1 2007-12-04 Ben Pfaff <blp@gnu.org>
3 * identifier.c (lex_id_match_n): New function.
4 (lex_id_match): Reimplement in terms of lex_id_match_n.
6 2007-11-24 Ben Pfaff <blp@gnu.org>
8 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
11 2007-11-24 Ben Pfaff <blp@gnu.org>
13 Fix warning reported when reading back system files that include
14 very long string variables. Thanks to Guido Gay <gay@irer.it> for
17 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
18 was written where `j' was meant.
20 2007-11-18 Ben Pfaff <blp@gnu.org>
22 Properly write variables that include a range of missing values to
23 system files. Thanks to Guido Gay <gay@irer.it> for reporting
26 * sys-file-writer.c (write_variable): Correctly calculate
27 n_missing_values field when writing variables that include a range
30 2007-11-10 Ben Pfaff <blp@gnu.org>
32 Cleanups and bug fixes devised while writing up documentation.
35 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
37 * dict-class.c: New file.
38 (dict_class_from_id): Move here.
39 (dict_class_to_name): Move here.
41 * dict-class.h: New file.
42 (enum dict_class): Move here. Change from consecutive integers to
43 consecutive bits, to make testing for any of multiple values
44 easier. Add new DC_ALL constant.
46 * dictionary.c (struct dictionary): Change `case_limit' from
48 (dict_get_vars): Make final argument an enum dict_class.
49 (dict_get_vars_mutable): Ditto.
50 (dict_get_case_limit): Change return value to casenumber.
51 (dict_set_case_limit): Change final argument to a casenumber.
52 (dict_unset_split_var): Add assertion. Rephrase slightly.
53 (dict_set_label): Use xstrndup to simplify.
55 * format.c (fmt_step_width): AHEX format also needs 2-byte
57 (fmt_set_style): Simplify assertions.
59 * missing-values.c (mv_add_num_range): Rename mv_add_range.
60 Simplify implementation.
61 (mv_has_value): Simplify implementation.
62 (mv_pop_value): Remove the first value, not the last, to avoid
63 having GET followed by SAVE reverse the order of missing values.
64 (mv_peek_value): Rename mv_get_value. Simplify assertion.
65 (mv_has_range): Simplify implementation.
66 (mv_peek_range): Rename mv_get_range.
67 (can_resize_string): Removed.
68 (mv_is_resizable): Use value_is_resizable.
69 (mv_resize): Use value_resize.
71 * short-names.h (SHORT_NAME_LEN): Move here.
73 * val-type.h: New file, for definitions related to type and width
74 of abstract values. Before, these definitions were mixed among
75 those related to "union value" and those related to variables.
76 (macro SYSMIS): Move here.
77 (macro LOWEST): Move here.
78 (macro HIGHEST): Move here.
79 (macro MAX_STRING); Move here.
80 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
81 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
82 All references updated.
83 (val_type_is_valid): New function. Replaces var_type_is_valid.
84 All references updated.
85 (val_type_from_width): New function. Replaces
86 var_type_from_width. All references updated.
88 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
90 (val_labs_can_set_width): Use value_is_resizable.
91 (val_labs_add): Simply return false if the value labels set is too
92 wide, instead of having undefined behavior.
93 (val_labs_replace): Ditto.
94 (val_labs_replace): Ditto.
95 (val_labs_first): Set iterator to null if iteration is complete.
96 (val_labs_first_sorted): Ditto.
97 (val_labs_done): Become a no-op if the iterator is null.
99 * value.c (value_is_resizable): New function.
100 (value_resize): New function.
102 * variable.c (var_get_dict_class): New function.
104 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
107 2007-11-08 Ben Pfaff <blp@gnu.org>
109 * data-in.c: Make formatted data parsing locale-independent.
110 (parse_number): Use c_strtod instead of strtod, to avoid
111 locale-specific behavior.
114 2007-11-06 Ben Pfaff <blp@gnu.org>
116 Patch #6256: add support for binary, 360 file formats. Reviewed
119 * data-in.c (struct data_in): Add `encoding' member.
120 (data_in): Add `encoding' parameter, and re-encode the data passed
121 in where appropriate. Update all callers to pass it in.
122 (parse_A): Implement EBCDIC recoding wart described in manual.
123 (parse_AHEX): Implement EBCDIC recoding.
125 * data-out.c (data_out_legacy): New function.
126 (data_out): Make into a wrapper around data_out_legacy.
128 * file-handle-def.c (struct file_handle): New member `encoding'.
129 (fh_create_file): Set encoding.
130 (fh_default_properties): Set default encoding.
131 (fh_get_legacy_encoding): New function.
133 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
134 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
135 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
136 (struct fh_properties): New member `encoding'.
138 2007-11-05 Ben Pfaff <blp@gnu.org>
140 Patch #6258. Reviewed by John Darrington.
142 * file-handle-def.c (fh_lock): Add comment that TYPE should be
143 marked with N_() in the caller. Added these markings to each
144 caller too. Should make i18n easier.
145 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
147 2007-11-03 Ben Pfaff <blp@gnu.org>
149 Allow output files to overwrite input files (bug #21280).
151 * file-handle-def.c: Separate locking of files for input and for
152 output, to allow both to take place at once. Also, distinguish a
153 file handle from the identity of the underlying file, because the
154 identity of a file changes over time and the file handle can't
155 represent two different identities.
156 (struct file_handle): Remove `next', `open_cnt', `deleted',
157 `type', `open_mode', `aux', `identity' members. Change `id' from
158 char array to char *. Add `ref_cnt' member.
159 (file_handle_from_ll) New function.
160 (file_handles) Removed.
161 (named_handles) New variable.
162 (fh_init) Initialize named_handles.
163 (fh_done) Remove name from all named_handles.
164 (free_handle) Updated for modified struct file_handle.
165 (unname_handle) New function.
166 (fh_ref) New function.
167 (fh_from_file_name) Removed.
168 (fh_unref) New function.
169 (fh_unname) New function.
170 (fh_from_id) Rewritten.
171 (create_handle) Updated for modified struct file_handle.
172 (fh_inline_file) Increment the handle's ref count.
173 (fh_create_file) Updated for modified struct file_handle.
174 (fh_create_scratch) Ditto.
179 (fh_is_open) Removed.
180 (fh_get_id) Updated for modified struct file_handle.
181 (fh_get_default_handle) Increment the handle's ref count.
182 (fh_set_default_handle) Handle ref counts.
183 (struct fh_lock) New structure.
184 (locks) New static var.
185 (fh_lock) New function.
186 (fh_unlock) New function.
187 (fh_lock_get_aux) New function.
188 (fh_lock_set_aux) New function.
189 (fh_is_locked) New function.
190 (make_key) New function.
191 (free_key) New function.
192 (compare_fh_locks) New function.
193 (hash_fh_lock) New function.
195 * file-handle-def.h (enum fh_access) New enum.
197 * file-name.c: Made file_identity the same in all supported
199 (struct file_identity): New `name' member.
200 (fn_get_identity): For a file that doesn't exist, get the
201 dev/inode of its directory plus its name. If even the directory
202 doesn't exist, just use its name. Merge the Windows
203 implementation into the Unix one.
204 (fn_compare_file_identities): Rewritten. Merge the Windows
205 implementation into the Unix one.
206 (fn_hash_identity): New function.
208 * make-file.c (struct replace_file): New structure.
209 (all_files): New static var.
210 (replace_file_start): New function.
211 (replace_file_commit): New function.
212 (replace_file_abort): New function.
213 (free_replace_file): New function.
214 (unlink_replace_files): New function.
216 * por-file-reader.c (struct pfm_reader): Add `lock' member.
217 (close_reader): Unlock file.
218 (pfm_open_reader): Lock file.
220 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
222 (pfm_open_writer): Lock file and prepare for its replacement.
223 (close_writer): Unlock file.
225 * scratch-handle.h (struct scratch_handle): Add unique_id so that
226 different generations of a scratch file can be distinguished.
228 * scratch-reader.c (scratch_reader_open): Verify that the file is
231 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
232 remove scratch_handle member.
233 (scratch_writer_open): Lock handle. Prepare to replace handle
234 data, instead of doing it immediately.
235 (scratch_writer_casewriter_destroy): Replace handle data and
238 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
239 (sfm_open_reader): Lock file.
240 (close_reader): Unlock file.
242 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
244 (sfm_open_writer): Lock file and prepare for its replacement.
245 (close_writer): Unlock file.
247 2007-11-02 Ben Pfaff <blp@gnu.org>
249 * data-out.c (output_number): Use isfinite (from C99) instead of
250 gsl_isfinite, since we now have universal support for it in
252 (output_infinite): Ditto.
253 [!HAVE_ROUND] (round): Remove definition, since we now have a
254 replacement in gnulib.
256 2007-10-27 John Darrington <john@darrington.wattle.id.au>
258 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
259 wasn't initialised correctly. Closes bugs #21330 and 21397
262 2007-10-08 Ben Pfaff <blp@gnu.org>
264 Fix bug #21280. Thanks to John Darrington for review.
266 * file-name.c (create_stream): New function.
268 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
269 file handle before creating the file, to ensure that we don't
270 truncate a file that we're reading. Make code easier to read by
273 * sys-file-write.c (sfm_open_writer): Ditto.
275 2007-10-01 Ben Pfaff <blp@gnu.org>
277 Fix bug #21192. Thanks to John Darrington for review.
279 * casereader.c (casereader_read): Decrement case_cnt before
280 calling the casereader's "read" member function, so that we
281 interact properly with lazy_casereader.
283 * datasheet.c: Add regression test for above bug fix.
284 (clone_datasheet): New function.
285 (lazy_callback): New function.
286 (check_datasheet_casereader): New function.
287 (check_datasheet): Check datasheet contents are reported correctly
288 through an ordinary casereader and a lazy casereader.
289 (clone_model): Use clone_datasheet.
291 2007-09-24 Ben Pfaff <blp@gnu.org>
293 Patch #6210. Reviewed by John Darrington.
294 * settings.c: Drop "long view", which is not needed any longer.
295 (static var long_view): Removed.
296 (force_long_view): Removed.
297 (get_viewwidth): Removed.
298 (init_viewport): Removed long_view reference.
300 2007-09-19 John Darrington <john@darrington.wattle.id.au>
302 * settings.c settings.h: Changed viewport's length and width to be
303 owned by the user interface which uses the data library. This allows
304 better abstraction, and makes dynamically adjustable dimensions easier.
306 2007-09-18 Ben Pfaff <blp@gnu.org>
308 * procedure.c (proc_extract_active_file_data): New function.
310 * lazy-casereader.h: New file.
312 * lazy-casereader.c: New file.
314 * casereader.c (casereader_dynamic_cast): New function.
316 2007-09-14 Ben Pfaff <blp@gnu.org>
318 * dictionary.c (dict_clone): Copy case indexes from cloned
319 dictionary. Fixes bug #21061. Reviewed by John Darrington.
321 2007-09-13 John Darrington <john@darrington.wattle.id.au>
323 * value.c value.h (create_value): New function.
325 2007-09-12 Ben Pfaff <blp@gnu.org>
327 Make it clear that translator casereader and casewriter translate
328 functions are supposed to destroy their input case.
330 * casereader-translator.c (struct casereader_translator): Change
331 input case parameter of translate member function from const
332 struct ccase * to struct ccase *.
333 (casereader_create_translator): Ditto, for translate parameter.
335 * casewriter-translator.c (struct casewriter_translator): Ditto.
336 (casewriter_create_translator): Ditto.
338 2007-08-27 John Darrington <john@darrington.wattle.id.au>
340 * sys-file-reader.c (read_display_parameters): Force display width
341 to 8, if the sys file says 0 (like SPSS does).
343 2007-08-12 Ben Pfaff <blp@gnu.org>
345 * dictionary.c (dict_dump): New function.
347 2007-08-12 Ben Pfaff <blp@gnu.org>
349 Drop dict_compactor in favor of using the new struct case_map.
351 * dictionary.c (struct copy_map): Removed.
352 (struct dict_compactor): Removed.
353 (dict_make_compactor): Removed.
354 (dict_compactor_compact): Removed.
355 (dict_compactor_destroy): Removed.
357 * procedure.c (struct data_set): Change `compactor' member to be a
359 (proc_open): Use case_map_to_compact_dict instead of
361 (proc_casereader_read): Use case_map_execute instead of
362 dict_compactor_compact.
363 (proc_commit): Use case_map_destroy instead of
364 dict_compactor_destroy.
366 * scratch-writer.c (struct scratch_writer): Change `compactor'
367 member to be a struct case_map *.
368 (scratch_writer_open): Use case_map_to_compact_dict instead of
370 (scratch_writer_casewriter_write): Use case_map_execute instead of
371 dict_compactor_compact.
373 2007-08-12 Ben Pfaff <blp@gnu.org>
375 * automake.mk: Add case-map.c, case-map.h.
377 * case-map.c: New file.
379 * case-map.h: New file.
381 2007-08-12 Ben Pfaff <blp@gnu.org>
383 * dictionary.c (dict_compact_values): Don't delete scratch
384 variables as well as compacting case indexes. Update all callers.
385 (dict_get_compacted_value_cnt): Rename dict_count_values and
386 change interface. Update all callers.
387 (dict_get_compacted_value_cnt): Remove.
388 (dict_compacting_would_shrink): Remove.
389 (dict_compacting_would_change): Remove.
390 (dict_make_compactor): Add new parameter. Update all callers.
392 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
393 instead of dict_count_values, changing an O(N) operation into
396 2007-08-12 Ben Pfaff <blp@gnu.org>
398 * casereader.c (casereader_read): Don't require cases read by a
399 casereader to be exactly the expected size: as long as they're big
402 2007-08-12 Ben Pfaff <blp@gnu.org>
404 Make casewriters keep track of the number of `union value's in
405 each case. This is useful for two reasons: casewriter_write can
406 then check that the case being written is large enough, and later
407 recipients of the casewriter can determine the size of the case.
409 * casewriter-translator.c (casewriter_create_translator): Add
412 * casewriter.c (struct casewriter): Add value_cnt member.
413 (casewriter_write): Check that the case passed in is big enough.
414 (casewriter_get_value_cnt): New function.
415 (casewriter_create): Add value_cnt parameter.
417 2007-08-09 Ben Pfaff <blp@gnu.org>
419 Fix bug reported by Jason Stover.
420 * settings.c: Move get_termcap_viewport and in particular the
421 #include for <curses.h> to the end of the file. curses.h
422 redefines bool on some systems (e.g. OpenBSD), which causes
423 disagreement between uses of bool before its inclusion and
425 Tested by Jason Stover.
427 2007-07-29 Ben Pfaff <blp@gnu.org>
429 Provisional fix for bug #18692 and bug #20161. Reviewed by John
432 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
433 argument (never "rb" or "wb") because SUSv3 says that only those
434 modes are defined, and glibc in fact rejects other modes.
436 Open portable files with fn_open so that they can be read from
437 pipes. Fix missing fh_close call to go along with fh_open.
438 Report an error if the file close reports an error.
439 * por-file-reader.c (close_reader): New function.
440 (por_file_casereader_destroy): Use close_reader.
441 (pfm_open_reader): Open file with fn_open.
443 2007-07-28 Ben Pfaff <blp@gnu.org>
445 Make PSPP able to read all the portable files I could find on the
446 web. Thanks to John Darrington for review. Bug #17620.
447 * por-file-reader.c (struct pfm_reader): New member `line_length'.
448 (error): Print file offset in hexadecimal.
449 (warning): New function.
450 (advance): Treat lines less than 80 bytes long as padded to 80
452 (pfm_open_reader): Call read_documents if we find an "E" record.
453 (convert_format): Convert invalid formats to the default format
454 instead of aborting reading the file.
455 (read_variables): Rename duplicate variable names instead of
456 aborting reading the file.
457 (read_value_label): Allow string variables of different widths to
458 be assigned value labels in the same record. Replace duplicate
459 value labels instead of aborting.
460 (read_documents): New function.
462 * por-file-writer.c (pfm_open_writer): Call write_documents if the
463 dictionary has documents.
464 (write_documents): New function.
466 2007-07-25 Ben Pfaff <blp@gnu.org>
468 Fix bugs related to bug #17213.
470 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
471 since the former is what config.h has. Include the needed ncurses
473 (static var echo) Rename to `do_echo' because the original name is
474 the same as an ncurses identifier.
475 (get_termcap_viewport) Use error instead of msg.
477 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
478 (fn_close): Don't close stdin, stdout, stderr.
480 2007-07-26 John Darrington <john@darrington.wattle.id.au>
482 * procedure.c procedure.h: Added callbacks which get invoked whenever
483 a dataset's transformation chain changes.
485 2007-07-24 Ben Pfaff <blp@gnu.org>
488 * sys-file-writer.c (write_variable_display_parameters): Use new
489 var_default_display_width function to choose display width of
490 segments after the first one in a given variable.
491 * variable.c (var_create): Use var_default_display_width to pick
492 new variable's display width.
493 (var_default_display_width): New function.
494 Reviewed by John Darrington.
496 2007-07-24 Ben Pfaff <blp@gnu.org>
499 * por-file-writer.c (write_variables): Write weight variable.
500 Reviewed by John Darrington.
502 2007-07-23 Ben Pfaff <blp@gnu.org>
504 Improvements to system file reader and writer.
506 First, move all detailed knowledge of very long strings into
507 sys-file-private.[ch], so that this nasty stuff can be isolated.
509 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
510 (EFFECTIVE_VLS_CHUNK): New macro.
511 (min_int): New function.
512 (max_int): New function.
513 (sfm_width_to_bytes): Rewrite.
514 (sfm_width_to_octs): New function.
515 (sfm_segment_alloc_width): New function.
516 (sfm_segment_alloc_bytes): New function.
517 (sfm_segment_used_bytes): New function.
518 (sfm_segment_offset): New function.
519 (sfm_segment_effective_offset): New function.
520 (sfm_dictionary_to_sfm_vars): New function.
522 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
523 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
524 (struct sfm_var): New structure.
526 Next, improvements to the system file reader.
528 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
529 type casenumber. Added `version_major', `version_minor',
532 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
533 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
534 Change `case_cnt' to type casenumber. Removed `has_vls'.
535 (struct sfm_var): Removed.
536 (sfm_open_reader): Don't warn on wrong case size if the file was
537 written by SPSS 13, which tends to get it wrong. Use
538 sfm_dictionary_to_sfm_vars.
539 (read_header): Always output system file info.
540 (read_variable_record): Simplify code for reading missing values.
541 (read_machine_int32_info): Save version numbers from system file
542 into info struct passed as new argument.
543 (read_long_string_map): Restructured to use new sys-file-private
545 (read_value_labels): Use size_overflow_p.
546 (sys_file_casereader_read): Get rid of distinction between fast
547 and slow paths. Use information provided by sys-file-primate's
548 struct sfm_var to simplify code.
549 (skip_whole_strings): New function.
550 (read_int32): Renamed read_int. Changed return value to int.
552 (read_flt64): Renamed read_float. Changed return value to
553 double. Updated all callers.
554 (int32_to_native): Removed. Changed callers to use
556 (flt64_to_double): Removed. Changed callers to use float_convert.
558 Finally, get rid of int32, flt64 terminology and types in system
559 file writer. The former wasn't very useful since a POSIX "int"
560 can hold the whole range of int32 and we generally didn't have a
561 need for it to be exactly-32-bits, just at-least-32-bits. The
562 latter was inconvenient because we had to assume that it could be
563 different from double and thereby convert special values SYSMIS,
564 HIGHEST, LOWEST to and from it in multiple places. Instead, now
565 we just use "int" and "double" in most places, and do conversions,
566 if necessary, very close to where we do I/O. This change meant
567 that the writer code couldn't represent records in the file as C
568 structs any longer, but that's no great loss. The code actually
569 seems to be more readable without them.
571 Simplify the compression buffering code: only buffer as much as
572 necessary, which is no more than eight 8-byte units at any given
575 * sys-file-writer.c (typedef flt64): Removed.
576 (macro second_lowest_flt64): Removed.
577 (struct sysfile_header): Removed.
578 (struct sysfile_variable): Removed.
579 (struct sfm_writer): Removed `needs_translation', `has_vls',
580 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
581 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
582 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
583 compression buffering by `opcodes', `opcode_cnt', `data',
584 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
585 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
586 data writer functions instead of structures.
587 (calc_oct_idx): New function.
588 (write_header): Use simple data writer functions instead of
590 (write_format_spec): Renamed write_format. New argument.
591 (write_variable_continuation_records): New function.
592 (write_variable): Use simple data writer functions instead of
593 structures. Use write_variable_continuation_records. Write
594 entire very long string instead of requiring caller to understand
596 (write_value_labels): Use simple data writer functions instead of
598 (write_documents): Ditto.
599 (write_variable_display_parameters): Use sys-file-private
600 functions to simplify. Use simple data writer functions instead
602 (write_vls_length_table): Use simple data writer functions instead
604 (write_longvar_table): Ditto.
605 (write_rec_7_34): Break into new functions
606 write_integer_info_record, write_float_info_record. Use simple
607 data writer functions instead of structures.
608 (buf_write): Removed.
609 (append_string_max): Removed.
610 (ensure_buf_space): Removed.
611 (sys_file_casewriter_write): Get rid of the distinction between
612 fast and slow paths, which didn't seem to be too useful. Use new
613 functions write_case_uncompressed, write_case_compressed.
614 (put_instruction): Removed.
615 (put_element): Removed.
616 (write_compressed_data): Removed.
617 (close_writer): Use flush_compressed. Only write case count to
618 system file if it will fit in the field.
619 (write_case_compressed): New function.
620 (write_case_uncompressed): New function.
621 (flush_compressed): New function.
622 (put_cmp_opcode): New function.
623 (put_cmp_number): New function.
624 (write_int): New function.
625 (convert_double_to_output_format): New function.
626 (write_float): New function.
627 (write_value): New function.
628 (write_string): New function.
629 (write_bytes): New function.
630 (write_zeros): New function.
631 (write_spaces): New function.
633 Reviewed by John Darrington.
635 2007-07-22 Ben Pfaff <blp@gnu.org>
637 Don't try to write very long strings to portable files. The
638 format does not support it.
640 * por-file-writer.c (MAX_POR_WIDTH): New macro.
641 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
642 (write_format): Add arg to take width to resize format to.
643 (write_value): Limit width of value written to MAX_POR_WIDTH.
644 (write_variables): Limit width of variable and its output formats
646 Reviewed by John Darrington.
648 2007-07-22 Ben Pfaff <blp@gnu.org>
650 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
651 local variable instead of literal 5.
652 Reviewed by John Darrington.
654 2007-07-22 Ben Pfaff <blp@gnu.org>
656 Fix problems with uniqueness of short names in system files with
657 very long string variables. Now a variable may have multiple
660 * automake.mk (src_data_libdata_a_SOURCES): Add new files
661 short-names.c, short-names.h.
663 * dictionary.c (dict_clone): Clone all the short names.
664 (compare_strings): Move into short-names.c.
665 (hash_strings): Ditto.
666 (set_var_short_name_suffix): Ditto.
667 (dict_assign_short_names): Ditto, rename short_names_assign,
668 change to assign all short names.
670 * por-file-writer.c (write_variables): Use short_names_assign
671 instead of dict_assign_short_names.
673 * short-names.c: New file.
675 * short-names.h: New file.
677 * sys-file-private.c (sfm_width_to_segments): New function.
679 * sys-file-reader.c (read_long_var_name_map): Save and restore all
680 the short names, not just the first one.
682 * sys-file-writer.c (cont_var_name): Removed.
683 (sfm_open_writer): Use short_names_assign instead of
684 dict_assign_short_names. Use unique short names assigned by
685 short_names_assign instead of those generated by cont_var_name.
687 * variable.c (struct variable): Remove `short_name' member,
688 replace by `short_names' and `short_name_cnt'.
689 (var_create) Initialize new members.
690 (var_get_short_name_cnt): New function.
691 (var_get_short_name): Now takes an index argument. Changed most
693 (var_set_short_name): Ditto.
694 (var_clear_short_name): Renamed var_clear_short_names, changed to
695 clear all short names.
697 Reviewed by John Darrington.
699 2007-07-22 Ben Pfaff <blp@gnu.org>
701 * variable.c (var_set_width): Use new var_set_width function.
703 * missing-values.c (mv_n_values): Drop assertion, which was not
706 * format.c (fmt_default_for_width): New function.
707 (fmt_resize): New function.
709 Reviewed by John Darrington.
711 2007-07-18 John Darrington <john@darrington.wattle.id.au>
713 * datasheet.c (datasheet_delete_columns): Added assertion to check
714 we're not deleting outside the range of the sheet.
717 * dictionary.c dictionary.h variable.c: Added the ability for string
718 variables to be resized.
720 * vardict.h: Added some prototypes (moved from dictionary.h) as
721 these should only be called by variable.c
724 2007-07-14 John Darrington <john@darrington.wattle.id.au>
726 * sfm-reader.c: Respect case_cnt field in file header.
728 2007-07-01 John Darrington <john@darrington.wattle.id.au>
730 * transformation.c transformation.h (trns_chain_execute): Changed the
731 signature (Patch #6057)
733 2007-06-10 Ben Pfaff <blp@gnu.org>
735 * casereader-filter.c (casereader_filter_destroy): Make sure to
736 write all the remaining excluded cases to the casewriter, if any.
738 * caseinit.c (init_list_destroy): Rewrite.
739 (init_list_clear): Ditto.
741 * casegrouper.c (casegrouper_get_next_group): Always set *reader
742 to null when returning false.
744 2007-06-06 Ben Pfaff <blp@gnu.org>
746 Actually implement the new procedure code and adapt all of its
747 clients to match. Also adapt all of the other case sources and
748 sinks in the tree and their clients to use the
749 casereader/casewriter infrastructure.
751 * automake.mk: Add and remove files.
753 * any-reader.c: Change into a casereader.
754 * por-file-reader.c: Ditto.
755 * scratch-reader.c: Ditto.
756 * sys-file-reader.c: Ditto.
758 * any-writer.c: Change into a casewriter.
759 * por-file-writer.c: Ditto.
760 * scratch-writer.c: Ditto.
761 * sys-file-writer.c: Ditto.
763 * procedure.c: Change to use casereader, casewriter, caseinit, and
764 other new infrastructure.
766 * scratch-handle.c: Adapt to new infrastructure.
768 * case-sink.c: Removed, now dead code.
769 * case-sink.h: Ditto.
770 * case-source.c: Ditto.
771 * case-source.h: Ditto.
772 * casefile-factory.c: Ditto.
773 * casefile-private.h: Ditto.
776 * casefilter.c: Ditto.
777 * casefilter.h: Ditto.
780 * fastfile-factory.c: Ditto.
781 * fastfile-factory.h: Ditto.
782 * storage-stream.c: Ditto.
783 * storage-stream.h: Ditto.
785 2007-06-06 Ben Pfaff <blp@gnu.org>
789 * automake.mk: Add new files.
791 * datasheet.c: New file.
793 * datasheet.h: New file.
795 2007-06-06 Ben Pfaff <blp@gnu.org>
797 Until now, the procedure code has provided a case to the
798 case_source, which has filled in the data values that come from
799 the active file. "Left" data values that don't come from the
800 active file naturally stay the same from case to case, because the
801 procedure code keeps using that same case.
803 One of the compromises that comes with the new procedure code is
804 that the active file allocates and provides its own case, which
805 the procedure code then has to resize to provide room for any
806 other variables that should go in the case and then fill in the
807 values of "left" variables. Then, when we're done with that case,
808 we have to save the values of "left" variables to copy into the
809 next case read from the active file.
811 The caseinit code helps with this.
813 * automake.mk: Add new files.
815 * caseinit.c: New file.
817 * caseinit.h: New file.
819 2007-06-06 Ben Pfaff <blp@gnu.org>
821 * value.h (value_cnt_from_width): New function.
823 * variable.c (var_get_value_cnt): Use new function.
825 2007-06-06 Ben Pfaff <blp@gnu.org>
827 Add casegrouper, to allow cases read from a given casereader to be
828 broken into groups, each of which has its own casereader.
829 Generally cases are grouped based on having equal values for some
832 * automake.mk: Add new files.
834 * casegrouper.c: New file.
836 * casegrouper.h: New file.
838 2007-06-06 Ben Pfaff <blp@gnu.org>
840 Add interface to lexicographical ordering of cases.
842 * automake.mk: Add new files.
844 * case-ordering.c: New file.
846 * case-ordering.h: New file.
848 2007-06-06 Ben Pfaff <blp@gnu.org>
850 Add casereaders and casewriters, the basis of the new data processing
851 implementation. A casereader is a uniform interface to reading cases
852 from a data source; a casewriter is a uniform interface to writing
853 cases to a data sink.
855 * automake.mk: Add new files.
857 * casereader-filter.c: New file.
859 * casereader-provider.h: New file.
861 * casereader-translator.c: New file.
863 * casereader.c: New file.
865 * casereader.h: New file.
867 * casewriter-provider.h: New file.
869 * casewriter-translator.c: New file.
871 * casewriter.c: New file.
873 * casewriter.h: New file.
875 2007-06-06 Ben Pfaff <blp@gnu.org>
877 "casewindow" data structure that extends the deque (from libpspp)
878 of cases with the ability to dump cases to disk if we get too many
881 * automake.mk: Add new files.
883 * casewindow.c: New file.
885 * casewindow.h: New file.
887 2007-06-06 Ben Pfaff <blp@gnu.org>
889 sparse_cases data structure that augments a sparse_array of cases
890 with the ability to dump cases to disk if we get too many cases in
893 * automake.mk: Add new files.
895 * sparse-cases.c: New file.
897 * sparse-cases.h: New file.
899 2007-06-06 Ben Pfaff <blp@gnu.org>
901 Adds a low-level on-disk case array data structure.
903 * automake.mk: Add new files.
905 * case-tmpfile.c: New file.
907 * case-tmpfile.h: New file.
909 2007-06-06 Ben Pfaff <blp@gnu.org>
911 In a couple of places we calculate the maximum number of cases to
912 keep in memory based on the user-defined workspace. Enable
913 centralizing the calculation through a new function.
915 * settings.c (get_workspace_cases): New function.
917 2007-06-06 Ben Pfaff <blp@gnu.org>
919 The casenumber type is defined in transformations.h, but case.h is
920 a more sensible place. Move it.
922 * case.h (CASENUMBER_MAX): New macro.
923 (typedef casenumber): Move here, from transformations.h.
925 2007-06-03 Ben Pfaff <blp@gnu.org>
927 Slightly generalize case_to_values and case_from_values functions.
929 * case.c (case_to_values): Rename case_copy_out, change interface.
930 (case_from_values): Rename case_copy_in, change interface.
932 * fastfile.c (fastfilereader_get_next_case): Update caller.
933 (write_case_to_disk): Ditto.
935 2007-06-02 Ben Pfaff <blp@gnu.org>
937 Clean up after a forgotten part of patch #5829.
939 * casedeque.h: Remove unused file.
941 * automake.mk: Remove casedeque.h from sources.
943 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
945 * category.c: Removed redundant #include
947 2007-05-06 Ben Pfaff <blp@gnu.org>
949 Abstract the documents within a dictionary a little better.
950 Thanks to John Darrington for suggestion, initial version, and
953 * dictionary.c (struct dictionary): Change `documents' member from
954 char * to struct string.
955 (dict_clear): Destroy struct string.
956 (dict_get_documents): Convert struct string to char *.
957 (dict_set_documents): Set struct string. Pad to 80-character
959 (dict_clear_documents): New function.
960 (dict_add_document_line): New function.
961 (dict_get_document_line_cnt): New function.
962 (dict_get_document_line): New function.
964 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
966 * sys-file-reader.c (read_documents): Use new document functions.
968 2007-04-19 John Darrington <john@darrington.wattle.id.au>
970 * sys-file-reader.c: When reading a system file which has no
971 long name table, automatically create one where the long names
972 are the lower case versions of the short names.
974 2007-04-22 Ben Pfaff <blp@gnu.org>
976 * dictionary.c (dict_set_split_vars): dict_destroy expects that
977 dict_clear will free most data related to the dictionary.
978 dict_clear does a decent job, except that dict_set_split_vars on
979 some systems won't actually free the dict's "split" member.
980 Instead, it'll allocate a 1-byte region. Fix this.
982 * value.c (value_copy): New function.
983 (value_set_missing): Ditto.
985 2007-04-22 John Darrington <john@darrington.wattle.id.au>
987 * Deleted existing category.h and moved cat-routines.h into
988 category.h Encapsulated struct cat_vals better.
990 2007-04-19 John Darrington <john@darrington.wattle.id.au>
992 * sys-file-reader.c: When reading a system file which has no
993 long name table, automatically create one where the long names
994 are the lower case versions of the short names.
996 2007-04-16 John Darrington <john@darrington.wattle.id.au>
998 * sys-file-reader.c: Some versions of Other Software seem to
999 produce system files with string variables' measure set to
1000 zero. We'll assume these are supposed to be nominal variables.
1002 2007-03-30 Ben Pfaff <blp@gnu.org>
1004 * procedure.c: Adapt to new deque data structure.
1006 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1007 Ben Pfaff <blp@gnu.org>
1009 * file-name.c: Mingw compatibility fixes.
1010 (fn_search_path): Use ISSLASH instead of comparing against '/'
1012 (fn_dir_name): Use dir_name from gnulib.
1013 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1014 (fn_get_identity): Use GetFullPathName instead of canonicalize
1015 from gnulib, because the latter does not fully support
1016 Windows-style path names. Use this implementation based on the
1017 detected presence of Windows instead of the absence of Unix, since
1018 the new implementation is Windows-specific.
1019 (fn_compare_file_identities): In Windows implementation, compare
1020 names case-insensitively.
1022 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1024 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1025 <stdlib.h> to have the same effect.
1027 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1029 * por-file-reader.c: Add missing _() around messages.
1031 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1033 * make-file.c: Include "mkstemp.h", without which linking on
1036 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1038 Reduce platform dependence.
1040 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1041 Everywhere we using this, we really should have just depended on
1042 the shell to expand tildes.
1043 (fn_search_path): Simplify, given that we don't do tilde expansion
1045 (fn_normalize): Removed. Caller changed to use the canonicalize
1047 (fn_get_cwd): Removed. Only user was fn_normalize.
1048 (fn_is_absolute): Really only test for absolute names.
1049 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1051 (fn_readlink): Removed, as it was only used fn_normalize.
1052 (fn_exists): Assume the stat function is available; gnulib does.
1053 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1055 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1057 * dictionary.c (dict_create_vector_assert): New function.
1059 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1061 * file-name.c (fn_normalize): Correct name of function
1062 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1063 for pointing this out.
1065 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1067 We are using a single member in struct file_handle, the "name"
1068 field, for more than one purpose. When it begins with '"', it's a
1069 file name; otherwise, it's a token that can be used to identify
1070 it. When that assertion fires, it's because we searched for the
1071 name case-sensitively as a file name (so that there was no match),
1072 and then we try to insert it case-insensitively as a token, which
1073 fails because duplicates aren't allowed.
1075 Solution: break the two purposes into two separate fields. This
1076 fixes the problem and likely makes the code easier to read too.
1078 Fixes bug #18922. Thanks to John Darrington for bug report and
1081 * file-handle-def.c (struct file_handle): New `id' member.
1082 (fh_from_name): Rename fh_from_id. Update all callers.
1083 (create_handle): New `id' parameter. Update all callers.
1084 (fh_create_file): Ditto.
1085 (fh_get_id): New function.
1087 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1089 * case.c (case_is_null): Change return type to bool.
1091 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1095 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1096 reference counting to share data. This makes it easy for
1097 valgrind, etc. to find accesses to cases that have been destroyed
1098 but have been kept around by another user's ref-count. This often
1099 happens when the data set is small enough to find in memory; if a
1100 bigger data set that would overflow to disk were used, then data
1101 corruption would occur.
1103 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1107 * case.c (case_unshare): Make it check internally whether the
1108 ref_cnt is greater than 1, so that the callers don't have to.
1109 Update callers not to check.
1111 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1113 Before, I was thinking that I might want to get rid of reference
1114 counting at some point. Now, I'm pretty sure that it's here to
1115 stay. Thus, because we have to store the value_cnt anyway for
1116 reference-counted cases, we might as well expose it to users.
1118 * case.c (case_get_value_cnt): New function.
1119 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1120 resize case if its size actually changed.
1122 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1123 instead of peeking inside struct case directly.
1124 (casefile_append): Ditto.
1126 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1128 Get rid of the inlines for the case functions, which made the
1129 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1130 the inlines didn't speed up "make check" at all, which is not a
1131 perfect benchmark but seems indicative.)
1133 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1134 definitions. Remove some assertions on nonnull pointers that were
1135 redundant with a pointer dereference soon after in the function.
1137 (struct case_data): Move definition here from case.h.
1141 (case_data_wr): Ditto.
1143 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1145 * automake.mk: Add casedeque.h to sources.
1147 * casedeque.h: New file.
1149 * procedure.c: (struct dataset) Change lag_count, lag_head,
1150 lag_queue member into single struct casedeque member. Update all
1151 users to use the casedeque instead.
1154 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1156 * procedure.c: Simplify lagged cases interface. Updated all
1157 clients--well, the only client--to use the simplified interface.
1158 (dataset_n_lag) Removed.
1159 (dataset_set_n_lag) Removed.
1160 (dataset_need_lag) New function.
1162 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1164 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1165 whenever appropriate, but only when the dataset/dictionary is in a
1168 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1170 * dictionary.c dictionary.h : Added callbacks for change of filter and
1171 split variables. Refactored some code to ensure that callbacks get
1172 invoked when appropriate.
1174 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1175 replace_dict callback occurs when permanent_dict replaces the current
1178 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1180 * dictionary.c dictionary.h : Added callback for when the weight
1181 variable of a dictionary changes.
1183 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1185 * dictionary.c dictionary.h : Added replace_source and replace_dict
1186 callbacks, and functions to deal with them.
1188 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1190 Simplify missing value handling.
1192 * missing-values.h (enum mv_class): New type.
1193 (enum mv_type): Moved definition into missing-values.c and renamed
1194 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1195 enums. Updated all uses.
1196 (struct missing_values): Changed type of `type' from `enum
1197 mv_type' to `int' because the definition is no longer exposed.
1199 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1200 parameter. Update all callers.
1201 (mv_is_num_missing): Ditto.
1202 (mv_is_str_missing): Ditto.
1203 (mv_is_value_user_missing): Removed. Changed callers to use
1204 mv_is_value_missing.
1205 (mv_is_num_user_missing): Removed. Changed callers to use
1207 (mv_is_str_user_missing): Removed. Changed callers to use
1209 (mv_is_value_system_missing): Removed. Changed callers to use
1210 mv_is_value_missing.
1211 (mv_set_type): Removed. Changed callers to use mv_clear.
1212 (mv_clear): New function.
1214 * variable.c (var_is_value_missing): Add new enum mv_class
1215 parameter. Update all callers.
1216 (var_is_num_missing): Ditto.
1217 (var_is_str_missing): Ditto.
1218 (var_is_value_user_missing): Removed. Changed callers to use
1219 var_is_value_missing.
1220 (var_is_num_user_missing): Removed. Changed callers to use
1222 (var_is_str_user_missing): Removed. Changed callers to use
1224 (var_is_value_system_missing): Removed. Changed callers to use
1225 var_is_value_missing.
1227 * casefilter.c (struct casefilter): Use enum mv_class in place of
1229 (casefilter_variable_missing): Adapt to new member.
1230 (casefilter_create): Change signature to take enum mv_class,
1233 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1235 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1237 * case-sink.c case-sink.h procedure.c procedure.h
1238 storage-stream.c: Now uses the factory.
1240 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1242 Make it possible to pull cases from the active file with a
1243 function call, instead of requiring indirection through a callback
1246 * case-source.h (struct case_source_class): Change ->read function
1247 to return a single case, instead of calling a callback function
1248 for each case. Change ->destroy function to return an error
1251 * case-source.c (free_case_source): Pass along the value returned
1252 by the case_source ->destroy function.
1254 * procedure.c (struct write_case_data): Removed.
1255 (struct dataset): Added some members to track procedure state.
1256 (procedure): Optimize the trivial case at this level.
1257 (internal_procedure): Re-implement in terms of proc_open,
1258 proc_read, proc_close.
1259 (proc_open) New function.
1260 (proc_read) New function.
1261 (proc_close) New function.
1262 (write_case) Moved into proc_read.
1263 (close_active_file) Moved closing of data source into proc_close.
1265 * storage-source.c: Rewrote to conform with modified
1266 case_source_class interface.
1268 * transformations.c (trns_chain_execute): Added argument to allow
1269 starting execution from an arbitrary transformation. Updated
1272 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1274 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1276 * sys-file-reader.c (read_display_parameters): Don't assume that
1277 MEASURE_* and ALIGN_* have the same values found in system files.
1279 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1281 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1282 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1283 the value of n_MEASURES, which was off by 1 (at least from my
1286 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1288 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1289 callbacks which are invoked when the dictionary or its
1290 variables are changed.
1292 * missing-values.c missing-values.h value-labels.c: Tidied up
1293 consistency checks, and made some of them return false
1294 instead of assert-failing.
1296 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1298 * calendar.c (calendar_days_in_month): New function.
1300 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1302 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1303 the value label's width.
1305 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1307 * sfm-private.h: Move contents into sys-file-writer.c, which is
1308 the only remaining user. Removed Borland C++-specific directives.
1310 * sys-file-reader.c: Clean up and rewrite entire file. The
1311 rewritten version is simpler and better abstracted, and should be
1312 easier to maintain and extend. It avoids using structures to read
1313 file data, which is prone to padding variations among compilers.
1314 It should also handle non-IEEE 754 system files, although I
1315 haven't been able to find any. It has been tested against many
1316 .sav files obtained from the Web and found to produce the same
1317 results as the earlier version of the code, or in some cases
1318 improved results. It is more tolerant of format variations found
1321 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1322 member, putting an enum integer_format in its place. New member
1323 `float_format'. Changed `compressed' member to type bool.
1325 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1327 * dictionary.c (dict_delete_consecutive_vars): New function.
1329 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1331 * file-name.c (fn_search_path): Remove prefix arg that was unused
1332 by any caller. Updated all callers.
1334 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1336 * format.c (fmt_dollar_template): Use user's decimal point
1337 character. Add assertion.
1339 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1341 * format.c (fmt_dollar_template): New function, based on
1342 dollar_format_template from var-type-dialog.c.
1344 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1346 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1349 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1351 Finish converting struct variable to an opaque type. In this
1352 phase, we add remaining setter and getter functions, convert the
1353 remaining PSPP code to use them, and do a bunch of cleanup. The
1354 resulting changes are pervasive but mostly trivial, and only the
1355 notable changes are logged.
1357 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1360 * case.c (case_data): Renamed case_data_idx.
1361 (case_num): Renamed case_num_idx.
1362 (case_str): Renamed case_str_idx.
1363 (case_data_rw): Renamed case_data_rw_idx.
1365 * case.h (case_data): New function with old name and an interface
1366 that takes a variable instead of an index, which is easier to
1367 use. Updated all callers to use the new interface, or to use the
1368 new *_idx function (see above).
1371 (case_data_rw): Ditto.
1373 * category.c (cat_stored_values_destroy): Changed interface to
1374 take a struct cat_vals * instead of a struct variable *.
1376 * dictionary.c (dict_clone): Use new vector_clone function.
1377 (dict_clear) Use new var_destroy function.
1378 (add_var) New function.
1379 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1380 (dict_create_var_assert) Rewrite in terms of add_var.
1381 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1382 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1383 (dict_lookup_var) Use new var_create, var_destroy functions.
1384 (dict_contains_var) Rewrite in terms of new vardict functionality.
1385 (set_var_dict_index) New function.
1386 (set_var_case_index) New function.
1387 (reindex_vars) New function.
1388 (dict_delete_var) Rewrite in terms of new vardict functionality.
1389 (dict_reorder_var) Ditto.
1390 (dict_reorder_vars) Ditto.
1391 (rename_var) New function.
1392 (dict_rename_var) Use rename_var.
1393 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1394 (dict_get_compacted_idx_to_fv) Rename
1395 dict_get_compacted_dict_index_to_case_index, update callers.
1396 (dict_create_vector) Use new vector_create function.
1397 (dict_clear_vectors) Use new vector_destroy function.
1398 (set_var_short_name_suffix) Move here from variable.c, renamed
1399 from var_set_short_name_suffix, make static, update caller.
1401 * sys-file-private.c: New file.
1402 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1403 width_to_bytes, update callers.
1405 * sys-file-private.h: New file. Later it will supplant
1406 sfm-private.h; for now it supplements it.
1407 (macro MIN_VERY_LONG_STRING) New macro.
1408 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1410 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1411 MAX_LONG_STRING was used before.
1413 * sys-file-writer.c: Ditto.
1415 * value-labels.c: Change the paradigm here to be that a null
1416 pointer is OK for a struct val_labs * in most cases; it just
1417 represents an empty set of value labels.
1418 (val_labs_copy) A copy of a null set is a null set.
1419 (val_labs_count) A null set has 0 labels.
1420 (val_labs_replace) Change return type to void. Rewrite for
1422 (val_labs_find) A null set does not contain the value.
1423 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1424 transposed argument order, updated all callers.
1426 * value.c: New file.
1427 (value_dup) Moved here from variable.c.
1428 (compare_values) Ditto.
1431 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1432 (macro MAX_LONG_STRING) Removed, because it was only interesting
1433 for system files, not for general code.
1434 (macro MAX_VERY_LONG_STRING) Ditto.
1435 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1436 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1438 * vardict.h: New file, for an interface between variables and
1441 * variable.c: A lot of functions were moved around, for better
1443 (struct variable) Move definition here, from variable.h.
1444 (var_type_adj) Removed--makes i18n hard.
1445 (var_type_noun) Ditto.
1446 (var_create) New function.
1447 (var_clone) New function.
1448 (var_destroy) New function.
1449 (var_set_name) Assert that variable is not in a dictionary.
1450 (compare_var_names) Rename compare_vars_by_name and fix a couple
1451 of callers who thought the args were strings.
1452 (hash_var_name) Rename hash_var_by_name.
1453 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1454 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1455 (var_is_very_long_string) Removed, because it was only interesting
1456 to system file code.
1457 (var_set_missing_values) Allow the argument to be the wrong width,
1458 as long as we can resize it. Simplify callers who were doing the
1459 resizing themselves.
1460 (var_get_value_labels) New function.
1461 (var_has_value_labels) New function.
1462 (var_set_value_labels) New function.
1463 (alloc_value_labels) New function.
1464 (var_add_value_label) New function.
1465 (var_replace_value_label) New function.
1466 (var_clear_value_labels) New function.
1467 (var_lookup_value_label) New function.
1468 (var_get_value_name) Moved here from variable.c, renamed from
1469 var_get_value_name, transposed argument order, updated all
1471 (var_to_string) Moved here, from variable-label.c.
1472 (var_set_leave) New function.
1473 (var_get_leave) New function.
1474 (var_must_leave) New function.
1475 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1476 set_var_short_name_suffix.
1477 (var_get_dict_index) New function.
1478 (var_get_case_index) New function.
1479 (var_get_obs_vals) New function.
1480 (var_set_obs_vals) New function.
1481 (var_has_obs_vals) New function.
1482 (var_get_vardict) New function.
1483 (var_set_vardict) New function.
1484 (var_has_vardict) New function.
1485 (var_clear_vardict) New function.
1486 (value_dup) Moved to value.c.
1487 (compare_values) Ditto.
1490 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1491 (enum ALPHA) Rename VAR_STRING, update all users.
1493 * vector.c: New file.
1494 (struct vector) Moved here, from variable.h.
1495 (check_widths) New function.
1496 (vector_create) New function.
1497 (vector_clone) New function.
1498 (vector_destroy) New function.
1499 (vector_get_name) New function.
1500 (vector_get_var) New function.
1501 (vector_get_var_cnt) New function.
1502 (compare_vector_ptrs_by_name) New function.
1504 * vector.h: New file.
1506 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1508 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1509 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1512 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1514 * value-labels.c (destroy_atoms): New function.
1515 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1517 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1519 Thanks to Jason Stover for pointing out this problem.
1521 * data-out.c (output_number): Use gsl_finite from GSL, which is
1522 portable, instead of isfinite, which is not.
1525 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1527 * variable.c variable.h (value_dup): New function.
1529 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1531 Start converting struct variable to an opaque type. In this
1532 phase, we add a bunch of setter and getter functions and convert
1533 most of the PSPP code to use them. The resulting changes are
1534 pervasive but mostly trivial, and only the notable changes are
1537 * format.c (fmt_equal): New function.
1539 * variable.c (var_type_is_valid): New function.
1540 (measure_is_valid) Moved here, from format.c.
1541 (alignment_is_valid) Moved here, from format.c.
1542 (var_get_name) New function.
1543 (var_set_name) New function.
1544 (width_to_type) New function.
1545 (var_get_type) New function.
1546 (var_get_width) New function.
1547 (var_set_width) New function.
1548 (var_is_numeric) New function.
1549 (var_is_alpha) New function.
1550 (var_is_short_string) New function.
1551 (var_is_long_string) New function.
1552 (var_is_very_long_string) New function.
1553 (var_get_missing_values) New function.
1554 (var_set_missing_values) New function.
1555 (var_clear_missing_values) New function.
1556 (var_has_missing_values) New function.
1557 (var_is_value_missing) New function.
1558 (var_is_num_missing) New function.
1559 (var_is_str_missing) New function.
1560 (var_is_value_user_missing) New function.
1561 (var_is_num_user_missing) New function.
1562 (var_is_str_user_missing) New function.
1563 (var_is_value_system_missing) New function.
1564 (var_get_print_format) New function.
1565 (var_set_print_format) New function.
1566 (var_get_write_format) New function.
1567 (var_set_write_format) New function.
1568 (var_set_both_formats) New function.
1569 (var_get_label) New function.
1570 (var_set_label) New function.
1571 (var_clear_label) New function.
1572 (var_has_label) New function.
1573 (var_get_measure) New function.
1574 (var_set_measure) New function.
1575 (var_get_display_width) New function.
1576 (var_set_display_width) New function.
1577 (var_get_alignment) New function.
1578 (var_set_alignment) New function.
1579 (var_get_value_cnt) New function.
1580 (var_get_leave) New function.
1581 (var_get_short_name) New function.
1583 * variable.h: (struct variable) Removed "type" and "nv" members;
1584 they are now computed from "width" where needed.
1586 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1588 * missing-values.c (mv_resize): Don't write beyond end of the
1589 allocated buffer when resizing a long string.
1591 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1593 Clean up identifier code: don't require identifier enumerations to
1594 be in a particular order; make better use of string library;
1595 expose less of the internals.
1597 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1598 change interface. Updated all callers.
1599 (lex_id_match) Change interface to use struct substring, update
1601 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1602 (global array keywords[]) Make static, change form. Update all
1603 users to use lex_id_name instead.
1604 (lex_is_keyword) New function.
1605 (lex_id_to_token) Change interface to use struct substring, update
1607 (lex_id_name) New function.
1609 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1610 lex_is_keyword instead.
1611 (T_LAST_KEYWORD) Removed.
1612 (T_N_KEYWORDS) Removed.
1614 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1616 * format.c: (fmt_date_template) Distinguish characters for which a
1617 space is output and any date delimiter is allowed on input, from
1618 those for which a space is output and only a space is allowed on
1619 input. The former is represented by X, the latter by a space.
1620 Also, drop distinction between h and H, changing the former to the
1623 * data-in.c: Completely rewrite internals to conform to SPSS input
1624 formats as closely as possible.
1625 (data_in) Changed external interface by replacing the structure
1626 that was used as a single argument by a set of arguments. Updated
1628 (data_in_finite_line) Removed. Converted all callers to use plain
1630 (data_in_get_integer_format) New function.
1631 (data_in_set_integer_format) New function.
1632 (data_in_get_float_format) New function.
1633 (data_in_set_float_format) New function.
1635 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1636 (struct data_in) Removed.
1638 * data-out.c: (output_date) Drop each component from the input as
1639 it is output, to allow us to drop the distinction between h (a
1640 count of hours) and H (the hour of day) template characters.
1641 Also, handle new X template character.
1642 (output_scientific) Follow more rational rule on when to drop
1643 fraction introduced between SPSS 13 and 15. Updated test case to
1646 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1648 Fix buffer overflow reported by John Darrington.
1650 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1651 realize that DIGITS is a count of nibbles, not of bytes.
1653 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1655 * calendar.c (calendar_offset_to_gregorian) Also return the
1656 year-of-day. Change callers to new interface.
1658 * data-out.c: Completely rewrite internals to conform to SPSS
1659 output formats as completely as possible.
1660 (data_out) Change interface to put input parameters before output
1661 parameters, for consistency with the style I now prefer. Update
1663 (data_out_get_integer_format) New public function.
1664 (data_out_set_integer_format) New public function.
1665 (data_out_get_float_format) New public function.
1666 (data_out_set_float_format) New public function.
1668 * data-out.h: New file. Move prototype for data_out here, from
1671 * format.c: (fmt_step_width) Use equality comparison instead of
1672 bitwise and, for clarity.
1673 (fmt_is_string) Ditto.
1674 (fmt_input_to_output) Fix categories that are translated to F
1677 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1679 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1680 casefile-private.h: Added casefilter to assist commands with missing
1683 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1685 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1687 * settings.c: (route_errors_to_terminal) New variable.
1688 (route_errors_to_listing) New variable.
1689 (get_error_routing_to_terminal) New function.
1690 (set_error_routing_to_terminal) New function.
1691 (get_error_routing_to_listing) New function.
1692 (set_error_routing_to_listing) New function.
1694 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1696 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1698 * format.c: Completely rewrite, to achieve better abstraction.
1699 Rewrite all references to formats in other files.
1701 * format.def: Rewrite and reorganize.
1703 * settings.c: Move everything related to custom currency formats
1704 into format.[ch], changing them in form, so as to group related
1705 code and definitions better. Changed all references to use the
1707 (static var decimal) Removed.
1708 (static var grouping) Removed.
1709 (static var cc) Removed.
1710 (get_decimal) Removed.
1711 (set_decimal) Removed.
1712 (get_grouping) Removed.
1713 (set_grouping) Removed.
1717 * settings.h: (macro CC_CNT) Removed.
1718 (macro CC_WIDTH) Removed.
1719 (struct custom_currency) Removed.
1721 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1723 * data-in.c (data_in): Use switch statement instead of table, to
1724 avoid dependence on the order of the FMT_* enums.
1726 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1728 * data-out.c: (num_to_string) Removed, because it was dead code.
1730 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1732 * data-in.c (parse_trailer): Fix error message.
1734 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1736 * format.c (fmt_is_binary): New function.
1738 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1740 * procedure.c procedure.h: Encapsulated the static data into a single
1743 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1745 * casefile.c (casereader_read_xfer): Always initialize the case,
1746 even on an error condition.
1748 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1750 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1751 entire function useless.
1753 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1755 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1756 casereader method casereader_clone.
1758 * procedure.c pransformations.h: Introduced new type casenum_t
1760 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1762 * variable.c: (width_to_bytes) Rephrase code for clarify.
1764 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1766 * format.c: (fmt_type_from_string) New function.
1767 (fmt_to_string) Include decimals in output if the format has
1768 decimals, even if the format type does not. This way, we can
1769 accurately reproduce incorrect formats in user output.
1770 (check_common_specifier) Make the check for a bad format type an
1771 assertion, so we get bug reports if they show up. Fix message.
1772 Check for decimal places with a format type that doesn't allow
1774 (check_input_specifier) Remove check for FMT_X, which has been
1776 (check_output_specifier) Ditto.
1778 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1780 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1781 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1782 (enum fmt_parse_flags) Removed.
1784 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1786 * casefile.c casefile.h: Converted to an abstract base class.
1787 * casefile-private.h fastfile.c fastfile.h: New files.
1788 * automake.mk procedure.c scratch-writer.c storage-stream.c
1790 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1792 * procedure.c (internal_procedure): Create sink_case with only as
1793 many values as the compacted dictionary.
1795 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1797 Remove "debugging" code that caused plenty of false positives and
1800 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1802 * case.c: Remove all references to `this' member.
1804 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1806 Fix link error noted by Jason Stover.
1808 * storage-stream.c: Include <assert.h>.
1810 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1812 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1813 ALL) and additional underlying system file issues.
1815 Thanks to John Darrington for review.
1817 First problem: var_hash points to variables not owned by the
1818 sys-file-reader, which the caller may free or modify. Use an
1819 array of sfm_vars instead, as done earlier (e.g. CVS version
1822 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1823 members and remove all code that references it. Add vars, var_cnt
1824 members. Remove fix_specials member, which was unused.
1825 (struct sfm_var) Remove name member, which was unused.
1826 (sfm_close_reader) Free vars member instead of var_hash.
1827 (compare_var_shortnames) Removed.
1828 (hash_var_shortname) Removed.
1829 (sfm_open_reader) Fill out vars array.
1830 (compare_var_index) Removed.
1831 (sfm_read_case) Use vars instead of var_hash.
1833 Second problem: we're confused about when we actually have very
1834 long strings, causing us to choose incorrectly between slow path
1835 and fast path in sfm_read_case.
1837 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1838 have very long strings, not when we have long variable names,
1839 which is an unrelated feature.
1841 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1843 * variable.h: Move var_set and variable parsing declarations to
1844 new header, src/language/lexer/variable-parser.h. Modified lots
1845 of files to include the new header.
1847 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1849 * value-labels.c (value_to_string): When there's no value label,
1850 format the variable according to its print format, instead of
1851 always effectively using A or F format.
1853 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1855 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1858 * format.c: Removed tortological assertion.
1860 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1862 Reform string library.
1864 * file-name.c (fn_interp_vars): Change interface to take a
1865 substring as input. Updated all users.
1867 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1869 * format.c (measure_is_valid): Really return false when m >=
1872 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1874 Implement random access to casefiles, for use in GUI.
1876 * casefile.c: (struct casereader) Add `random', `file_ofs',
1877 `buffer_ofs' members.
1878 (casefile_get_random_reader) New function.
1879 (read_open_file) Break part into new function
1880 seek_and_fill_buffer().
1881 (fill_buffer) Update buffer_ofs, file_ofs.
1882 (casereader_seek) New function.
1884 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1886 * settings.c: Added call to i18n{done, init}.
1888 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1890 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1893 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1895 * casefile.c: Convert many uses of `int' to `bool'.
1897 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1899 * transformations.c (trns_chain_destroy): Destroy chain's trns
1900 member, to fix memory leak.
1902 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1904 * storage-stream.c (storage_source_decapsulate): Destroy case
1905 source to fix memory leak.
1907 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1909 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1910 case passed as argument instead of initializing the argument as a
1911 case. Fixes memory leak that showed up in
1912 tests/command/aggregate.sh with scratch files.
1914 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1916 * procedure.c (proc_done): Destroy default_dict, to fix memory
1919 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1921 Simplify procedure_with_splits().
1923 * procedure.c (struct split_aux_data): Removed case_count member.
1924 (procedure_with_splits) Don't initialize case_count.
1925 (split_procedure_case_func) Check whether prev_case is null
1926 instead of case_count.
1927 (split_procedure_end_func) Ditto.
1929 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1931 * case.c (case_move): Do nothing if dst and src are the same
1933 (case_try_create) Merge two similar cases.
1934 (case_copy) Unshare only if data must be actually copied.
1936 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1938 * data-in.c data-out.c dictionary.c sys-file-reader.c
1939 sys-file-writer.c variable.c variable.h: Reworked very long string
1940 support for better encapsulation.
1942 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1944 * value-labels.c (val_labs_can_set_width): New function.
1945 (val_labs_set_width) Clear labels if increasing width to long
1947 (val_labs_destroy) Remove unneeded test for null.
1949 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1951 * value-labels.h: Remove unneeded dependency on variable.h.
1953 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1955 Get rid of `char *c' member in union value, for cleanliness.
1957 * value.h: (union value) Remove `c' member.
1959 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1961 Make dictionary compacting functions a little more general.
1963 * sys-file-writer.c (sfm_open_writer): Use
1964 dict_compacting_would_change().
1965 (does_dict_need_translation) Removed.
1967 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1969 Make dictionary compacting functions a little more general.
1971 * dictionary.c (dict_needs_compaction): Rename
1972 dict_compacting_would_shrink(). Update all callers.
1973 (dict_compacting_would_change) New function.
1975 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1977 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1978 inverted return value (!).
1980 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1982 Continue reforming procedure execution.
1984 * procedure.c: Search and replace "vfm" by "proc". Notably:
1985 (static var vfm_source) Rename proc_source. Update all
1987 (static var vfm_sink) Rename proc_sink. Update all references.
1989 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1991 Continue reforming procedure execution. In this phase, remove
1992 PROCESS IF, which was deprecated anyway and can be easily
1993 simulated with TEMPORARY followed by SELECT IF.
1995 * procedure.c: (open_active_file) Don't call
1996 add_process_if_trns().
1997 (discard_variables) Get rid of redundant call to
1998 proc_cancel_all_transformations().
1999 (add_process_if_trns) Removed.
2000 (process_if_trns_proc) Removed.
2001 (process_if_trns_free) Removed.
2003 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2005 Continue reforming procedure execution. In this phase, add
2006 `const' to the case passed to procedure()'s callback.
2008 Updated all users of procedure() as well.
2010 * procedure.c: (struct write_case_data) Add "const" to ccase
2011 parameter for case_func member.
2012 (procedure) Add "const" to ccase parameter for proc_func
2014 (multipass_case_func) Make ccase parameter const.
2015 (internal_procedure) Add "const" to ccase parameter for case_func
2017 (split_procedure_case_func) Make ccase parameter const.
2018 (multipass_split_case_func) Make ccase parameter const.
2020 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2022 Continue reforming procedure execution. In this phase, get rid of
2023 the output code for SPLIT FILE groups in procedure.c, which really
2024 shouldn't be doing any output. Move it into the individual
2025 procedures instead. This also adds some flexibility.
2027 Updated many users of procedure_with_splits() and
2028 multipass_procedure_with_splits() to call
2029 output_split_file_values() and to deal with increased use of
2032 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2033 parameter to begin_func member.
2034 (procedure_with_splits) Add "const struct ccase *" parameter to
2035 begin_func parameter. Make ccase parameter const in proc_func
2037 (split_procedure_case_func) Don't dump split file group. Pass
2039 (dump_splits) Moved to language/dictionary/split-file.c as
2040 output_split_file_values().
2041 (struct multipass_split_aux_data) Add "const struct ccase *"
2042 parameter to split_func member.
2043 (multipass_procedure_with_splits) Add "const struct ccase *"
2044 parameter to split_func parameter.
2045 (multipass_split_case_func) Save new SPLIT FILE case before
2047 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2049 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2051 Continue reforming procedure execution. Change
2052 internal_procedure() so that it calls open_active_file() and
2053 close_active_file(), which isolates most of the actual procedure
2056 * procedure.c: (struct write_case_data) Rename `proc_func' member
2057 to `case_func' and update all references.
2058 (procedure) Rewrite as one-line wrapper around
2059 internal_procedure().
2060 (struct multipass_aux_data) New.
2061 (multipass_callback) Renamed multipass_case_func(). Use struct
2062 multipass_aux_data as auxiliary data.
2063 (multipass_end_func) New function.
2064 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2065 that uses multipass_case_func, multipass_end_func.
2066 (internal_procedure) Add `end_func' argument. Move optimization
2067 of trivial case in here. Move call to open_active_file() and
2068 close_active_file() in here. Now assert that vfm_source is
2070 (procedure_with_splits_callback) Rename
2071 split_procedure_case_func().
2072 (split_procedure_end_func) New function.
2073 (multipass_split_callback) Rename multipass_split_case_func.
2074 (multipass_split_end_func) New function.
2075 (discard_variables) No need to test for nonnull vfm_source.
2077 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2079 Continue reforming procedure execution. Get rid of unused member.
2081 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2083 (write_case) Don't increment cases_analyzed.
2085 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2087 Continue reforming procedure execution. In this phase, move
2088 procedure.c and procedure.h from src to src/data. Update
2089 makefiles and #includes accordingly.
2091 * procedure.c: Moved here from src/.
2093 * procedure.h: Moved here from src/.
2095 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2097 Continue reforming procedure execution. In this phase, get rid of
2098 many global variables, consolidating procedure execution in
2099 procedure.c. Encapsulate transformations in new "struct
2100 trns_chain". Also, change implementation of N OF CASES, FILTER,
2101 and PROCESS IF from special cases to transformations.
2103 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2106 * dictionary.c: (global variable default_dict) Move to
2109 * variable.h: (TRNS_*) Move to transformations.h.
2110 (struct transformation) Move to transformations.c.
2112 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2114 * sys-file-reader.c: Fixed invalid read problems.
2116 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2118 * storage-stream.c: Add missing function comments.
2120 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2122 Continue reforming procedure execution. In this phase, add some
2123 new, needed functionality to storage-stream.
2125 * storage-stream.c: (storage_source_decapsulate) New function.
2127 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2129 * variable.c (width_to_bytes): Declarations must precede
2130 statements for C90 compliance.
2132 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2134 * data-out.c, data-in.c, variable.c, variable.h: New functions
2135 copy_mangle and copy_demangle for reading/writing cases; emulates the
2136 way SPSS deals with strings > 255 bytes.
2138 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2139 subtype 14 needed for strings longer than 255 bytes.
2141 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2142 instead of literal values. Also fixed some constness issues.
2144 * format.h: Constness
2146 * sfm-private.h: Renamed the case_size identifier, since I discovered
2147 that SPSS's respect for this variable is very nominal.
2149 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2151 Change case limit type from int to size_t.
2153 * dictionary.c: (struct dictionary) Change type of case_limit
2155 (dict_get_case_limit) Change return type.
2156 (dict_set_case_limit) Change parameter type.
2158 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2160 * variable.h: (struct variable) Rename `reinit' member as `leave'
2161 and invert sense. Fix up all references.
2163 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2165 Continue reforming procedure execution. In this phase, break
2166 procedure.c into multiple files.
2168 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2170 * case-sink.c: New file.
2172 * case-sink.h: New file.
2174 * case-source.c: New file.
2176 * case-source.h: New file.
2178 * storage-stream.c: New file.
2180 * storage-stream.h: New file.
2182 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2184 * variable.h: (struct variable) Remove `init' member and all
2185 references to it from other files. It was initialized in several
2186 places, but nothing really ever used it for anything worthwhile.
2187 Thanks to Jason Stover for pointing out how confusing this
2190 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2192 Continue reforming error message support. In this phase, get rid
2193 of message "titles" and put the message text in `struct error'.
2194 Now `struct error' encapsulates a message more properly.
2196 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2197 Format message ourselves.
2199 * data-in.c: (vdls_error) Ditto.
2201 * por-file-reader.c: (error) Ditto.
2203 * sys-file-reader.c: (corrupt_msg) Ditto.
2205 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2207 GNU standards require "file name" instead of "filename" in
2208 documentation. It's nice for our code to follow the convention
2211 * casefile.c: (struct casefile) Rename `filename' member to
2212 `file_name'. Updated all references.
2214 * file-name.c: [!unix] (struct file_identity) Rename
2215 normalized_filename member to normalized_file_name. Updated all
2218 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2220 We don't really support anything but Unix-like environments well,
2221 so we might as well de-obfuscate by writing directory and path
2222 separators explicitly.
2224 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2226 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2227 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2229 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2232 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2234 GNU standards require "file name" instead of "filename" in
2235 documentation. It's nice for our code to follow the convention
2238 * filename.c: Rename to file-name.c.
2240 * filename.h: Rename to file-name.h. Update all inclusions.
2241 Update header guards.
2243 * automake.mk: Update file names.
2245 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2247 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2249 (fn_basename) Removed (dead code).
2250 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2251 (fn_special_p) Renamed fn_is_special(), all references updated.
2252 (fn_exists_p) Renamed fn_exists(), all references updated.
2254 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2256 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2257 Also, now it only tilde-expands file names, not paths.
2258 (fn_search_path) Tilde-expand one directory at a time.
2260 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2262 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2263 get rid of non-Unixlike version of the code, which has probably
2265 (fn_prepend_dir) Removed (dead code).
2267 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2268 (macro PATH_SEPARATOR_STRING) New.
2269 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2271 Continue reforming error message support. In this phase, we get
2272 rid of VM() and the other msg() support for "verbosity", replacing
2273 it by a new function verbose_msg().
2275 * filename.c: (fn_search_path) Use verbose_msg() instead of
2278 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2280 * sfm-private.h: Get rid of #defines after #error, which makes no
2283 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2285 Get rid of our own int32 type in favor of the standard int32_t
2288 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2291 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2293 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2295 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2297 Remove ill-considered file routines that are no longer used.
2299 * filename.c: (fn_open_ext) Removed.
2300 (fn_close_ext) Removed.
2302 * filename.h: (struct file_ext) Removed.
2304 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2306 * variable.c (var_is_valid_name): Move declarations before code
2309 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2311 * filename.ch (fn_interp_vars): Fixed small buglet.
2313 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2315 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2316 so as to modify the string inline. Thus makeing it easier to
2317 destroy the results when no longer needed.
2319 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2321 * category.c (cat_stored_values_destroy): Fixed memory leak.
2323 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2325 Add some missing frees. Thanks to John Darrington for reporting
2328 * any-writer.c (any_writer_close): Free writer.
2330 * any-reader.c (any_reader_close): Free reader.
2332 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2334 * por-file-reader.c: (error) Mark as NO_RETURN.
2336 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2338 * settings.c: Changed default value of scompress to true.
2340 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2342 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2344 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2347 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2348 here from pref.h.orig.
2350 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2352 * sys-file-reader.c: Fixed bug reading compressed files.
2354 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2356 * Numerous renames. See src/ChangeLog for details.
2358 * Moved files from src directory