1 2007-11-24 Ben Pfaff <blp@gnu.org>
3 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
6 2007-11-24 Ben Pfaff <blp@gnu.org>
8 Fix warning reported when reading back system files that include
9 very long string variables. Thanks to Guido Gay <gay@irer.it> for
12 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
13 was written where `j' was meant.
15 2007-11-18 Ben Pfaff <blp@gnu.org>
17 Properly write variables that include a range of missing values to
18 system files. Thanks to Guido Gay <gay@irer.it> for reporting
21 * sys-file-writer.c (write_variable): Correctly calculate
22 n_missing_values field when writing variables that include a range
25 2007-11-10 Ben Pfaff <blp@gnu.org>
27 Cleanups and bug fixes devised while writing up documentation.
30 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
32 * dict-class.c: New file.
33 (dict_class_from_id): Move here.
34 (dict_class_to_name): Move here.
36 * dict-class.h: New file.
37 (enum dict_class): Move here. Change from consecutive integers to
38 consecutive bits, to make testing for any of multiple values
39 easier. Add new DC_ALL constant.
41 * dictionary.c (struct dictionary): Change `case_limit' from
43 (dict_get_vars): Make final argument an enum dict_class.
44 (dict_get_vars_mutable): Ditto.
45 (dict_get_case_limit): Change return value to casenumber.
46 (dict_set_case_limit): Change final argument to a casenumber.
47 (dict_unset_split_var): Add assertion. Rephrase slightly.
48 (dict_set_label): Use xstrndup to simplify.
50 * format.c (fmt_step_width): AHEX format also needs 2-byte
52 (fmt_set_style): Simplify assertions.
54 * missing-values.c (mv_add_num_range): Rename mv_add_range.
55 Simplify implementation.
56 (mv_has_value): Simplify implementation.
57 (mv_pop_value): Remove the first value, not the last, to avoid
58 having GET followed by SAVE reverse the order of missing values.
59 (mv_peek_value): Rename mv_get_value. Simplify assertion.
60 (mv_has_range): Simplify implementation.
61 (mv_peek_range): Rename mv_get_range.
62 (can_resize_string): Removed.
63 (mv_is_resizable): Use value_is_resizable.
64 (mv_resize): Use value_resize.
66 * short-names.h (SHORT_NAME_LEN): Move here.
68 * val-type.h: New file, for definitions related to type and width
69 of abstract values. Before, these definitions were mixed among
70 those related to "union value" and those related to variables.
71 (macro SYSMIS): Move here.
72 (macro LOWEST): Move here.
73 (macro HIGHEST): Move here.
74 (macro MAX_STRING); Move here.
75 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
76 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
77 All references updated.
78 (val_type_is_valid): New function. Replaces var_type_is_valid.
79 All references updated.
80 (val_type_from_width): New function. Replaces
81 var_type_from_width. All references updated.
83 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
85 (val_labs_can_set_width): Use value_is_resizable.
86 (val_labs_add): Simply return false if the value labels set is too
87 wide, instead of having undefined behavior.
88 (val_labs_replace): Ditto.
89 (val_labs_replace): Ditto.
90 (val_labs_first): Set iterator to null if iteration is complete.
91 (val_labs_first_sorted): Ditto.
92 (val_labs_done): Become a no-op if the iterator is null.
94 * value.c (value_is_resizable): New function.
95 (value_resize): New function.
97 * variable.c (var_get_dict_class): New function.
99 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
102 2007-11-08 Ben Pfaff <blp@gnu.org>
104 * data-in.c: Make formatted data parsing locale-independent.
105 (parse_number): Use c_strtod instead of strtod, to avoid
106 locale-specific behavior.
109 2007-11-06 Ben Pfaff <blp@gnu.org>
111 Patch #6256: add support for binary, 360 file formats. Reviewed
114 * data-in.c (struct data_in): Add `encoding' member.
115 (data_in): Add `encoding' parameter, and re-encode the data passed
116 in where appropriate. Update all callers to pass it in.
117 (parse_A): Implement EBCDIC recoding wart described in manual.
118 (parse_AHEX): Implement EBCDIC recoding.
120 * data-out.c (data_out_legacy): New function.
121 (data_out): Make into a wrapper around data_out_legacy.
123 * file-handle-def.c (struct file_handle): New member `encoding'.
124 (fh_create_file): Set encoding.
125 (fh_default_properties): Set default encoding.
126 (fh_get_legacy_encoding): New function.
128 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
129 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
130 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
131 (struct fh_properties): New member `encoding'.
133 2007-11-05 Ben Pfaff <blp@gnu.org>
135 Patch #6258. Reviewed by John Darrington.
137 * file-handle-def.c (fh_lock): Add comment that TYPE should be
138 marked with N_() in the caller. Added these markings to each
139 caller too. Should make i18n easier.
140 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
142 2007-11-03 Ben Pfaff <blp@gnu.org>
144 Allow output files to overwrite input files (bug #21280).
146 * file-handle-def.c: Separate locking of files for input and for
147 output, to allow both to take place at once. Also, distinguish a
148 file handle from the identity of the underlying file, because the
149 identity of a file changes over time and the file handle can't
150 represent two different identities.
151 (struct file_handle): Remove `next', `open_cnt', `deleted',
152 `type', `open_mode', `aux', `identity' members. Change `id' from
153 char array to char *. Add `ref_cnt' member.
154 (file_handle_from_ll) New function.
155 (file_handles) Removed.
156 (named_handles) New variable.
157 (fh_init) Initialize named_handles.
158 (fh_done) Remove name from all named_handles.
159 (free_handle) Updated for modified struct file_handle.
160 (unname_handle) New function.
161 (fh_ref) New function.
162 (fh_from_file_name) Removed.
163 (fh_unref) New function.
164 (fh_unname) New function.
165 (fh_from_id) Rewritten.
166 (create_handle) Updated for modified struct file_handle.
167 (fh_inline_file) Increment the handle's ref count.
168 (fh_create_file) Updated for modified struct file_handle.
169 (fh_create_scratch) Ditto.
174 (fh_is_open) Removed.
175 (fh_get_id) Updated for modified struct file_handle.
176 (fh_get_default_handle) Increment the handle's ref count.
177 (fh_set_default_handle) Handle ref counts.
178 (struct fh_lock) New structure.
179 (locks) New static var.
180 (fh_lock) New function.
181 (fh_unlock) New function.
182 (fh_lock_get_aux) New function.
183 (fh_lock_set_aux) New function.
184 (fh_is_locked) New function.
185 (make_key) New function.
186 (free_key) New function.
187 (compare_fh_locks) New function.
188 (hash_fh_lock) New function.
190 * file-handle-def.h (enum fh_access) New enum.
192 * file-name.c: Made file_identity the same in all supported
194 (struct file_identity): New `name' member.
195 (fn_get_identity): For a file that doesn't exist, get the
196 dev/inode of its directory plus its name. If even the directory
197 doesn't exist, just use its name. Merge the Windows
198 implementation into the Unix one.
199 (fn_compare_file_identities): Rewritten. Merge the Windows
200 implementation into the Unix one.
201 (fn_hash_identity): New function.
203 * make-file.c (struct replace_file): New structure.
204 (all_files): New static var.
205 (replace_file_start): New function.
206 (replace_file_commit): New function.
207 (replace_file_abort): New function.
208 (free_replace_file): New function.
209 (unlink_replace_files): New function.
211 * por-file-reader.c (struct pfm_reader): Add `lock' member.
212 (close_reader): Unlock file.
213 (pfm_open_reader): Lock file.
215 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
217 (pfm_open_writer): Lock file and prepare for its replacement.
218 (close_writer): Unlock file.
220 * scratch-handle.h (struct scratch_handle): Add unique_id so that
221 different generations of a scratch file can be distinguished.
223 * scratch-reader.c (scratch_reader_open): Verify that the file is
226 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
227 remove scratch_handle member.
228 (scratch_writer_open): Lock handle. Prepare to replace handle
229 data, instead of doing it immediately.
230 (scratch_writer_casewriter_destroy): Replace handle data and
233 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
234 (sfm_open_reader): Lock file.
235 (close_reader): Unlock file.
237 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
239 (sfm_open_writer): Lock file and prepare for its replacement.
240 (close_writer): Unlock file.
242 2007-11-02 Ben Pfaff <blp@gnu.org>
244 * data-out.c (output_number): Use isfinite (from C99) instead of
245 gsl_isfinite, since we now have universal support for it in
247 (output_infinite): Ditto.
248 [!HAVE_ROUND] (round): Remove definition, since we now have a
249 replacement in gnulib.
251 2007-10-27 John Darrington <john@darrington.wattle.id.au>
253 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
254 wasn't initialised correctly. Closes bugs #21330 and 21397
257 2007-10-08 Ben Pfaff <blp@gnu.org>
259 Fix bug #21280. Thanks to John Darrington for review.
261 * file-name.c (create_stream): New function.
263 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
264 file handle before creating the file, to ensure that we don't
265 truncate a file that we're reading. Make code easier to read by
268 * sys-file-write.c (sfm_open_writer): Ditto.
270 2007-10-01 Ben Pfaff <blp@gnu.org>
272 Fix bug #21192. Thanks to John Darrington for review.
274 * casereader.c (casereader_read): Decrement case_cnt before
275 calling the casereader's "read" member function, so that we
276 interact properly with lazy_casereader.
278 * datasheet.c: Add regression test for above bug fix.
279 (clone_datasheet): New function.
280 (lazy_callback): New function.
281 (check_datasheet_casereader): New function.
282 (check_datasheet): Check datasheet contents are reported correctly
283 through an ordinary casereader and a lazy casereader.
284 (clone_model): Use clone_datasheet.
286 2007-09-24 Ben Pfaff <blp@gnu.org>
288 Patch #6210. Reviewed by John Darrington.
289 * settings.c: Drop "long view", which is not needed any longer.
290 (static var long_view): Removed.
291 (force_long_view): Removed.
292 (get_viewwidth): Removed.
293 (init_viewport): Removed long_view reference.
295 2007-09-19 John Darrington <john@darrington.wattle.id.au>
297 * settings.c settings.h: Changed viewport's length and width to be
298 owned by the user interface which uses the data library. This allows
299 better abstraction, and makes dynamically adjustable dimensions easier.
301 2007-09-18 Ben Pfaff <blp@gnu.org>
303 * procedure.c (proc_extract_active_file_data): New function.
305 * lazy-casereader.h: New file.
307 * lazy-casereader.c: New file.
309 * casereader.c (casereader_dynamic_cast): New function.
311 2007-09-14 Ben Pfaff <blp@gnu.org>
313 * dictionary.c (dict_clone): Copy case indexes from cloned
314 dictionary. Fixes bug #21061. Reviewed by John Darrington.
316 2007-09-13 John Darrington <john@darrington.wattle.id.au>
318 * value.c value.h (create_value): New function.
320 2007-09-12 Ben Pfaff <blp@gnu.org>
322 Make it clear that translator casereader and casewriter translate
323 functions are supposed to destroy their input case.
325 * casereader-translator.c (struct casereader_translator): Change
326 input case parameter of translate member function from const
327 struct ccase * to struct ccase *.
328 (casereader_create_translator): Ditto, for translate parameter.
330 * casewriter-translator.c (struct casewriter_translator): Ditto.
331 (casewriter_create_translator): Ditto.
333 2007-08-27 John Darrington <john@darrington.wattle.id.au>
335 * sys-file-reader.c (read_display_parameters): Force display width
336 to 8, if the sys file says 0 (like SPSS does).
338 2007-08-12 Ben Pfaff <blp@gnu.org>
340 * dictionary.c (dict_dump): New function.
342 2007-08-12 Ben Pfaff <blp@gnu.org>
344 Drop dict_compactor in favor of using the new struct case_map.
346 * dictionary.c (struct copy_map): Removed.
347 (struct dict_compactor): Removed.
348 (dict_make_compactor): Removed.
349 (dict_compactor_compact): Removed.
350 (dict_compactor_destroy): Removed.
352 * procedure.c (struct data_set): Change `compactor' member to be a
354 (proc_open): Use case_map_to_compact_dict instead of
356 (proc_casereader_read): Use case_map_execute instead of
357 dict_compactor_compact.
358 (proc_commit): Use case_map_destroy instead of
359 dict_compactor_destroy.
361 * scratch-writer.c (struct scratch_writer): Change `compactor'
362 member to be a struct case_map *.
363 (scratch_writer_open): Use case_map_to_compact_dict instead of
365 (scratch_writer_casewriter_write): Use case_map_execute instead of
366 dict_compactor_compact.
368 2007-08-12 Ben Pfaff <blp@gnu.org>
370 * automake.mk: Add case-map.c, case-map.h.
372 * case-map.c: New file.
374 * case-map.h: New file.
376 2007-08-12 Ben Pfaff <blp@gnu.org>
378 * dictionary.c (dict_compact_values): Don't delete scratch
379 variables as well as compacting case indexes. Update all callers.
380 (dict_get_compacted_value_cnt): Rename dict_count_values and
381 change interface. Update all callers.
382 (dict_get_compacted_value_cnt): Remove.
383 (dict_compacting_would_shrink): Remove.
384 (dict_compacting_would_change): Remove.
385 (dict_make_compactor): Add new parameter. Update all callers.
387 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
388 instead of dict_count_values, changing an O(N) operation into
391 2007-08-12 Ben Pfaff <blp@gnu.org>
393 * casereader.c (casereader_read): Don't require cases read by a
394 casereader to be exactly the expected size: as long as they're big
397 2007-08-12 Ben Pfaff <blp@gnu.org>
399 Make casewriters keep track of the number of `union value's in
400 each case. This is useful for two reasons: casewriter_write can
401 then check that the case being written is large enough, and later
402 recipients of the casewriter can determine the size of the case.
404 * casewriter-translator.c (casewriter_create_translator): Add
407 * casewriter.c (struct casewriter): Add value_cnt member.
408 (casewriter_write): Check that the case passed in is big enough.
409 (casewriter_get_value_cnt): New function.
410 (casewriter_create): Add value_cnt parameter.
412 2007-08-09 Ben Pfaff <blp@gnu.org>
414 Fix bug reported by Jason Stover.
415 * settings.c: Move get_termcap_viewport and in particular the
416 #include for <curses.h> to the end of the file. curses.h
417 redefines bool on some systems (e.g. OpenBSD), which causes
418 disagreement between uses of bool before its inclusion and
420 Tested by Jason Stover.
422 2007-07-29 Ben Pfaff <blp@gnu.org>
424 Provisional fix for bug #18692 and bug #20161. Reviewed by John
427 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
428 argument (never "rb" or "wb") because SUSv3 says that only those
429 modes are defined, and glibc in fact rejects other modes.
431 Open portable files with fn_open so that they can be read from
432 pipes. Fix missing fh_close call to go along with fh_open.
433 Report an error if the file close reports an error.
434 * por-file-reader.c (close_reader): New function.
435 (por_file_casereader_destroy): Use close_reader.
436 (pfm_open_reader): Open file with fn_open.
438 2007-07-28 Ben Pfaff <blp@gnu.org>
440 Make PSPP able to read all the portable files I could find on the
441 web. Thanks to John Darrington for review. Bug #17620.
442 * por-file-reader.c (struct pfm_reader): New member `line_length'.
443 (error): Print file offset in hexadecimal.
444 (warning): New function.
445 (advance): Treat lines less than 80 bytes long as padded to 80
447 (pfm_open_reader): Call read_documents if we find an "E" record.
448 (convert_format): Convert invalid formats to the default format
449 instead of aborting reading the file.
450 (read_variables): Rename duplicate variable names instead of
451 aborting reading the file.
452 (read_value_label): Allow string variables of different widths to
453 be assigned value labels in the same record. Replace duplicate
454 value labels instead of aborting.
455 (read_documents): New function.
457 * por-file-writer.c (pfm_open_writer): Call write_documents if the
458 dictionary has documents.
459 (write_documents): New function.
461 2007-07-25 Ben Pfaff <blp@gnu.org>
463 Fix bugs related to bug #17213.
465 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
466 since the former is what config.h has. Include the needed ncurses
468 (static var echo) Rename to `do_echo' because the original name is
469 the same as an ncurses identifier.
470 (get_termcap_viewport) Use error instead of msg.
472 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
473 (fn_close): Don't close stdin, stdout, stderr.
475 2007-07-26 John Darrington <john@darrington.wattle.id.au>
477 * procedure.c procedure.h: Added callbacks which get invoked whenever
478 a dataset's transformation chain changes.
480 2007-07-24 Ben Pfaff <blp@gnu.org>
483 * sys-file-writer.c (write_variable_display_parameters): Use new
484 var_default_display_width function to choose display width of
485 segments after the first one in a given variable.
486 * variable.c (var_create): Use var_default_display_width to pick
487 new variable's display width.
488 (var_default_display_width): New function.
489 Reviewed by John Darrington.
491 2007-07-24 Ben Pfaff <blp@gnu.org>
494 * por-file-writer.c (write_variables): Write weight variable.
495 Reviewed by John Darrington.
497 2007-07-23 Ben Pfaff <blp@gnu.org>
499 Improvements to system file reader and writer.
501 First, move all detailed knowledge of very long strings into
502 sys-file-private.[ch], so that this nasty stuff can be isolated.
504 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
505 (EFFECTIVE_VLS_CHUNK): New macro.
506 (min_int): New function.
507 (max_int): New function.
508 (sfm_width_to_bytes): Rewrite.
509 (sfm_width_to_octs): New function.
510 (sfm_segment_alloc_width): New function.
511 (sfm_segment_alloc_bytes): New function.
512 (sfm_segment_used_bytes): New function.
513 (sfm_segment_offset): New function.
514 (sfm_segment_effective_offset): New function.
515 (sfm_dictionary_to_sfm_vars): New function.
517 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
518 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
519 (struct sfm_var): New structure.
521 Next, improvements to the system file reader.
523 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
524 type casenumber. Added `version_major', `version_minor',
527 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
528 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
529 Change `case_cnt' to type casenumber. Removed `has_vls'.
530 (struct sfm_var): Removed.
531 (sfm_open_reader): Don't warn on wrong case size if the file was
532 written by SPSS 13, which tends to get it wrong. Use
533 sfm_dictionary_to_sfm_vars.
534 (read_header): Always output system file info.
535 (read_variable_record): Simplify code for reading missing values.
536 (read_machine_int32_info): Save version numbers from system file
537 into info struct passed as new argument.
538 (read_long_string_map): Restructured to use new sys-file-private
540 (read_value_labels): Use size_overflow_p.
541 (sys_file_casereader_read): Get rid of distinction between fast
542 and slow paths. Use information provided by sys-file-primate's
543 struct sfm_var to simplify code.
544 (skip_whole_strings): New function.
545 (read_int32): Renamed read_int. Changed return value to int.
547 (read_flt64): Renamed read_float. Changed return value to
548 double. Updated all callers.
549 (int32_to_native): Removed. Changed callers to use
551 (flt64_to_double): Removed. Changed callers to use float_convert.
553 Finally, get rid of int32, flt64 terminology and types in system
554 file writer. The former wasn't very useful since a POSIX "int"
555 can hold the whole range of int32 and we generally didn't have a
556 need for it to be exactly-32-bits, just at-least-32-bits. The
557 latter was inconvenient because we had to assume that it could be
558 different from double and thereby convert special values SYSMIS,
559 HIGHEST, LOWEST to and from it in multiple places. Instead, now
560 we just use "int" and "double" in most places, and do conversions,
561 if necessary, very close to where we do I/O. This change meant
562 that the writer code couldn't represent records in the file as C
563 structs any longer, but that's no great loss. The code actually
564 seems to be more readable without them.
566 Simplify the compression buffering code: only buffer as much as
567 necessary, which is no more than eight 8-byte units at any given
570 * sys-file-writer.c (typedef flt64): Removed.
571 (macro second_lowest_flt64): Removed.
572 (struct sysfile_header): Removed.
573 (struct sysfile_variable): Removed.
574 (struct sfm_writer): Removed `needs_translation', `has_vls',
575 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
576 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
577 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
578 compression buffering by `opcodes', `opcode_cnt', `data',
579 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
580 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
581 data writer functions instead of structures.
582 (calc_oct_idx): New function.
583 (write_header): Use simple data writer functions instead of
585 (write_format_spec): Renamed write_format. New argument.
586 (write_variable_continuation_records): New function.
587 (write_variable): Use simple data writer functions instead of
588 structures. Use write_variable_continuation_records. Write
589 entire very long string instead of requiring caller to understand
591 (write_value_labels): Use simple data writer functions instead of
593 (write_documents): Ditto.
594 (write_variable_display_parameters): Use sys-file-private
595 functions to simplify. Use simple data writer functions instead
597 (write_vls_length_table): Use simple data writer functions instead
599 (write_longvar_table): Ditto.
600 (write_rec_7_34): Break into new functions
601 write_integer_info_record, write_float_info_record. Use simple
602 data writer functions instead of structures.
603 (buf_write): Removed.
604 (append_string_max): Removed.
605 (ensure_buf_space): Removed.
606 (sys_file_casewriter_write): Get rid of the distinction between
607 fast and slow paths, which didn't seem to be too useful. Use new
608 functions write_case_uncompressed, write_case_compressed.
609 (put_instruction): Removed.
610 (put_element): Removed.
611 (write_compressed_data): Removed.
612 (close_writer): Use flush_compressed. Only write case count to
613 system file if it will fit in the field.
614 (write_case_compressed): New function.
615 (write_case_uncompressed): New function.
616 (flush_compressed): New function.
617 (put_cmp_opcode): New function.
618 (put_cmp_number): New function.
619 (write_int): New function.
620 (convert_double_to_output_format): New function.
621 (write_float): New function.
622 (write_value): New function.
623 (write_string): New function.
624 (write_bytes): New function.
625 (write_zeros): New function.
626 (write_spaces): New function.
628 Reviewed by John Darrington.
630 2007-07-22 Ben Pfaff <blp@gnu.org>
632 Don't try to write very long strings to portable files. The
633 format does not support it.
635 * por-file-writer.c (MAX_POR_WIDTH): New macro.
636 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
637 (write_format): Add arg to take width to resize format to.
638 (write_value): Limit width of value written to MAX_POR_WIDTH.
639 (write_variables): Limit width of variable and its output formats
641 Reviewed by John Darrington.
643 2007-07-22 Ben Pfaff <blp@gnu.org>
645 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
646 local variable instead of literal 5.
647 Reviewed by John Darrington.
649 2007-07-22 Ben Pfaff <blp@gnu.org>
651 Fix problems with uniqueness of short names in system files with
652 very long string variables. Now a variable may have multiple
655 * automake.mk (src_data_libdata_a_SOURCES): Add new files
656 short-names.c, short-names.h.
658 * dictionary.c (dict_clone): Clone all the short names.
659 (compare_strings): Move into short-names.c.
660 (hash_strings): Ditto.
661 (set_var_short_name_suffix): Ditto.
662 (dict_assign_short_names): Ditto, rename short_names_assign,
663 change to assign all short names.
665 * por-file-writer.c (write_variables): Use short_names_assign
666 instead of dict_assign_short_names.
668 * short-names.c: New file.
670 * short-names.h: New file.
672 * sys-file-private.c (sfm_width_to_segments): New function.
674 * sys-file-reader.c (read_long_var_name_map): Save and restore all
675 the short names, not just the first one.
677 * sys-file-writer.c (cont_var_name): Removed.
678 (sfm_open_writer): Use short_names_assign instead of
679 dict_assign_short_names. Use unique short names assigned by
680 short_names_assign instead of those generated by cont_var_name.
682 * variable.c (struct variable): Remove `short_name' member,
683 replace by `short_names' and `short_name_cnt'.
684 (var_create) Initialize new members.
685 (var_get_short_name_cnt): New function.
686 (var_get_short_name): Now takes an index argument. Changed most
688 (var_set_short_name): Ditto.
689 (var_clear_short_name): Renamed var_clear_short_names, changed to
690 clear all short names.
692 Reviewed by John Darrington.
694 2007-07-22 Ben Pfaff <blp@gnu.org>
696 * variable.c (var_set_width): Use new var_set_width function.
698 * missing-values.c (mv_n_values): Drop assertion, which was not
701 * format.c (fmt_default_for_width): New function.
702 (fmt_resize): New function.
704 Reviewed by John Darrington.
706 2007-07-18 John Darrington <john@darrington.wattle.id.au>
708 * datasheet.c (datasheet_delete_columns): Added assertion to check
709 we're not deleting outside the range of the sheet.
712 * dictionary.c dictionary.h variable.c: Added the ability for string
713 variables to be resized.
715 * vardict.h: Added some prototypes (moved from dictionary.h) as
716 these should only be called by variable.c
719 2007-07-14 John Darrington <john@darrington.wattle.id.au>
721 * sfm-reader.c: Respect case_cnt field in file header.
723 2007-07-01 John Darrington <john@darrington.wattle.id.au>
725 * transformation.c transformation.h (trns_chain_execute): Changed the
726 signature (Patch #6057)
728 2007-06-10 Ben Pfaff <blp@gnu.org>
730 * casereader-filter.c (casereader_filter_destroy): Make sure to
731 write all the remaining excluded cases to the casewriter, if any.
733 * caseinit.c (init_list_destroy): Rewrite.
734 (init_list_clear): Ditto.
736 * casegrouper.c (casegrouper_get_next_group): Always set *reader
737 to null when returning false.
739 2007-06-06 Ben Pfaff <blp@gnu.org>
741 Actually implement the new procedure code and adapt all of its
742 clients to match. Also adapt all of the other case sources and
743 sinks in the tree and their clients to use the
744 casereader/casewriter infrastructure.
746 * automake.mk: Add and remove files.
748 * any-reader.c: Change into a casereader.
749 * por-file-reader.c: Ditto.
750 * scratch-reader.c: Ditto.
751 * sys-file-reader.c: Ditto.
753 * any-writer.c: Change into a casewriter.
754 * por-file-writer.c: Ditto.
755 * scratch-writer.c: Ditto.
756 * sys-file-writer.c: Ditto.
758 * procedure.c: Change to use casereader, casewriter, caseinit, and
759 other new infrastructure.
761 * scratch-handle.c: Adapt to new infrastructure.
763 * case-sink.c: Removed, now dead code.
764 * case-sink.h: Ditto.
765 * case-source.c: Ditto.
766 * case-source.h: Ditto.
767 * casefile-factory.c: Ditto.
768 * casefile-private.h: Ditto.
771 * casefilter.c: Ditto.
772 * casefilter.h: Ditto.
775 * fastfile-factory.c: Ditto.
776 * fastfile-factory.h: Ditto.
777 * storage-stream.c: Ditto.
778 * storage-stream.h: Ditto.
780 2007-06-06 Ben Pfaff <blp@gnu.org>
784 * automake.mk: Add new files.
786 * datasheet.c: New file.
788 * datasheet.h: New file.
790 2007-06-06 Ben Pfaff <blp@gnu.org>
792 Until now, the procedure code has provided a case to the
793 case_source, which has filled in the data values that come from
794 the active file. "Left" data values that don't come from the
795 active file naturally stay the same from case to case, because the
796 procedure code keeps using that same case.
798 One of the compromises that comes with the new procedure code is
799 that the active file allocates and provides its own case, which
800 the procedure code then has to resize to provide room for any
801 other variables that should go in the case and then fill in the
802 values of "left" variables. Then, when we're done with that case,
803 we have to save the values of "left" variables to copy into the
804 next case read from the active file.
806 The caseinit code helps with this.
808 * automake.mk: Add new files.
810 * caseinit.c: New file.
812 * caseinit.h: New file.
814 2007-06-06 Ben Pfaff <blp@gnu.org>
816 * value.h (value_cnt_from_width): New function.
818 * variable.c (var_get_value_cnt): Use new function.
820 2007-06-06 Ben Pfaff <blp@gnu.org>
822 Add casegrouper, to allow cases read from a given casereader to be
823 broken into groups, each of which has its own casereader.
824 Generally cases are grouped based on having equal values for some
827 * automake.mk: Add new files.
829 * casegrouper.c: New file.
831 * casegrouper.h: New file.
833 2007-06-06 Ben Pfaff <blp@gnu.org>
835 Add interface to lexicographical ordering of cases.
837 * automake.mk: Add new files.
839 * case-ordering.c: New file.
841 * case-ordering.h: New file.
843 2007-06-06 Ben Pfaff <blp@gnu.org>
845 Add casereaders and casewriters, the basis of the new data processing
846 implementation. A casereader is a uniform interface to reading cases
847 from a data source; a casewriter is a uniform interface to writing
848 cases to a data sink.
850 * automake.mk: Add new files.
852 * casereader-filter.c: New file.
854 * casereader-provider.h: New file.
856 * casereader-translator.c: New file.
858 * casereader.c: New file.
860 * casereader.h: New file.
862 * casewriter-provider.h: New file.
864 * casewriter-translator.c: New file.
866 * casewriter.c: New file.
868 * casewriter.h: New file.
870 2007-06-06 Ben Pfaff <blp@gnu.org>
872 "casewindow" data structure that extends the deque (from libpspp)
873 of cases with the ability to dump cases to disk if we get too many
876 * automake.mk: Add new files.
878 * casewindow.c: New file.
880 * casewindow.h: New file.
882 2007-06-06 Ben Pfaff <blp@gnu.org>
884 sparse_cases data structure that augments a sparse_array of cases
885 with the ability to dump cases to disk if we get too many cases in
888 * automake.mk: Add new files.
890 * sparse-cases.c: New file.
892 * sparse-cases.h: New file.
894 2007-06-06 Ben Pfaff <blp@gnu.org>
896 Adds a low-level on-disk case array data structure.
898 * automake.mk: Add new files.
900 * case-tmpfile.c: New file.
902 * case-tmpfile.h: New file.
904 2007-06-06 Ben Pfaff <blp@gnu.org>
906 In a couple of places we calculate the maximum number of cases to
907 keep in memory based on the user-defined workspace. Enable
908 centralizing the calculation through a new function.
910 * settings.c (get_workspace_cases): New function.
912 2007-06-06 Ben Pfaff <blp@gnu.org>
914 The casenumber type is defined in transformations.h, but case.h is
915 a more sensible place. Move it.
917 * case.h (CASENUMBER_MAX): New macro.
918 (typedef casenumber): Move here, from transformations.h.
920 2007-06-03 Ben Pfaff <blp@gnu.org>
922 Slightly generalize case_to_values and case_from_values functions.
924 * case.c (case_to_values): Rename case_copy_out, change interface.
925 (case_from_values): Rename case_copy_in, change interface.
927 * fastfile.c (fastfilereader_get_next_case): Update caller.
928 (write_case_to_disk): Ditto.
930 2007-06-02 Ben Pfaff <blp@gnu.org>
932 Clean up after a forgotten part of patch #5829.
934 * casedeque.h: Remove unused file.
936 * automake.mk: Remove casedeque.h from sources.
938 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
940 * category.c: Removed redundant #include
942 2007-05-06 Ben Pfaff <blp@gnu.org>
944 Abstract the documents within a dictionary a little better.
945 Thanks to John Darrington for suggestion, initial version, and
948 * dictionary.c (struct dictionary): Change `documents' member from
949 char * to struct string.
950 (dict_clear): Destroy struct string.
951 (dict_get_documents): Convert struct string to char *.
952 (dict_set_documents): Set struct string. Pad to 80-character
954 (dict_clear_documents): New function.
955 (dict_add_document_line): New function.
956 (dict_get_document_line_cnt): New function.
957 (dict_get_document_line): New function.
959 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
961 * sys-file-reader.c (read_documents): Use new document functions.
963 2007-04-19 John Darrington <john@darrington.wattle.id.au>
965 * sys-file-reader.c: When reading a system file which has no
966 long name table, automatically create one where the long names
967 are the lower case versions of the short names.
969 2007-04-22 Ben Pfaff <blp@gnu.org>
971 * dictionary.c (dict_set_split_vars): dict_destroy expects that
972 dict_clear will free most data related to the dictionary.
973 dict_clear does a decent job, except that dict_set_split_vars on
974 some systems won't actually free the dict's "split" member.
975 Instead, it'll allocate a 1-byte region. Fix this.
977 * value.c (value_copy): New function.
978 (value_set_missing): Ditto.
980 2007-04-22 John Darrington <john@darrington.wattle.id.au>
982 * Deleted existing category.h and moved cat-routines.h into
983 category.h Encapsulated struct cat_vals better.
985 2007-04-19 John Darrington <john@darrington.wattle.id.au>
987 * sys-file-reader.c: When reading a system file which has no
988 long name table, automatically create one where the long names
989 are the lower case versions of the short names.
991 2007-04-16 John Darrington <john@darrington.wattle.id.au>
993 * sys-file-reader.c: Some versions of Other Software seem to
994 produce system files with string variables' measure set to
995 zero. We'll assume these are supposed to be nominal variables.
997 2007-03-30 Ben Pfaff <blp@gnu.org>
999 * procedure.c: Adapt to new deque data structure.
1001 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1002 Ben Pfaff <blp@gnu.org>
1004 * file-name.c: Mingw compatibility fixes.
1005 (fn_search_path): Use ISSLASH instead of comparing against '/'
1007 (fn_dir_name): Use dir_name from gnulib.
1008 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1009 (fn_get_identity): Use GetFullPathName instead of canonicalize
1010 from gnulib, because the latter does not fully support
1011 Windows-style path names. Use this implementation based on the
1012 detected presence of Windows instead of the absence of Unix, since
1013 the new implementation is Windows-specific.
1014 (fn_compare_file_identities): In Windows implementation, compare
1015 names case-insensitively.
1017 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1019 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1020 <stdlib.h> to have the same effect.
1022 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1024 * por-file-reader.c: Add missing _() around messages.
1026 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1028 * make-file.c: Include "mkstemp.h", without which linking on
1031 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1033 Reduce platform dependence.
1035 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1036 Everywhere we using this, we really should have just depended on
1037 the shell to expand tildes.
1038 (fn_search_path): Simplify, given that we don't do tilde expansion
1040 (fn_normalize): Removed. Caller changed to use the canonicalize
1042 (fn_get_cwd): Removed. Only user was fn_normalize.
1043 (fn_is_absolute): Really only test for absolute names.
1044 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1046 (fn_readlink): Removed, as it was only used fn_normalize.
1047 (fn_exists): Assume the stat function is available; gnulib does.
1048 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1050 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1052 * dictionary.c (dict_create_vector_assert): New function.
1054 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1056 * file-name.c (fn_normalize): Correct name of function
1057 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1058 for pointing this out.
1060 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1062 We are using a single member in struct file_handle, the "name"
1063 field, for more than one purpose. When it begins with '"', it's a
1064 file name; otherwise, it's a token that can be used to identify
1065 it. When that assertion fires, it's because we searched for the
1066 name case-sensitively as a file name (so that there was no match),
1067 and then we try to insert it case-insensitively as a token, which
1068 fails because duplicates aren't allowed.
1070 Solution: break the two purposes into two separate fields. This
1071 fixes the problem and likely makes the code easier to read too.
1073 Fixes bug #18922. Thanks to John Darrington for bug report and
1076 * file-handle-def.c (struct file_handle): New `id' member.
1077 (fh_from_name): Rename fh_from_id. Update all callers.
1078 (create_handle): New `id' parameter. Update all callers.
1079 (fh_create_file): Ditto.
1080 (fh_get_id): New function.
1082 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1084 * case.c (case_is_null): Change return type to bool.
1086 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1090 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1091 reference counting to share data. This makes it easy for
1092 valgrind, etc. to find accesses to cases that have been destroyed
1093 but have been kept around by another user's ref-count. This often
1094 happens when the data set is small enough to find in memory; if a
1095 bigger data set that would overflow to disk were used, then data
1096 corruption would occur.
1098 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1102 * case.c (case_unshare): Make it check internally whether the
1103 ref_cnt is greater than 1, so that the callers don't have to.
1104 Update callers not to check.
1106 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1108 Before, I was thinking that I might want to get rid of reference
1109 counting at some point. Now, I'm pretty sure that it's here to
1110 stay. Thus, because we have to store the value_cnt anyway for
1111 reference-counted cases, we might as well expose it to users.
1113 * case.c (case_get_value_cnt): New function.
1114 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1115 resize case if its size actually changed.
1117 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1118 instead of peeking inside struct case directly.
1119 (casefile_append): Ditto.
1121 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1123 Get rid of the inlines for the case functions, which made the
1124 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1125 the inlines didn't speed up "make check" at all, which is not a
1126 perfect benchmark but seems indicative.)
1128 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1129 definitions. Remove some assertions on nonnull pointers that were
1130 redundant with a pointer dereference soon after in the function.
1132 (struct case_data): Move definition here from case.h.
1136 (case_data_wr): Ditto.
1138 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1140 * automake.mk: Add casedeque.h to sources.
1142 * casedeque.h: New file.
1144 * procedure.c: (struct dataset) Change lag_count, lag_head,
1145 lag_queue member into single struct casedeque member. Update all
1146 users to use the casedeque instead.
1149 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1151 * procedure.c: Simplify lagged cases interface. Updated all
1152 clients--well, the only client--to use the simplified interface.
1153 (dataset_n_lag) Removed.
1154 (dataset_set_n_lag) Removed.
1155 (dataset_need_lag) New function.
1157 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1159 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1160 whenever appropriate, but only when the dataset/dictionary is in a
1163 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1165 * dictionary.c dictionary.h : Added callbacks for change of filter and
1166 split variables. Refactored some code to ensure that callbacks get
1167 invoked when appropriate.
1169 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1170 replace_dict callback occurs when permanent_dict replaces the current
1173 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1175 * dictionary.c dictionary.h : Added callback for when the weight
1176 variable of a dictionary changes.
1178 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1180 * dictionary.c dictionary.h : Added replace_source and replace_dict
1181 callbacks, and functions to deal with them.
1183 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1185 Simplify missing value handling.
1187 * missing-values.h (enum mv_class): New type.
1188 (enum mv_type): Moved definition into missing-values.c and renamed
1189 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1190 enums. Updated all uses.
1191 (struct missing_values): Changed type of `type' from `enum
1192 mv_type' to `int' because the definition is no longer exposed.
1194 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1195 parameter. Update all callers.
1196 (mv_is_num_missing): Ditto.
1197 (mv_is_str_missing): Ditto.
1198 (mv_is_value_user_missing): Removed. Changed callers to use
1199 mv_is_value_missing.
1200 (mv_is_num_user_missing): Removed. Changed callers to use
1202 (mv_is_str_user_missing): Removed. Changed callers to use
1204 (mv_is_value_system_missing): Removed. Changed callers to use
1205 mv_is_value_missing.
1206 (mv_set_type): Removed. Changed callers to use mv_clear.
1207 (mv_clear): New function.
1209 * variable.c (var_is_value_missing): Add new enum mv_class
1210 parameter. Update all callers.
1211 (var_is_num_missing): Ditto.
1212 (var_is_str_missing): Ditto.
1213 (var_is_value_user_missing): Removed. Changed callers to use
1214 var_is_value_missing.
1215 (var_is_num_user_missing): Removed. Changed callers to use
1217 (var_is_str_user_missing): Removed. Changed callers to use
1219 (var_is_value_system_missing): Removed. Changed callers to use
1220 var_is_value_missing.
1222 * casefilter.c (struct casefilter): Use enum mv_class in place of
1224 (casefilter_variable_missing): Adapt to new member.
1225 (casefilter_create): Change signature to take enum mv_class,
1228 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1230 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1232 * case-sink.c case-sink.h procedure.c procedure.h
1233 storage-stream.c: Now uses the factory.
1235 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1237 Make it possible to pull cases from the active file with a
1238 function call, instead of requiring indirection through a callback
1241 * case-source.h (struct case_source_class): Change ->read function
1242 to return a single case, instead of calling a callback function
1243 for each case. Change ->destroy function to return an error
1246 * case-source.c (free_case_source): Pass along the value returned
1247 by the case_source ->destroy function.
1249 * procedure.c (struct write_case_data): Removed.
1250 (struct dataset): Added some members to track procedure state.
1251 (procedure): Optimize the trivial case at this level.
1252 (internal_procedure): Re-implement in terms of proc_open,
1253 proc_read, proc_close.
1254 (proc_open) New function.
1255 (proc_read) New function.
1256 (proc_close) New function.
1257 (write_case) Moved into proc_read.
1258 (close_active_file) Moved closing of data source into proc_close.
1260 * storage-source.c: Rewrote to conform with modified
1261 case_source_class interface.
1263 * transformations.c (trns_chain_execute): Added argument to allow
1264 starting execution from an arbitrary transformation. Updated
1267 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1269 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1271 * sys-file-reader.c (read_display_parameters): Don't assume that
1272 MEASURE_* and ALIGN_* have the same values found in system files.
1274 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1276 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1277 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1278 the value of n_MEASURES, which was off by 1 (at least from my
1281 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1283 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1284 callbacks which are invoked when the dictionary or its
1285 variables are changed.
1287 * missing-values.c missing-values.h value-labels.c: Tidied up
1288 consistency checks, and made some of them return false
1289 instead of assert-failing.
1291 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1293 * calendar.c (calendar_days_in_month): New function.
1295 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1297 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1298 the value label's width.
1300 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1302 * sfm-private.h: Move contents into sys-file-writer.c, which is
1303 the only remaining user. Removed Borland C++-specific directives.
1305 * sys-file-reader.c: Clean up and rewrite entire file. The
1306 rewritten version is simpler and better abstracted, and should be
1307 easier to maintain and extend. It avoids using structures to read
1308 file data, which is prone to padding variations among compilers.
1309 It should also handle non-IEEE 754 system files, although I
1310 haven't been able to find any. It has been tested against many
1311 .sav files obtained from the Web and found to produce the same
1312 results as the earlier version of the code, or in some cases
1313 improved results. It is more tolerant of format variations found
1316 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1317 member, putting an enum integer_format in its place. New member
1318 `float_format'. Changed `compressed' member to type bool.
1320 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1322 * dictionary.c (dict_delete_consecutive_vars): New function.
1324 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1326 * file-name.c (fn_search_path): Remove prefix arg that was unused
1327 by any caller. Updated all callers.
1329 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1331 * format.c (fmt_dollar_template): Use user's decimal point
1332 character. Add assertion.
1334 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1336 * format.c (fmt_dollar_template): New function, based on
1337 dollar_format_template from var-type-dialog.c.
1339 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1341 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1344 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1346 Finish converting struct variable to an opaque type. In this
1347 phase, we add remaining setter and getter functions, convert the
1348 remaining PSPP code to use them, and do a bunch of cleanup. The
1349 resulting changes are pervasive but mostly trivial, and only the
1350 notable changes are logged.
1352 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1355 * case.c (case_data): Renamed case_data_idx.
1356 (case_num): Renamed case_num_idx.
1357 (case_str): Renamed case_str_idx.
1358 (case_data_rw): Renamed case_data_rw_idx.
1360 * case.h (case_data): New function with old name and an interface
1361 that takes a variable instead of an index, which is easier to
1362 use. Updated all callers to use the new interface, or to use the
1363 new *_idx function (see above).
1366 (case_data_rw): Ditto.
1368 * category.c (cat_stored_values_destroy): Changed interface to
1369 take a struct cat_vals * instead of a struct variable *.
1371 * dictionary.c (dict_clone): Use new vector_clone function.
1372 (dict_clear) Use new var_destroy function.
1373 (add_var) New function.
1374 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1375 (dict_create_var_assert) Rewrite in terms of add_var.
1376 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1377 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1378 (dict_lookup_var) Use new var_create, var_destroy functions.
1379 (dict_contains_var) Rewrite in terms of new vardict functionality.
1380 (set_var_dict_index) New function.
1381 (set_var_case_index) New function.
1382 (reindex_vars) New function.
1383 (dict_delete_var) Rewrite in terms of new vardict functionality.
1384 (dict_reorder_var) Ditto.
1385 (dict_reorder_vars) Ditto.
1386 (rename_var) New function.
1387 (dict_rename_var) Use rename_var.
1388 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1389 (dict_get_compacted_idx_to_fv) Rename
1390 dict_get_compacted_dict_index_to_case_index, update callers.
1391 (dict_create_vector) Use new vector_create function.
1392 (dict_clear_vectors) Use new vector_destroy function.
1393 (set_var_short_name_suffix) Move here from variable.c, renamed
1394 from var_set_short_name_suffix, make static, update caller.
1396 * sys-file-private.c: New file.
1397 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1398 width_to_bytes, update callers.
1400 * sys-file-private.h: New file. Later it will supplant
1401 sfm-private.h; for now it supplements it.
1402 (macro MIN_VERY_LONG_STRING) New macro.
1403 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1405 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1406 MAX_LONG_STRING was used before.
1408 * sys-file-writer.c: Ditto.
1410 * value-labels.c: Change the paradigm here to be that a null
1411 pointer is OK for a struct val_labs * in most cases; it just
1412 represents an empty set of value labels.
1413 (val_labs_copy) A copy of a null set is a null set.
1414 (val_labs_count) A null set has 0 labels.
1415 (val_labs_replace) Change return type to void. Rewrite for
1417 (val_labs_find) A null set does not contain the value.
1418 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1419 transposed argument order, updated all callers.
1421 * value.c: New file.
1422 (value_dup) Moved here from variable.c.
1423 (compare_values) Ditto.
1426 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1427 (macro MAX_LONG_STRING) Removed, because it was only interesting
1428 for system files, not for general code.
1429 (macro MAX_VERY_LONG_STRING) Ditto.
1430 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1431 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1433 * vardict.h: New file, for an interface between variables and
1436 * variable.c: A lot of functions were moved around, for better
1438 (struct variable) Move definition here, from variable.h.
1439 (var_type_adj) Removed--makes i18n hard.
1440 (var_type_noun) Ditto.
1441 (var_create) New function.
1442 (var_clone) New function.
1443 (var_destroy) New function.
1444 (var_set_name) Assert that variable is not in a dictionary.
1445 (compare_var_names) Rename compare_vars_by_name and fix a couple
1446 of callers who thought the args were strings.
1447 (hash_var_name) Rename hash_var_by_name.
1448 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1449 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1450 (var_is_very_long_string) Removed, because it was only interesting
1451 to system file code.
1452 (var_set_missing_values) Allow the argument to be the wrong width,
1453 as long as we can resize it. Simplify callers who were doing the
1454 resizing themselves.
1455 (var_get_value_labels) New function.
1456 (var_has_value_labels) New function.
1457 (var_set_value_labels) New function.
1458 (alloc_value_labels) New function.
1459 (var_add_value_label) New function.
1460 (var_replace_value_label) New function.
1461 (var_clear_value_labels) New function.
1462 (var_lookup_value_label) New function.
1463 (var_get_value_name) Moved here from variable.c, renamed from
1464 var_get_value_name, transposed argument order, updated all
1466 (var_to_string) Moved here, from variable-label.c.
1467 (var_set_leave) New function.
1468 (var_get_leave) New function.
1469 (var_must_leave) New function.
1470 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1471 set_var_short_name_suffix.
1472 (var_get_dict_index) New function.
1473 (var_get_case_index) New function.
1474 (var_get_obs_vals) New function.
1475 (var_set_obs_vals) New function.
1476 (var_has_obs_vals) New function.
1477 (var_get_vardict) New function.
1478 (var_set_vardict) New function.
1479 (var_has_vardict) New function.
1480 (var_clear_vardict) New function.
1481 (value_dup) Moved to value.c.
1482 (compare_values) Ditto.
1485 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1486 (enum ALPHA) Rename VAR_STRING, update all users.
1488 * vector.c: New file.
1489 (struct vector) Moved here, from variable.h.
1490 (check_widths) New function.
1491 (vector_create) New function.
1492 (vector_clone) New function.
1493 (vector_destroy) New function.
1494 (vector_get_name) New function.
1495 (vector_get_var) New function.
1496 (vector_get_var_cnt) New function.
1497 (compare_vector_ptrs_by_name) New function.
1499 * vector.h: New file.
1501 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1503 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1504 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1507 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1509 * value-labels.c (destroy_atoms): New function.
1510 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1512 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1514 Thanks to Jason Stover for pointing out this problem.
1516 * data-out.c (output_number): Use gsl_finite from GSL, which is
1517 portable, instead of isfinite, which is not.
1520 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1522 * variable.c variable.h (value_dup): New function.
1524 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1526 Start converting struct variable to an opaque type. In this
1527 phase, we add a bunch of setter and getter functions and convert
1528 most of the PSPP code to use them. The resulting changes are
1529 pervasive but mostly trivial, and only the notable changes are
1532 * format.c (fmt_equal): New function.
1534 * variable.c (var_type_is_valid): New function.
1535 (measure_is_valid) Moved here, from format.c.
1536 (alignment_is_valid) Moved here, from format.c.
1537 (var_get_name) New function.
1538 (var_set_name) New function.
1539 (width_to_type) New function.
1540 (var_get_type) New function.
1541 (var_get_width) New function.
1542 (var_set_width) New function.
1543 (var_is_numeric) New function.
1544 (var_is_alpha) New function.
1545 (var_is_short_string) New function.
1546 (var_is_long_string) New function.
1547 (var_is_very_long_string) New function.
1548 (var_get_missing_values) New function.
1549 (var_set_missing_values) New function.
1550 (var_clear_missing_values) New function.
1551 (var_has_missing_values) New function.
1552 (var_is_value_missing) New function.
1553 (var_is_num_missing) New function.
1554 (var_is_str_missing) New function.
1555 (var_is_value_user_missing) New function.
1556 (var_is_num_user_missing) New function.
1557 (var_is_str_user_missing) New function.
1558 (var_is_value_system_missing) New function.
1559 (var_get_print_format) New function.
1560 (var_set_print_format) New function.
1561 (var_get_write_format) New function.
1562 (var_set_write_format) New function.
1563 (var_set_both_formats) New function.
1564 (var_get_label) New function.
1565 (var_set_label) New function.
1566 (var_clear_label) New function.
1567 (var_has_label) New function.
1568 (var_get_measure) New function.
1569 (var_set_measure) New function.
1570 (var_get_display_width) New function.
1571 (var_set_display_width) New function.
1572 (var_get_alignment) New function.
1573 (var_set_alignment) New function.
1574 (var_get_value_cnt) New function.
1575 (var_get_leave) New function.
1576 (var_get_short_name) New function.
1578 * variable.h: (struct variable) Removed "type" and "nv" members;
1579 they are now computed from "width" where needed.
1581 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1583 * missing-values.c (mv_resize): Don't write beyond end of the
1584 allocated buffer when resizing a long string.
1586 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1588 Clean up identifier code: don't require identifier enumerations to
1589 be in a particular order; make better use of string library;
1590 expose less of the internals.
1592 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1593 change interface. Updated all callers.
1594 (lex_id_match) Change interface to use struct substring, update
1596 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1597 (global array keywords[]) Make static, change form. Update all
1598 users to use lex_id_name instead.
1599 (lex_is_keyword) New function.
1600 (lex_id_to_token) Change interface to use struct substring, update
1602 (lex_id_name) New function.
1604 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1605 lex_is_keyword instead.
1606 (T_LAST_KEYWORD) Removed.
1607 (T_N_KEYWORDS) Removed.
1609 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1611 * format.c: (fmt_date_template) Distinguish characters for which a
1612 space is output and any date delimiter is allowed on input, from
1613 those for which a space is output and only a space is allowed on
1614 input. The former is represented by X, the latter by a space.
1615 Also, drop distinction between h and H, changing the former to the
1618 * data-in.c: Completely rewrite internals to conform to SPSS input
1619 formats as closely as possible.
1620 (data_in) Changed external interface by replacing the structure
1621 that was used as a single argument by a set of arguments. Updated
1623 (data_in_finite_line) Removed. Converted all callers to use plain
1625 (data_in_get_integer_format) New function.
1626 (data_in_set_integer_format) New function.
1627 (data_in_get_float_format) New function.
1628 (data_in_set_float_format) New function.
1630 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1631 (struct data_in) Removed.
1633 * data-out.c: (output_date) Drop each component from the input as
1634 it is output, to allow us to drop the distinction between h (a
1635 count of hours) and H (the hour of day) template characters.
1636 Also, handle new X template character.
1637 (output_scientific) Follow more rational rule on when to drop
1638 fraction introduced between SPSS 13 and 15. Updated test case to
1641 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1643 Fix buffer overflow reported by John Darrington.
1645 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1646 realize that DIGITS is a count of nibbles, not of bytes.
1648 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1650 * calendar.c (calendar_offset_to_gregorian) Also return the
1651 year-of-day. Change callers to new interface.
1653 * data-out.c: Completely rewrite internals to conform to SPSS
1654 output formats as completely as possible.
1655 (data_out) Change interface to put input parameters before output
1656 parameters, for consistency with the style I now prefer. Update
1658 (data_out_get_integer_format) New public function.
1659 (data_out_set_integer_format) New public function.
1660 (data_out_get_float_format) New public function.
1661 (data_out_set_float_format) New public function.
1663 * data-out.h: New file. Move prototype for data_out here, from
1666 * format.c: (fmt_step_width) Use equality comparison instead of
1667 bitwise and, for clarity.
1668 (fmt_is_string) Ditto.
1669 (fmt_input_to_output) Fix categories that are translated to F
1672 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1674 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1675 casefile-private.h: Added casefilter to assist commands with missing
1678 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1680 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1682 * settings.c: (route_errors_to_terminal) New variable.
1683 (route_errors_to_listing) New variable.
1684 (get_error_routing_to_terminal) New function.
1685 (set_error_routing_to_terminal) New function.
1686 (get_error_routing_to_listing) New function.
1687 (set_error_routing_to_listing) New function.
1689 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1691 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1693 * format.c: Completely rewrite, to achieve better abstraction.
1694 Rewrite all references to formats in other files.
1696 * format.def: Rewrite and reorganize.
1698 * settings.c: Move everything related to custom currency formats
1699 into format.[ch], changing them in form, so as to group related
1700 code and definitions better. Changed all references to use the
1702 (static var decimal) Removed.
1703 (static var grouping) Removed.
1704 (static var cc) Removed.
1705 (get_decimal) Removed.
1706 (set_decimal) Removed.
1707 (get_grouping) Removed.
1708 (set_grouping) Removed.
1712 * settings.h: (macro CC_CNT) Removed.
1713 (macro CC_WIDTH) Removed.
1714 (struct custom_currency) Removed.
1716 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1718 * data-in.c (data_in): Use switch statement instead of table, to
1719 avoid dependence on the order of the FMT_* enums.
1721 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1723 * data-out.c: (num_to_string) Removed, because it was dead code.
1725 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1727 * data-in.c (parse_trailer): Fix error message.
1729 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1731 * format.c (fmt_is_binary): New function.
1733 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1735 * procedure.c procedure.h: Encapsulated the static data into a single
1738 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1740 * casefile.c (casereader_read_xfer): Always initialize the case,
1741 even on an error condition.
1743 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1745 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1746 entire function useless.
1748 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1750 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1751 casereader method casereader_clone.
1753 * procedure.c pransformations.h: Introduced new type casenum_t
1755 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1757 * variable.c: (width_to_bytes) Rephrase code for clarify.
1759 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1761 * format.c: (fmt_type_from_string) New function.
1762 (fmt_to_string) Include decimals in output if the format has
1763 decimals, even if the format type does not. This way, we can
1764 accurately reproduce incorrect formats in user output.
1765 (check_common_specifier) Make the check for a bad format type an
1766 assertion, so we get bug reports if they show up. Fix message.
1767 Check for decimal places with a format type that doesn't allow
1769 (check_input_specifier) Remove check for FMT_X, which has been
1771 (check_output_specifier) Ditto.
1773 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1775 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1776 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1777 (enum fmt_parse_flags) Removed.
1779 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1781 * casefile.c casefile.h: Converted to an abstract base class.
1782 * casefile-private.h fastfile.c fastfile.h: New files.
1783 * automake.mk procedure.c scratch-writer.c storage-stream.c
1785 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1787 * procedure.c (internal_procedure): Create sink_case with only as
1788 many values as the compacted dictionary.
1790 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1792 Remove "debugging" code that caused plenty of false positives and
1795 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1797 * case.c: Remove all references to `this' member.
1799 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1801 Fix link error noted by Jason Stover.
1803 * storage-stream.c: Include <assert.h>.
1805 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1807 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1808 ALL) and additional underlying system file issues.
1810 Thanks to John Darrington for review.
1812 First problem: var_hash points to variables not owned by the
1813 sys-file-reader, which the caller may free or modify. Use an
1814 array of sfm_vars instead, as done earlier (e.g. CVS version
1817 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1818 members and remove all code that references it. Add vars, var_cnt
1819 members. Remove fix_specials member, which was unused.
1820 (struct sfm_var) Remove name member, which was unused.
1821 (sfm_close_reader) Free vars member instead of var_hash.
1822 (compare_var_shortnames) Removed.
1823 (hash_var_shortname) Removed.
1824 (sfm_open_reader) Fill out vars array.
1825 (compare_var_index) Removed.
1826 (sfm_read_case) Use vars instead of var_hash.
1828 Second problem: we're confused about when we actually have very
1829 long strings, causing us to choose incorrectly between slow path
1830 and fast path in sfm_read_case.
1832 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1833 have very long strings, not when we have long variable names,
1834 which is an unrelated feature.
1836 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1838 * variable.h: Move var_set and variable parsing declarations to
1839 new header, src/language/lexer/variable-parser.h. Modified lots
1840 of files to include the new header.
1842 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1844 * value-labels.c (value_to_string): When there's no value label,
1845 format the variable according to its print format, instead of
1846 always effectively using A or F format.
1848 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1850 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1853 * format.c: Removed tortological assertion.
1855 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1857 Reform string library.
1859 * file-name.c (fn_interp_vars): Change interface to take a
1860 substring as input. Updated all users.
1862 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1864 * format.c (measure_is_valid): Really return false when m >=
1867 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1869 Implement random access to casefiles, for use in GUI.
1871 * casefile.c: (struct casereader) Add `random', `file_ofs',
1872 `buffer_ofs' members.
1873 (casefile_get_random_reader) New function.
1874 (read_open_file) Break part into new function
1875 seek_and_fill_buffer().
1876 (fill_buffer) Update buffer_ofs, file_ofs.
1877 (casereader_seek) New function.
1879 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1881 * settings.c: Added call to i18n{done, init}.
1883 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1885 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1888 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1890 * casefile.c: Convert many uses of `int' to `bool'.
1892 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1894 * transformations.c (trns_chain_destroy): Destroy chain's trns
1895 member, to fix memory leak.
1897 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1899 * storage-stream.c (storage_source_decapsulate): Destroy case
1900 source to fix memory leak.
1902 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1904 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1905 case passed as argument instead of initializing the argument as a
1906 case. Fixes memory leak that showed up in
1907 tests/command/aggregate.sh with scratch files.
1909 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1911 * procedure.c (proc_done): Destroy default_dict, to fix memory
1914 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1916 Simplify procedure_with_splits().
1918 * procedure.c (struct split_aux_data): Removed case_count member.
1919 (procedure_with_splits) Don't initialize case_count.
1920 (split_procedure_case_func) Check whether prev_case is null
1921 instead of case_count.
1922 (split_procedure_end_func) Ditto.
1924 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1926 * case.c (case_move): Do nothing if dst and src are the same
1928 (case_try_create) Merge two similar cases.
1929 (case_copy) Unshare only if data must be actually copied.
1931 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1933 * data-in.c data-out.c dictionary.c sys-file-reader.c
1934 sys-file-writer.c variable.c variable.h: Reworked very long string
1935 support for better encapsulation.
1937 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1939 * value-labels.c (val_labs_can_set_width): New function.
1940 (val_labs_set_width) Clear labels if increasing width to long
1942 (val_labs_destroy) Remove unneeded test for null.
1944 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1946 * value-labels.h: Remove unneeded dependency on variable.h.
1948 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1950 Get rid of `char *c' member in union value, for cleanliness.
1952 * value.h: (union value) Remove `c' member.
1954 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1956 Make dictionary compacting functions a little more general.
1958 * sys-file-writer.c (sfm_open_writer): Use
1959 dict_compacting_would_change().
1960 (does_dict_need_translation) Removed.
1962 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1964 Make dictionary compacting functions a little more general.
1966 * dictionary.c (dict_needs_compaction): Rename
1967 dict_compacting_would_shrink(). Update all callers.
1968 (dict_compacting_would_change) New function.
1970 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1972 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1973 inverted return value (!).
1975 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1977 Continue reforming procedure execution.
1979 * procedure.c: Search and replace "vfm" by "proc". Notably:
1980 (static var vfm_source) Rename proc_source. Update all
1982 (static var vfm_sink) Rename proc_sink. Update all references.
1984 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1986 Continue reforming procedure execution. In this phase, remove
1987 PROCESS IF, which was deprecated anyway and can be easily
1988 simulated with TEMPORARY followed by SELECT IF.
1990 * procedure.c: (open_active_file) Don't call
1991 add_process_if_trns().
1992 (discard_variables) Get rid of redundant call to
1993 proc_cancel_all_transformations().
1994 (add_process_if_trns) Removed.
1995 (process_if_trns_proc) Removed.
1996 (process_if_trns_free) Removed.
1998 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2000 Continue reforming procedure execution. In this phase, add
2001 `const' to the case passed to procedure()'s callback.
2003 Updated all users of procedure() as well.
2005 * procedure.c: (struct write_case_data) Add "const" to ccase
2006 parameter for case_func member.
2007 (procedure) Add "const" to ccase parameter for proc_func
2009 (multipass_case_func) Make ccase parameter const.
2010 (internal_procedure) Add "const" to ccase parameter for case_func
2012 (split_procedure_case_func) Make ccase parameter const.
2013 (multipass_split_case_func) Make ccase parameter const.
2015 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2017 Continue reforming procedure execution. In this phase, get rid of
2018 the output code for SPLIT FILE groups in procedure.c, which really
2019 shouldn't be doing any output. Move it into the individual
2020 procedures instead. This also adds some flexibility.
2022 Updated many users of procedure_with_splits() and
2023 multipass_procedure_with_splits() to call
2024 output_split_file_values() and to deal with increased use of
2027 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2028 parameter to begin_func member.
2029 (procedure_with_splits) Add "const struct ccase *" parameter to
2030 begin_func parameter. Make ccase parameter const in proc_func
2032 (split_procedure_case_func) Don't dump split file group. Pass
2034 (dump_splits) Moved to language/dictionary/split-file.c as
2035 output_split_file_values().
2036 (struct multipass_split_aux_data) Add "const struct ccase *"
2037 parameter to split_func member.
2038 (multipass_procedure_with_splits) Add "const struct ccase *"
2039 parameter to split_func parameter.
2040 (multipass_split_case_func) Save new SPLIT FILE case before
2042 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2044 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2046 Continue reforming procedure execution. Change
2047 internal_procedure() so that it calls open_active_file() and
2048 close_active_file(), which isolates most of the actual procedure
2051 * procedure.c: (struct write_case_data) Rename `proc_func' member
2052 to `case_func' and update all references.
2053 (procedure) Rewrite as one-line wrapper around
2054 internal_procedure().
2055 (struct multipass_aux_data) New.
2056 (multipass_callback) Renamed multipass_case_func(). Use struct
2057 multipass_aux_data as auxiliary data.
2058 (multipass_end_func) New function.
2059 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2060 that uses multipass_case_func, multipass_end_func.
2061 (internal_procedure) Add `end_func' argument. Move optimization
2062 of trivial case in here. Move call to open_active_file() and
2063 close_active_file() in here. Now assert that vfm_source is
2065 (procedure_with_splits_callback) Rename
2066 split_procedure_case_func().
2067 (split_procedure_end_func) New function.
2068 (multipass_split_callback) Rename multipass_split_case_func.
2069 (multipass_split_end_func) New function.
2070 (discard_variables) No need to test for nonnull vfm_source.
2072 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2074 Continue reforming procedure execution. Get rid of unused member.
2076 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2078 (write_case) Don't increment cases_analyzed.
2080 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2082 Continue reforming procedure execution. In this phase, move
2083 procedure.c and procedure.h from src to src/data. Update
2084 makefiles and #includes accordingly.
2086 * procedure.c: Moved here from src/.
2088 * procedure.h: Moved here from src/.
2090 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2092 Continue reforming procedure execution. In this phase, get rid of
2093 many global variables, consolidating procedure execution in
2094 procedure.c. Encapsulate transformations in new "struct
2095 trns_chain". Also, change implementation of N OF CASES, FILTER,
2096 and PROCESS IF from special cases to transformations.
2098 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2101 * dictionary.c: (global variable default_dict) Move to
2104 * variable.h: (TRNS_*) Move to transformations.h.
2105 (struct transformation) Move to transformations.c.
2107 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2109 * sys-file-reader.c: Fixed invalid read problems.
2111 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2113 * storage-stream.c: Add missing function comments.
2115 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2117 Continue reforming procedure execution. In this phase, add some
2118 new, needed functionality to storage-stream.
2120 * storage-stream.c: (storage_source_decapsulate) New function.
2122 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2124 * variable.c (width_to_bytes): Declarations must precede
2125 statements for C90 compliance.
2127 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2129 * data-out.c, data-in.c, variable.c, variable.h: New functions
2130 copy_mangle and copy_demangle for reading/writing cases; emulates the
2131 way SPSS deals with strings > 255 bytes.
2133 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2134 subtype 14 needed for strings longer than 255 bytes.
2136 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2137 instead of literal values. Also fixed some constness issues.
2139 * format.h: Constness
2141 * sfm-private.h: Renamed the case_size identifier, since I discovered
2142 that SPSS's respect for this variable is very nominal.
2144 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2146 Change case limit type from int to size_t.
2148 * dictionary.c: (struct dictionary) Change type of case_limit
2150 (dict_get_case_limit) Change return type.
2151 (dict_set_case_limit) Change parameter type.
2153 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2155 * variable.h: (struct variable) Rename `reinit' member as `leave'
2156 and invert sense. Fix up all references.
2158 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2160 Continue reforming procedure execution. In this phase, break
2161 procedure.c into multiple files.
2163 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2165 * case-sink.c: New file.
2167 * case-sink.h: New file.
2169 * case-source.c: New file.
2171 * case-source.h: New file.
2173 * storage-stream.c: New file.
2175 * storage-stream.h: New file.
2177 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2179 * variable.h: (struct variable) Remove `init' member and all
2180 references to it from other files. It was initialized in several
2181 places, but nothing really ever used it for anything worthwhile.
2182 Thanks to Jason Stover for pointing out how confusing this
2185 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2187 Continue reforming error message support. In this phase, get rid
2188 of message "titles" and put the message text in `struct error'.
2189 Now `struct error' encapsulates a message more properly.
2191 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2192 Format message ourselves.
2194 * data-in.c: (vdls_error) Ditto.
2196 * por-file-reader.c: (error) Ditto.
2198 * sys-file-reader.c: (corrupt_msg) Ditto.
2200 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2202 GNU standards require "file name" instead of "filename" in
2203 documentation. It's nice for our code to follow the convention
2206 * casefile.c: (struct casefile) Rename `filename' member to
2207 `file_name'. Updated all references.
2209 * file-name.c: [!unix] (struct file_identity) Rename
2210 normalized_filename member to normalized_file_name. Updated all
2213 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2215 We don't really support anything but Unix-like environments well,
2216 so we might as well de-obfuscate by writing directory and path
2217 separators explicitly.
2219 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2221 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2222 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2224 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2227 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2229 GNU standards require "file name" instead of "filename" in
2230 documentation. It's nice for our code to follow the convention
2233 * filename.c: Rename to file-name.c.
2235 * filename.h: Rename to file-name.h. Update all inclusions.
2236 Update header guards.
2238 * automake.mk: Update file names.
2240 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2242 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2244 (fn_basename) Removed (dead code).
2245 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2246 (fn_special_p) Renamed fn_is_special(), all references updated.
2247 (fn_exists_p) Renamed fn_exists(), all references updated.
2249 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2251 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2252 Also, now it only tilde-expands file names, not paths.
2253 (fn_search_path) Tilde-expand one directory at a time.
2255 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2257 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2258 get rid of non-Unixlike version of the code, which has probably
2260 (fn_prepend_dir) Removed (dead code).
2262 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2263 (macro PATH_SEPARATOR_STRING) New.
2264 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2266 Continue reforming error message support. In this phase, we get
2267 rid of VM() and the other msg() support for "verbosity", replacing
2268 it by a new function verbose_msg().
2270 * filename.c: (fn_search_path) Use verbose_msg() instead of
2273 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2275 * sfm-private.h: Get rid of #defines after #error, which makes no
2278 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2280 Get rid of our own int32 type in favor of the standard int32_t
2283 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2286 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2288 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2290 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2292 Remove ill-considered file routines that are no longer used.
2294 * filename.c: (fn_open_ext) Removed.
2295 (fn_close_ext) Removed.
2297 * filename.h: (struct file_ext) Removed.
2299 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2301 * variable.c (var_is_valid_name): Move declarations before code
2304 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2306 * filename.ch (fn_interp_vars): Fixed small buglet.
2308 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2310 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2311 so as to modify the string inline. Thus makeing it easier to
2312 destroy the results when no longer needed.
2314 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2316 * category.c (cat_stored_values_destroy): Fixed memory leak.
2318 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2320 Add some missing frees. Thanks to John Darrington for reporting
2323 * any-writer.c (any_writer_close): Free writer.
2325 * any-reader.c (any_reader_close): Free reader.
2327 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2329 * por-file-reader.c: (error) Mark as NO_RETURN.
2331 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2333 * settings.c: Changed default value of scompress to true.
2335 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2337 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2339 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2342 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2343 here from pref.h.orig.
2345 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2347 * sys-file-reader.c: Fixed bug reading compressed files.
2349 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2351 * Numerous renames. See src/ChangeLog for details.
2353 * Moved files from src directory