250 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2251 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2253 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2256 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2258 GNU standards require "file name" instead of "filename" in
2259 documentation. It's nice for our code to follow the convention
2262 * filename.c: Rename to file-name.c.
2264 * filename.h: Rename to file-name.h. Update all inclusions.
2265 Update header guards.
2267 * automake.mk: Update file names.
2269 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2271 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2273 (fn_basename) Removed (dead code).
2274 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2275 (fn_special_p) Renamed fn_is_special(), all references updated.
2276 (fn_exists_p) Renamed fn_exists(), all references updated.
2278 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2280 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2281 Also, now it only tilde-expands file names, not paths.
2282 (fn_search_path) Tilde-expand one directory at a time.
2284 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2286 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2287 get rid of non-Unixlike version of the code, which has probably
2289 (fn_prepend_dir) Removed (dead code).
2291 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2292 (macro PATH_SEPARATOR_STRING) New.
2293 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2295 Continue reforming error message support. In this phase, we get
2296 rid of VM() and the other msg() support for "verbosity", replacing
2297 it by a new function verbose_msg().
2299 * filename.c: (fn_search_path) Use verbose_msg() instead of
2302 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2304 * sfm-private.h: Get rid of #defines after #error, which makes no
2307 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2309 Get rid of our own int32 type in favor of the standard int32_t
2312 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2315 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2317 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2319 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2321 Remove ill-considered file routines that are no longer used.
2323 * filename.c: (fn_open_ext) Removed.
2324 (fn_close_ext) Removed.
2326 * filename.h: (struct file_ext) Removed.
2328 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2330 * variable.c (var_is_valid_name): Move declarations before code
2333 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2335 * filename.ch (fn_interp_vars): Fixed small buglet.
2337 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2339 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2340 so as to modify the string inline. Thus makeing it easier to
2341 destroy the results when no longer needed.
2343 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2345 * category.c (cat_stored_values_destroy): Fixed memory leak.
2347 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2349 Add some missing frees. Thanks to John Darrington for reporting
2352 * any-writer.c (any_writer_close): Free writer.
2354 * any-reader.c (any_reader_close): Free reader.
2356 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2358 * por-file-reader.c: (error) Mark as NO_RETURN.
2360 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2362 * settings.c: Changed default value of scompress to true.
2364 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2366 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2368 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2371 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2372 here from pref.h.orig.
2374 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2376 * sys-file-reader.c: Fixed bug reading compressed files.
2378 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2380 * Numerous renames. See src/ChangeLog for details.
2382 * Moved files from src directory