1 2007-11-18 Ben Pfaff <blp@gnu.org>
3 Properly write variables that include a range of missing values to
4 system files. Thanks to Guido Gay <gay@irer.it> for reporting
7 * sys-file-writer.c (write_variable): Correctly calculate
8 n_missing_values field when writing variables that include a range
11 2007-11-10 Ben Pfaff <blp@gnu.org>
13 Cleanups and bug fixes devised while writing up documentation.
16 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
18 * dict-class.c: New file.
19 (dict_class_from_id): Move here.
20 (dict_class_to_name): Move here.
22 * dict-class.h: New file.
23 (enum dict_class): Move here. Change from consecutive integers to
24 consecutive bits, to make testing for any of multiple values
25 easier. Add new DC_ALL constant.
27 * dictionary.c (struct dictionary): Change `case_limit' from
29 (dict_get_vars): Make final argument an enum dict_class.
30 (dict_get_vars_mutable): Ditto.
31 (dict_get_case_limit): Change return value to casenumber.
32 (dict_set_case_limit): Change final argument to a casenumber.
33 (dict_unset_split_var): Add assertion. Rephrase slightly.
34 (dict_set_label): Use xstrndup to simplify.
36 * format.c (fmt_step_width): AHEX format also needs 2-byte
38 (fmt_set_style): Simplify assertions.
40 * missing-values.c (mv_add_num_range): Rename mv_add_range.
41 Simplify implementation.
42 (mv_has_value): Simplify implementation.
43 (mv_pop_value): Remove the first value, not the last, to avoid
44 having GET followed by SAVE reverse the order of missing values.
45 (mv_peek_value): Rename mv_get_value. Simplify assertion.
46 (mv_has_range): Simplify implementation.
47 (mv_peek_range): Rename mv_get_range.
48 (can_resize_string): Removed.
49 (mv_is_resizable): Use value_is_resizable.
50 (mv_resize): Use value_resize.
52 * short-names.h (SHORT_NAME_LEN): Move here.
54 * val-type.h: New file, for definitions related to type and width
55 of abstract values. Before, these definitions were mixed among
56 those related to "union value" and those related to variables.
57 (macro SYSMIS): Move here.
58 (macro LOWEST): Move here.
59 (macro HIGHEST): Move here.
60 (macro MAX_STRING); Move here.
61 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
62 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
63 All references updated.
64 (val_type_is_valid): New function. Replaces var_type_is_valid.
65 All references updated.
66 (val_type_from_width): New function. Replaces
67 var_type_from_width. All references updated.
69 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
71 (val_labs_can_set_width): Use value_is_resizable.
72 (val_labs_add): Simply return false if the value labels set is too
73 wide, instead of having undefined behavior.
74 (val_labs_replace): Ditto.
75 (val_labs_replace): Ditto.
76 (val_labs_first): Set iterator to null if iteration is complete.
77 (val_labs_first_sorted): Ditto.
78 (val_labs_done): Become a no-op if the iterator is null.
80 * value.c (value_is_resizable): New function.
81 (value_resize): New function.
83 * variable.c (var_get_dict_class): New function.
85 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
88 2007-11-08 Ben Pfaff <blp@gnu.org>
90 * data-in.c: Make formatted data parsing locale-independent.
91 (parse_number): Use c_strtod instead of strtod, to avoid
92 locale-specific behavior.
95 2007-11-06 Ben Pfaff <blp@gnu.org>
97 Patch #6256: add support for binary, 360 file formats. Reviewed
100 * data-in.c (struct data_in): Add `encoding' member.
101 (data_in): Add `encoding' parameter, and re-encode the data passed
102 in where appropriate. Update all callers to pass it in.
103 (parse_A): Implement EBCDIC recoding wart described in manual.
104 (parse_AHEX): Implement EBCDIC recoding.
106 * data-out.c (data_out_legacy): New function.
107 (data_out): Make into a wrapper around data_out_legacy.
109 * file-handle-def.c (struct file_handle): New member `encoding'.
110 (fh_create_file): Set encoding.
111 (fh_default_properties): Set default encoding.
112 (fh_get_legacy_encoding): New function.
114 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
115 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
116 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
117 (struct fh_properties): New member `encoding'.
119 2007-11-05 Ben Pfaff <blp@gnu.org>
121 Patch #6258. Reviewed by John Darrington.
123 * file-handle-def.c (fh_lock): Add comment that TYPE should be
124 marked with N_() in the caller. Added these markings to each
125 caller too. Should make i18n easier.
126 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
128 2007-11-03 Ben Pfaff <blp@gnu.org>
130 Allow output files to overwrite input files (bug #21280).
132 * file-handle-def.c: Separate locking of files for input and for
133 output, to allow both to take place at once. Also, distinguish a
134 file handle from the identity of the underlying file, because the
135 identity of a file changes over time and the file handle can't
136 represent two different identities.
137 (struct file_handle): Remove `next', `open_cnt', `deleted',
138 `type', `open_mode', `aux', `identity' members. Change `id' from
139 char array to char *. Add `ref_cnt' member.
140 (file_handle_from_ll) New function.
141 (file_handles) Removed.
142 (named_handles) New variable.
143 (fh_init) Initialize named_handles.
144 (fh_done) Remove name from all named_handles.
145 (free_handle) Updated for modified struct file_handle.
146 (unname_handle) New function.
147 (fh_ref) New function.
148 (fh_from_file_name) Removed.
149 (fh_unref) New function.
150 (fh_unname) New function.
151 (fh_from_id) Rewritten.
152 (create_handle) Updated for modified struct file_handle.
153 (fh_inline_file) Increment the handle's ref count.
154 (fh_create_file) Updated for modified struct file_handle.
155 (fh_create_scratch) Ditto.
160 (fh_is_open) Removed.
161 (fh_get_id) Updated for modified struct file_handle.
162 (fh_get_default_handle) Increment the handle's ref count.
163 (fh_set_default_handle) Handle ref counts.
164 (struct fh_lock) New structure.
165 (locks) New static var.
166 (fh_lock) New function.
167 (fh_unlock) New function.
168 (fh_lock_get_aux) New function.
169 (fh_lock_set_aux) New function.
170 (fh_is_locked) New function.
171 (make_key) New function.
172 (free_key) New function.
173 (compare_fh_locks) New function.
174 (hash_fh_lock) New function.
176 * file-handle-def.h (enum fh_access) New enum.
178 * file-name.c: Made file_identity the same in all supported
180 (struct file_identity): New `name' member.
181 (fn_get_identity): For a file that doesn't exist, get the
182 dev/inode of its directory plus its name. If even the directory
183 doesn't exist, just use its name. Merge the Windows
184 implementation into the Unix one.
185 (fn_compare_file_identities): Rewritten. Merge the Windows
186 implementation into the Unix one.
187 (fn_hash_identity): New function.
189 * make-file.c (struct replace_file): New structure.
190 (all_files): New static var.
191 (replace_file_start): New function.
192 (replace_file_commit): New function.
193 (replace_file_abort): New function.
194 (free_replace_file): New function.
195 (unlink_replace_files): New function.
197 * por-file-reader.c (struct pfm_reader): Add `lock' member.
198 (close_reader): Unlock file.
199 (pfm_open_reader): Lock file.
201 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
203 (pfm_open_writer): Lock file and prepare for its replacement.
204 (close_writer): Unlock file.
206 * scratch-handle.h (struct scratch_handle): Add unique_id so that
207 different generations of a scratch file can be distinguished.
209 * scratch-reader.c (scratch_reader_open): Verify that the file is
212 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
213 remove scratch_handle member.
214 (scratch_writer_open): Lock handle. Prepare to replace handle
215 data, instead of doing it immediately.
216 (scratch_writer_casewriter_destroy): Replace handle data and
219 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
220 (sfm_open_reader): Lock file.
221 (close_reader): Unlock file.
223 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
225 (sfm_open_writer): Lock file and prepare for its replacement.
226 (close_writer): Unlock file.
228 2007-11-02 Ben Pfaff <blp@gnu.org>
230 * data-out.c (output_number): Use isfinite (from C99) instead of
231 gsl_isfinite, since we now have universal support for it in
233 (output_infinite): Ditto.
234 [!HAVE_ROUND] (round): Remove definition, since we now have a
235 replacement in gnulib.
237 2007-10-27 John Darrington <john@darrington.wattle.id.au>
239 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
240 wasn't initialised correctly. Closes bugs #21330 and 21397
243 2007-10-08 Ben Pfaff <blp@gnu.org>
245 Fix bug #21280. Thanks to John Darrington for review.
247 * file-name.c (create_stream): New function.
249 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
250 file handle before creating the file, to ensure that we don't
251 truncate a file that we're reading. Make code easier to read by
254 * sys-file-write.c (sfm_open_writer): Ditto.
256 2007-10-01 Ben Pfaff <blp@gnu.org>
258 Fix bug #21192. Thanks to John Darrington for review.
260 * casereader.c (casereader_read): Decrement case_cnt before
261 calling the casereader's "read" member function, so that we
262 interact properly with lazy_casereader.
264 * datasheet.c: Add regression test for above bug fix.
265 (clone_datasheet): New function.
266 (lazy_callback): New function.
267 (check_datasheet_casereader): New function.
268 (check_datasheet): Check datasheet contents are reported correctly
269 through an ordinary casereader and a lazy casereader.
270 (clone_model): Use clone_datasheet.
272 2007-09-24 Ben Pfaff <blp@gnu.org>
274 Patch #6210. Reviewed by John Darrington.
275 * settings.c: Drop "long view", which is not needed any longer.
276 (static var long_view): Removed.
277 (force_long_view): Removed.
278 (get_viewwidth): Removed.
279 (init_viewport): Removed long_view reference.
281 2007-09-19 John Darrington <john@darrington.wattle.id.au>
283 * settings.c settings.h: Changed viewport's length and width to be
284 owned by the user interface which uses the data library. This allows
285 better abstraction, and makes dynamically adjustable dimensions easier.
287 2007-09-18 Ben Pfaff <blp@gnu.org>
289 * procedure.c (proc_extract_active_file_data): New function.
291 * lazy-casereader.h: New file.
293 * lazy-casereader.c: New file.
295 * casereader.c (casereader_dynamic_cast): New function.
297 2007-09-14 Ben Pfaff <blp@gnu.org>
299 * dictionary.c (dict_clone): Copy case indexes from cloned
300 dictionary. Fixes bug #21061. Reviewed by John Darrington.
302 2007-09-13 John Darrington <john@darrington.wattle.id.au>
304 * value.c value.h (create_value): New function.
306 2007-09-12 Ben Pfaff <blp@gnu.org>
308 Make it clear that translator casereader and casewriter translate
309 functions are supposed to destroy their input case.
311 * casereader-translator.c (struct casereader_translator): Change
312 input case parameter of translate member function from const
313 struct ccase * to struct ccase *.
314 (casereader_create_translator): Ditto, for translate parameter.
316 * casewriter-translator.c (struct casewriter_translator): Ditto.
317 (casewriter_create_translator): Ditto.
319 2007-08-27 John Darrington <john@darrington.wattle.id.au>
321 * sys-file-reader.c (read_display_parameters): Force display width
322 to 8, if the sys file says 0 (like SPSS does).
324 2007-08-12 Ben Pfaff <blp@gnu.org>
326 * dictionary.c (dict_dump): New function.
328 2007-08-12 Ben Pfaff <blp@gnu.org>
330 Drop dict_compactor in favor of using the new struct case_map.
332 * dictionary.c (struct copy_map): Removed.
333 (struct dict_compactor): Removed.
334 (dict_make_compactor): Removed.
335 (dict_compactor_compact): Removed.
336 (dict_compactor_destroy): Removed.
338 * procedure.c (struct data_set): Change `compactor' member to be a
340 (proc_open): Use case_map_to_compact_dict instead of
342 (proc_casereader_read): Use case_map_execute instead of
343 dict_compactor_compact.
344 (proc_commit): Use case_map_destroy instead of
345 dict_compactor_destroy.
347 * scratch-writer.c (struct scratch_writer): Change `compactor'
348 member to be a struct case_map *.
349 (scratch_writer_open): Use case_map_to_compact_dict instead of
351 (scratch_writer_casewriter_write): Use case_map_execute instead of
352 dict_compactor_compact.
354 2007-08-12 Ben Pfaff <blp@gnu.org>
356 * automake.mk: Add case-map.c, case-map.h.
358 * case-map.c: New file.
360 * case-map.h: New file.
362 2007-08-12 Ben Pfaff <blp@gnu.org>
364 * dictionary.c (dict_compact_values): Don't delete scratch
365 variables as well as compacting case indexes. Update all callers.
366 (dict_get_compacted_value_cnt): Rename dict_count_values and
367 change interface. Update all callers.
368 (dict_get_compacted_value_cnt): Remove.
369 (dict_compacting_would_shrink): Remove.
370 (dict_compacting_would_change): Remove.
371 (dict_make_compactor): Add new parameter. Update all callers.
373 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
374 instead of dict_count_values, changing an O(N) operation into
377 2007-08-12 Ben Pfaff <blp@gnu.org>
379 * casereader.c (casereader_read): Don't require cases read by a
380 casereader to be exactly the expected size: as long as they're big
383 2007-08-12 Ben Pfaff <blp@gnu.org>
385 Make casewriters keep track of the number of `union value's in
386 each case. This is useful for two reasons: casewriter_write can
387 then check that the case being written is large enough, and later
388 recipients of the casewriter can determine the size of the case.
390 * casewriter-translator.c (casewriter_create_translator): Add
393 * casewriter.c (struct casewriter): Add value_cnt member.
394 (casewriter_write): Check that the case passed in is big enough.
395 (casewriter_get_value_cnt): New function.
396 (casewriter_create): Add value_cnt parameter.
398 2007-08-09 Ben Pfaff <blp@gnu.org>
400 Fix bug reported by Jason Stover.
401 * settings.c: Move get_termcap_viewport and in particular the
402 #include for <curses.h> to the end of the file. curses.h
403 redefines bool on some systems (e.g. OpenBSD), which causes
404 disagreement between uses of bool before its inclusion and
406 Tested by Jason Stover.
408 2007-07-29 Ben Pfaff <blp@gnu.org>
410 Provisional fix for bug #18692 and bug #20161. Reviewed by John
413 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
414 argument (never "rb" or "wb") because SUSv3 says that only those
415 modes are defined, and glibc in fact rejects other modes.
417 Open portable files with fn_open so that they can be read from
418 pipes. Fix missing fh_close call to go along with fh_open.
419 Report an error if the file close reports an error.
420 * por-file-reader.c (close_reader): New function.
421 (por_file_casereader_destroy): Use close_reader.
422 (pfm_open_reader): Open file with fn_open.
424 2007-07-28 Ben Pfaff <blp@gnu.org>
426 Make PSPP able to read all the portable files I could find on the
427 web. Thanks to John Darrington for review. Bug #17620.
428 * por-file-reader.c (struct pfm_reader): New member `line_length'.
429 (error): Print file offset in hexadecimal.
430 (warning): New function.
431 (advance): Treat lines less than 80 bytes long as padded to 80
433 (pfm_open_reader): Call read_documents if we find an "E" record.
434 (convert_format): Convert invalid formats to the default format
435 instead of aborting reading the file.
436 (read_variables): Rename duplicate variable names instead of
437 aborting reading the file.
438 (read_value_label): Allow string variables of different widths to
439 be assigned value labels in the same record. Replace duplicate
440 value labels instead of aborting.
441 (read_documents): New function.
443 * por-file-writer.c (pfm_open_writer): Call write_documents if the
444 dictionary has documents.
445 (write_documents): New function.
447 2007-07-25 Ben Pfaff <blp@gnu.org>
449 Fix bugs related to bug #17213.
451 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
452 since the former is what config.h has. Include the needed ncurses
454 (static var echo) Rename to `do_echo' because the original name is
455 the same as an ncurses identifier.
456 (get_termcap_viewport) Use error instead of msg.
458 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
459 (fn_close): Don't close stdin, stdout, stderr.
461 2007-07-26 John Darrington <john@darrington.wattle.id.au>
463 * procedure.c procedure.h: Added callbacks which get invoked whenever
464 a dataset's transformation chain changes.
466 2007-07-24 Ben Pfaff <blp@gnu.org>
469 * sys-file-writer.c (write_variable_display_parameters): Use new
470 var_default_display_width function to choose display width of
471 segments after the first one in a given variable.
472 * variable.c (var_create): Use var_default_display_width to pick
473 new variable's display width.
474 (var_default_display_width): New function.
475 Reviewed by John Darrington.
477 2007-07-24 Ben Pfaff <blp@gnu.org>
480 * por-file-writer.c (write_variables): Write weight variable.
481 Reviewed by John Darrington.
483 2007-07-23 Ben Pfaff <blp@gnu.org>
485 Improvements to system file reader and writer.
487 First, move all detailed knowledge of very long strings into
488 sys-file-private.[ch], so that this nasty stuff can be isolated.
490 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
491 (EFFECTIVE_VLS_CHUNK): New macro.
492 (min_int): New function.
493 (max_int): New function.
494 (sfm_width_to_bytes): Rewrite.
495 (sfm_width_to_octs): New function.
496 (sfm_segment_alloc_width): New function.
497 (sfm_segment_alloc_bytes): New function.
498 (sfm_segment_used_bytes): New function.
499 (sfm_segment_offset): New function.
500 (sfm_segment_effective_offset): New function.
501 (sfm_dictionary_to_sfm_vars): New function.
503 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
504 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
505 (struct sfm_var): New structure.
507 Next, improvements to the system file reader.
509 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
510 type casenumber. Added `version_major', `version_minor',
513 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
514 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
515 Change `case_cnt' to type casenumber. Removed `has_vls'.
516 (struct sfm_var): Removed.
517 (sfm_open_reader): Don't warn on wrong case size if the file was
518 written by SPSS 13, which tends to get it wrong. Use
519 sfm_dictionary_to_sfm_vars.
520 (read_header): Always output system file info.
521 (read_variable_record): Simplify code for reading missing values.
522 (read_machine_int32_info): Save version numbers from system file
523 into info struct passed as new argument.
524 (read_long_string_map): Restructured to use new sys-file-private
526 (read_value_labels): Use size_overflow_p.
527 (sys_file_casereader_read): Get rid of distinction between fast
528 and slow paths. Use information provided by sys-file-primate's
529 struct sfm_var to simplify code.
530 (skip_whole_strings): New function.
531 (read_int32): Renamed read_int. Changed return value to int.
533 (read_flt64): Renamed read_float. Changed return value to
534 double. Updated all callers.
535 (int32_to_native): Removed. Changed callers to use
537 (flt64_to_double): Removed. Changed callers to use float_convert.
539 Finally, get rid of int32, flt64 terminology and types in system
540 file writer. The former wasn't very useful since a POSIX "int"
541 can hold the whole range of int32 and we generally didn't have a
542 need for it to be exactly-32-bits, just at-least-32-bits. The
543 latter was inconvenient because we had to assume that it could be
544 different from double and thereby convert special values SYSMIS,
545 HIGHEST, LOWEST to and from it in multiple places. Instead, now
546 we just use "int" and "double" in most places, and do conversions,
547 if necessary, very close to where we do I/O. This change meant
548 that the writer code couldn't represent records in the file as C
549 structs any longer, but that's no great loss. The code actually
550 seems to be more readable without them.
552 Simplify the compression buffering code: only buffer as much as
553 necessary, which is no more than eight 8-byte units at any given
556 * sys-file-writer.c (typedef flt64): Removed.
557 (macro second_lowest_flt64): Removed.
558 (struct sysfile_header): Removed.
559 (struct sysfile_variable): Removed.
560 (struct sfm_writer): Removed `needs_translation', `has_vls',
561 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
562 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
563 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
564 compression buffering by `opcodes', `opcode_cnt', `data',
565 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
566 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
567 data writer functions instead of structures.
568 (calc_oct_idx): New function.
569 (write_header): Use simple data writer functions instead of
571 (write_format_spec): Renamed write_format. New argument.
572 (write_variable_continuation_records): New function.
573 (write_variable): Use simple data writer functions instead of
574 structures. Use write_variable_continuation_records. Write
575 entire very long string instead of requiring caller to understand
577 (write_value_labels): Use simple data writer functions instead of
579 (write_documents): Ditto.
580 (write_variable_display_parameters): Use sys-file-private
581 functions to simplify. Use simple data writer functions instead
583 (write_vls_length_table): Use simple data writer functions instead
585 (write_longvar_table): Ditto.
586 (write_rec_7_34): Break into new functions
587 write_integer_info_record, write_float_info_record. Use simple
588 data writer functions instead of structures.
589 (buf_write): Removed.
590 (append_string_max): Removed.
591 (ensure_buf_space): Removed.
592 (sys_file_casewriter_write): Get rid of the distinction between
593 fast and slow paths, which didn't seem to be too useful. Use new
594 functions write_case_uncompressed, write_case_compressed.
595 (put_instruction): Removed.
596 (put_element): Removed.
597 (write_compressed_data): Removed.
598 (close_writer): Use flush_compressed. Only write case count to
599 system file if it will fit in the field.
600 (write_case_compressed): New function.
601 (write_case_uncompressed): New function.
602 (flush_compressed): New function.
603 (put_cmp_opcode): New function.
604 (put_cmp_number): New function.
605 (write_int): New function.
606 (convert_double_to_output_format): New function.
607 (write_float): New function.
608 (write_value): New function.
609 (write_string): New function.
610 (write_bytes): New function.
611 (write_zeros): New function.
612 (write_spaces): New function.
614 Reviewed by John Darrington.
616 2007-07-22 Ben Pfaff <blp@gnu.org>
618 Don't try to write very long strings to portable files. The
619 format does not support it.
621 * por-file-writer.c (MAX_POR_WIDTH): New macro.
622 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
623 (write_format): Add arg to take width to resize format to.
624 (write_value): Limit width of value written to MAX_POR_WIDTH.
625 (write_variables): Limit width of variable and its output formats
627 Reviewed by John Darrington.
629 2007-07-22 Ben Pfaff <blp@gnu.org>
631 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
632 local variable instead of literal 5.
633 Reviewed by John Darrington.
635 2007-07-22 Ben Pfaff <blp@gnu.org>
637 Fix problems with uniqueness of short names in system files with
638 very long string variables. Now a variable may have multiple
641 * automake.mk (src_data_libdata_a_SOURCES): Add new files
642 short-names.c, short-names.h.
644 * dictionary.c (dict_clone): Clone all the short names.
645 (compare_strings): Move into short-names.c.
646 (hash_strings): Ditto.
647 (set_var_short_name_suffix): Ditto.
648 (dict_assign_short_names): Ditto, rename short_names_assign,
649 change to assign all short names.
651 * por-file-writer.c (write_variables): Use short_names_assign
652 instead of dict_assign_short_names.
654 * short-names.c: New file.
656 * short-names.h: New file.
658 * sys-file-private.c (sfm_width_to_segments): New function.
660 * sys-file-reader.c (read_long_var_name_map): Save and restore all
661 the short names, not just the first one.
663 * sys-file-writer.c (cont_var_name): Removed.
664 (sfm_open_writer): Use short_names_assign instead of
665 dict_assign_short_names. Use unique short names assigned by
666 short_names_assign instead of those generated by cont_var_name.
668 * variable.c (struct variable): Remove `short_name' member,
669 replace by `short_names' and `short_name_cnt'.
670 (var_create) Initialize new members.
671 (var_get_short_name_cnt): New function.
672 (var_get_short_name): Now takes an index argument. Changed most
674 (var_set_short_name): Ditto.
675 (var_clear_short_name): Renamed var_clear_short_names, changed to
676 clear all short names.
678 Reviewed by John Darrington.
680 2007-07-22 Ben Pfaff <blp@gnu.org>
682 * variable.c (var_set_width): Use new var_set_width function.
684 * missing-values.c (mv_n_values): Drop assertion, which was not
687 * format.c (fmt_default_for_width): New function.
688 (fmt_resize): New function.
690 Reviewed by John Darrington.
692 2007-07-18 John Darrington <john@darrington.wattle.id.au>
694 * datasheet.c (datasheet_delete_columns): Added assertion to check
695 we're not deleting outside the range of the sheet.
698 * dictionary.c dictionary.h variable.c: Added the ability for string
699 variables to be resized.
701 * vardict.h: Added some prototypes (moved from dictionary.h) as
702 these should only be called by variable.c
705 2007-07-14 John Darrington <john@darrington.wattle.id.au>
707 * sfm-reader.c: Respect case_cnt field in file header.
709 2007-07-01 John Darrington <john@darrington.wattle.id.au>
711 * transformation.c transformation.h (trns_chain_execute): Changed the
712 signature (Patch #6057)
714 2007-06-10 Ben Pfaff <blp@gnu.org>
716 * casereader-filter.c (casereader_filter_destroy): Make sure to
717 write all the remaining excluded cases to the casewriter, if any.
719 * caseinit.c (init_list_destroy): Rewrite.
720 (init_list_clear): Ditto.
722 * casegrouper.c (casegrouper_get_next_group): Always set *reader
723 to null when returning false.
725 2007-06-06 Ben Pfaff <blp@gnu.org>
727 Actually implement the new procedure code and adapt all of its
728 clients to match. Also adapt all of the other case sources and
729 sinks in the tree and their clients to use the
730 casereader/casewriter infrastructure.
732 * automake.mk: Add and remove files.
734 * any-reader.c: Change into a casereader.
735 * por-file-reader.c: Ditto.
736 * scratch-reader.c: Ditto.
737 * sys-file-reader.c: Ditto.
739 * any-writer.c: Change into a casewriter.
740 * por-file-writer.c: Ditto.
741 * scratch-writer.c: Ditto.
742 * sys-file-writer.c: Ditto.
744 * procedure.c: Change to use casereader, casewriter, caseinit, and
745 other new infrastructure.
747 * scratch-handle.c: Adapt to new infrastructure.
749 * case-sink.c: Removed, now dead code.
750 * case-sink.h: Ditto.
751 * case-source.c: Ditto.
752 * case-source.h: Ditto.
753 * casefile-factory.c: Ditto.
754 * casefile-private.h: Ditto.
757 * casefilter.c: Ditto.
758 * casefilter.h: Ditto.
761 * fastfile-factory.c: Ditto.
762 * fastfile-factory.h: Ditto.
763 * storage-stream.c: Ditto.
764 * storage-stream.h: Ditto.
766 2007-06-06 Ben Pfaff <blp@gnu.org>
770 * automake.mk: Add new files.
772 * datasheet.c: New file.
774 * datasheet.h: New file.
776 2007-06-06 Ben Pfaff <blp@gnu.org>
778 Until now, the procedure code has provided a case to the
779 case_source, which has filled in the data values that come from
780 the active file. "Left" data values that don't come from the
781 active file naturally stay the same from case to case, because the
782 procedure code keeps using that same case.
784 One of the compromises that comes with the new procedure code is
785 that the active file allocates and provides its own case, which
786 the procedure code then has to resize to provide room for any
787 other variables that should go in the case and then fill in the
788 values of "left" variables. Then, when we're done with that case,
789 we have to save the values of "left" variables to copy into the
790 next case read from the active file.
792 The caseinit code helps with this.
794 * automake.mk: Add new files.
796 * caseinit.c: New file.
798 * caseinit.h: New file.
800 2007-06-06 Ben Pfaff <blp@gnu.org>
802 * value.h (value_cnt_from_width): New function.
804 * variable.c (var_get_value_cnt): Use new function.
806 2007-06-06 Ben Pfaff <blp@gnu.org>
808 Add casegrouper, to allow cases read from a given casereader to be
809 broken into groups, each of which has its own casereader.
810 Generally cases are grouped based on having equal values for some
813 * automake.mk: Add new files.
815 * casegrouper.c: New file.
817 * casegrouper.h: New file.
819 2007-06-06 Ben Pfaff <blp@gnu.org>
821 Add interface to lexicographical ordering of cases.
823 * automake.mk: Add new files.
825 * case-ordering.c: New file.
827 * case-ordering.h: New file.
829 2007-06-06 Ben Pfaff <blp@gnu.org>
831 Add casereaders and casewriters, the basis of the new data processing
832 implementation. A casereader is a uniform interface to reading cases
833 from a data source; a casewriter is a uniform interface to writing
834 cases to a data sink.
836 * automake.mk: Add new files.
838 * casereader-filter.c: New file.
840 * casereader-provider.h: New file.
842 * casereader-translator.c: New file.
844 * casereader.c: New file.
846 * casereader.h: New file.
848 * casewriter-provider.h: New file.
850 * casewriter-translator.c: New file.
852 * casewriter.c: New file.
854 * casewriter.h: New file.
856 2007-06-06 Ben Pfaff <blp@gnu.org>
858 "casewindow" data structure that extends the deque (from libpspp)
859 of cases with the ability to dump cases to disk if we get too many
862 * automake.mk: Add new files.
864 * casewindow.c: New file.
866 * casewindow.h: New file.
868 2007-06-06 Ben Pfaff <blp@gnu.org>
870 sparse_cases data structure that augments a sparse_array of cases
871 with the ability to dump cases to disk if we get too many cases in
874 * automake.mk: Add new files.
876 * sparse-cases.c: New file.
878 * sparse-cases.h: New file.
880 2007-06-06 Ben Pfaff <blp@gnu.org>
882 Adds a low-level on-disk case array data structure.
884 * automake.mk: Add new files.
886 * case-tmpfile.c: New file.
888 * case-tmpfile.h: New file.
890 2007-06-06 Ben Pfaff <blp@gnu.org>
892 In a couple of places we calculate the maximum number of cases to
893 keep in memory based on the user-defined workspace. Enable
894 centralizing the calculation through a new function.
896 * settings.c (get_workspace_cases): New function.
898 2007-06-06 Ben Pfaff <blp@gnu.org>
900 The casenumber type is defined in transformations.h, but case.h is
901 a more sensible place. Move it.
903 * case.h (CASENUMBER_MAX): New macro.
904 (typedef casenumber): Move here, from transformations.h.
906 2007-06-03 Ben Pfaff <blp@gnu.org>
908 Slightly generalize case_to_values and case_from_values functions.
910 * case.c (case_to_values): Rename case_copy_out, change interface.
911 (case_from_values): Rename case_copy_in, change interface.
913 * fastfile.c (fastfilereader_get_next_case): Update caller.
914 (write_case_to_disk): Ditto.
916 2007-06-02 Ben Pfaff <blp@gnu.org>
918 Clean up after a forgotten part of patch #5829.
920 * casedeque.h: Remove unused file.
922 * automake.mk: Remove casedeque.h from sources.
924 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
926 * category.c: Removed redundant #include
928 2007-05-06 Ben Pfaff <blp@gnu.org>
930 Abstract the documents within a dictionary a little better.
931 Thanks to John Darrington for suggestion, initial version, and
934 * dictionary.c (struct dictionary): Change `documents' member from
935 char * to struct string.
936 (dict_clear): Destroy struct string.
937 (dict_get_documents): Convert struct string to char *.
938 (dict_set_documents): Set struct string. Pad to 80-character
940 (dict_clear_documents): New function.
941 (dict_add_document_line): New function.
942 (dict_get_document_line_cnt): New function.
943 (dict_get_document_line): New function.
945 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
947 * sys-file-reader.c (read_documents): Use new document functions.
949 2007-04-19 John Darrington <john@darrington.wattle.id.au>
951 * sys-file-reader.c: When reading a system file which has no
952 long name table, automatically create one where the long names
953 are the lower case versions of the short names.
955 2007-04-22 Ben Pfaff <blp@gnu.org>
957 * dictionary.c (dict_set_split_vars): dict_destroy expects that
958 dict_clear will free most data related to the dictionary.
959 dict_clear does a decent job, except that dict_set_split_vars on
960 some systems won't actually free the dict's "split" member.
961 Instead, it'll allocate a 1-byte region. Fix this.
963 * value.c (value_copy): New function.
964 (value_set_missing): Ditto.
966 2007-04-22 John Darrington <john@darrington.wattle.id.au>
968 * Deleted existing category.h and moved cat-routines.h into
969 category.h Encapsulated struct cat_vals better.
971 2007-04-19 John Darrington <john@darrington.wattle.id.au>
973 * sys-file-reader.c: When reading a system file which has no
974 long name table, automatically create one where the long names
975 are the lower case versions of the short names.
977 2007-04-16 John Darrington <john@darrington.wattle.id.au>
979 * sys-file-reader.c: Some versions of Other Software seem to
980 produce system files with string variables' measure set to
981 zero. We'll assume these are supposed to be nominal variables.
983 2007-03-30 Ben Pfaff <blp@gnu.org>
985 * procedure.c: Adapt to new deque data structure.
987 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
988 Ben Pfaff <blp@gnu.org>
990 * file-name.c: Mingw compatibility fixes.
991 (fn_search_path): Use ISSLASH instead of comparing against '/'
993 (fn_dir_name): Use dir_name from gnulib.
994 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
995 (fn_get_identity): Use GetFullPathName instead of canonicalize
996 from gnulib, because the latter does not fully support
997 Windows-style path names. Use this implementation based on the
998 detected presence of Windows instead of the absence of Unix, since
999 the new implementation is Windows-specific.
1000 (fn_compare_file_identities): In Windows implementation, compare
1001 names case-insensitively.
1003 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1005 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1006 <stdlib.h> to have the same effect.
1008 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1010 * por-file-reader.c: Add missing _() around messages.
1012 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1014 * make-file.c: Include "mkstemp.h", without which linking on
1017 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1019 Reduce platform dependence.
1021 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1022 Everywhere we using this, we really should have just depended on
1023 the shell to expand tildes.
1024 (fn_search_path): Simplify, given that we don't do tilde expansion
1026 (fn_normalize): Removed. Caller changed to use the canonicalize
1028 (fn_get_cwd): Removed. Only user was fn_normalize.
1029 (fn_is_absolute): Really only test for absolute names.
1030 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1032 (fn_readlink): Removed, as it was only used fn_normalize.
1033 (fn_exists): Assume the stat function is available; gnulib does.
1034 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1036 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1038 * dictionary.c (dict_create_vector_assert): New function.
1040 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1042 * file-name.c (fn_normalize): Correct name of function
1043 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1044 for pointing this out.
1046 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1048 We are using a single member in struct file_handle, the "name"
1049 field, for more than one purpose. When it begins with '"', it's a
1050 file name; otherwise, it's a token that can be used to identify
1051 it. When that assertion fires, it's because we searched for the
1052 name case-sensitively as a file name (so that there was no match),
1053 and then we try to insert it case-insensitively as a token, which
1054 fails because duplicates aren't allowed.
1056 Solution: break the two purposes into two separate fields. This
1057 fixes the problem and likely makes the code easier to read too.
1059 Fixes bug #18922. Thanks to John Darrington for bug report and
1062 * file-handle-def.c (struct file_handle): New `id' member.
1063 (fh_from_name): Rename fh_from_id. Update all callers.
1064 (create_handle): New `id' parameter. Update all callers.
1065 (fh_create_file): Ditto.
1066 (fh_get_id): New function.
1068 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1070 * case.c (case_is_null): Change return type to bool.
1072 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1076 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1077 reference counting to share data. This makes it easy for
1078 valgrind, etc. to find accesses to cases that have been destroyed
1079 but have been kept around by another user's ref-count. This often
1080 happens when the data set is small enough to find in memory; if a
1081 bigger data set that would overflow to disk were used, then data
1082 corruption would occur.
1084 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1088 * case.c (case_unshare): Make it check internally whether the
1089 ref_cnt is greater than 1, so that the callers don't have to.
1090 Update callers not to check.
1092 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1094 Before, I was thinking that I might want to get rid of reference
1095 counting at some point. Now, I'm pretty sure that it's here to
1096 stay. Thus, because we have to store the value_cnt anyway for
1097 reference-counted cases, we might as well expose it to users.
1099 * case.c (case_get_value_cnt): New function.
1100 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1101 resize case if its size actually changed.
1103 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1104 instead of peeking inside struct case directly.
1105 (casefile_append): Ditto.
1107 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1109 Get rid of the inlines for the case functions, which made the
1110 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1111 the inlines didn't speed up "make check" at all, which is not a
1112 perfect benchmark but seems indicative.)
1114 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1115 definitions. Remove some assertions on nonnull pointers that were
1116 redundant with a pointer dereference soon after in the function.
1118 (struct case_data): Move definition here from case.h.
1122 (case_data_wr): Ditto.
1124 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1126 * automake.mk: Add casedeque.h to sources.
1128 * casedeque.h: New file.
1130 * procedure.c: (struct dataset) Change lag_count, lag_head,
1131 lag_queue member into single struct casedeque member. Update all
1132 users to use the casedeque instead.
1135 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1137 * procedure.c: Simplify lagged cases interface. Updated all
1138 clients--well, the only client--to use the simplified interface.
1139 (dataset_n_lag) Removed.
1140 (dataset_set_n_lag) Removed.
1141 (dataset_need_lag) New function.
1143 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1145 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1146 whenever appropriate, but only when the dataset/dictionary is in a
1149 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1151 * dictionary.c dictionary.h : Added callbacks for change of filter and
1152 split variables. Refactored some code to ensure that callbacks get
1153 invoked when appropriate.
1155 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1156 replace_dict callback occurs when permanent_dict replaces the current
1159 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1161 * dictionary.c dictionary.h : Added callback for when the weight
1162 variable of a dictionary changes.
1164 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1166 * dictionary.c dictionary.h : Added replace_source and replace_dict
1167 callbacks, and functions to deal with them.
1169 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1171 Simplify missing value handling.
1173 * missing-values.h (enum mv_class): New type.
1174 (enum mv_type): Moved definition into missing-values.c and renamed
1175 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1176 enums. Updated all uses.
1177 (struct missing_values): Changed type of `type' from `enum
1178 mv_type' to `int' because the definition is no longer exposed.
1180 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1181 parameter. Update all callers.
1182 (mv_is_num_missing): Ditto.
1183 (mv_is_str_missing): Ditto.
1184 (mv_is_value_user_missing): Removed. Changed callers to use
1185 mv_is_value_missing.
1186 (mv_is_num_user_missing): Removed. Changed callers to use
1188 (mv_is_str_user_missing): Removed. Changed callers to use
1190 (mv_is_value_system_missing): Removed. Changed callers to use
1191 mv_is_value_missing.
1192 (mv_set_type): Removed. Changed callers to use mv_clear.
1193 (mv_clear): New function.
1195 * variable.c (var_is_value_missing): Add new enum mv_class
1196 parameter. Update all callers.
1197 (var_is_num_missing): Ditto.
1198 (var_is_str_missing): Ditto.
1199 (var_is_value_user_missing): Removed. Changed callers to use
1200 var_is_value_missing.
1201 (var_is_num_user_missing): Removed. Changed callers to use
1203 (var_is_str_user_missing): Removed. Changed callers to use
1205 (var_is_value_system_missing): Removed. Changed callers to use
1206 var_is_value_missing.
1208 * casefilter.c (struct casefilter): Use enum mv_class in place of
1210 (casefilter_variable_missing): Adapt to new member.
1211 (casefilter_create): Change signature to take enum mv_class,
1214 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1216 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1218 * case-sink.c case-sink.h procedure.c procedure.h
1219 storage-stream.c: Now uses the factory.
1221 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1223 Make it possible to pull cases from the active file with a
1224 function call, instead of requiring indirection through a callback
1227 * case-source.h (struct case_source_class): Change ->read function
1228 to return a single case, instead of calling a callback function
1229 for each case. Change ->destroy function to return an error
1232 * case-source.c (free_case_source): Pass along the value returned
1233 by the case_source ->destroy function.
1235 * procedure.c (struct write_case_data): Removed.
1236 (struct dataset): Added some members to track procedure state.
1237 (procedure): Optimize the trivial case at this level.
1238 (internal_procedure): Re-implement in terms of proc_open,
1239 proc_read, proc_close.
1240 (proc_open) New function.
1241 (proc_read) New function.
1242 (proc_close) New function.
1243 (write_case) Moved into proc_read.
1244 (close_active_file) Moved closing of data source into proc_close.
1246 * storage-source.c: Rewrote to conform with modified
1247 case_source_class interface.
1249 * transformations.c (trns_chain_execute): Added argument to allow
1250 starting execution from an arbitrary transformation. Updated
1253 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1255 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1257 * sys-file-reader.c (read_display_parameters): Don't assume that
1258 MEASURE_* and ALIGN_* have the same values found in system files.
1260 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1262 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1263 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1264 the value of n_MEASURES, which was off by 1 (at least from my
1267 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1269 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1270 callbacks which are invoked when the dictionary or its
1271 variables are changed.
1273 * missing-values.c missing-values.h value-labels.c: Tidied up
1274 consistency checks, and made some of them return false
1275 instead of assert-failing.
1277 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1279 * calendar.c (calendar_days_in_month): New function.
1281 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1283 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1284 the value label's width.
1286 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1288 * sfm-private.h: Move contents into sys-file-writer.c, which is
1289 the only remaining user. Removed Borland C++-specific directives.
1291 * sys-file-reader.c: Clean up and rewrite entire file. The
1292 rewritten version is simpler and better abstracted, and should be
1293 easier to maintain and extend. It avoids using structures to read
1294 file data, which is prone to padding variations among compilers.
1295 It should also handle non-IEEE 754 system files, although I
1296 haven't been able to find any. It has been tested against many
1297 .sav files obtained from the Web and found to produce the same
1298 results as the earlier version of the code, or in some cases
1299 improved results. It is more tolerant of format variations found
1302 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1303 member, putting an enum integer_format in its place. New member
1304 `float_format'. Changed `compressed' member to type bool.
1306 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1308 * dictionary.c (dict_delete_consecutive_vars): New function.
1310 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1312 * file-name.c (fn_search_path): Remove prefix arg that was unused
1313 by any caller. Updated all callers.
1315 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1317 * format.c (fmt_dollar_template): Use user's decimal point
1318 character. Add assertion.
1320 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1322 * format.c (fmt_dollar_template): New function, based on
1323 dollar_format_template from var-type-dialog.c.
1325 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1327 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1330 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1332 Finish converting struct variable to an opaque type. In this
1333 phase, we add remaining setter and getter functions, convert the
1334 remaining PSPP code to use them, and do a bunch of cleanup. The
1335 resulting changes are pervasive but mostly trivial, and only the
1336 notable changes are logged.
1338 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1341 * case.c (case_data): Renamed case_data_idx.
1342 (case_num): Renamed case_num_idx.
1343 (case_str): Renamed case_str_idx.
1344 (case_data_rw): Renamed case_data_rw_idx.
1346 * case.h (case_data): New function with old name and an interface
1347 that takes a variable instead of an index, which is easier to
1348 use. Updated all callers to use the new interface, or to use the
1349 new *_idx function (see above).
1352 (case_data_rw): Ditto.
1354 * category.c (cat_stored_values_destroy): Changed interface to
1355 take a struct cat_vals * instead of a struct variable *.
1357 * dictionary.c (dict_clone): Use new vector_clone function.
1358 (dict_clear) Use new var_destroy function.
1359 (add_var) New function.
1360 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1361 (dict_create_var_assert) Rewrite in terms of add_var.
1362 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1363 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1364 (dict_lookup_var) Use new var_create, var_destroy functions.
1365 (dict_contains_var) Rewrite in terms of new vardict functionality.
1366 (set_var_dict_index) New function.
1367 (set_var_case_index) New function.
1368 (reindex_vars) New function.
1369 (dict_delete_var) Rewrite in terms of new vardict functionality.
1370 (dict_reorder_var) Ditto.
1371 (dict_reorder_vars) Ditto.
1372 (rename_var) New function.
1373 (dict_rename_var) Use rename_var.
1374 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1375 (dict_get_compacted_idx_to_fv) Rename
1376 dict_get_compacted_dict_index_to_case_index, update callers.
1377 (dict_create_vector) Use new vector_create function.
1378 (dict_clear_vectors) Use new vector_destroy function.
1379 (set_var_short_name_suffix) Move here from variable.c, renamed
1380 from var_set_short_name_suffix, make static, update caller.
1382 * sys-file-private.c: New file.
1383 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1384 width_to_bytes, update callers.
1386 * sys-file-private.h: New file. Later it will supplant
1387 sfm-private.h; for now it supplements it.
1388 (macro MIN_VERY_LONG_STRING) New macro.
1389 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1391 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1392 MAX_LONG_STRING was used before.
1394 * sys-file-writer.c: Ditto.
1396 * value-labels.c: Change the paradigm here to be that a null
1397 pointer is OK for a struct val_labs * in most cases; it just
1398 represents an empty set of value labels.
1399 (val_labs_copy) A copy of a null set is a null set.
1400 (val_labs_count) A null set has 0 labels.
1401 (val_labs_replace) Change return type to void. Rewrite for
1403 (val_labs_find) A null set does not contain the value.
1404 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1405 transposed argument order, updated all callers.
1407 * value.c: New file.
1408 (value_dup) Moved here from variable.c.
1409 (compare_values) Ditto.
1412 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1413 (macro MAX_LONG_STRING) Removed, because it was only interesting
1414 for system files, not for general code.
1415 (macro MAX_VERY_LONG_STRING) Ditto.
1416 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1417 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1419 * vardict.h: New file, for an interface between variables and
1422 * variable.c: A lot of functions were moved around, for better
1424 (struct variable) Move definition here, from variable.h.
1425 (var_type_adj) Removed--makes i18n hard.
1426 (var_type_noun) Ditto.
1427 (var_create) New function.
1428 (var_clone) New function.
1429 (var_destroy) New function.
1430 (var_set_name) Assert that variable is not in a dictionary.
1431 (compare_var_names) Rename compare_vars_by_name and fix a couple
1432 of callers who thought the args were strings.
1433 (hash_var_name) Rename hash_var_by_name.
1434 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1435 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1436 (var_is_very_long_string) Removed, because it was only interesting
1437 to system file code.
1438 (var_set_missing_values) Allow the argument to be the wrong width,
1439 as long as we can resize it. Simplify callers who were doing the
1440 resizing themselves.
1441 (var_get_value_labels) New function.
1442 (var_has_value_labels) New function.
1443 (var_set_value_labels) New function.
1444 (alloc_value_labels) New function.
1445 (var_add_value_label) New function.
1446 (var_replace_value_label) New function.
1447 (var_clear_value_labels) New function.
1448 (var_lookup_value_label) New function.
1449 (var_get_value_name) Moved here from variable.c, renamed from
1450 var_get_value_name, transposed argument order, updated all
1452 (var_to_string) Moved here, from variable-label.c.
1453 (var_set_leave) New function.
1454 (var_get_leave) New function.
1455 (var_must_leave) New function.
1456 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1457 set_var_short_name_suffix.
1458 (var_get_dict_index) New function.
1459 (var_get_case_index) New function.
1460 (var_get_obs_vals) New function.
1461 (var_set_obs_vals) New function.
1462 (var_has_obs_vals) New function.
1463 (var_get_vardict) New function.
1464 (var_set_vardict) New function.
1465 (var_has_vardict) New function.
1466 (var_clear_vardict) New function.
1467 (value_dup) Moved to value.c.
1468 (compare_values) Ditto.
1471 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1472 (enum ALPHA) Rename VAR_STRING, update all users.
1474 * vector.c: New file.
1475 (struct vector) Moved here, from variable.h.
1476 (check_widths) New function.
1477 (vector_create) New function.
1478 (vector_clone) New function.
1479 (vector_destroy) New function.
1480 (vector_get_name) New function.
1481 (vector_get_var) New function.
1482 (vector_get_var_cnt) New function.
1483 (compare_vector_ptrs_by_name) New function.
1485 * vector.h: New file.
1487 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1489 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1490 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1493 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1495 * value-labels.c (destroy_atoms): New function.
1496 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1498 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1500 Thanks to Jason Stover for pointing out this problem.
1502 * data-out.c (output_number): Use gsl_finite from GSL, which is
1503 portable, instead of isfinite, which is not.
1506 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1508 * variable.c variable.h (value_dup): New function.
1510 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1512 Start converting struct variable to an opaque type. In this
1513 phase, we add a bunch of setter and getter functions and convert
1514 most of the PSPP code to use them. The resulting changes are
1515 pervasive but mostly trivial, and only the notable changes are
1518 * format.c (fmt_equal): New function.
1520 * variable.c (var_type_is_valid): New function.
1521 (measure_is_valid) Moved here, from format.c.
1522 (alignment_is_valid) Moved here, from format.c.
1523 (var_get_name) New function.
1524 (var_set_name) New function.
1525 (width_to_type) New function.
1526 (var_get_type) New function.
1527 (var_get_width) New function.
1528 (var_set_width) New function.
1529 (var_is_numeric) New function.
1530 (var_is_alpha) New function.
1531 (var_is_short_string) New function.
1532 (var_is_long_string) New function.
1533 (var_is_very_long_string) New function.
1534 (var_get_missing_values) New function.
1535 (var_set_missing_values) New function.
1536 (var_clear_missing_values) New function.
1537 (var_has_missing_values) New function.
1538 (var_is_value_missing) New function.
1539 (var_is_num_missing) New function.
1540 (var_is_str_missing) New function.
1541 (var_is_value_user_missing) New function.
1542 (var_is_num_user_missing) New function.
1543 (var_is_str_user_missing) New function.
1544 (var_is_value_system_missing) New function.
1545 (var_get_print_format) New function.
1546 (var_set_print_format) New function.
1547 (var_get_write_format) New function.
1548 (var_set_write_format) New function.
1549 (var_set_both_formats) New function.
1550 (var_get_label) New function.
1551 (var_set_label) New function.
1552 (var_clear_label) New function.
1553 (var_has_label) New function.
1554 (var_get_measure) New function.
1555 (var_set_measure) New function.
1556 (var_get_display_width) New function.
1557 (var_set_display_width) New function.
1558 (var_get_alignment) New function.
1559 (var_set_alignment) New function.
1560 (var_get_value_cnt) New function.
1561 (var_get_leave) New function.
1562 (var_get_short_name) New function.
1564 * variable.h: (struct variable) Removed "type" and "nv" members;
1565 they are now computed from "width" where needed.
1567 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1569 * missing-values.c (mv_resize): Don't write beyond end of the
1570 allocated buffer when resizing a long string.
1572 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1574 Clean up identifier code: don't require identifier enumerations to
1575 be in a particular order; make better use of string library;
1576 expose less of the internals.
1578 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1579 change interface. Updated all callers.
1580 (lex_id_match) Change interface to use struct substring, update
1582 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1583 (global array keywords[]) Make static, change form. Update all
1584 users to use lex_id_name instead.
1585 (lex_is_keyword) New function.
1586 (lex_id_to_token) Change interface to use struct substring, update
1588 (lex_id_name) New function.
1590 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1591 lex_is_keyword instead.
1592 (T_LAST_KEYWORD) Removed.
1593 (T_N_KEYWORDS) Removed.
1595 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1597 * format.c: (fmt_date_template) Distinguish characters for which a
1598 space is output and any date delimiter is allowed on input, from
1599 those for which a space is output and only a space is allowed on
1600 input. The former is represented by X, the latter by a space.
1601 Also, drop distinction between h and H, changing the former to the
1604 * data-in.c: Completely rewrite internals to conform to SPSS input
1605 formats as closely as possible.
1606 (data_in) Changed external interface by replacing the structure
1607 that was used as a single argument by a set of arguments. Updated
1609 (data_in_finite_line) Removed. Converted all callers to use plain
1611 (data_in_get_integer_format) New function.
1612 (data_in_set_integer_format) New function.
1613 (data_in_get_float_format) New function.
1614 (data_in_set_float_format) New function.
1616 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1617 (struct data_in) Removed.
1619 * data-out.c: (output_date) Drop each component from the input as
1620 it is output, to allow us to drop the distinction between h (a
1621 count of hours) and H (the hour of day) template characters.
1622 Also, handle new X template character.
1623 (output_scientific) Follow more rational rule on when to drop
1624 fraction introduced between SPSS 13 and 15. Updated test case to
1627 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1629 Fix buffer overflow reported by John Darrington.
1631 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1632 realize that DIGITS is a count of nibbles, not of bytes.
1634 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1636 * calendar.c (calendar_offset_to_gregorian) Also return the
1637 year-of-day. Change callers to new interface.
1639 * data-out.c: Completely rewrite internals to conform to SPSS
1640 output formats as completely as possible.
1641 (data_out) Change interface to put input parameters before output
1642 parameters, for consistency with the style I now prefer. Update
1644 (data_out_get_integer_format) New public function.
1645 (data_out_set_integer_format) New public function.
1646 (data_out_get_float_format) New public function.
1647 (data_out_set_float_format) New public function.
1649 * data-out.h: New file. Move prototype for data_out here, from
1652 * format.c: (fmt_step_width) Use equality comparison instead of
1653 bitwise and, for clarity.
1654 (fmt_is_string) Ditto.
1655 (fmt_input_to_output) Fix categories that are translated to F
1658 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1660 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1661 casefile-private.h: Added casefilter to assist commands with missing
1664 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1666 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1668 * settings.c: (route_errors_to_terminal) New variable.
1669 (route_errors_to_listing) New variable.
1670 (get_error_routing_to_terminal) New function.
1671 (set_error_routing_to_terminal) New function.
1672 (get_error_routing_to_listing) New function.
1673 (set_error_routing_to_listing) New function.
1675 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1677 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1679 * format.c: Completely rewrite, to achieve better abstraction.
1680 Rewrite all references to formats in other files.
1682 * format.def: Rewrite and reorganize.
1684 * settings.c: Move everything related to custom currency formats
1685 into format.[ch], changing them in form, so as to group related
1686 code and definitions better. Changed all references to use the
1688 (static var decimal) Removed.
1689 (static var grouping) Removed.
1690 (static var cc) Removed.
1691 (get_decimal) Removed.
1692 (set_decimal) Removed.
1693 (get_grouping) Removed.
1694 (set_grouping) Removed.
1698 * settings.h: (macro CC_CNT) Removed.
1699 (macro CC_WIDTH) Removed.
1700 (struct custom_currency) Removed.
1702 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1704 * data-in.c (data_in): Use switch statement instead of table, to
1705 avoid dependence on the order of the FMT_* enums.
1707 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1709 * data-out.c: (num_to_string) Removed, because it was dead code.
1711 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1713 * data-in.c (parse_trailer): Fix error message.
1715 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1717 * format.c (fmt_is_binary): New function.
1719 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1721 * procedure.c procedure.h: Encapsulated the static data into a single
1724 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1726 * casefile.c (casereader_read_xfer): Always initialize the case,
1727 even on an error condition.
1729 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1731 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1732 entire function useless.
1734 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1736 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1737 casereader method casereader_clone.
1739 * procedure.c pransformations.h: Introduced new type casenum_t
1741 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1743 * variable.c: (width_to_bytes) Rephrase code for clarify.
1745 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1747 * format.c: (fmt_type_from_string) New function.
1748 (fmt_to_string) Include decimals in output if the format has
1749 decimals, even if the format type does not. This way, we can
1750 accurately reproduce incorrect formats in user output.
1751 (check_common_specifier) Make the check for a bad format type an
1752 assertion, so we get bug reports if they show up. Fix message.
1753 Check for decimal places with a format type that doesn't allow
1755 (check_input_specifier) Remove check for FMT_X, which has been
1757 (check_output_specifier) Ditto.
1759 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1761 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1762 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1763 (enum fmt_parse_flags) Removed.
1765 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1767 * casefile.c casefile.h: Converted to an abstract base class.
1768 * casefile-private.h fastfile.c fastfile.h: New files.
1769 * automake.mk procedure.c scratch-writer.c storage-stream.c
1771 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1773 * procedure.c (internal_procedure): Create sink_case with only as
1774 many values as the compacted dictionary.
1776 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1778 Remove "debugging" code that caused plenty of false positives and
1781 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1783 * case.c: Remove all references to `this' member.
1785 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1787 Fix link error noted by Jason Stover.
1789 * storage-stream.c: Include <assert.h>.
1791 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1793 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1794 ALL) and additional underlying system file issues.
1796 Thanks to John Darrington for review.
1798 First problem: var_hash points to variables not owned by the
1799 sys-file-reader, which the caller may free or modify. Use an
1800 array of sfm_vars instead, as done earlier (e.g. CVS version
1803 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1804 members and remove all code that references it. Add vars, var_cnt
1805 members. Remove fix_specials member, which was unused.
1806 (struct sfm_var) Remove name member, which was unused.
1807 (sfm_close_reader) Free vars member instead of var_hash.
1808 (compare_var_shortnames) Removed.
1809 (hash_var_shortname) Removed.
1810 (sfm_open_reader) Fill out vars array.
1811 (compare_var_index) Removed.
1812 (sfm_read_case) Use vars instead of var_hash.
1814 Second problem: we're confused about when we actually have very
1815 long strings, causing us to choose incorrectly between slow path
1816 and fast path in sfm_read_case.
1818 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1819 have very long strings, not when we have long variable names,
1820 which is an unrelated feature.
1822 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1824 * variable.h: Move var_set and variable parsing declarations to
1825 new header, src/language/lexer/variable-parser.h. Modified lots
1826 of files to include the new header.
1828 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1830 * value-labels.c (value_to_string): When there's no value label,
1831 format the variable according to its print format, instead of
1832 always effectively using A or F format.
1834 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1836 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1839 * format.c: Removed tortological assertion.
1841 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1843 Reform string library.
1845 * file-name.c (fn_interp_vars): Change interface to take a
1846 substring as input. Updated all users.
1848 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1850 * format.c (measure_is_valid): Really return false when m >=
1853 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1855 Implement random access to casefiles, for use in GUI.
1857 * casefile.c: (struct casereader) Add `random', `file_ofs',
1858 `buffer_ofs' members.
1859 (casefile_get_random_reader) New function.
1860 (read_open_file) Break part into new function
1861 seek_and_fill_buffer().
1862 (fill_buffer) Update buffer_ofs, file_ofs.
1863 (casereader_seek) New function.
1865 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1867 * settings.c: Added call to i18n{done, init}.
1869 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1871 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1874 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1876 * casefile.c: Convert many uses of `int' to `bool'.
1878 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1880 * transformations.c (trns_chain_destroy): Destroy chain's trns
1881 member, to fix memory leak.
1883 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1885 * storage-stream.c (storage_source_decapsulate): Destroy case
1886 source to fix memory leak.
1888 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1890 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1891 case passed as argument instead of initializing the argument as a
1892 case. Fixes memory leak that showed up in
1893 tests/command/aggregate.sh with scratch files.
1895 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1897 * procedure.c (proc_done): Destroy default_dict, to fix memory
1900 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1902 Simplify procedure_with_splits().
1904 * procedure.c (struct split_aux_data): Removed case_count member.
1905 (procedure_with_splits) Don't initialize case_count.
1906 (split_procedure_case_func) Check whether prev_case is null
1907 instead of case_count.
1908 (split_procedure_end_func) Ditto.
1910 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1912 * case.c (case_move): Do nothing if dst and src are the same
1914 (case_try_create) Merge two similar cases.
1915 (case_copy) Unshare only if data must be actually copied.
1917 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1919 * data-in.c data-out.c dictionary.c sys-file-reader.c
1920 sys-file-writer.c variable.c variable.h: Reworked very long string
1921 support for better encapsulation.
1923 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1925 * value-labels.c (val_labs_can_set_width): New function.
1926 (val_labs_set_width) Clear labels if increasing width to long
1928 (val_labs_destroy) Remove unneeded test for null.
1930 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1932 * value-labels.h: Remove unneeded dependency on variable.h.
1934 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1936 Get rid of `char *c' member in union value, for cleanliness.
1938 * value.h: (union value) Remove `c' member.
1940 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1942 Make dictionary compacting functions a little more general.
1944 * sys-file-writer.c (sfm_open_writer): Use
1945 dict_compacting_would_change().
1946 (does_dict_need_translation) Removed.
1948 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1950 Make dictionary compacting functions a little more general.
1952 * dictionary.c (dict_needs_compaction): Rename
1953 dict_compacting_would_shrink(). Update all callers.
1954 (dict_compacting_would_change) New function.
1956 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1958 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1959 inverted return value (!).
1961 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1963 Continue reforming procedure execution.
1965 * procedure.c: Search and replace "vfm" by "proc". Notably:
1966 (static var vfm_source) Rename proc_source. Update all
1968 (static var vfm_sink) Rename proc_sink. Update all references.
1970 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1972 Continue reforming procedure execution. In this phase, remove
1973 PROCESS IF, which was deprecated anyway and can be easily
1974 simulated with TEMPORARY followed by SELECT IF.
1976 * procedure.c: (open_active_file) Don't call
1977 add_process_if_trns().
1978 (discard_variables) Get rid of redundant call to
1979 proc_cancel_all_transformations().
1980 (add_process_if_trns) Removed.
1981 (process_if_trns_proc) Removed.
1982 (process_if_trns_free) Removed.
1984 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1986 Continue reforming procedure execution. In this phase, add
1987 `const' to the case passed to procedure()'s callback.
1989 Updated all users of procedure() as well.
1991 * procedure.c: (struct write_case_data) Add "const" to ccase
1992 parameter for case_func member.
1993 (procedure) Add "const" to ccase parameter for proc_func
1995 (multipass_case_func) Make ccase parameter const.
1996 (internal_procedure) Add "const" to ccase parameter for case_func
1998 (split_procedure_case_func) Make ccase parameter const.
1999 (multipass_split_case_func) Make ccase parameter const.
2001 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2003 Continue reforming procedure execution. In this phase, get rid of
2004 the output code for SPLIT FILE groups in procedure.c, which really
2005 shouldn't be doing any output. Move it into the individual
2006 procedures instead. This also adds some flexibility.
2008 Updated many users of procedure_with_splits() and
2009 multipass_procedure_with_splits() to call
2010 output_split_file_values() and to deal with increased use of
2013 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2014 parameter to begin_func member.
2015 (procedure_with_splits) Add "const struct ccase *" parameter to
2016 begin_func parameter. Make ccase parameter const in proc_func
2018 (split_procedure_case_func) Don't dump split file group. Pass
2020 (dump_splits) Moved to language/dictionary/split-file.c as
2021 output_split_file_values().
2022 (struct multipass_split_aux_data) Add "const struct ccase *"
2023 parameter to split_func member.
2024 (multipass_procedure_with_splits) Add "const struct ccase *"
2025 parameter to split_func parameter.
2026 (multipass_split_case_func) Save new SPLIT FILE case before
2028 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2030 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2032 Continue reforming procedure execution. Change
2033 internal_procedure() so that it calls open_active_file() and
2034 close_active_file(), which isolates most of the actual procedure
2037 * procedure.c: (struct write_case_data) Rename `proc_func' member
2038 to `case_func' and update all references.
2039 (procedure) Rewrite as one-line wrapper around
2040 internal_procedure().
2041 (struct multipass_aux_data) New.
2042 (multipass_callback) Renamed multipass_case_func(). Use struct
2043 multipass_aux_data as auxiliary data.
2044 (multipass_end_func) New function.
2045 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2046 that uses multipass_case_func, multipass_end_func.
2047 (internal_procedure) Add `end_func' argument. Move optimization
2048 of trivial case in here. Move call to open_active_file() and
2049 close_active_file() in here. Now assert that vfm_source is
2051 (procedure_with_splits_callback) Rename
2052 split_procedure_case_func().
2053 (split_procedure_end_func) New function.
2054 (multipass_split_callback) Rename multipass_split_case_func.
2055 (multipass_split_end_func) New function.
2056 (discard_variables) No need to test for nonnull vfm_source.
2058 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2060 Continue reforming procedure execution. Get rid of unused member.
2062 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2064 (write_case) Don't increment cases_analyzed.
2066 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2068 Continue reforming procedure execution. In this phase, move
2069 procedure.c and procedure.h from src to src/data. Update
2070 makefiles and #includes accordingly.
2072 * procedure.c: Moved here from src/.
2074 * procedure.h: Moved here from src/.
2076 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2078 Continue reforming procedure execution. In this phase, get rid of
2079 many global variables, consolidating procedure execution in
2080 procedure.c. Encapsulate transformations in new "struct
2081 trns_chain". Also, change implementation of N OF CASES, FILTER,
2082 and PROCESS IF from special cases to transformations.
2084 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2087 * dictionary.c: (global variable default_dict) Move to
2090 * variable.h: (TRNS_*) Move to transformations.h.
2091 (struct transformation) Move to transformations.c.
2093 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2095 * sys-file-reader.c: Fixed invalid read problems.
2097 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2099 * storage-stream.c: Add missing function comments.
2101 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2103 Continue reforming procedure execution. In this phase, add some
2104 new, needed functionality to storage-stream.
2106 * storage-stream.c: (storage_source_decapsulate) New function.
2108 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2110 * variable.c (width_to_bytes): Declarations must precede
2111 statements for C90 compliance.
2113 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2115 * data-out.c, data-in.c, variable.c, variable.h: New functions
2116 copy_mangle and copy_demangle for reading/writing cases; emulates the
2117 way SPSS deals with strings > 255 bytes.
2119 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2120 subtype 14 needed for strings longer than 255 bytes.
2122 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2123 instead of literal values. Also fixed some constness issues.
2125 * format.h: Constness
2127 * sfm-private.h: Renamed the case_size identifier, since I discovered
2128 that SPSS's respect for this variable is very nominal.
2130 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2132 Change case limit type from int to size_t.
2134 * dictionary.c: (struct dictionary) Change type of case_limit
2136 (dict_get_case_limit) Change return type.
2137 (dict_set_case_limit) Change parameter type.
2139 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2141 * variable.h: (struct variable) Rename `reinit' member as `leave'
2142 and invert sense. Fix up all references.
2144 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2146 Continue reforming procedure execution. In this phase, break
2147 procedure.c into multiple files.
2149 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2151 * case-sink.c: New file.
2153 * case-sink.h: New file.
2155 * case-source.c: New file.
2157 * case-source.h: New file.
2159 * storage-stream.c: New file.
2161 * storage-stream.h: New file.
2163 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2165 * variable.h: (struct variable) Remove `init' member and all
2166 references to it from other files. It was initialized in several
2167 places, but nothing really ever used it for anything worthwhile.
2168 Thanks to Jason Stover for pointing out how confusing this
2171 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2173 Continue reforming error message support. In this phase, get rid
2174 of message "titles" and put the message text in `struct error'.
2175 Now `struct error' encapsulates a message more properly.
2177 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2178 Format message ourselves.
2180 * data-in.c: (vdls_error) Ditto.
2182 * por-file-reader.c: (error) Ditto.
2184 * sys-file-reader.c: (corrupt_msg) Ditto.
2186 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2188 GNU standards require "file name" instead of "filename" in
2189 documentation. It's nice for our code to follow the convention
2192 * casefile.c: (struct casefile) Rename `filename' member to
2193 `file_name'. Updated all references.
2195 * file-name.c: [!unix] (struct file_identity) Rename
2196 normalized_filename member to normalized_file_name. Updated all
2199 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2201 We don't really support anything but Unix-like environments well,
2202 so we might as well de-obfuscate by writing directory and path
2203 separators explicitly.
2205 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2207 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2208 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2210 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2213 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2215 GNU standards require "file name" instead of "filename" in
2216 documentation. It's nice for our code to follow the convention
2219 * filename.c: Rename to file-name.c.
2221 * filename.h: Rename to file-name.h. Update all inclusions.
2222 Update header guards.
2224 * automake.mk: Update file names.
2226 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2228 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2230 (fn_basename) Removed (dead code).
2231 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2232 (fn_special_p) Renamed fn_is_special(), all references updated.
2233 (fn_exists_p) Renamed fn_exists(), all references updated.
2235 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2237 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2238 Also, now it only tilde-expands file names, not paths.
2239 (fn_search_path) Tilde-expand one directory at a time.
2241 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2243 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2244 get rid of non-Unixlike version of the code, which has probably
2246 (fn_prepend_dir) Removed (dead code).
2248 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2249 (macro PATH_SEPARATOR_STRING) New.
2250 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2252 Continue reforming error message support. In this phase, we get
2253 rid of VM() and the other msg() support for "verbosity", replacing
2254 it by a new function verbose_msg().
2256 * filename.c: (fn_search_path) Use verbose_msg() instead of
2259 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2261 * sfm-private.h: Get rid of #defines after #error, which makes no
2264 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2266 Get rid of our own int32 type in favor of the standard int32_t
2269 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2272 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2274 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2276 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2278 Remove ill-considered file routines that are no longer used.
2280 * filename.c: (fn_open_ext) Removed.
2281 (fn_close_ext) Removed.
2283 * filename.h: (struct file_ext) Removed.
2285 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2287 * variable.c (var_is_valid_name): Move declarations before code
2290 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2292 * filename.ch (fn_interp_vars): Fixed small buglet.
2294 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2296 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2297 so as to modify the string inline. Thus makeing it easier to
2298 destroy the results when no longer needed.
2300 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2302 * category.c (cat_stored_values_destroy): Fixed memory leak.
2304 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2306 Add some missing frees. Thanks to John Darrington for reporting
2309 * any-writer.c (any_writer_close): Free writer.
2311 * any-reader.c (any_reader_close): Free reader.
2313 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2315 * por-file-reader.c: (error) Mark as NO_RETURN.
2317 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2319 * settings.c: Changed default value of scompress to true.
2321 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2323 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2325 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2328 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2329 here from pref.h.orig.
2331 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2333 * sys-file-reader.c: Fixed bug reading compressed files.
2335 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2337 * Numerous renames. See src/ChangeLog for details.
2339 * Moved files from src directory