1 2007-11-24 Ben Pfaff <blp@gnu.org>
3 Fix warning reported when reading back system files that include
4 very long string variables. Thanks to Guido Gay <gay@irer.it> for
7 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
8 was written where `j' was meant.
10 2007-11-18 Ben Pfaff <blp@gnu.org>
12 Properly write variables that include a range of missing values to
13 system files. Thanks to Guido Gay <gay@irer.it> for reporting
16 * sys-file-writer.c (write_variable): Correctly calculate
17 n_missing_values field when writing variables that include a range
20 2007-11-10 Ben Pfaff <blp@gnu.org>
22 Cleanups and bug fixes devised while writing up documentation.
25 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
27 * dict-class.c: New file.
28 (dict_class_from_id): Move here.
29 (dict_class_to_name): Move here.
31 * dict-class.h: New file.
32 (enum dict_class): Move here. Change from consecutive integers to
33 consecutive bits, to make testing for any of multiple values
34 easier. Add new DC_ALL constant.
36 * dictionary.c (struct dictionary): Change `case_limit' from
38 (dict_get_vars): Make final argument an enum dict_class.
39 (dict_get_vars_mutable): Ditto.
40 (dict_get_case_limit): Change return value to casenumber.
41 (dict_set_case_limit): Change final argument to a casenumber.
42 (dict_unset_split_var): Add assertion. Rephrase slightly.
43 (dict_set_label): Use xstrndup to simplify.
45 * format.c (fmt_step_width): AHEX format also needs 2-byte
47 (fmt_set_style): Simplify assertions.
49 * missing-values.c (mv_add_num_range): Rename mv_add_range.
50 Simplify implementation.
51 (mv_has_value): Simplify implementation.
52 (mv_pop_value): Remove the first value, not the last, to avoid
53 having GET followed by SAVE reverse the order of missing values.
54 (mv_peek_value): Rename mv_get_value. Simplify assertion.
55 (mv_has_range): Simplify implementation.
56 (mv_peek_range): Rename mv_get_range.
57 (can_resize_string): Removed.
58 (mv_is_resizable): Use value_is_resizable.
59 (mv_resize): Use value_resize.
61 * short-names.h (SHORT_NAME_LEN): Move here.
63 * val-type.h: New file, for definitions related to type and width
64 of abstract values. Before, these definitions were mixed among
65 those related to "union value" and those related to variables.
66 (macro SYSMIS): Move here.
67 (macro LOWEST): Move here.
68 (macro HIGHEST): Move here.
69 (macro MAX_STRING); Move here.
70 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
71 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
72 All references updated.
73 (val_type_is_valid): New function. Replaces var_type_is_valid.
74 All references updated.
75 (val_type_from_width): New function. Replaces
76 var_type_from_width. All references updated.
78 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
80 (val_labs_can_set_width): Use value_is_resizable.
81 (val_labs_add): Simply return false if the value labels set is too
82 wide, instead of having undefined behavior.
83 (val_labs_replace): Ditto.
84 (val_labs_replace): Ditto.
85 (val_labs_first): Set iterator to null if iteration is complete.
86 (val_labs_first_sorted): Ditto.
87 (val_labs_done): Become a no-op if the iterator is null.
89 * value.c (value_is_resizable): New function.
90 (value_resize): New function.
92 * variable.c (var_get_dict_class): New function.
94 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
97 2007-11-08 Ben Pfaff <blp@gnu.org>
99 * data-in.c: Make formatted data parsing locale-independent.
100 (parse_number): Use c_strtod instead of strtod, to avoid
101 locale-specific behavior.
104 2007-11-06 Ben Pfaff <blp@gnu.org>
106 Patch #6256: add support for binary, 360 file formats. Reviewed
109 * data-in.c (struct data_in): Add `encoding' member.
110 (data_in): Add `encoding' parameter, and re-encode the data passed
111 in where appropriate. Update all callers to pass it in.
112 (parse_A): Implement EBCDIC recoding wart described in manual.
113 (parse_AHEX): Implement EBCDIC recoding.
115 * data-out.c (data_out_legacy): New function.
116 (data_out): Make into a wrapper around data_out_legacy.
118 * file-handle-def.c (struct file_handle): New member `encoding'.
119 (fh_create_file): Set encoding.
120 (fh_default_properties): Set default encoding.
121 (fh_get_legacy_encoding): New function.
123 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
124 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
125 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
126 (struct fh_properties): New member `encoding'.
128 2007-11-05 Ben Pfaff <blp@gnu.org>
130 Patch #6258. Reviewed by John Darrington.
132 * file-handle-def.c (fh_lock): Add comment that TYPE should be
133 marked with N_() in the caller. Added these markings to each
134 caller too. Should make i18n easier.
135 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
137 2007-11-03 Ben Pfaff <blp@gnu.org>
139 Allow output files to overwrite input files (bug #21280).
141 * file-handle-def.c: Separate locking of files for input and for
142 output, to allow both to take place at once. Also, distinguish a
143 file handle from the identity of the underlying file, because the
144 identity of a file changes over time and the file handle can't
145 represent two different identities.
146 (struct file_handle): Remove `next', `open_cnt', `deleted',
147 `type', `open_mode', `aux', `identity' members. Change `id' from
148 char array to char *. Add `ref_cnt' member.
149 (file_handle_from_ll) New function.
150 (file_handles) Removed.
151 (named_handles) New variable.
152 (fh_init) Initialize named_handles.
153 (fh_done) Remove name from all named_handles.
154 (free_handle) Updated for modified struct file_handle.
155 (unname_handle) New function.
156 (fh_ref) New function.
157 (fh_from_file_name) Removed.
158 (fh_unref) New function.
159 (fh_unname) New function.
160 (fh_from_id) Rewritten.
161 (create_handle) Updated for modified struct file_handle.
162 (fh_inline_file) Increment the handle's ref count.
163 (fh_create_file) Updated for modified struct file_handle.
164 (fh_create_scratch) Ditto.
169 (fh_is_open) Removed.
170 (fh_get_id) Updated for modified struct file_handle.
171 (fh_get_default_handle) Increment the handle's ref count.
172 (fh_set_default_handle) Handle ref counts.
173 (struct fh_lock) New structure.
174 (locks) New static var.
175 (fh_lock) New function.
176 (fh_unlock) New function.
177 (fh_lock_get_aux) New function.
178 (fh_lock_set_aux) New function.
179 (fh_is_locked) New function.
180 (make_key) New function.
181 (free_key) New function.
182 (compare_fh_locks) New function.
183 (hash_fh_lock) New function.
185 * file-handle-def.h (enum fh_access) New enum.
187 * file-name.c: Made file_identity the same in all supported
189 (struct file_identity): New `name' member.
190 (fn_get_identity): For a file that doesn't exist, get the
191 dev/inode of its directory plus its name. If even the directory
192 doesn't exist, just use its name. Merge the Windows
193 implementation into the Unix one.
194 (fn_compare_file_identities): Rewritten. Merge the Windows
195 implementation into the Unix one.
196 (fn_hash_identity): New function.
198 * make-file.c (struct replace_file): New structure.
199 (all_files): New static var.
200 (replace_file_start): New function.
201 (replace_file_commit): New function.
202 (replace_file_abort): New function.
203 (free_replace_file): New function.
204 (unlink_replace_files): New function.
206 * por-file-reader.c (struct pfm_reader): Add `lock' member.
207 (close_reader): Unlock file.
208 (pfm_open_reader): Lock file.
210 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
212 (pfm_open_writer): Lock file and prepare for its replacement.
213 (close_writer): Unlock file.
215 * scratch-handle.h (struct scratch_handle): Add unique_id so that
216 different generations of a scratch file can be distinguished.
218 * scratch-reader.c (scratch_reader_open): Verify that the file is
221 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
222 remove scratch_handle member.
223 (scratch_writer_open): Lock handle. Prepare to replace handle
224 data, instead of doing it immediately.
225 (scratch_writer_casewriter_destroy): Replace handle data and
228 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
229 (sfm_open_reader): Lock file.
230 (close_reader): Unlock file.
232 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
234 (sfm_open_writer): Lock file and prepare for its replacement.
235 (close_writer): Unlock file.
237 2007-11-02 Ben Pfaff <blp@gnu.org>
239 * data-out.c (output_number): Use isfinite (from C99) instead of
240 gsl_isfinite, since we now have universal support for it in
242 (output_infinite): Ditto.
243 [!HAVE_ROUND] (round): Remove definition, since we now have a
244 replacement in gnulib.
246 2007-10-27 John Darrington <john@darrington.wattle.id.au>
248 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
249 wasn't initialised correctly. Closes bugs #21330 and 21397
252 2007-10-08 Ben Pfaff <blp@gnu.org>
254 Fix bug #21280. Thanks to John Darrington for review.
256 * file-name.c (create_stream): New function.
258 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
259 file handle before creating the file, to ensure that we don't
260 truncate a file that we're reading. Make code easier to read by
263 * sys-file-write.c (sfm_open_writer): Ditto.
265 2007-10-01 Ben Pfaff <blp@gnu.org>
267 Fix bug #21192. Thanks to John Darrington for review.
269 * casereader.c (casereader_read): Decrement case_cnt before
270 calling the casereader's "read" member function, so that we
271 interact properly with lazy_casereader.
273 * datasheet.c: Add regression test for above bug fix.
274 (clone_datasheet): New function.
275 (lazy_callback): New function.
276 (check_datasheet_casereader): New function.
277 (check_datasheet): Check datasheet contents are reported correctly
278 through an ordinary casereader and a lazy casereader.
279 (clone_model): Use clone_datasheet.
281 2007-09-24 Ben Pfaff <blp@gnu.org>
283 Patch #6210. Reviewed by John Darrington.
284 * settings.c: Drop "long view", which is not needed any longer.
285 (static var long_view): Removed.
286 (force_long_view): Removed.
287 (get_viewwidth): Removed.
288 (init_viewport): Removed long_view reference.
290 2007-09-19 John Darrington <john@darrington.wattle.id.au>
292 * settings.c settings.h: Changed viewport's length and width to be
293 owned by the user interface which uses the data library. This allows
294 better abstraction, and makes dynamically adjustable dimensions easier.
296 2007-09-18 Ben Pfaff <blp@gnu.org>
298 * procedure.c (proc_extract_active_file_data): New function.
300 * lazy-casereader.h: New file.
302 * lazy-casereader.c: New file.
304 * casereader.c (casereader_dynamic_cast): New function.
306 2007-09-14 Ben Pfaff <blp@gnu.org>
308 * dictionary.c (dict_clone): Copy case indexes from cloned
309 dictionary. Fixes bug #21061. Reviewed by John Darrington.
311 2007-09-13 John Darrington <john@darrington.wattle.id.au>
313 * value.c value.h (create_value): New function.
315 2007-09-12 Ben Pfaff <blp@gnu.org>
317 Make it clear that translator casereader and casewriter translate
318 functions are supposed to destroy their input case.
320 * casereader-translator.c (struct casereader_translator): Change
321 input case parameter of translate member function from const
322 struct ccase * to struct ccase *.
323 (casereader_create_translator): Ditto, for translate parameter.
325 * casewriter-translator.c (struct casewriter_translator): Ditto.
326 (casewriter_create_translator): Ditto.
328 2007-08-27 John Darrington <john@darrington.wattle.id.au>
330 * sys-file-reader.c (read_display_parameters): Force display width
331 to 8, if the sys file says 0 (like SPSS does).
333 2007-08-12 Ben Pfaff <blp@gnu.org>
335 * dictionary.c (dict_dump): New function.
337 2007-08-12 Ben Pfaff <blp@gnu.org>
339 Drop dict_compactor in favor of using the new struct case_map.
341 * dictionary.c (struct copy_map): Removed.
342 (struct dict_compactor): Removed.
343 (dict_make_compactor): Removed.
344 (dict_compactor_compact): Removed.
345 (dict_compactor_destroy): Removed.
347 * procedure.c (struct data_set): Change `compactor' member to be a
349 (proc_open): Use case_map_to_compact_dict instead of
351 (proc_casereader_read): Use case_map_execute instead of
352 dict_compactor_compact.
353 (proc_commit): Use case_map_destroy instead of
354 dict_compactor_destroy.
356 * scratch-writer.c (struct scratch_writer): Change `compactor'
357 member to be a struct case_map *.
358 (scratch_writer_open): Use case_map_to_compact_dict instead of
360 (scratch_writer_casewriter_write): Use case_map_execute instead of
361 dict_compactor_compact.
363 2007-08-12 Ben Pfaff <blp@gnu.org>
365 * automake.mk: Add case-map.c, case-map.h.
367 * case-map.c: New file.
369 * case-map.h: New file.
371 2007-08-12 Ben Pfaff <blp@gnu.org>
373 * dictionary.c (dict_compact_values): Don't delete scratch
374 variables as well as compacting case indexes. Update all callers.
375 (dict_get_compacted_value_cnt): Rename dict_count_values and
376 change interface. Update all callers.
377 (dict_get_compacted_value_cnt): Remove.
378 (dict_compacting_would_shrink): Remove.
379 (dict_compacting_would_change): Remove.
380 (dict_make_compactor): Add new parameter. Update all callers.
382 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
383 instead of dict_count_values, changing an O(N) operation into
386 2007-08-12 Ben Pfaff <blp@gnu.org>
388 * casereader.c (casereader_read): Don't require cases read by a
389 casereader to be exactly the expected size: as long as they're big
392 2007-08-12 Ben Pfaff <blp@gnu.org>
394 Make casewriters keep track of the number of `union value's in
395 each case. This is useful for two reasons: casewriter_write can
396 then check that the case being written is large enough, and later
397 recipients of the casewriter can determine the size of the case.
399 * casewriter-translator.c (casewriter_create_translator): Add
402 * casewriter.c (struct casewriter): Add value_cnt member.
403 (casewriter_write): Check that the case passed in is big enough.
404 (casewriter_get_value_cnt): New function.
405 (casewriter_create): Add value_cnt parameter.
407 2007-08-09 Ben Pfaff <blp@gnu.org>
409 Fix bug reported by Jason Stover.
410 * settings.c: Move get_termcap_viewport and in particular the
411 #include for <curses.h> to the end of the file. curses.h
412 redefines bool on some systems (e.g. OpenBSD), which causes
413 disagreement between uses of bool before its inclusion and
415 Tested by Jason Stover.
417 2007-07-29 Ben Pfaff <blp@gnu.org>
419 Provisional fix for bug #18692 and bug #20161. Reviewed by John
422 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
423 argument (never "rb" or "wb") because SUSv3 says that only those
424 modes are defined, and glibc in fact rejects other modes.
426 Open portable files with fn_open so that they can be read from
427 pipes. Fix missing fh_close call to go along with fh_open.
428 Report an error if the file close reports an error.
429 * por-file-reader.c (close_reader): New function.
430 (por_file_casereader_destroy): Use close_reader.
431 (pfm_open_reader): Open file with fn_open.
433 2007-07-28 Ben Pfaff <blp@gnu.org>
435 Make PSPP able to read all the portable files I could find on the
436 web. Thanks to John Darrington for review. Bug #17620.
437 * por-file-reader.c (struct pfm_reader): New member `line_length'.
438 (error): Print file offset in hexadecimal.
439 (warning): New function.
440 (advance): Treat lines less than 80 bytes long as padded to 80
442 (pfm_open_reader): Call read_documents if we find an "E" record.
443 (convert_format): Convert invalid formats to the default format
444 instead of aborting reading the file.
445 (read_variables): Rename duplicate variable names instead of
446 aborting reading the file.
447 (read_value_label): Allow string variables of different widths to
448 be assigned value labels in the same record. Replace duplicate
449 value labels instead of aborting.
450 (read_documents): New function.
452 * por-file-writer.c (pfm_open_writer): Call write_documents if the
453 dictionary has documents.
454 (write_documents): New function.
456 2007-07-25 Ben Pfaff <blp@gnu.org>
458 Fix bugs related to bug #17213.
460 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
461 since the former is what config.h has. Include the needed ncurses
463 (static var echo) Rename to `do_echo' because the original name is
464 the same as an ncurses identifier.
465 (get_termcap_viewport) Use error instead of msg.
467 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
468 (fn_close): Don't close stdin, stdout, stderr.
470 2007-07-26 John Darrington <john@darrington.wattle.id.au>
472 * procedure.c procedure.h: Added callbacks which get invoked whenever
473 a dataset's transformation chain changes.
475 2007-07-24 Ben Pfaff <blp@gnu.org>
478 * sys-file-writer.c (write_variable_display_parameters): Use new
479 var_default_display_width function to choose display width of
480 segments after the first one in a given variable.
481 * variable.c (var_create): Use var_default_display_width to pick
482 new variable's display width.
483 (var_default_display_width): New function.
484 Reviewed by John Darrington.
486 2007-07-24 Ben Pfaff <blp@gnu.org>
489 * por-file-writer.c (write_variables): Write weight variable.
490 Reviewed by John Darrington.
492 2007-07-23 Ben Pfaff <blp@gnu.org>
494 Improvements to system file reader and writer.
496 First, move all detailed knowledge of very long strings into
497 sys-file-private.[ch], so that this nasty stuff can be isolated.
499 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
500 (EFFECTIVE_VLS_CHUNK): New macro.
501 (min_int): New function.
502 (max_int): New function.
503 (sfm_width_to_bytes): Rewrite.
504 (sfm_width_to_octs): New function.
505 (sfm_segment_alloc_width): New function.
506 (sfm_segment_alloc_bytes): New function.
507 (sfm_segment_used_bytes): New function.
508 (sfm_segment_offset): New function.
509 (sfm_segment_effective_offset): New function.
510 (sfm_dictionary_to_sfm_vars): New function.
512 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
513 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
514 (struct sfm_var): New structure.
516 Next, improvements to the system file reader.
518 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
519 type casenumber. Added `version_major', `version_minor',
522 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
523 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
524 Change `case_cnt' to type casenumber. Removed `has_vls'.
525 (struct sfm_var): Removed.
526 (sfm_open_reader): Don't warn on wrong case size if the file was
527 written by SPSS 13, which tends to get it wrong. Use
528 sfm_dictionary_to_sfm_vars.
529 (read_header): Always output system file info.
530 (read_variable_record): Simplify code for reading missing values.
531 (read_machine_int32_info): Save version numbers from system file
532 into info struct passed as new argument.
533 (read_long_string_map): Restructured to use new sys-file-private
535 (read_value_labels): Use size_overflow_p.
536 (sys_file_casereader_read): Get rid of distinction between fast
537 and slow paths. Use information provided by sys-file-primate's
538 struct sfm_var to simplify code.
539 (skip_whole_strings): New function.
540 (read_int32): Renamed read_int. Changed return value to int.
542 (read_flt64): Renamed read_float. Changed return value to
543 double. Updated all callers.
544 (int32_to_native): Removed. Changed callers to use
546 (flt64_to_double): Removed. Changed callers to use float_convert.
548 Finally, get rid of int32, flt64 terminology and types in system
549 file writer. The former wasn't very useful since a POSIX "int"
550 can hold the whole range of int32 and we generally didn't have a
551 need for it to be exactly-32-bits, just at-least-32-bits. The
552 latter was inconvenient because we had to assume that it could be
553 different from double and thereby convert special values SYSMIS,
554 HIGHEST, LOWEST to and from it in multiple places. Instead, now
555 we just use "int" and "double" in most places, and do conversions,
556 if necessary, very close to where we do I/O. This change meant
557 that the writer code couldn't represent records in the file as C
558 structs any longer, but that's no great loss. The code actually
559 seems to be more readable without them.
561 Simplify the compression buffering code: only buffer as much as
562 necessary, which is no more than eight 8-byte units at any given
565 * sys-file-writer.c (typedef flt64): Removed.
566 (macro second_lowest_flt64): Removed.
567 (struct sysfile_header): Removed.
568 (struct sysfile_variable): Removed.
569 (struct sfm_writer): Removed `needs_translation', `has_vls',
570 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
571 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
572 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
573 compression buffering by `opcodes', `opcode_cnt', `data',
574 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
575 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
576 data writer functions instead of structures.
577 (calc_oct_idx): New function.
578 (write_header): Use simple data writer functions instead of
580 (write_format_spec): Renamed write_format. New argument.
581 (write_variable_continuation_records): New function.
582 (write_variable): Use simple data writer functions instead of
583 structures. Use write_variable_continuation_records. Write
584 entire very long string instead of requiring caller to understand
586 (write_value_labels): Use simple data writer functions instead of
588 (write_documents): Ditto.
589 (write_variable_display_parameters): Use sys-file-private
590 functions to simplify. Use simple data writer functions instead
592 (write_vls_length_table): Use simple data writer functions instead
594 (write_longvar_table): Ditto.
595 (write_rec_7_34): Break into new functions
596 write_integer_info_record, write_float_info_record. Use simple
597 data writer functions instead of structures.
598 (buf_write): Removed.
599 (append_string_max): Removed.
600 (ensure_buf_space): Removed.
601 (sys_file_casewriter_write): Get rid of the distinction between
602 fast and slow paths, which didn't seem to be too useful. Use new
603 functions write_case_uncompressed, write_case_compressed.
604 (put_instruction): Removed.
605 (put_element): Removed.
606 (write_compressed_data): Removed.
607 (close_writer): Use flush_compressed. Only write case count to
608 system file if it will fit in the field.
609 (write_case_compressed): New function.
610 (write_case_uncompressed): New function.
611 (flush_compressed): New function.
612 (put_cmp_opcode): New function.
613 (put_cmp_number): New function.
614 (write_int): New function.
615 (convert_double_to_output_format): New function.
616 (write_float): New function.
617 (write_value): New function.
618 (write_string): New function.
619 (write_bytes): New function.
620 (write_zeros): New function.
621 (write_spaces): New function.
623 Reviewed by John Darrington.
625 2007-07-22 Ben Pfaff <blp@gnu.org>
627 Don't try to write very long strings to portable files. The
628 format does not support it.
630 * por-file-writer.c (MAX_POR_WIDTH): New macro.
631 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
632 (write_format): Add arg to take width to resize format to.
633 (write_value): Limit width of value written to MAX_POR_WIDTH.
634 (write_variables): Limit width of variable and its output formats
636 Reviewed by John Darrington.
638 2007-07-22 Ben Pfaff <blp@gnu.org>
640 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
641 local variable instead of literal 5.
642 Reviewed by John Darrington.
644 2007-07-22 Ben Pfaff <blp@gnu.org>
646 Fix problems with uniqueness of short names in system files with
647 very long string variables. Now a variable may have multiple
650 * automake.mk (src_data_libdata_a_SOURCES): Add new files
651 short-names.c, short-names.h.
653 * dictionary.c (dict_clone): Clone all the short names.
654 (compare_strings): Move into short-names.c.
655 (hash_strings): Ditto.
656 (set_var_short_name_suffix): Ditto.
657 (dict_assign_short_names): Ditto, rename short_names_assign,
658 change to assign all short names.
660 * por-file-writer.c (write_variables): Use short_names_assign
661 instead of dict_assign_short_names.
663 * short-names.c: New file.
665 * short-names.h: New file.
667 * sys-file-private.c (sfm_width_to_segments): New function.
669 * sys-file-reader.c (read_long_var_name_map): Save and restore all
670 the short names, not just the first one.
672 * sys-file-writer.c (cont_var_name): Removed.
673 (sfm_open_writer): Use short_names_assign instead of
674 dict_assign_short_names. Use unique short names assigned by
675 short_names_assign instead of those generated by cont_var_name.
677 * variable.c (struct variable): Remove `short_name' member,
678 replace by `short_names' and `short_name_cnt'.
679 (var_create) Initialize new members.
680 (var_get_short_name_cnt): New function.
681 (var_get_short_name): Now takes an index argument. Changed most
683 (var_set_short_name): Ditto.
684 (var_clear_short_name): Renamed var_clear_short_names, changed to
685 clear all short names.
687 Reviewed by John Darrington.
689 2007-07-22 Ben Pfaff <blp@gnu.org>
691 * variable.c (var_set_width): Use new var_set_width function.
693 * missing-values.c (mv_n_values): Drop assertion, which was not
696 * format.c (fmt_default_for_width): New function.
697 (fmt_resize): New function.
699 Reviewed by John Darrington.
701 2007-07-18 John Darrington <john@darrington.wattle.id.au>
703 * datasheet.c (datasheet_delete_columns): Added assertion to check
704 we're not deleting outside the range of the sheet.
707 * dictionary.c dictionary.h variable.c: Added the ability for string
708 variables to be resized.
710 * vardict.h: Added some prototypes (moved from dictionary.h) as
711 these should only be called by variable.c
714 2007-07-14 John Darrington <john@darrington.wattle.id.au>
716 * sfm-reader.c: Respect case_cnt field in file header.
718 2007-07-01 John Darrington <john@darrington.wattle.id.au>
720 * transformation.c transformation.h (trns_chain_execute): Changed the
721 signature (Patch #6057)
723 2007-06-10 Ben Pfaff <blp@gnu.org>
725 * casereader-filter.c (casereader_filter_destroy): Make sure to
726 write all the remaining excluded cases to the casewriter, if any.
728 * caseinit.c (init_list_destroy): Rewrite.
729 (init_list_clear): Ditto.
731 * casegrouper.c (casegrouper_get_next_group): Always set *reader
732 to null when returning false.
734 2007-06-06 Ben Pfaff <blp@gnu.org>
736 Actually implement the new procedure code and adapt all of its
737 clients to match. Also adapt all of the other case sources and
738 sinks in the tree and their clients to use the
739 casereader/casewriter infrastructure.
741 * automake.mk: Add and remove files.
743 * any-reader.c: Change into a casereader.
744 * por-file-reader.c: Ditto.
745 * scratch-reader.c: Ditto.
746 * sys-file-reader.c: Ditto.
748 * any-writer.c: Change into a casewriter.
749 * por-file-writer.c: Ditto.
750 * scratch-writer.c: Ditto.
751 * sys-file-writer.c: Ditto.
753 * procedure.c: Change to use casereader, casewriter, caseinit, and
754 other new infrastructure.
756 * scratch-handle.c: Adapt to new infrastructure.
758 * case-sink.c: Removed, now dead code.
759 * case-sink.h: Ditto.
760 * case-source.c: Ditto.
761 * case-source.h: Ditto.
762 * casefile-factory.c: Ditto.
763 * casefile-private.h: Ditto.
766 * casefilter.c: Ditto.
767 * casefilter.h: Ditto.
770 * fastfile-factory.c: Ditto.
771 * fastfile-factory.h: Ditto.
772 * storage-stream.c: Ditto.
773 * storage-stream.h: Ditto.
775 2007-06-06 Ben Pfaff <blp@gnu.org>
779 * automake.mk: Add new files.
781 * datasheet.c: New file.
783 * datasheet.h: New file.
785 2007-06-06 Ben Pfaff <blp@gnu.org>
787 Until now, the procedure code has provided a case to the
788 case_source, which has filled in the data values that come from
789 the active file. "Left" data values that don't come from the
790 active file naturally stay the same from case to case, because the
791 procedure code keeps using that same case.
793 One of the compromises that comes with the new procedure code is
794 that the active file allocates and provides its own case, which
795 the procedure code then has to resize to provide room for any
796 other variables that should go in the case and then fill in the
797 values of "left" variables. Then, when we're done with that case,
798 we have to save the values of "left" variables to copy into the
799 next case read from the active file.
801 The caseinit code helps with this.
803 * automake.mk: Add new files.
805 * caseinit.c: New file.
807 * caseinit.h: New file.
809 2007-06-06 Ben Pfaff <blp@gnu.org>
811 * value.h (value_cnt_from_width): New function.
813 * variable.c (var_get_value_cnt): Use new function.
815 2007-06-06 Ben Pfaff <blp@gnu.org>
817 Add casegrouper, to allow cases read from a given casereader to be
818 broken into groups, each of which has its own casereader.
819 Generally cases are grouped based on having equal values for some
822 * automake.mk: Add new files.
824 * casegrouper.c: New file.
826 * casegrouper.h: New file.
828 2007-06-06 Ben Pfaff <blp@gnu.org>
830 Add interface to lexicographical ordering of cases.
832 * automake.mk: Add new files.
834 * case-ordering.c: New file.
836 * case-ordering.h: New file.
838 2007-06-06 Ben Pfaff <blp@gnu.org>
840 Add casereaders and casewriters, the basis of the new data processing
841 implementation. A casereader is a uniform interface to reading cases
842 from a data source; a casewriter is a uniform interface to writing
843 cases to a data sink.
845 * automake.mk: Add new files.
847 * casereader-filter.c: New file.
849 * casereader-provider.h: New file.
851 * casereader-translator.c: New file.
853 * casereader.c: New file.
855 * casereader.h: New file.
857 * casewriter-provider.h: New file.
859 * casewriter-translator.c: New file.
861 * casewriter.c: New file.
863 * casewriter.h: New file.
865 2007-06-06 Ben Pfaff <blp@gnu.org>
867 "casewindow" data structure that extends the deque (from libpspp)
868 of cases with the ability to dump cases to disk if we get too many
871 * automake.mk: Add new files.
873 * casewindow.c: New file.
875 * casewindow.h: New file.
877 2007-06-06 Ben Pfaff <blp@gnu.org>
879 sparse_cases data structure that augments a sparse_array of cases
880 with the ability to dump cases to disk if we get too many cases in
883 * automake.mk: Add new files.
885 * sparse-cases.c: New file.
887 * sparse-cases.h: New file.
889 2007-06-06 Ben Pfaff <blp@gnu.org>
891 Adds a low-level on-disk case array data structure.
893 * automake.mk: Add new files.
895 * case-tmpfile.c: New file.
897 * case-tmpfile.h: New file.
899 2007-06-06 Ben Pfaff <blp@gnu.org>
901 In a couple of places we calculate the maximum number of cases to
902 keep in memory based on the user-defined workspace. Enable
903 centralizing the calculation through a new function.
905 * settings.c (get_workspace_cases): New function.
907 2007-06-06 Ben Pfaff <blp@gnu.org>
909 The casenumber type is defined in transformations.h, but case.h is
910 a more sensible place. Move it.
912 * case.h (CASENUMBER_MAX): New macro.
913 (typedef casenumber): Move here, from transformations.h.
915 2007-06-03 Ben Pfaff <blp@gnu.org>
917 Slightly generalize case_to_values and case_from_values functions.
919 * case.c (case_to_values): Rename case_copy_out, change interface.
920 (case_from_values): Rename case_copy_in, change interface.
922 * fastfile.c (fastfilereader_get_next_case): Update caller.
923 (write_case_to_disk): Ditto.
925 2007-06-02 Ben Pfaff <blp@gnu.org>
927 Clean up after a forgotten part of patch #5829.
929 * casedeque.h: Remove unused file.
931 * automake.mk: Remove casedeque.h from sources.
933 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
935 * category.c: Removed redundant #include
937 2007-05-06 Ben Pfaff <blp@gnu.org>
939 Abstract the documents within a dictionary a little better.
940 Thanks to John Darrington for suggestion, initial version, and
943 * dictionary.c (struct dictionary): Change `documents' member from
944 char * to struct string.
945 (dict_clear): Destroy struct string.
946 (dict_get_documents): Convert struct string to char *.
947 (dict_set_documents): Set struct string. Pad to 80-character
949 (dict_clear_documents): New function.
950 (dict_add_document_line): New function.
951 (dict_get_document_line_cnt): New function.
952 (dict_get_document_line): New function.
954 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
956 * sys-file-reader.c (read_documents): Use new document functions.
958 2007-04-19 John Darrington <john@darrington.wattle.id.au>
960 * sys-file-reader.c: When reading a system file which has no
961 long name table, automatically create one where the long names
962 are the lower case versions of the short names.
964 2007-04-22 Ben Pfaff <blp@gnu.org>
966 * dictionary.c (dict_set_split_vars): dict_destroy expects that
967 dict_clear will free most data related to the dictionary.
968 dict_clear does a decent job, except that dict_set_split_vars on
969 some systems won't actually free the dict's "split" member.
970 Instead, it'll allocate a 1-byte region. Fix this.
972 * value.c (value_copy): New function.
973 (value_set_missing): Ditto.
975 2007-04-22 John Darrington <john@darrington.wattle.id.au>
977 * Deleted existing category.h and moved cat-routines.h into
978 category.h Encapsulated struct cat_vals better.
980 2007-04-19 John Darrington <john@darrington.wattle.id.au>
982 * sys-file-reader.c: When reading a system file which has no
983 long name table, automatically create one where the long names
984 are the lower case versions of the short names.
986 2007-04-16 John Darrington <john@darrington.wattle.id.au>
988 * sys-file-reader.c: Some versions of Other Software seem to
989 produce system files with string variables' measure set to
990 zero. We'll assume these are supposed to be nominal variables.
992 2007-03-30 Ben Pfaff <blp@gnu.org>
994 * procedure.c: Adapt to new deque data structure.
996 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
997 Ben Pfaff <blp@gnu.org>
999 * file-name.c: Mingw compatibility fixes.
1000 (fn_search_path): Use ISSLASH instead of comparing against '/'
1002 (fn_dir_name): Use dir_name from gnulib.
1003 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1004 (fn_get_identity): Use GetFullPathName instead of canonicalize
1005 from gnulib, because the latter does not fully support
1006 Windows-style path names. Use this implementation based on the
1007 detected presence of Windows instead of the absence of Unix, since
1008 the new implementation is Windows-specific.
1009 (fn_compare_file_identities): In Windows implementation, compare
1010 names case-insensitively.
1012 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1014 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1015 <stdlib.h> to have the same effect.
1017 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1019 * por-file-reader.c: Add missing _() around messages.
1021 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1023 * make-file.c: Include "mkstemp.h", without which linking on
1026 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1028 Reduce platform dependence.
1030 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1031 Everywhere we using this, we really should have just depended on
1032 the shell to expand tildes.
1033 (fn_search_path): Simplify, given that we don't do tilde expansion
1035 (fn_normalize): Removed. Caller changed to use the canonicalize
1037 (fn_get_cwd): Removed. Only user was fn_normalize.
1038 (fn_is_absolute): Really only test for absolute names.
1039 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1041 (fn_readlink): Removed, as it was only used fn_normalize.
1042 (fn_exists): Assume the stat function is available; gnulib does.
1043 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1045 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1047 * dictionary.c (dict_create_vector_assert): New function.
1049 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1051 * file-name.c (fn_normalize): Correct name of function
1052 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1053 for pointing this out.
1055 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1057 We are using a single member in struct file_handle, the "name"
1058 field, for more than one purpose. When it begins with '"', it's a
1059 file name; otherwise, it's a token that can be used to identify
1060 it. When that assertion fires, it's because we searched for the
1061 name case-sensitively as a file name (so that there was no match),
1062 and then we try to insert it case-insensitively as a token, which
1063 fails because duplicates aren't allowed.
1065 Solution: break the two purposes into two separate fields. This
1066 fixes the problem and likely makes the code easier to read too.
1068 Fixes bug #18922. Thanks to John Darrington for bug report and
1071 * file-handle-def.c (struct file_handle): New `id' member.
1072 (fh_from_name): Rename fh_from_id. Update all callers.
1073 (create_handle): New `id' parameter. Update all callers.
1074 (fh_create_file): Ditto.
1075 (fh_get_id): New function.
1077 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1079 * case.c (case_is_null): Change return type to bool.
1081 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1085 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1086 reference counting to share data. This makes it easy for
1087 valgrind, etc. to find accesses to cases that have been destroyed
1088 but have been kept around by another user's ref-count. This often
1089 happens when the data set is small enough to find in memory; if a
1090 bigger data set that would overflow to disk were used, then data
1091 corruption would occur.
1093 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1097 * case.c (case_unshare): Make it check internally whether the
1098 ref_cnt is greater than 1, so that the callers don't have to.
1099 Update callers not to check.
1101 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1103 Before, I was thinking that I might want to get rid of reference
1104 counting at some point. Now, I'm pretty sure that it's here to
1105 stay. Thus, because we have to store the value_cnt anyway for
1106 reference-counted cases, we might as well expose it to users.
1108 * case.c (case_get_value_cnt): New function.
1109 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1110 resize case if its size actually changed.
1112 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1113 instead of peeking inside struct case directly.
1114 (casefile_append): Ditto.
1116 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1118 Get rid of the inlines for the case functions, which made the
1119 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1120 the inlines didn't speed up "make check" at all, which is not a
1121 perfect benchmark but seems indicative.)
1123 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1124 definitions. Remove some assertions on nonnull pointers that were
1125 redundant with a pointer dereference soon after in the function.
1127 (struct case_data): Move definition here from case.h.
1131 (case_data_wr): Ditto.
1133 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1135 * automake.mk: Add casedeque.h to sources.
1137 * casedeque.h: New file.
1139 * procedure.c: (struct dataset) Change lag_count, lag_head,
1140 lag_queue member into single struct casedeque member. Update all
1141 users to use the casedeque instead.
1144 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1146 * procedure.c: Simplify lagged cases interface. Updated all
1147 clients--well, the only client--to use the simplified interface.
1148 (dataset_n_lag) Removed.
1149 (dataset_set_n_lag) Removed.
1150 (dataset_need_lag) New function.
1152 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1154 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1155 whenever appropriate, but only when the dataset/dictionary is in a
1158 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1160 * dictionary.c dictionary.h : Added callbacks for change of filter and
1161 split variables. Refactored some code to ensure that callbacks get
1162 invoked when appropriate.
1164 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1165 replace_dict callback occurs when permanent_dict replaces the current
1168 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1170 * dictionary.c dictionary.h : Added callback for when the weight
1171 variable of a dictionary changes.
1173 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1175 * dictionary.c dictionary.h : Added replace_source and replace_dict
1176 callbacks, and functions to deal with them.
1178 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1180 Simplify missing value handling.
1182 * missing-values.h (enum mv_class): New type.
1183 (enum mv_type): Moved definition into missing-values.c and renamed
1184 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1185 enums. Updated all uses.
1186 (struct missing_values): Changed type of `type' from `enum
1187 mv_type' to `int' because the definition is no longer exposed.
1189 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1190 parameter. Update all callers.
1191 (mv_is_num_missing): Ditto.
1192 (mv_is_str_missing): Ditto.
1193 (mv_is_value_user_missing): Removed. Changed callers to use
1194 mv_is_value_missing.
1195 (mv_is_num_user_missing): Removed. Changed callers to use
1197 (mv_is_str_user_missing): Removed. Changed callers to use
1199 (mv_is_value_system_missing): Removed. Changed callers to use
1200 mv_is_value_missing.
1201 (mv_set_type): Removed. Changed callers to use mv_clear.
1202 (mv_clear): New function.
1204 * variable.c (var_is_value_missing): Add new enum mv_class
1205 parameter. Update all callers.
1206 (var_is_num_missing): Ditto.
1207 (var_is_str_missing): Ditto.
1208 (var_is_value_user_missing): Removed. Changed callers to use
1209 var_is_value_missing.
1210 (var_is_num_user_missing): Removed. Changed callers to use
1212 (var_is_str_user_missing): Removed. Changed callers to use
1214 (var_is_value_system_missing): Removed. Changed callers to use
1215 var_is_value_missing.
1217 * casefilter.c (struct casefilter): Use enum mv_class in place of
1219 (casefilter_variable_missing): Adapt to new member.
1220 (casefilter_create): Change signature to take enum mv_class,
1223 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1225 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1227 * case-sink.c case-sink.h procedure.c procedure.h
1228 storage-stream.c: Now uses the factory.
1230 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1232 Make it possible to pull cases from the active file with a
1233 function call, instead of requiring indirection through a callback
1236 * case-source.h (struct case_source_class): Change ->read function
1237 to return a single case, instead of calling a callback function
1238 for each case. Change ->destroy function to return an error
1241 * case-source.c (free_case_source): Pass along the value returned
1242 by the case_source ->destroy function.
1244 * procedure.c (struct write_case_data): Removed.
1245 (struct dataset): Added some members to track procedure state.
1246 (procedure): Optimize the trivial case at this level.
1247 (internal_procedure): Re-implement in terms of proc_open,
1248 proc_read, proc_close.
1249 (proc_open) New function.
1250 (proc_read) New function.
1251 (proc_close) New function.
1252 (write_case) Moved into proc_read.
1253 (close_active_file) Moved closing of data source into proc_close.
1255 * storage-source.c: Rewrote to conform with modified
1256 case_source_class interface.
1258 * transformations.c (trns_chain_execute): Added argument to allow
1259 starting execution from an arbitrary transformation. Updated
1262 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1264 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1266 * sys-file-reader.c (read_display_parameters): Don't assume that
1267 MEASURE_* and ALIGN_* have the same values found in system files.
1269 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1271 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1272 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1273 the value of n_MEASURES, which was off by 1 (at least from my
1276 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1278 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1279 callbacks which are invoked when the dictionary or its
1280 variables are changed.
1282 * missing-values.c missing-values.h value-labels.c: Tidied up
1283 consistency checks, and made some of them return false
1284 instead of assert-failing.
1286 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1288 * calendar.c (calendar_days_in_month): New function.
1290 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1292 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1293 the value label's width.
1295 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1297 * sfm-private.h: Move contents into sys-file-writer.c, which is
1298 the only remaining user. Removed Borland C++-specific directives.
1300 * sys-file-reader.c: Clean up and rewrite entire file. The
1301 rewritten version is simpler and better abstracted, and should be
1302 easier to maintain and extend. It avoids using structures to read
1303 file data, which is prone to padding variations among compilers.
1304 It should also handle non-IEEE 754 system files, although I
1305 haven't been able to find any. It has been tested against many
1306 .sav files obtained from the Web and found to produce the same
1307 results as the earlier version of the code, or in some cases
1308 improved results. It is more tolerant of format variations found
1311 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1312 member, putting an enum integer_format in its place. New member
1313 `float_format'. Changed `compressed' member to type bool.
1315 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1317 * dictionary.c (dict_delete_consecutive_vars): New function.
1319 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1321 * file-name.c (fn_search_path): Remove prefix arg that was unused
1322 by any caller. Updated all callers.
1324 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1326 * format.c (fmt_dollar_template): Use user's decimal point
1327 character. Add assertion.
1329 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1331 * format.c (fmt_dollar_template): New function, based on
1332 dollar_format_template from var-type-dialog.c.
1334 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1336 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1339 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1341 Finish converting struct variable to an opaque type. In this
1342 phase, we add remaining setter and getter functions, convert the
1343 remaining PSPP code to use them, and do a bunch of cleanup. The
1344 resulting changes are pervasive but mostly trivial, and only the
1345 notable changes are logged.
1347 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1350 * case.c (case_data): Renamed case_data_idx.
1351 (case_num): Renamed case_num_idx.
1352 (case_str): Renamed case_str_idx.
1353 (case_data_rw): Renamed case_data_rw_idx.
1355 * case.h (case_data): New function with old name and an interface
1356 that takes a variable instead of an index, which is easier to
1357 use. Updated all callers to use the new interface, or to use the
1358 new *_idx function (see above).
1361 (case_data_rw): Ditto.
1363 * category.c (cat_stored_values_destroy): Changed interface to
1364 take a struct cat_vals * instead of a struct variable *.
1366 * dictionary.c (dict_clone): Use new vector_clone function.
1367 (dict_clear) Use new var_destroy function.
1368 (add_var) New function.
1369 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1370 (dict_create_var_assert) Rewrite in terms of add_var.
1371 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1372 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1373 (dict_lookup_var) Use new var_create, var_destroy functions.
1374 (dict_contains_var) Rewrite in terms of new vardict functionality.
1375 (set_var_dict_index) New function.
1376 (set_var_case_index) New function.
1377 (reindex_vars) New function.
1378 (dict_delete_var) Rewrite in terms of new vardict functionality.
1379 (dict_reorder_var) Ditto.
1380 (dict_reorder_vars) Ditto.
1381 (rename_var) New function.
1382 (dict_rename_var) Use rename_var.
1383 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1384 (dict_get_compacted_idx_to_fv) Rename
1385 dict_get_compacted_dict_index_to_case_index, update callers.
1386 (dict_create_vector) Use new vector_create function.
1387 (dict_clear_vectors) Use new vector_destroy function.
1388 (set_var_short_name_suffix) Move here from variable.c, renamed
1389 from var_set_short_name_suffix, make static, update caller.
1391 * sys-file-private.c: New file.
1392 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1393 width_to_bytes, update callers.
1395 * sys-file-private.h: New file. Later it will supplant
1396 sfm-private.h; for now it supplements it.
1397 (macro MIN_VERY_LONG_STRING) New macro.
1398 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1400 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1401 MAX_LONG_STRING was used before.
1403 * sys-file-writer.c: Ditto.
1405 * value-labels.c: Change the paradigm here to be that a null
1406 pointer is OK for a struct val_labs * in most cases; it just
1407 represents an empty set of value labels.
1408 (val_labs_copy) A copy of a null set is a null set.
1409 (val_labs_count) A null set has 0 labels.
1410 (val_labs_replace) Change return type to void. Rewrite for
1412 (val_labs_find) A null set does not contain the value.
1413 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1414 transposed argument order, updated all callers.
1416 * value.c: New file.
1417 (value_dup) Moved here from variable.c.
1418 (compare_values) Ditto.
1421 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1422 (macro MAX_LONG_STRING) Removed, because it was only interesting
1423 for system files, not for general code.
1424 (macro MAX_VERY_LONG_STRING) Ditto.
1425 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1426 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1428 * vardict.h: New file, for an interface between variables and
1431 * variable.c: A lot of functions were moved around, for better
1433 (struct variable) Move definition here, from variable.h.
1434 (var_type_adj) Removed--makes i18n hard.
1435 (var_type_noun) Ditto.
1436 (var_create) New function.
1437 (var_clone) New function.
1438 (var_destroy) New function.
1439 (var_set_name) Assert that variable is not in a dictionary.
1440 (compare_var_names) Rename compare_vars_by_name and fix a couple
1441 of callers who thought the args were strings.
1442 (hash_var_name) Rename hash_var_by_name.
1443 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1444 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1445 (var_is_very_long_string) Removed, because it was only interesting
1446 to system file code.
1447 (var_set_missing_values) Allow the argument to be the wrong width,
1448 as long as we can resize it. Simplify callers who were doing the
1449 resizing themselves.
1450 (var_get_value_labels) New function.
1451 (var_has_value_labels) New function.
1452 (var_set_value_labels) New function.
1453 (alloc_value_labels) New function.
1454 (var_add_value_label) New function.
1455 (var_replace_value_label) New function.
1456 (var_clear_value_labels) New function.
1457 (var_lookup_value_label) New function.
1458 (var_get_value_name) Moved here from variable.c, renamed from
1459 var_get_value_name, transposed argument order, updated all
1461 (var_to_string) Moved here, from variable-label.c.
1462 (var_set_leave) New function.
1463 (var_get_leave) New function.
1464 (var_must_leave) New function.
1465 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1466 set_var_short_name_suffix.
1467 (var_get_dict_index) New function.
1468 (var_get_case_index) New function.
1469 (var_get_obs_vals) New function.
1470 (var_set_obs_vals) New function.
1471 (var_has_obs_vals) New function.
1472 (var_get_vardict) New function.
1473 (var_set_vardict) New function.
1474 (var_has_vardict) New function.
1475 (var_clear_vardict) New function.
1476 (value_dup) Moved to value.c.
1477 (compare_values) Ditto.
1480 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1481 (enum ALPHA) Rename VAR_STRING, update all users.
1483 * vector.c: New file.
1484 (struct vector) Moved here, from variable.h.
1485 (check_widths) New function.
1486 (vector_create) New function.
1487 (vector_clone) New function.
1488 (vector_destroy) New function.
1489 (vector_get_name) New function.
1490 (vector_get_var) New function.
1491 (vector_get_var_cnt) New function.
1492 (compare_vector_ptrs_by_name) New function.
1494 * vector.h: New file.
1496 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1498 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1499 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1502 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1504 * value-labels.c (destroy_atoms): New function.
1505 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1507 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1509 Thanks to Jason Stover for pointing out this problem.
1511 * data-out.c (output_number): Use gsl_finite from GSL, which is
1512 portable, instead of isfinite, which is not.
1515 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1517 * variable.c variable.h (value_dup): New function.
1519 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1521 Start converting struct variable to an opaque type. In this
1522 phase, we add a bunch of setter and getter functions and convert
1523 most of the PSPP code to use them. The resulting changes are
1524 pervasive but mostly trivial, and only the notable changes are
1527 * format.c (fmt_equal): New function.
1529 * variable.c (var_type_is_valid): New function.
1530 (measure_is_valid) Moved here, from format.c.
1531 (alignment_is_valid) Moved here, from format.c.
1532 (var_get_name) New function.
1533 (var_set_name) New function.
1534 (width_to_type) New function.
1535 (var_get_type) New function.
1536 (var_get_width) New function.
1537 (var_set_width) New function.
1538 (var_is_numeric) New function.
1539 (var_is_alpha) New function.
1540 (var_is_short_string) New function.
1541 (var_is_long_string) New function.
1542 (var_is_very_long_string) New function.
1543 (var_get_missing_values) New function.
1544 (var_set_missing_values) New function.
1545 (var_clear_missing_values) New function.
1546 (var_has_missing_values) New function.
1547 (var_is_value_missing) New function.
1548 (var_is_num_missing) New function.
1549 (var_is_str_missing) New function.
1550 (var_is_value_user_missing) New function.
1551 (var_is_num_user_missing) New function.
1552 (var_is_str_user_missing) New function.
1553 (var_is_value_system_missing) New function.
1554 (var_get_print_format) New function.
1555 (var_set_print_format) New function.
1556 (var_get_write_format) New function.
1557 (var_set_write_format) New function.
1558 (var_set_both_formats) New function.
1559 (var_get_label) New function.
1560 (var_set_label) New function.
1561 (var_clear_label) New function.
1562 (var_has_label) New function.
1563 (var_get_measure) New function.
1564 (var_set_measure) New function.
1565 (var_get_display_width) New function.
1566 (var_set_display_width) New function.
1567 (var_get_alignment) New function.
1568 (var_set_alignment) New function.
1569 (var_get_value_cnt) New function.
1570 (var_get_leave) New function.
1571 (var_get_short_name) New function.
1573 * variable.h: (struct variable) Removed "type" and "nv" members;
1574 they are now computed from "width" where needed.
1576 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1578 * missing-values.c (mv_resize): Don't write beyond end of the
1579 allocated buffer when resizing a long string.
1581 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1583 Clean up identifier code: don't require identifier enumerations to
1584 be in a particular order; make better use of string library;
1585 expose less of the internals.
1587 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1588 change interface. Updated all callers.
1589 (lex_id_match) Change interface to use struct substring, update
1591 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1592 (global array keywords[]) Make static, change form. Update all
1593 users to use lex_id_name instead.
1594 (lex_is_keyword) New function.
1595 (lex_id_to_token) Change interface to use struct substring, update
1597 (lex_id_name) New function.
1599 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1600 lex_is_keyword instead.
1601 (T_LAST_KEYWORD) Removed.
1602 (T_N_KEYWORDS) Removed.
1604 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1606 * format.c: (fmt_date_template) Distinguish characters for which a
1607 space is output and any date delimiter is allowed on input, from
1608 those for which a space is output and only a space is allowed on
1609 input. The former is represented by X, the latter by a space.
1610 Also, drop distinction between h and H, changing the former to the
1613 * data-in.c: Completely rewrite internals to conform to SPSS input
1614 formats as closely as possible.
1615 (data_in) Changed external interface by replacing the structure
1616 that was used as a single argument by a set of arguments. Updated
1618 (data_in_finite_line) Removed. Converted all callers to use plain
1620 (data_in_get_integer_format) New function.
1621 (data_in_set_integer_format) New function.
1622 (data_in_get_float_format) New function.
1623 (data_in_set_float_format) New function.
1625 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1626 (struct data_in) Removed.
1628 * data-out.c: (output_date) Drop each component from the input as
1629 it is output, to allow us to drop the distinction between h (a
1630 count of hours) and H (the hour of day) template characters.
1631 Also, handle new X template character.
1632 (output_scientific) Follow more rational rule on when to drop
1633 fraction introduced between SPSS 13 and 15. Updated test case to
1636 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1638 Fix buffer overflow reported by John Darrington.
1640 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1641 realize that DIGITS is a count of nibbles, not of bytes.
1643 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1645 * calendar.c (calendar_offset_to_gregorian) Also return the
1646 year-of-day. Change callers to new interface.
1648 * data-out.c: Completely rewrite internals to conform to SPSS
1649 output formats as completely as possible.
1650 (data_out) Change interface to put input parameters before output
1651 parameters, for consistency with the style I now prefer. Update
1653 (data_out_get_integer_format) New public function.
1654 (data_out_set_integer_format) New public function.
1655 (data_out_get_float_format) New public function.
1656 (data_out_set_float_format) New public function.
1658 * data-out.h: New file. Move prototype for data_out here, from
1661 * format.c: (fmt_step_width) Use equality comparison instead of
1662 bitwise and, for clarity.
1663 (fmt_is_string) Ditto.
1664 (fmt_input_to_output) Fix categories that are translated to F
1667 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1669 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1670 casefile-private.h: Added casefilter to assist commands with missing
1673 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1675 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1677 * settings.c: (route_errors_to_terminal) New variable.
1678 (route_errors_to_listing) New variable.
1679 (get_error_routing_to_terminal) New function.
1680 (set_error_routing_to_terminal) New function.
1681 (get_error_routing_to_listing) New function.
1682 (set_error_routing_to_listing) New function.
1684 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1686 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1688 * format.c: Completely rewrite, to achieve better abstraction.
1689 Rewrite all references to formats in other files.
1691 * format.def: Rewrite and reorganize.
1693 * settings.c: Move everything related to custom currency formats
1694 into format.[ch], changing them in form, so as to group related
1695 code and definitions better. Changed all references to use the
1697 (static var decimal) Removed.
1698 (static var grouping) Removed.
1699 (static var cc) Removed.
1700 (get_decimal) Removed.
1701 (set_decimal) Removed.
1702 (get_grouping) Removed.
1703 (set_grouping) Removed.
1707 * settings.h: (macro CC_CNT) Removed.
1708 (macro CC_WIDTH) Removed.
1709 (struct custom_currency) Removed.
1711 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1713 * data-in.c (data_in): Use switch statement instead of table, to
1714 avoid dependence on the order of the FMT_* enums.
1716 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1718 * data-out.c: (num_to_string) Removed, because it was dead code.
1720 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1722 * data-in.c (parse_trailer): Fix error message.
1724 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1726 * format.c (fmt_is_binary): New function.
1728 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1730 * procedure.c procedure.h: Encapsulated the static data into a single
1733 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1735 * casefile.c (casereader_read_xfer): Always initialize the case,
1736 even on an error condition.
1738 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1740 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1741 entire function useless.
1743 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1745 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1746 casereader method casereader_clone.
1748 * procedure.c pransformations.h: Introduced new type casenum_t
1750 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1752 * variable.c: (width_to_bytes) Rephrase code for clarify.
1754 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1756 * format.c: (fmt_type_from_string) New function.
1757 (fmt_to_string) Include decimals in output if the format has
1758 decimals, even if the format type does not. This way, we can
1759 accurately reproduce incorrect formats in user output.
1760 (check_common_specifier) Make the check for a bad format type an
1761 assertion, so we get bug reports if they show up. Fix message.
1762 Check for decimal places with a format type that doesn't allow
1764 (check_input_specifier) Remove check for FMT_X, which has been
1766 (check_output_specifier) Ditto.
1768 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1770 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1771 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1772 (enum fmt_parse_flags) Removed.
1774 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1776 * casefile.c casefile.h: Converted to an abstract base class.
1777 * casefile-private.h fastfile.c fastfile.h: New files.
1778 * automake.mk procedure.c scratch-writer.c storage-stream.c
1780 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1782 * procedure.c (internal_procedure): Create sink_case with only as
1783 many values as the compacted dictionary.
1785 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1787 Remove "debugging" code that caused plenty of false positives and
1790 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1792 * case.c: Remove all references to `this' member.
1794 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1796 Fix link error noted by Jason Stover.
1798 * storage-stream.c: Include <assert.h>.
1800 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1802 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1803 ALL) and additional underlying system file issues.
1805 Thanks to John Darrington for review.
1807 First problem: var_hash points to variables not owned by the
1808 sys-file-reader, which the caller may free or modify. Use an
1809 array of sfm_vars instead, as done earlier (e.g. CVS version
1812 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1813 members and remove all code that references it. Add vars, var_cnt
1814 members. Remove fix_specials member, which was unused.
1815 (struct sfm_var) Remove name member, which was unused.
1816 (sfm_close_reader) Free vars member instead of var_hash.
1817 (compare_var_shortnames) Removed.
1818 (hash_var_shortname) Removed.
1819 (sfm_open_reader) Fill out vars array.
1820 (compare_var_index) Removed.
1821 (sfm_read_case) Use vars instead of var_hash.
1823 Second problem: we're confused about when we actually have very
1824 long strings, causing us to choose incorrectly between slow path
1825 and fast path in sfm_read_case.
1827 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1828 have very long strings, not when we have long variable names,
1829 which is an unrelated feature.
1831 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1833 * variable.h: Move var_set and variable parsing declarations to
1834 new header, src/language/lexer/variable-parser.h. Modified lots
1835 of files to include the new header.
1837 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1839 * value-labels.c (value_to_string): When there's no value label,
1840 format the variable according to its print format, instead of
1841 always effectively using A or F format.
1843 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1845 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1848 * format.c: Removed tortological assertion.
1850 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1852 Reform string library.
1854 * file-name.c (fn_interp_vars): Change interface to take a
1855 substring as input. Updated all users.
1857 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1859 * format.c (measure_is_valid): Really return false when m >=
1862 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1864 Implement random access to casefiles, for use in GUI.
1866 * casefile.c: (struct casereader) Add `random', `file_ofs',
1867 `buffer_ofs' members.
1868 (casefile_get_random_reader) New function.
1869 (read_open_file) Break part into new function
1870 seek_and_fill_buffer().
1871 (fill_buffer) Update buffer_ofs, file_ofs.
1872 (casereader_seek) New function.
1874 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1876 * settings.c: Added call to i18n{done, init}.
1878 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1880 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1883 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1885 * casefile.c: Convert many uses of `int' to `bool'.
1887 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1889 * transformations.c (trns_chain_destroy): Destroy chain's trns
1890 member, to fix memory leak.
1892 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1894 * storage-stream.c (storage_source_decapsulate): Destroy case
1895 source to fix memory leak.
1897 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1899 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1900 case passed as argument instead of initializing the argument as a
1901 case. Fixes memory leak that showed up in
1902 tests/command/aggregate.sh with scratch files.
1904 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1906 * procedure.c (proc_done): Destroy default_dict, to fix memory
1909 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1911 Simplify procedure_with_splits().
1913 * procedure.c (struct split_aux_data): Removed case_count member.
1914 (procedure_with_splits) Don't initialize case_count.
1915 (split_procedure_case_func) Check whether prev_case is null
1916 instead of case_count.
1917 (split_procedure_end_func) Ditto.
1919 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1921 * case.c (case_move): Do nothing if dst and src are the same
1923 (case_try_create) Merge two similar cases.
1924 (case_copy) Unshare only if data must be actually copied.
1926 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1928 * data-in.c data-out.c dictionary.c sys-file-reader.c
1929 sys-file-writer.c variable.c variable.h: Reworked very long string
1930 support for better encapsulation.
1932 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1934 * value-labels.c (val_labs_can_set_width): New function.
1935 (val_labs_set_width) Clear labels if increasing width to long
1937 (val_labs_destroy) Remove unneeded test for null.
1939 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1941 * value-labels.h: Remove unneeded dependency on variable.h.
1943 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1945 Get rid of `char *c' member in union value, for cleanliness.
1947 * value.h: (union value) Remove `c' member.
1949 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1951 Make dictionary compacting functions a little more general.
1953 * sys-file-writer.c (sfm_open_writer): Use
1954 dict_compacting_would_change().
1955 (does_dict_need_translation) Removed.
1957 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1959 Make dictionary compacting functions a little more general.
1961 * dictionary.c (dict_needs_compaction): Rename
1962 dict_compacting_would_shrink(). Update all callers.
1963 (dict_compacting_would_change) New function.
1965 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1967 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1968 inverted return value (!).
1970 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1972 Continue reforming procedure execution.
1974 * procedure.c: Search and replace "vfm" by "proc". Notably:
1975 (static var vfm_source) Rename proc_source. Update all
1977 (static var vfm_sink) Rename proc_sink. Update all references.
1979 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1981 Continue reforming procedure execution. In this phase, remove
1982 PROCESS IF, which was deprecated anyway and can be easily
1983 simulated with TEMPORARY followed by SELECT IF.
1985 * procedure.c: (open_active_file) Don't call
1986 add_process_if_trns().
1987 (discard_variables) Get rid of redundant call to
1988 proc_cancel_all_transformations().
1989 (add_process_if_trns) Removed.
1990 (process_if_trns_proc) Removed.
1991 (process_if_trns_free) Removed.
1993 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1995 Continue reforming procedure execution. In this phase, add
1996 `const' to the case passed to procedure()'s callback.
1998 Updated all users of procedure() as well.
2000 * procedure.c: (struct write_case_data) Add "const" to ccase
2001 parameter for case_func member.
2002 (procedure) Add "const" to ccase parameter for proc_func
2004 (multipass_case_func) Make ccase parameter const.
2005 (internal_procedure) Add "const" to ccase parameter for case_func
2007 (split_procedure_case_func) Make ccase parameter const.
2008 (multipass_split_case_func) Make ccase parameter const.
2010 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2012 Continue reforming procedure execution. In this phase, get rid of
2013 the output code for SPLIT FILE groups in procedure.c, which really
2014 shouldn't be doing any output. Move it into the individual
2015 procedures instead. This also adds some flexibility.
2017 Updated many users of procedure_with_splits() and
2018 multipass_procedure_with_splits() to call
2019 output_split_file_values() and to deal with increased use of
2022 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2023 parameter to begin_func member.
2024 (procedure_with_splits) Add "const struct ccase *" parameter to
2025 begin_func parameter. Make ccase parameter const in proc_func
2027 (split_procedure_case_func) Don't dump split file group. Pass
2029 (dump_splits) Moved to language/dictionary/split-file.c as
2030 output_split_file_values().
2031 (struct multipass_split_aux_data) Add "const struct ccase *"
2032 parameter to split_func member.
2033 (multipass_procedure_with_splits) Add "const struct ccase *"
2034 parameter to split_func parameter.
2035 (multipass_split_case_func) Save new SPLIT FILE case before
2037 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2039 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2041 Continue reforming procedure execution. Change
2042 internal_procedure() so that it calls open_active_file() and
2043 close_active_file(), which isolates most of the actual procedure
2046 * procedure.c: (struct write_case_data) Rename `proc_func' member
2047 to `case_func' and update all references.
2048 (procedure) Rewrite as one-line wrapper around
2049 internal_procedure().
2050 (struct multipass_aux_data) New.
2051 (multipass_callback) Renamed multipass_case_func(). Use struct
2052 multipass_aux_data as auxiliary data.
2053 (multipass_end_func) New function.
2054 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2055 that uses multipass_case_func, multipass_end_func.
2056 (internal_procedure) Add `end_func' argument. Move optimization
2057 of trivial case in here. Move call to open_active_file() and
2058 close_active_file() in here. Now assert that vfm_source is
2060 (procedure_with_splits_callback) Rename
2061 split_procedure_case_func().
2062 (split_procedure_end_func) New function.
2063 (multipass_split_callback) Rename multipass_split_case_func.
2064 (multipass_split_end_func) New function.
2065 (discard_variables) No need to test for nonnull vfm_source.
2067 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2069 Continue reforming procedure execution. Get rid of unused member.
2071 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2073 (write_case) Don't increment cases_analyzed.
2075 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2077 Continue reforming procedure execution. In this phase, move
2078 procedure.c and procedure.h from src to src/data. Update
2079 makefiles and #includes accordingly.
2081 * procedure.c: Moved here from src/.
2083 * procedure.h: Moved here from src/.
2085 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2087 Continue reforming procedure execution. In this phase, get rid of
2088 many global variables, consolidating procedure execution in
2089 procedure.c. Encapsulate transformations in new "struct
2090 trns_chain". Also, change implementation of N OF CASES, FILTER,
2091 and PROCESS IF from special cases to transformations.
2093 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2096 * dictionary.c: (global variable default_dict) Move to
2099 * variable.h: (TRNS_*) Move to transformations.h.
2100 (struct transformation) Move to transformations.c.
2102 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2104 * sys-file-reader.c: Fixed invalid read problems.
2106 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2108 * storage-stream.c: Add missing function comments.
2110 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2112 Continue reforming procedure execution. In this phase, add some
2113 new, needed functionality to storage-stream.
2115 * storage-stream.c: (storage_source_decapsulate) New function.
2117 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2119 * variable.c (width_to_bytes): Declarations must precede
2120 statements for C90 compliance.
2122 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2124 * data-out.c, data-in.c, variable.c, variable.h: New functions
2125 copy_mangle and copy_demangle for reading/writing cases; emulates the
2126 way SPSS deals with strings > 255 bytes.
2128 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2129 subtype 14 needed for strings longer than 255 bytes.
2131 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2132 instead of literal values. Also fixed some constness issues.
2134 * format.h: Constness
2136 * sfm-private.h: Renamed the case_size identifier, since I discovered
2137 that SPSS's respect for this variable is very nominal.
2139 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2141 Change case limit type from int to size_t.
2143 * dictionary.c: (struct dictionary) Change type of case_limit
2145 (dict_get_case_limit) Change return type.
2146 (dict_set_case_limit) Change parameter type.
2148 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2150 * variable.h: (struct variable) Rename `reinit' member as `leave'
2151 and invert sense. Fix up all references.
2153 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2155 Continue reforming procedure execution. In this phase, break
2156 procedure.c into multiple files.
2158 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2160 * case-sink.c: New file.
2162 * case-sink.h: New file.
2164 * case-source.c: New file.
2166 * case-source.h: New file.
2168 * storage-stream.c: New file.
2170 * storage-stream.h: New file.
2172 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2174 * variable.h: (struct variable) Remove `init' member and all
2175 references to it from other files. It was initialized in several
2176 places, but nothing really ever used it for anything worthwhile.
2177 Thanks to Jason Stover for pointing out how confusing this
2180 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2182 Continue reforming error message support. In this phase, get rid
2183 of message "titles" and put the message text in `struct error'.
2184 Now `struct error' encapsulates a message more properly.
2186 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2187 Format message ourselves.
2189 * data-in.c: (vdls_error) Ditto.
2191 * por-file-reader.c: (error) Ditto.
2193 * sys-file-reader.c: (corrupt_msg) Ditto.
2195 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2197 GNU standards require "file name" instead of "filename" in
2198 documentation. It's nice for our code to follow the convention
2201 * casefile.c: (struct casefile) Rename `filename' member to
2202 `file_name'. Updated all references.
2204 * file-name.c: [!unix] (struct file_identity) Rename
2205 normalized_filename member to normalized_file_name. Updated all
2208 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2210 We don't really support anything but Unix-like environments well,
2211 so we might as well de-obfuscate by writing directory and path
2212 separators explicitly.
2214 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2216 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2217 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2219 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2222 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2224 GNU standards require "file name" instead of "filename" in
2225 documentation. It's nice for our code to follow the convention
2228 * filename.c: Rename to file-name.c.
2230 * filename.h: Rename to file-name.h. Update all inclusions.
2231 Update header guards.
2233 * automake.mk: Update file names.
2235 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2237 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2239 (fn_basename) Removed (dead code).
2240 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2241 (fn_special_p) Renamed fn_is_special(), all references updated.
2242 (fn_exists_p) Renamed fn_exists(), all references updated.
2244 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2246 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2247 Also, now it only tilde-expands file names, not paths.
2248 (fn_search_path) Tilde-expand one directory at a time.
2250 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2252 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2253 get rid of non-Unixlike version of the code, which has probably
2255 (fn_prepend_dir) Removed (dead code).
2257 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2258 (macro PATH_SEPARATOR_STRING) New.
2259 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2261 Continue reforming error message support. In this phase, we get
2262 rid of VM() and the other msg() support for "verbosity", replacing
2263 it by a new function verbose_msg().
2265 * filename.c: (fn_search_path) Use verbose_msg() instead of
2268 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2270 * sfm-private.h: Get rid of #defines after #error, which makes no
2273 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2275 Get rid of our own int32 type in favor of the standard int32_t
2278 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2281 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2283 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2285 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2287 Remove ill-considered file routines that are no longer used.
2289 * filename.c: (fn_open_ext) Removed.
2290 (fn_close_ext) Removed.
2292 * filename.h: (struct file_ext) Removed.
2294 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2296 * variable.c (var_is_valid_name): Move declarations before code
2299 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2301 * filename.ch (fn_interp_vars): Fixed small buglet.
2303 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2305 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2306 so as to modify the string inline. Thus makeing it easier to
2307 destroy the results when no longer needed.
2309 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2311 * category.c (cat_stored_values_destroy): Fixed memory leak.
2313 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2315 Add some missing frees. Thanks to John Darrington for reporting
2318 * any-writer.c (any_writer_close): Free writer.
2320 * any-reader.c (any_reader_close): Free reader.
2322 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2324 * por-file-reader.c: (error) Mark as NO_RETURN.
2326 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2328 * settings.c: Changed default value of scompress to true.
2330 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2332 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2334 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2337 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2338 here from pref.h.orig.
2340 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2342 * sys-file-reader.c: Fixed bug reading compressed files.
2344 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2346 * Numerous renames. See src/ChangeLog for details.
2348 * Moved files from src directory