1 2007-11-10 Ben Pfaff <blp@gnu.org>
3 Cleanups and bug fixes devised while writing up documentation.
6 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
8 * dict-class.c: New file.
9 (dict_class_from_id): Move here.
10 (dict_class_to_name): Move here.
12 * dict-class.h: New file.
13 (enum dict_class): Move here. Change from consecutive integers to
14 consecutive bits, to make testing for any of multiple values
15 easier. Add new DC_ALL constant.
17 * dictionary.c (struct dictionary): Change `case_limit' from
19 (dict_get_vars): Make final argument an enum dict_class.
20 (dict_get_vars_mutable): Ditto.
21 (dict_get_case_limit): Change return value to casenumber.
22 (dict_set_case_limit): Change final argument to a casenumber.
23 (dict_unset_split_var): Add assertion. Rephrase slightly.
24 (dict_set_label): Use xstrndup to simplify.
26 * format.c (fmt_step_width): AHEX format also needs 2-byte
28 (fmt_set_style): Simplify assertions.
30 * missing-values.c (mv_add_num_range): Rename mv_add_range.
31 Simplify implementation.
32 (mv_has_value): Simplify implementation.
33 (mv_pop_value): Remove the first value, not the last, to avoid
34 having GET followed by SAVE reverse the order of missing values.
35 (mv_peek_value): Rename mv_get_value. Simplify assertion.
36 (mv_has_range): Simplify implementation.
37 (mv_peek_range): Rename mv_get_range.
38 (can_resize_string): Removed.
39 (mv_is_resizable): Use value_is_resizable.
40 (mv_resize): Use value_resize.
42 * short-names.h (SHORT_NAME_LEN): Move here.
44 * val-type.h: New file, for definitions related to type and width
45 of abstract values. Before, these definitions were mixed among
46 those related to "union value" and those related to variables.
47 (macro SYSMIS): Move here.
48 (macro LOWEST): Move here.
49 (macro HIGHEST): Move here.
50 (macro MAX_STRING); Move here.
51 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
52 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
53 All references updated.
54 (val_type_is_valid): New function. Replaces var_type_is_valid.
55 All references updated.
56 (val_type_from_width): New function. Replaces
57 var_type_from_width. All references updated.
59 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
61 (val_labs_can_set_width): Use value_is_resizable.
62 (val_labs_add): Simply return false if the value labels set is too
63 wide, instead of having undefined behavior.
64 (val_labs_replace): Ditto.
65 (val_labs_replace): Ditto.
66 (val_labs_first): Set iterator to null if iteration is complete.
67 (val_labs_first_sorted): Ditto.
68 (val_labs_done): Become a no-op if the iterator is null.
70 * value.c (value_is_resizable): New function.
71 (value_resize): New function.
73 * variable.c (var_get_dict_class): New function.
75 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
78 2007-11-08 Ben Pfaff <blp@gnu.org>
80 * data-in.c: Make formatted data parsing locale-independent.
81 (parse_number): Use c_strtod instead of strtod, to avoid
82 locale-specific behavior.
85 2007-11-06 Ben Pfaff <blp@gnu.org>
87 Patch #6256: add support for binary, 360 file formats. Reviewed
90 * data-in.c (struct data_in): Add `encoding' member.
91 (data_in): Add `encoding' parameter, and re-encode the data passed
92 in where appropriate. Update all callers to pass it in.
93 (parse_A): Implement EBCDIC recoding wart described in manual.
94 (parse_AHEX): Implement EBCDIC recoding.
96 * data-out.c (data_out_legacy): New function.
97 (data_out): Make into a wrapper around data_out_legacy.
99 * file-handle-def.c (struct file_handle): New member `encoding'.
100 (fh_create_file): Set encoding.
101 (fh_default_properties): Set default encoding.
102 (fh_get_legacy_encoding): New function.
104 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
105 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
106 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
107 (struct fh_properties): New member `encoding'.
109 2007-11-05 Ben Pfaff <blp@gnu.org>
111 Patch #6258. Reviewed by John Darrington.
113 * file-handle-def.c (fh_lock): Add comment that TYPE should be
114 marked with N_() in the caller. Added these markings to each
115 caller too. Should make i18n easier.
116 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
118 2007-11-03 Ben Pfaff <blp@gnu.org>
120 Allow output files to overwrite input files (bug #21280).
122 * file-handle-def.c: Separate locking of files for input and for
123 output, to allow both to take place at once. Also, distinguish a
124 file handle from the identity of the underlying file, because the
125 identity of a file changes over time and the file handle can't
126 represent two different identities.
127 (struct file_handle): Remove `next', `open_cnt', `deleted',
128 `type', `open_mode', `aux', `identity' members. Change `id' from
129 char array to char *. Add `ref_cnt' member.
130 (file_handle_from_ll) New function.
131 (file_handles) Removed.
132 (named_handles) New variable.
133 (fh_init) Initialize named_handles.
134 (fh_done) Remove name from all named_handles.
135 (free_handle) Updated for modified struct file_handle.
136 (unname_handle) New function.
137 (fh_ref) New function.
138 (fh_from_file_name) Removed.
139 (fh_unref) New function.
140 (fh_unname) New function.
141 (fh_from_id) Rewritten.
142 (create_handle) Updated for modified struct file_handle.
143 (fh_inline_file) Increment the handle's ref count.
144 (fh_create_file) Updated for modified struct file_handle.
145 (fh_create_scratch) Ditto.
150 (fh_is_open) Removed.
151 (fh_get_id) Updated for modified struct file_handle.
152 (fh_get_default_handle) Increment the handle's ref count.
153 (fh_set_default_handle) Handle ref counts.
154 (struct fh_lock) New structure.
155 (locks) New static var.
156 (fh_lock) New function.
157 (fh_unlock) New function.
158 (fh_lock_get_aux) New function.
159 (fh_lock_set_aux) New function.
160 (fh_is_locked) New function.
161 (make_key) New function.
162 (free_key) New function.
163 (compare_fh_locks) New function.
164 (hash_fh_lock) New function.
166 * file-handle-def.h (enum fh_access) New enum.
168 * file-name.c: Made file_identity the same in all supported
170 (struct file_identity): New `name' member.
171 (fn_get_identity): For a file that doesn't exist, get the
172 dev/inode of its directory plus its name. If even the directory
173 doesn't exist, just use its name. Merge the Windows
174 implementation into the Unix one.
175 (fn_compare_file_identities): Rewritten. Merge the Windows
176 implementation into the Unix one.
177 (fn_hash_identity): New function.
179 * make-file.c (struct replace_file): New structure.
180 (all_files): New static var.
181 (replace_file_start): New function.
182 (replace_file_commit): New function.
183 (replace_file_abort): New function.
184 (free_replace_file): New function.
185 (unlink_replace_files): New function.
187 * por-file-reader.c (struct pfm_reader): Add `lock' member.
188 (close_reader): Unlock file.
189 (pfm_open_reader): Lock file.
191 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
193 (pfm_open_writer): Lock file and prepare for its replacement.
194 (close_writer): Unlock file.
196 * scratch-handle.h (struct scratch_handle): Add unique_id so that
197 different generations of a scratch file can be distinguished.
199 * scratch-reader.c (scratch_reader_open): Verify that the file is
202 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
203 remove scratch_handle member.
204 (scratch_writer_open): Lock handle. Prepare to replace handle
205 data, instead of doing it immediately.
206 (scratch_writer_casewriter_destroy): Replace handle data and
209 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
210 (sfm_open_reader): Lock file.
211 (close_reader): Unlock file.
213 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
215 (sfm_open_writer): Lock file and prepare for its replacement.
216 (close_writer): Unlock file.
218 2007-11-02 Ben Pfaff <blp@gnu.org>
220 * data-out.c (output_number): Use isfinite (from C99) instead of
221 gsl_isfinite, since we now have universal support for it in
223 (output_infinite): Ditto.
224 [!HAVE_ROUND] (round): Remove definition, since we now have a
225 replacement in gnulib.
227 2007-10-27 John Darrington <john@darrington.wattle.id.au>
229 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
230 wasn't initialised correctly. Closes bugs #21330 and 21397
233 2007-10-08 Ben Pfaff <blp@gnu.org>
235 Fix bug #21280. Thanks to John Darrington for review.
237 * file-name.c (create_stream): New function.
239 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
240 file handle before creating the file, to ensure that we don't
241 truncate a file that we're reading. Make code easier to read by
244 * sys-file-write.c (sfm_open_writer): Ditto.
246 2007-10-01 Ben Pfaff <blp@gnu.org>
248 Fix bug #21192. Thanks to John Darrington for review.
250 * casereader.c (casereader_read): Decrement case_cnt before
251 calling the casereader's "read" member function, so that we
252 interact properly with lazy_casereader.
254 * datasheet.c: Add regression test for above bug fix.
255 (clone_datasheet): New function.
256 (lazy_callback): New function.
257 (check_datasheet_casereader): New function.
258 (check_datasheet): Check datasheet contents are reported correctly
259 through an ordinary casereader and a lazy casereader.
260 (clone_model): Use clone_datasheet.
262 2007-09-24 Ben Pfaff <blp@gnu.org>
264 Patch #6210. Reviewed by John Darrington.
265 * settings.c: Drop "long view", which is not needed any longer.
266 (static var long_view): Removed.
267 (force_long_view): Removed.
268 (get_viewwidth): Removed.
269 (init_viewport): Removed long_view reference.
271 2007-09-19 John Darrington <john@darrington.wattle.id.au>
273 * settings.c settings.h: Changed viewport's length and width to be
274 owned by the user interface which uses the data library. This allows
275 better abstraction, and makes dynamically adjustable dimensions easier.
277 2007-09-18 Ben Pfaff <blp@gnu.org>
279 * procedure.c (proc_extract_active_file_data): New function.
281 * lazy-casereader.h: New file.
283 * lazy-casereader.c: New file.
285 * casereader.c (casereader_dynamic_cast): New function.
287 2007-09-14 Ben Pfaff <blp@gnu.org>
289 * dictionary.c (dict_clone): Copy case indexes from cloned
290 dictionary. Fixes bug #21061. Reviewed by John Darrington.
292 2007-09-13 John Darrington <john@darrington.wattle.id.au>
294 * value.c value.h (create_value): New function.
296 2007-09-12 Ben Pfaff <blp@gnu.org>
298 Make it clear that translator casereader and casewriter translate
299 functions are supposed to destroy their input case.
301 * casereader-translator.c (struct casereader_translator): Change
302 input case parameter of translate member function from const
303 struct ccase * to struct ccase *.
304 (casereader_create_translator): Ditto, for translate parameter.
306 * casewriter-translator.c (struct casewriter_translator): Ditto.
307 (casewriter_create_translator): Ditto.
309 2007-08-27 John Darrington <john@darrington.wattle.id.au>
311 * sys-file-reader.c (read_display_parameters): Force display width
312 to 8, if the sys file says 0 (like SPSS does).
314 2007-08-12 Ben Pfaff <blp@gnu.org>
316 * dictionary.c (dict_dump): New function.
318 2007-08-12 Ben Pfaff <blp@gnu.org>
320 Drop dict_compactor in favor of using the new struct case_map.
322 * dictionary.c (struct copy_map): Removed.
323 (struct dict_compactor): Removed.
324 (dict_make_compactor): Removed.
325 (dict_compactor_compact): Removed.
326 (dict_compactor_destroy): Removed.
328 * procedure.c (struct data_set): Change `compactor' member to be a
330 (proc_open): Use case_map_to_compact_dict instead of
332 (proc_casereader_read): Use case_map_execute instead of
333 dict_compactor_compact.
334 (proc_commit): Use case_map_destroy instead of
335 dict_compactor_destroy.
337 * scratch-writer.c (struct scratch_writer): Change `compactor'
338 member to be a struct case_map *.
339 (scratch_writer_open): Use case_map_to_compact_dict instead of
341 (scratch_writer_casewriter_write): Use case_map_execute instead of
342 dict_compactor_compact.
344 2007-08-12 Ben Pfaff <blp@gnu.org>
346 * automake.mk: Add case-map.c, case-map.h.
348 * case-map.c: New file.
350 * case-map.h: New file.
352 2007-08-12 Ben Pfaff <blp@gnu.org>
354 * dictionary.c (dict_compact_values): Don't delete scratch
355 variables as well as compacting case indexes. Update all callers.
356 (dict_get_compacted_value_cnt): Rename dict_count_values and
357 change interface. Update all callers.
358 (dict_get_compacted_value_cnt): Remove.
359 (dict_compacting_would_shrink): Remove.
360 (dict_compacting_would_change): Remove.
361 (dict_make_compactor): Add new parameter. Update all callers.
363 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
364 instead of dict_count_values, changing an O(N) operation into
367 2007-08-12 Ben Pfaff <blp@gnu.org>
369 * casereader.c (casereader_read): Don't require cases read by a
370 casereader to be exactly the expected size: as long as they're big
373 2007-08-12 Ben Pfaff <blp@gnu.org>
375 Make casewriters keep track of the number of `union value's in
376 each case. This is useful for two reasons: casewriter_write can
377 then check that the case being written is large enough, and later
378 recipients of the casewriter can determine the size of the case.
380 * casewriter-translator.c (casewriter_create_translator): Add
383 * casewriter.c (struct casewriter): Add value_cnt member.
384 (casewriter_write): Check that the case passed in is big enough.
385 (casewriter_get_value_cnt): New function.
386 (casewriter_create): Add value_cnt parameter.
388 2007-08-09 Ben Pfaff <blp@gnu.org>
390 Fix bug reported by Jason Stover.
391 * settings.c: Move get_termcap_viewport and in particular the
392 #include for <curses.h> to the end of the file. curses.h
393 redefines bool on some systems (e.g. OpenBSD), which causes
394 disagreement between uses of bool before its inclusion and
396 Tested by Jason Stover.
398 2007-07-29 Ben Pfaff <blp@gnu.org>
400 Provisional fix for bug #18692 and bug #20161. Reviewed by John
403 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
404 argument (never "rb" or "wb") because SUSv3 says that only those
405 modes are defined, and glibc in fact rejects other modes.
407 Open portable files with fn_open so that they can be read from
408 pipes. Fix missing fh_close call to go along with fh_open.
409 Report an error if the file close reports an error.
410 * por-file-reader.c (close_reader): New function.
411 (por_file_casereader_destroy): Use close_reader.
412 (pfm_open_reader): Open file with fn_open.
414 2007-07-28 Ben Pfaff <blp@gnu.org>
416 Make PSPP able to read all the portable files I could find on the
417 web. Thanks to John Darrington for review. Bug #17620.
418 * por-file-reader.c (struct pfm_reader): New member `line_length'.
419 (error): Print file offset in hexadecimal.
420 (warning): New function.
421 (advance): Treat lines less than 80 bytes long as padded to 80
423 (pfm_open_reader): Call read_documents if we find an "E" record.
424 (convert_format): Convert invalid formats to the default format
425 instead of aborting reading the file.
426 (read_variables): Rename duplicate variable names instead of
427 aborting reading the file.
428 (read_value_label): Allow string variables of different widths to
429 be assigned value labels in the same record. Replace duplicate
430 value labels instead of aborting.
431 (read_documents): New function.
433 * por-file-writer.c (pfm_open_writer): Call write_documents if the
434 dictionary has documents.
435 (write_documents): New function.
437 2007-07-25 Ben Pfaff <blp@gnu.org>
439 Fix bugs related to bug #17213.
441 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
442 since the former is what config.h has. Include the needed ncurses
444 (static var echo) Rename to `do_echo' because the original name is
445 the same as an ncurses identifier.
446 (get_termcap_viewport) Use error instead of msg.
448 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
449 (fn_close): Don't close stdin, stdout, stderr.
451 2007-07-26 John Darrington <john@darrington.wattle.id.au>
453 * procedure.c procedure.h: Added callbacks which get invoked whenever
454 a dataset's transformation chain changes.
456 2007-07-24 Ben Pfaff <blp@gnu.org>
459 * sys-file-writer.c (write_variable_display_parameters): Use new
460 var_default_display_width function to choose display width of
461 segments after the first one in a given variable.
462 * variable.c (var_create): Use var_default_display_width to pick
463 new variable's display width.
464 (var_default_display_width): New function.
465 Reviewed by John Darrington.
467 2007-07-24 Ben Pfaff <blp@gnu.org>
470 * por-file-writer.c (write_variables): Write weight variable.
471 Reviewed by John Darrington.
473 2007-07-23 Ben Pfaff <blp@gnu.org>
475 Improvements to system file reader and writer.
477 First, move all detailed knowledge of very long strings into
478 sys-file-private.[ch], so that this nasty stuff can be isolated.
480 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
481 (EFFECTIVE_VLS_CHUNK): New macro.
482 (min_int): New function.
483 (max_int): New function.
484 (sfm_width_to_bytes): Rewrite.
485 (sfm_width_to_octs): New function.
486 (sfm_segment_alloc_width): New function.
487 (sfm_segment_alloc_bytes): New function.
488 (sfm_segment_used_bytes): New function.
489 (sfm_segment_offset): New function.
490 (sfm_segment_effective_offset): New function.
491 (sfm_dictionary_to_sfm_vars): New function.
493 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
494 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
495 (struct sfm_var): New structure.
497 Next, improvements to the system file reader.
499 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
500 type casenumber. Added `version_major', `version_minor',
503 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
504 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
505 Change `case_cnt' to type casenumber. Removed `has_vls'.
506 (struct sfm_var): Removed.
507 (sfm_open_reader): Don't warn on wrong case size if the file was
508 written by SPSS 13, which tends to get it wrong. Use
509 sfm_dictionary_to_sfm_vars.
510 (read_header): Always output system file info.
511 (read_variable_record): Simplify code for reading missing values.
512 (read_machine_int32_info): Save version numbers from system file
513 into info struct passed as new argument.
514 (read_long_string_map): Restructured to use new sys-file-private
516 (read_value_labels): Use size_overflow_p.
517 (sys_file_casereader_read): Get rid of distinction between fast
518 and slow paths. Use information provided by sys-file-primate's
519 struct sfm_var to simplify code.
520 (skip_whole_strings): New function.
521 (read_int32): Renamed read_int. Changed return value to int.
523 (read_flt64): Renamed read_float. Changed return value to
524 double. Updated all callers.
525 (int32_to_native): Removed. Changed callers to use
527 (flt64_to_double): Removed. Changed callers to use float_convert.
529 Finally, get rid of int32, flt64 terminology and types in system
530 file writer. The former wasn't very useful since a POSIX "int"
531 can hold the whole range of int32 and we generally didn't have a
532 need for it to be exactly-32-bits, just at-least-32-bits. The
533 latter was inconvenient because we had to assume that it could be
534 different from double and thereby convert special values SYSMIS,
535 HIGHEST, LOWEST to and from it in multiple places. Instead, now
536 we just use "int" and "double" in most places, and do conversions,
537 if necessary, very close to where we do I/O. This change meant
538 that the writer code couldn't represent records in the file as C
539 structs any longer, but that's no great loss. The code actually
540 seems to be more readable without them.
542 Simplify the compression buffering code: only buffer as much as
543 necessary, which is no more than eight 8-byte units at any given
546 * sys-file-writer.c (typedef flt64): Removed.
547 (macro second_lowest_flt64): Removed.
548 (struct sysfile_header): Removed.
549 (struct sysfile_variable): Removed.
550 (struct sfm_writer): Removed `needs_translation', `has_vls',
551 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
552 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
553 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
554 compression buffering by `opcodes', `opcode_cnt', `data',
555 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
556 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
557 data writer functions instead of structures.
558 (calc_oct_idx): New function.
559 (write_header): Use simple data writer functions instead of
561 (write_format_spec): Renamed write_format. New argument.
562 (write_variable_continuation_records): New function.
563 (write_variable): Use simple data writer functions instead of
564 structures. Use write_variable_continuation_records. Write
565 entire very long string instead of requiring caller to understand
567 (write_value_labels): Use simple data writer functions instead of
569 (write_documents): Ditto.
570 (write_variable_display_parameters): Use sys-file-private
571 functions to simplify. Use simple data writer functions instead
573 (write_vls_length_table): Use simple data writer functions instead
575 (write_longvar_table): Ditto.
576 (write_rec_7_34): Break into new functions
577 write_integer_info_record, write_float_info_record. Use simple
578 data writer functions instead of structures.
579 (buf_write): Removed.
580 (append_string_max): Removed.
581 (ensure_buf_space): Removed.
582 (sys_file_casewriter_write): Get rid of the distinction between
583 fast and slow paths, which didn't seem to be too useful. Use new
584 functions write_case_uncompressed, write_case_compressed.
585 (put_instruction): Removed.
586 (put_element): Removed.
587 (write_compressed_data): Removed.
588 (close_writer): Use flush_compressed. Only write case count to
589 system file if it will fit in the field.
590 (write_case_compressed): New function.
591 (write_case_uncompressed): New function.
592 (flush_compressed): New function.
593 (put_cmp_opcode): New function.
594 (put_cmp_number): New function.
595 (write_int): New function.
596 (convert_double_to_output_format): New function.
597 (write_float): New function.
598 (write_value): New function.
599 (write_string): New function.
600 (write_bytes): New function.
601 (write_zeros): New function.
602 (write_spaces): New function.
604 Reviewed by John Darrington.
606 2007-07-22 Ben Pfaff <blp@gnu.org>
608 Don't try to write very long strings to portable files. The
609 format does not support it.
611 * por-file-writer.c (MAX_POR_WIDTH): New macro.
612 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
613 (write_format): Add arg to take width to resize format to.
614 (write_value): Limit width of value written to MAX_POR_WIDTH.
615 (write_variables): Limit width of variable and its output formats
617 Reviewed by John Darrington.
619 2007-07-22 Ben Pfaff <blp@gnu.org>
621 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
622 local variable instead of literal 5.
623 Reviewed by John Darrington.
625 2007-07-22 Ben Pfaff <blp@gnu.org>
627 Fix problems with uniqueness of short names in system files with
628 very long string variables. Now a variable may have multiple
631 * automake.mk (src_data_libdata_a_SOURCES): Add new files
632 short-names.c, short-names.h.
634 * dictionary.c (dict_clone): Clone all the short names.
635 (compare_strings): Move into short-names.c.
636 (hash_strings): Ditto.
637 (set_var_short_name_suffix): Ditto.
638 (dict_assign_short_names): Ditto, rename short_names_assign,
639 change to assign all short names.
641 * por-file-writer.c (write_variables): Use short_names_assign
642 instead of dict_assign_short_names.
644 * short-names.c: New file.
646 * short-names.h: New file.
648 * sys-file-private.c (sfm_width_to_segments): New function.
650 * sys-file-reader.c (read_long_var_name_map): Save and restore all
651 the short names, not just the first one.
653 * sys-file-writer.c (cont_var_name): Removed.
654 (sfm_open_writer): Use short_names_assign instead of
655 dict_assign_short_names. Use unique short names assigned by
656 short_names_assign instead of those generated by cont_var_name.
658 * variable.c (struct variable): Remove `short_name' member,
659 replace by `short_names' and `short_name_cnt'.
660 (var_create) Initialize new members.
661 (var_get_short_name_cnt): New function.
662 (var_get_short_name): Now takes an index argument. Changed most
664 (var_set_short_name): Ditto.
665 (var_clear_short_name): Renamed var_clear_short_names, changed to
666 clear all short names.
668 Reviewed by John Darrington.
670 2007-07-22 Ben Pfaff <blp@gnu.org>
672 * variable.c (var_set_width): Use new var_set_width function.
674 * missing-values.c (mv_n_values): Drop assertion, which was not
677 * format.c (fmt_default_for_width): New function.
678 (fmt_resize): New function.
680 Reviewed by John Darrington.
682 2007-07-18 John Darrington <john@darrington.wattle.id.au>
684 * datasheet.c (datasheet_delete_columns): Added assertion to check
685 we're not deleting outside the range of the sheet.
688 * dictionary.c dictionary.h variable.c: Added the ability for string
689 variables to be resized.
691 * vardict.h: Added some prototypes (moved from dictionary.h) as
692 these should only be called by variable.c
695 2007-07-14 John Darrington <john@darrington.wattle.id.au>
697 * sfm-reader.c: Respect case_cnt field in file header.
699 2007-07-01 John Darrington <john@darrington.wattle.id.au>
701 * transformation.c transformation.h (trns_chain_execute): Changed the
702 signature (Patch #6057)
704 2007-06-10 Ben Pfaff <blp@gnu.org>
706 * casereader-filter.c (casereader_filter_destroy): Make sure to
707 write all the remaining excluded cases to the casewriter, if any.
709 * caseinit.c (init_list_destroy): Rewrite.
710 (init_list_clear): Ditto.
712 * casegrouper.c (casegrouper_get_next_group): Always set *reader
713 to null when returning false.
715 2007-06-06 Ben Pfaff <blp@gnu.org>
717 Actually implement the new procedure code and adapt all of its
718 clients to match. Also adapt all of the other case sources and
719 sinks in the tree and their clients to use the
720 casereader/casewriter infrastructure.
722 * automake.mk: Add and remove files.
724 * any-reader.c: Change into a casereader.
725 * por-file-reader.c: Ditto.
726 * scratch-reader.c: Ditto.
727 * sys-file-reader.c: Ditto.
729 * any-writer.c: Change into a casewriter.
730 * por-file-writer.c: Ditto.
731 * scratch-writer.c: Ditto.
732 * sys-file-writer.c: Ditto.
734 * procedure.c: Change to use casereader, casewriter, caseinit, and
735 other new infrastructure.
737 * scratch-handle.c: Adapt to new infrastructure.
739 * case-sink.c: Removed, now dead code.
740 * case-sink.h: Ditto.
741 * case-source.c: Ditto.
742 * case-source.h: Ditto.
743 * casefile-factory.c: Ditto.
744 * casefile-private.h: Ditto.
747 * casefilter.c: Ditto.
748 * casefilter.h: Ditto.
751 * fastfile-factory.c: Ditto.
752 * fastfile-factory.h: Ditto.
753 * storage-stream.c: Ditto.
754 * storage-stream.h: Ditto.
756 2007-06-06 Ben Pfaff <blp@gnu.org>
760 * automake.mk: Add new files.
762 * datasheet.c: New file.
764 * datasheet.h: New file.
766 2007-06-06 Ben Pfaff <blp@gnu.org>
768 Until now, the procedure code has provided a case to the
769 case_source, which has filled in the data values that come from
770 the active file. "Left" data values that don't come from the
771 active file naturally stay the same from case to case, because the
772 procedure code keeps using that same case.
774 One of the compromises that comes with the new procedure code is
775 that the active file allocates and provides its own case, which
776 the procedure code then has to resize to provide room for any
777 other variables that should go in the case and then fill in the
778 values of "left" variables. Then, when we're done with that case,
779 we have to save the values of "left" variables to copy into the
780 next case read from the active file.
782 The caseinit code helps with this.
784 * automake.mk: Add new files.
786 * caseinit.c: New file.
788 * caseinit.h: New file.
790 2007-06-06 Ben Pfaff <blp@gnu.org>
792 * value.h (value_cnt_from_width): New function.
794 * variable.c (var_get_value_cnt): Use new function.
796 2007-06-06 Ben Pfaff <blp@gnu.org>
798 Add casegrouper, to allow cases read from a given casereader to be
799 broken into groups, each of which has its own casereader.
800 Generally cases are grouped based on having equal values for some
803 * automake.mk: Add new files.
805 * casegrouper.c: New file.
807 * casegrouper.h: New file.
809 2007-06-06 Ben Pfaff <blp@gnu.org>
811 Add interface to lexicographical ordering of cases.
813 * automake.mk: Add new files.
815 * case-ordering.c: New file.
817 * case-ordering.h: New file.
819 2007-06-06 Ben Pfaff <blp@gnu.org>
821 Add casereaders and casewriters, the basis of the new data processing
822 implementation. A casereader is a uniform interface to reading cases
823 from a data source; a casewriter is a uniform interface to writing
824 cases to a data sink.
826 * automake.mk: Add new files.
828 * casereader-filter.c: New file.
830 * casereader-provider.h: New file.
832 * casereader-translator.c: New file.
834 * casereader.c: New file.
836 * casereader.h: New file.
838 * casewriter-provider.h: New file.
840 * casewriter-translator.c: New file.
842 * casewriter.c: New file.
844 * casewriter.h: New file.
846 2007-06-06 Ben Pfaff <blp@gnu.org>
848 "casewindow" data structure that extends the deque (from libpspp)
849 of cases with the ability to dump cases to disk if we get too many
852 * automake.mk: Add new files.
854 * casewindow.c: New file.
856 * casewindow.h: New file.
858 2007-06-06 Ben Pfaff <blp@gnu.org>
860 sparse_cases data structure that augments a sparse_array of cases
861 with the ability to dump cases to disk if we get too many cases in
864 * automake.mk: Add new files.
866 * sparse-cases.c: New file.
868 * sparse-cases.h: New file.
870 2007-06-06 Ben Pfaff <blp@gnu.org>
872 Adds a low-level on-disk case array data structure.
874 * automake.mk: Add new files.
876 * case-tmpfile.c: New file.
878 * case-tmpfile.h: New file.
880 2007-06-06 Ben Pfaff <blp@gnu.org>
882 In a couple of places we calculate the maximum number of cases to
883 keep in memory based on the user-defined workspace. Enable
884 centralizing the calculation through a new function.
886 * settings.c (get_workspace_cases): New function.
888 2007-06-06 Ben Pfaff <blp@gnu.org>
890 The casenumber type is defined in transformations.h, but case.h is
891 a more sensible place. Move it.
893 * case.h (CASENUMBER_MAX): New macro.
894 (typedef casenumber): Move here, from transformations.h.
896 2007-06-03 Ben Pfaff <blp@gnu.org>
898 Slightly generalize case_to_values and case_from_values functions.
900 * case.c (case_to_values): Rename case_copy_out, change interface.
901 (case_from_values): Rename case_copy_in, change interface.
903 * fastfile.c (fastfilereader_get_next_case): Update caller.
904 (write_case_to_disk): Ditto.
906 2007-06-02 Ben Pfaff <blp@gnu.org>
908 Clean up after a forgotten part of patch #5829.
910 * casedeque.h: Remove unused file.
912 * automake.mk: Remove casedeque.h from sources.
914 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
916 * category.c: Removed redundant #include
918 2007-05-06 Ben Pfaff <blp@gnu.org>
920 Abstract the documents within a dictionary a little better.
921 Thanks to John Darrington for suggestion, initial version, and
924 * dictionary.c (struct dictionary): Change `documents' member from
925 char * to struct string.
926 (dict_clear): Destroy struct string.
927 (dict_get_documents): Convert struct string to char *.
928 (dict_set_documents): Set struct string. Pad to 80-character
930 (dict_clear_documents): New function.
931 (dict_add_document_line): New function.
932 (dict_get_document_line_cnt): New function.
933 (dict_get_document_line): New function.
935 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
937 * sys-file-reader.c (read_documents): Use new document functions.
939 2007-04-19 John Darrington <john@darrington.wattle.id.au>
941 * sys-file-reader.c: When reading a system file which has no
942 long name table, automatically create one where the long names
943 are the lower case versions of the short names.
945 2007-04-22 Ben Pfaff <blp@gnu.org>
947 * dictionary.c (dict_set_split_vars): dict_destroy expects that
948 dict_clear will free most data related to the dictionary.
949 dict_clear does a decent job, except that dict_set_split_vars on
950 some systems won't actually free the dict's "split" member.
951 Instead, it'll allocate a 1-byte region. Fix this.
953 * value.c (value_copy): New function.
954 (value_set_missing): Ditto.
956 2007-04-22 John Darrington <john@darrington.wattle.id.au>
958 * Deleted existing category.h and moved cat-routines.h into
959 category.h Encapsulated struct cat_vals better.
961 2007-04-19 John Darrington <john@darrington.wattle.id.au>
963 * sys-file-reader.c: When reading a system file which has no
964 long name table, automatically create one where the long names
965 are the lower case versions of the short names.
967 2007-04-16 John Darrington <john@darrington.wattle.id.au>
969 * sys-file-reader.c: Some versions of Other Software seem to
970 produce system files with string variables' measure set to
971 zero. We'll assume these are supposed to be nominal variables.
973 2007-03-30 Ben Pfaff <blp@gnu.org>
975 * procedure.c: Adapt to new deque data structure.
977 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
978 Ben Pfaff <blp@gnu.org>
980 * file-name.c: Mingw compatibility fixes.
981 (fn_search_path): Use ISSLASH instead of comparing against '/'
983 (fn_dir_name): Use dir_name from gnulib.
984 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
985 (fn_get_identity): Use GetFullPathName instead of canonicalize
986 from gnulib, because the latter does not fully support
987 Windows-style path names. Use this implementation based on the
988 detected presence of Windows instead of the absence of Unix, since
989 the new implementation is Windows-specific.
990 (fn_compare_file_identities): In Windows implementation, compare
991 names case-insensitively.
993 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
995 * make-file.c: Don't include mkstemp.h, because gnulib now causes
996 <stdlib.h> to have the same effect.
998 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1000 * por-file-reader.c: Add missing _() around messages.
1002 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1004 * make-file.c: Include "mkstemp.h", without which linking on
1007 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1009 Reduce platform dependence.
1011 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1012 Everywhere we using this, we really should have just depended on
1013 the shell to expand tildes.
1014 (fn_search_path): Simplify, given that we don't do tilde expansion
1016 (fn_normalize): Removed. Caller changed to use the canonicalize
1018 (fn_get_cwd): Removed. Only user was fn_normalize.
1019 (fn_is_absolute): Really only test for absolute names.
1020 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1022 (fn_readlink): Removed, as it was only used fn_normalize.
1023 (fn_exists): Assume the stat function is available; gnulib does.
1024 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1026 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1028 * dictionary.c (dict_create_vector_assert): New function.
1030 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1032 * file-name.c (fn_normalize): Correct name of function
1033 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1034 for pointing this out.
1036 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1038 We are using a single member in struct file_handle, the "name"
1039 field, for more than one purpose. When it begins with '"', it's a
1040 file name; otherwise, it's a token that can be used to identify
1041 it. When that assertion fires, it's because we searched for the
1042 name case-sensitively as a file name (so that there was no match),
1043 and then we try to insert it case-insensitively as a token, which
1044 fails because duplicates aren't allowed.
1046 Solution: break the two purposes into two separate fields. This
1047 fixes the problem and likely makes the code easier to read too.
1049 Fixes bug #18922. Thanks to John Darrington for bug report and
1052 * file-handle-def.c (struct file_handle): New `id' member.
1053 (fh_from_name): Rename fh_from_id. Update all callers.
1054 (create_handle): New `id' parameter. Update all callers.
1055 (fh_create_file): Ditto.
1056 (fh_get_id): New function.
1058 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1060 * case.c (case_is_null): Change return type to bool.
1062 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1066 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1067 reference counting to share data. This makes it easy for
1068 valgrind, etc. to find accesses to cases that have been destroyed
1069 but have been kept around by another user's ref-count. This often
1070 happens when the data set is small enough to find in memory; if a
1071 bigger data set that would overflow to disk were used, then data
1072 corruption would occur.
1074 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1078 * case.c (case_unshare): Make it check internally whether the
1079 ref_cnt is greater than 1, so that the callers don't have to.
1080 Update callers not to check.
1082 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1084 Before, I was thinking that I might want to get rid of reference
1085 counting at some point. Now, I'm pretty sure that it's here to
1086 stay. Thus, because we have to store the value_cnt anyway for
1087 reference-counted cases, we might as well expose it to users.
1089 * case.c (case_get_value_cnt): New function.
1090 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1091 resize case if its size actually changed.
1093 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1094 instead of peeking inside struct case directly.
1095 (casefile_append): Ditto.
1097 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1099 Get rid of the inlines for the case functions, which made the
1100 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1101 the inlines didn't speed up "make check" at all, which is not a
1102 perfect benchmark but seems indicative.)
1104 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1105 definitions. Remove some assertions on nonnull pointers that were
1106 redundant with a pointer dereference soon after in the function.
1108 (struct case_data): Move definition here from case.h.
1112 (case_data_wr): Ditto.
1114 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1116 * automake.mk: Add casedeque.h to sources.
1118 * casedeque.h: New file.
1120 * procedure.c: (struct dataset) Change lag_count, lag_head,
1121 lag_queue member into single struct casedeque member. Update all
1122 users to use the casedeque instead.
1125 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1127 * procedure.c: Simplify lagged cases interface. Updated all
1128 clients--well, the only client--to use the simplified interface.
1129 (dataset_n_lag) Removed.
1130 (dataset_set_n_lag) Removed.
1131 (dataset_need_lag) New function.
1133 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1135 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1136 whenever appropriate, but only when the dataset/dictionary is in a
1139 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1141 * dictionary.c dictionary.h : Added callbacks for change of filter and
1142 split variables. Refactored some code to ensure that callbacks get
1143 invoked when appropriate.
1145 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1146 replace_dict callback occurs when permanent_dict replaces the current
1149 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1151 * dictionary.c dictionary.h : Added callback for when the weight
1152 variable of a dictionary changes.
1154 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1156 * dictionary.c dictionary.h : Added replace_source and replace_dict
1157 callbacks, and functions to deal with them.
1159 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1161 Simplify missing value handling.
1163 * missing-values.h (enum mv_class): New type.
1164 (enum mv_type): Moved definition into missing-values.c and renamed
1165 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1166 enums. Updated all uses.
1167 (struct missing_values): Changed type of `type' from `enum
1168 mv_type' to `int' because the definition is no longer exposed.
1170 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1171 parameter. Update all callers.
1172 (mv_is_num_missing): Ditto.
1173 (mv_is_str_missing): Ditto.
1174 (mv_is_value_user_missing): Removed. Changed callers to use
1175 mv_is_value_missing.
1176 (mv_is_num_user_missing): Removed. Changed callers to use
1178 (mv_is_str_user_missing): Removed. Changed callers to use
1180 (mv_is_value_system_missing): Removed. Changed callers to use
1181 mv_is_value_missing.
1182 (mv_set_type): Removed. Changed callers to use mv_clear.
1183 (mv_clear): New function.
1185 * variable.c (var_is_value_missing): Add new enum mv_class
1186 parameter. Update all callers.
1187 (var_is_num_missing): Ditto.
1188 (var_is_str_missing): Ditto.
1189 (var_is_value_user_missing): Removed. Changed callers to use
1190 var_is_value_missing.
1191 (var_is_num_user_missing): Removed. Changed callers to use
1193 (var_is_str_user_missing): Removed. Changed callers to use
1195 (var_is_value_system_missing): Removed. Changed callers to use
1196 var_is_value_missing.
1198 * casefilter.c (struct casefilter): Use enum mv_class in place of
1200 (casefilter_variable_missing): Adapt to new member.
1201 (casefilter_create): Change signature to take enum mv_class,
1204 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1206 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1208 * case-sink.c case-sink.h procedure.c procedure.h
1209 storage-stream.c: Now uses the factory.
1211 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1213 Make it possible to pull cases from the active file with a
1214 function call, instead of requiring indirection through a callback
1217 * case-source.h (struct case_source_class): Change ->read function
1218 to return a single case, instead of calling a callback function
1219 for each case. Change ->destroy function to return an error
1222 * case-source.c (free_case_source): Pass along the value returned
1223 by the case_source ->destroy function.
1225 * procedure.c (struct write_case_data): Removed.
1226 (struct dataset): Added some members to track procedure state.
1227 (procedure): Optimize the trivial case at this level.
1228 (internal_procedure): Re-implement in terms of proc_open,
1229 proc_read, proc_close.
1230 (proc_open) New function.
1231 (proc_read) New function.
1232 (proc_close) New function.
1233 (write_case) Moved into proc_read.
1234 (close_active_file) Moved closing of data source into proc_close.
1236 * storage-source.c: Rewrote to conform with modified
1237 case_source_class interface.
1239 * transformations.c (trns_chain_execute): Added argument to allow
1240 starting execution from an arbitrary transformation. Updated
1243 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1245 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1247 * sys-file-reader.c (read_display_parameters): Don't assume that
1248 MEASURE_* and ALIGN_* have the same values found in system files.
1250 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1252 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1253 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1254 the value of n_MEASURES, which was off by 1 (at least from my
1257 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1259 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1260 callbacks which are invoked when the dictionary or its
1261 variables are changed.
1263 * missing-values.c missing-values.h value-labels.c: Tidied up
1264 consistency checks, and made some of them return false
1265 instead of assert-failing.
1267 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1269 * calendar.c (calendar_days_in_month): New function.
1271 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1273 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1274 the value label's width.
1276 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1278 * sfm-private.h: Move contents into sys-file-writer.c, which is
1279 the only remaining user. Removed Borland C++-specific directives.
1281 * sys-file-reader.c: Clean up and rewrite entire file. The
1282 rewritten version is simpler and better abstracted, and should be
1283 easier to maintain and extend. It avoids using structures to read
1284 file data, which is prone to padding variations among compilers.
1285 It should also handle non-IEEE 754 system files, although I
1286 haven't been able to find any. It has been tested against many
1287 .sav files obtained from the Web and found to produce the same
1288 results as the earlier version of the code, or in some cases
1289 improved results. It is more tolerant of format variations found
1292 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1293 member, putting an enum integer_format in its place. New member
1294 `float_format'. Changed `compressed' member to type bool.
1296 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1298 * dictionary.c (dict_delete_consecutive_vars): New function.
1300 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1302 * file-name.c (fn_search_path): Remove prefix arg that was unused
1303 by any caller. Updated all callers.
1305 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1307 * format.c (fmt_dollar_template): Use user's decimal point
1308 character. Add assertion.
1310 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1312 * format.c (fmt_dollar_template): New function, based on
1313 dollar_format_template from var-type-dialog.c.
1315 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1317 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1320 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1322 Finish converting struct variable to an opaque type. In this
1323 phase, we add remaining setter and getter functions, convert the
1324 remaining PSPP code to use them, and do a bunch of cleanup. The
1325 resulting changes are pervasive but mostly trivial, and only the
1326 notable changes are logged.
1328 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1331 * case.c (case_data): Renamed case_data_idx.
1332 (case_num): Renamed case_num_idx.
1333 (case_str): Renamed case_str_idx.
1334 (case_data_rw): Renamed case_data_rw_idx.
1336 * case.h (case_data): New function with old name and an interface
1337 that takes a variable instead of an index, which is easier to
1338 use. Updated all callers to use the new interface, or to use the
1339 new *_idx function (see above).
1342 (case_data_rw): Ditto.
1344 * category.c (cat_stored_values_destroy): Changed interface to
1345 take a struct cat_vals * instead of a struct variable *.
1347 * dictionary.c (dict_clone): Use new vector_clone function.
1348 (dict_clear) Use new var_destroy function.
1349 (add_var) New function.
1350 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1351 (dict_create_var_assert) Rewrite in terms of add_var.
1352 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1353 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1354 (dict_lookup_var) Use new var_create, var_destroy functions.
1355 (dict_contains_var) Rewrite in terms of new vardict functionality.
1356 (set_var_dict_index) New function.
1357 (set_var_case_index) New function.
1358 (reindex_vars) New function.
1359 (dict_delete_var) Rewrite in terms of new vardict functionality.
1360 (dict_reorder_var) Ditto.
1361 (dict_reorder_vars) Ditto.
1362 (rename_var) New function.
1363 (dict_rename_var) Use rename_var.
1364 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1365 (dict_get_compacted_idx_to_fv) Rename
1366 dict_get_compacted_dict_index_to_case_index, update callers.
1367 (dict_create_vector) Use new vector_create function.
1368 (dict_clear_vectors) Use new vector_destroy function.
1369 (set_var_short_name_suffix) Move here from variable.c, renamed
1370 from var_set_short_name_suffix, make static, update caller.
1372 * sys-file-private.c: New file.
1373 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1374 width_to_bytes, update callers.
1376 * sys-file-private.h: New file. Later it will supplant
1377 sfm-private.h; for now it supplements it.
1378 (macro MIN_VERY_LONG_STRING) New macro.
1379 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1381 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1382 MAX_LONG_STRING was used before.
1384 * sys-file-writer.c: Ditto.
1386 * value-labels.c: Change the paradigm here to be that a null
1387 pointer is OK for a struct val_labs * in most cases; it just
1388 represents an empty set of value labels.
1389 (val_labs_copy) A copy of a null set is a null set.
1390 (val_labs_count) A null set has 0 labels.
1391 (val_labs_replace) Change return type to void. Rewrite for
1393 (val_labs_find) A null set does not contain the value.
1394 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1395 transposed argument order, updated all callers.
1397 * value.c: New file.
1398 (value_dup) Moved here from variable.c.
1399 (compare_values) Ditto.
1402 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1403 (macro MAX_LONG_STRING) Removed, because it was only interesting
1404 for system files, not for general code.
1405 (macro MAX_VERY_LONG_STRING) Ditto.
1406 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1407 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1409 * vardict.h: New file, for an interface between variables and
1412 * variable.c: A lot of functions were moved around, for better
1414 (struct variable) Move definition here, from variable.h.
1415 (var_type_adj) Removed--makes i18n hard.
1416 (var_type_noun) Ditto.
1417 (var_create) New function.
1418 (var_clone) New function.
1419 (var_destroy) New function.
1420 (var_set_name) Assert that variable is not in a dictionary.
1421 (compare_var_names) Rename compare_vars_by_name and fix a couple
1422 of callers who thought the args were strings.
1423 (hash_var_name) Rename hash_var_by_name.
1424 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1425 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1426 (var_is_very_long_string) Removed, because it was only interesting
1427 to system file code.
1428 (var_set_missing_values) Allow the argument to be the wrong width,
1429 as long as we can resize it. Simplify callers who were doing the
1430 resizing themselves.
1431 (var_get_value_labels) New function.
1432 (var_has_value_labels) New function.
1433 (var_set_value_labels) New function.
1434 (alloc_value_labels) New function.
1435 (var_add_value_label) New function.
1436 (var_replace_value_label) New function.
1437 (var_clear_value_labels) New function.
1438 (var_lookup_value_label) New function.
1439 (var_get_value_name) Moved here from variable.c, renamed from
1440 var_get_value_name, transposed argument order, updated all
1442 (var_to_string) Moved here, from variable-label.c.
1443 (var_set_leave) New function.
1444 (var_get_leave) New function.
1445 (var_must_leave) New function.
1446 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1447 set_var_short_name_suffix.
1448 (var_get_dict_index) New function.
1449 (var_get_case_index) New function.
1450 (var_get_obs_vals) New function.
1451 (var_set_obs_vals) New function.
1452 (var_has_obs_vals) New function.
1453 (var_get_vardict) New function.
1454 (var_set_vardict) New function.
1455 (var_has_vardict) New function.
1456 (var_clear_vardict) New function.
1457 (value_dup) Moved to value.c.
1458 (compare_values) Ditto.
1461 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1462 (enum ALPHA) Rename VAR_STRING, update all users.
1464 * vector.c: New file.
1465 (struct vector) Moved here, from variable.h.
1466 (check_widths) New function.
1467 (vector_create) New function.
1468 (vector_clone) New function.
1469 (vector_destroy) New function.
1470 (vector_get_name) New function.
1471 (vector_get_var) New function.
1472 (vector_get_var_cnt) New function.
1473 (compare_vector_ptrs_by_name) New function.
1475 * vector.h: New file.
1477 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1479 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1480 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1483 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1485 * value-labels.c (destroy_atoms): New function.
1486 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1488 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1490 Thanks to Jason Stover for pointing out this problem.
1492 * data-out.c (output_number): Use gsl_finite from GSL, which is
1493 portable, instead of isfinite, which is not.
1496 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1498 * variable.c variable.h (value_dup): New function.
1500 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1502 Start converting struct variable to an opaque type. In this
1503 phase, we add a bunch of setter and getter functions and convert
1504 most of the PSPP code to use them. The resulting changes are
1505 pervasive but mostly trivial, and only the notable changes are
1508 * format.c (fmt_equal): New function.
1510 * variable.c (var_type_is_valid): New function.
1511 (measure_is_valid) Moved here, from format.c.
1512 (alignment_is_valid) Moved here, from format.c.
1513 (var_get_name) New function.
1514 (var_set_name) New function.
1515 (width_to_type) New function.
1516 (var_get_type) New function.
1517 (var_get_width) New function.
1518 (var_set_width) New function.
1519 (var_is_numeric) New function.
1520 (var_is_alpha) New function.
1521 (var_is_short_string) New function.
1522 (var_is_long_string) New function.
1523 (var_is_very_long_string) New function.
1524 (var_get_missing_values) New function.
1525 (var_set_missing_values) New function.
1526 (var_clear_missing_values) New function.
1527 (var_has_missing_values) New function.
1528 (var_is_value_missing) New function.
1529 (var_is_num_missing) New function.
1530 (var_is_str_missing) New function.
1531 (var_is_value_user_missing) New function.
1532 (var_is_num_user_missing) New function.
1533 (var_is_str_user_missing) New function.
1534 (var_is_value_system_missing) New function.
1535 (var_get_print_format) New function.
1536 (var_set_print_format) New function.
1537 (var_get_write_format) New function.
1538 (var_set_write_format) New function.
1539 (var_set_both_formats) New function.
1540 (var_get_label) New function.
1541 (var_set_label) New function.
1542 (var_clear_label) New function.
1543 (var_has_label) New function.
1544 (var_get_measure) New function.
1545 (var_set_measure) New function.
1546 (var_get_display_width) New function.
1547 (var_set_display_width) New function.
1548 (var_get_alignment) New function.
1549 (var_set_alignment) New function.
1550 (var_get_value_cnt) New function.
1551 (var_get_leave) New function.
1552 (var_get_short_name) New function.
1554 * variable.h: (struct variable) Removed "type" and "nv" members;
1555 they are now computed from "width" where needed.
1557 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1559 * missing-values.c (mv_resize): Don't write beyond end of the
1560 allocated buffer when resizing a long string.
1562 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1564 Clean up identifier code: don't require identifier enumerations to
1565 be in a particular order; make better use of string library;
1566 expose less of the internals.
1568 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1569 change interface. Updated all callers.
1570 (lex_id_match) Change interface to use struct substring, update
1572 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1573 (global array keywords[]) Make static, change form. Update all
1574 users to use lex_id_name instead.
1575 (lex_is_keyword) New function.
1576 (lex_id_to_token) Change interface to use struct substring, update
1578 (lex_id_name) New function.
1580 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1581 lex_is_keyword instead.
1582 (T_LAST_KEYWORD) Removed.
1583 (T_N_KEYWORDS) Removed.
1585 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1587 * format.c: (fmt_date_template) Distinguish characters for which a
1588 space is output and any date delimiter is allowed on input, from
1589 those for which a space is output and only a space is allowed on
1590 input. The former is represented by X, the latter by a space.
1591 Also, drop distinction between h and H, changing the former to the
1594 * data-in.c: Completely rewrite internals to conform to SPSS input
1595 formats as closely as possible.
1596 (data_in) Changed external interface by replacing the structure
1597 that was used as a single argument by a set of arguments. Updated
1599 (data_in_finite_line) Removed. Converted all callers to use plain
1601 (data_in_get_integer_format) New function.
1602 (data_in_set_integer_format) New function.
1603 (data_in_get_float_format) New function.
1604 (data_in_set_float_format) New function.
1606 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1607 (struct data_in) Removed.
1609 * data-out.c: (output_date) Drop each component from the input as
1610 it is output, to allow us to drop the distinction between h (a
1611 count of hours) and H (the hour of day) template characters.
1612 Also, handle new X template character.
1613 (output_scientific) Follow more rational rule on when to drop
1614 fraction introduced between SPSS 13 and 15. Updated test case to
1617 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1619 Fix buffer overflow reported by John Darrington.
1621 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1622 realize that DIGITS is a count of nibbles, not of bytes.
1624 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1626 * calendar.c (calendar_offset_to_gregorian) Also return the
1627 year-of-day. Change callers to new interface.
1629 * data-out.c: Completely rewrite internals to conform to SPSS
1630 output formats as completely as possible.
1631 (data_out) Change interface to put input parameters before output
1632 parameters, for consistency with the style I now prefer. Update
1634 (data_out_get_integer_format) New public function.
1635 (data_out_set_integer_format) New public function.
1636 (data_out_get_float_format) New public function.
1637 (data_out_set_float_format) New public function.
1639 * data-out.h: New file. Move prototype for data_out here, from
1642 * format.c: (fmt_step_width) Use equality comparison instead of
1643 bitwise and, for clarity.
1644 (fmt_is_string) Ditto.
1645 (fmt_input_to_output) Fix categories that are translated to F
1648 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1650 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1651 casefile-private.h: Added casefilter to assist commands with missing
1654 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1656 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1658 * settings.c: (route_errors_to_terminal) New variable.
1659 (route_errors_to_listing) New variable.
1660 (get_error_routing_to_terminal) New function.
1661 (set_error_routing_to_terminal) New function.
1662 (get_error_routing_to_listing) New function.
1663 (set_error_routing_to_listing) New function.
1665 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1667 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1669 * format.c: Completely rewrite, to achieve better abstraction.
1670 Rewrite all references to formats in other files.
1672 * format.def: Rewrite and reorganize.
1674 * settings.c: Move everything related to custom currency formats
1675 into format.[ch], changing them in form, so as to group related
1676 code and definitions better. Changed all references to use the
1678 (static var decimal) Removed.
1679 (static var grouping) Removed.
1680 (static var cc) Removed.
1681 (get_decimal) Removed.
1682 (set_decimal) Removed.
1683 (get_grouping) Removed.
1684 (set_grouping) Removed.
1688 * settings.h: (macro CC_CNT) Removed.
1689 (macro CC_WIDTH) Removed.
1690 (struct custom_currency) Removed.
1692 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1694 * data-in.c (data_in): Use switch statement instead of table, to
1695 avoid dependence on the order of the FMT_* enums.
1697 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1699 * data-out.c: (num_to_string) Removed, because it was dead code.
1701 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1703 * data-in.c (parse_trailer): Fix error message.
1705 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1707 * format.c (fmt_is_binary): New function.
1709 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1711 * procedure.c procedure.h: Encapsulated the static data into a single
1714 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1716 * casefile.c (casereader_read_xfer): Always initialize the case,
1717 even on an error condition.
1719 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1721 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1722 entire function useless.
1724 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1726 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1727 casereader method casereader_clone.
1729 * procedure.c pransformations.h: Introduced new type casenum_t
1731 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1733 * variable.c: (width_to_bytes) Rephrase code for clarify.
1735 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1737 * format.c: (fmt_type_from_string) New function.
1738 (fmt_to_string) Include decimals in output if the format has
1739 decimals, even if the format type does not. This way, we can
1740 accurately reproduce incorrect formats in user output.
1741 (check_common_specifier) Make the check for a bad format type an
1742 assertion, so we get bug reports if they show up. Fix message.
1743 Check for decimal places with a format type that doesn't allow
1745 (check_input_specifier) Remove check for FMT_X, which has been
1747 (check_output_specifier) Ditto.
1749 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1751 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1752 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1753 (enum fmt_parse_flags) Removed.
1755 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1757 * casefile.c casefile.h: Converted to an abstract base class.
1758 * casefile-private.h fastfile.c fastfile.h: New files.
1759 * automake.mk procedure.c scratch-writer.c storage-stream.c
1761 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1763 * procedure.c (internal_procedure): Create sink_case with only as
1764 many values as the compacted dictionary.
1766 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1768 Remove "debugging" code that caused plenty of false positives and
1771 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1773 * case.c: Remove all references to `this' member.
1775 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1777 Fix link error noted by Jason Stover.
1779 * storage-stream.c: Include <assert.h>.
1781 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1783 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1784 ALL) and additional underlying system file issues.
1786 Thanks to John Darrington for review.
1788 First problem: var_hash points to variables not owned by the
1789 sys-file-reader, which the caller may free or modify. Use an
1790 array of sfm_vars instead, as done earlier (e.g. CVS version
1793 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1794 members and remove all code that references it. Add vars, var_cnt
1795 members. Remove fix_specials member, which was unused.
1796 (struct sfm_var) Remove name member, which was unused.
1797 (sfm_close_reader) Free vars member instead of var_hash.
1798 (compare_var_shortnames) Removed.
1799 (hash_var_shortname) Removed.
1800 (sfm_open_reader) Fill out vars array.
1801 (compare_var_index) Removed.
1802 (sfm_read_case) Use vars instead of var_hash.
1804 Second problem: we're confused about when we actually have very
1805 long strings, causing us to choose incorrectly between slow path
1806 and fast path in sfm_read_case.
1808 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1809 have very long strings, not when we have long variable names,
1810 which is an unrelated feature.
1812 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1814 * variable.h: Move var_set and variable parsing declarations to
1815 new header, src/language/lexer/variable-parser.h. Modified lots
1816 of files to include the new header.
1818 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1820 * value-labels.c (value_to_string): When there's no value label,
1821 format the variable according to its print format, instead of
1822 always effectively using A or F format.
1824 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1826 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1829 * format.c: Removed tortological assertion.
1831 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1833 Reform string library.
1835 * file-name.c (fn_interp_vars): Change interface to take a
1836 substring as input. Updated all users.
1838 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1840 * format.c (measure_is_valid): Really return false when m >=
1843 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1845 Implement random access to casefiles, for use in GUI.
1847 * casefile.c: (struct casereader) Add `random', `file_ofs',
1848 `buffer_ofs' members.
1849 (casefile_get_random_reader) New function.
1850 (read_open_file) Break part into new function
1851 seek_and_fill_buffer().
1852 (fill_buffer) Update buffer_ofs, file_ofs.
1853 (casereader_seek) New function.
1855 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1857 * settings.c: Added call to i18n{done, init}.
1859 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1861 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1864 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1866 * casefile.c: Convert many uses of `int' to `bool'.
1868 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1870 * transformations.c (trns_chain_destroy): Destroy chain's trns
1871 member, to fix memory leak.
1873 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1875 * storage-stream.c (storage_source_decapsulate): Destroy case
1876 source to fix memory leak.
1878 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1880 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1881 case passed as argument instead of initializing the argument as a
1882 case. Fixes memory leak that showed up in
1883 tests/command/aggregate.sh with scratch files.
1885 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1887 * procedure.c (proc_done): Destroy default_dict, to fix memory
1890 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1892 Simplify procedure_with_splits().
1894 * procedure.c (struct split_aux_data): Removed case_count member.
1895 (procedure_with_splits) Don't initialize case_count.
1896 (split_procedure_case_func) Check whether prev_case is null
1897 instead of case_count.
1898 (split_procedure_end_func) Ditto.
1900 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1902 * case.c (case_move): Do nothing if dst and src are the same
1904 (case_try_create) Merge two similar cases.
1905 (case_copy) Unshare only if data must be actually copied.
1907 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1909 * data-in.c data-out.c dictionary.c sys-file-reader.c
1910 sys-file-writer.c variable.c variable.h: Reworked very long string
1911 support for better encapsulation.
1913 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1915 * value-labels.c (val_labs_can_set_width): New function.
1916 (val_labs_set_width) Clear labels if increasing width to long
1918 (val_labs_destroy) Remove unneeded test for null.
1920 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1922 * value-labels.h: Remove unneeded dependency on variable.h.
1924 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1926 Get rid of `char *c' member in union value, for cleanliness.
1928 * value.h: (union value) Remove `c' member.
1930 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1932 Make dictionary compacting functions a little more general.
1934 * sys-file-writer.c (sfm_open_writer): Use
1935 dict_compacting_would_change().
1936 (does_dict_need_translation) Removed.
1938 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1940 Make dictionary compacting functions a little more general.
1942 * dictionary.c (dict_needs_compaction): Rename
1943 dict_compacting_would_shrink(). Update all callers.
1944 (dict_compacting_would_change) New function.
1946 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1948 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1949 inverted return value (!).
1951 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1953 Continue reforming procedure execution.
1955 * procedure.c: Search and replace "vfm" by "proc". Notably:
1956 (static var vfm_source) Rename proc_source. Update all
1958 (static var vfm_sink) Rename proc_sink. Update all references.
1960 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1962 Continue reforming procedure execution. In this phase, remove
1963 PROCESS IF, which was deprecated anyway and can be easily
1964 simulated with TEMPORARY followed by SELECT IF.
1966 * procedure.c: (open_active_file) Don't call
1967 add_process_if_trns().
1968 (discard_variables) Get rid of redundant call to
1969 proc_cancel_all_transformations().
1970 (add_process_if_trns) Removed.
1971 (process_if_trns_proc) Removed.
1972 (process_if_trns_free) Removed.
1974 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1976 Continue reforming procedure execution. In this phase, add
1977 `const' to the case passed to procedure()'s callback.
1979 Updated all users of procedure() as well.
1981 * procedure.c: (struct write_case_data) Add "const" to ccase
1982 parameter for case_func member.
1983 (procedure) Add "const" to ccase parameter for proc_func
1985 (multipass_case_func) Make ccase parameter const.
1986 (internal_procedure) Add "const" to ccase parameter for case_func
1988 (split_procedure_case_func) Make ccase parameter const.
1989 (multipass_split_case_func) Make ccase parameter const.
1991 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
1993 Continue reforming procedure execution. In this phase, get rid of
1994 the output code for SPLIT FILE groups in procedure.c, which really
1995 shouldn't be doing any output. Move it into the individual
1996 procedures instead. This also adds some flexibility.
1998 Updated many users of procedure_with_splits() and
1999 multipass_procedure_with_splits() to call
2000 output_split_file_values() and to deal with increased use of
2003 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2004 parameter to begin_func member.
2005 (procedure_with_splits) Add "const struct ccase *" parameter to
2006 begin_func parameter. Make ccase parameter const in proc_func
2008 (split_procedure_case_func) Don't dump split file group. Pass
2010 (dump_splits) Moved to language/dictionary/split-file.c as
2011 output_split_file_values().
2012 (struct multipass_split_aux_data) Add "const struct ccase *"
2013 parameter to split_func member.
2014 (multipass_procedure_with_splits) Add "const struct ccase *"
2015 parameter to split_func parameter.
2016 (multipass_split_case_func) Save new SPLIT FILE case before
2018 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2020 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2022 Continue reforming procedure execution. Change
2023 internal_procedure() so that it calls open_active_file() and
2024 close_active_file(), which isolates most of the actual procedure
2027 * procedure.c: (struct write_case_data) Rename `proc_func' member
2028 to `case_func' and update all references.
2029 (procedure) Rewrite as one-line wrapper around
2030 internal_procedure().
2031 (struct multipass_aux_data) New.
2032 (multipass_callback) Renamed multipass_case_func(). Use struct
2033 multipass_aux_data as auxiliary data.
2034 (multipass_end_func) New function.
2035 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2036 that uses multipass_case_func, multipass_end_func.
2037 (internal_procedure) Add `end_func' argument. Move optimization
2038 of trivial case in here. Move call to open_active_file() and
2039 close_active_file() in here. Now assert that vfm_source is
2041 (procedure_with_splits_callback) Rename
2042 split_procedure_case_func().
2043 (split_procedure_end_func) New function.
2044 (multipass_split_callback) Rename multipass_split_case_func.
2045 (multipass_split_end_func) New function.
2046 (discard_variables) No need to test for nonnull vfm_source.
2048 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2050 Continue reforming procedure execution. Get rid of unused member.
2052 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2054 (write_case) Don't increment cases_analyzed.
2056 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2058 Continue reforming procedure execution. In this phase, move
2059 procedure.c and procedure.h from src to src/data. Update
2060 makefiles and #includes accordingly.
2062 * procedure.c: Moved here from src/.
2064 * procedure.h: Moved here from src/.
2066 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2068 Continue reforming procedure execution. In this phase, get rid of
2069 many global variables, consolidating procedure execution in
2070 procedure.c. Encapsulate transformations in new "struct
2071 trns_chain". Also, change implementation of N OF CASES, FILTER,
2072 and PROCESS IF from special cases to transformations.
2074 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2077 * dictionary.c: (global variable default_dict) Move to
2080 * variable.h: (TRNS_*) Move to transformations.h.
2081 (struct transformation) Move to transformations.c.
2083 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2085 * sys-file-reader.c: Fixed invalid read problems.
2087 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2089 * storage-stream.c: Add missing function comments.
2091 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2093 Continue reforming procedure execution. In this phase, add some
2094 new, needed functionality to storage-stream.
2096 * storage-stream.c: (storage_source_decapsulate) New function.
2098 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2100 * variable.c (width_to_bytes): Declarations must precede
2101 statements for C90 compliance.
2103 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2105 * data-out.c, data-in.c, variable.c, variable.h: New functions
2106 copy_mangle and copy_demangle for reading/writing cases; emulates the
2107 way SPSS deals with strings > 255 bytes.
2109 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2110 subtype 14 needed for strings longer than 255 bytes.
2112 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2113 instead of literal values. Also fixed some constness issues.
2115 * format.h: Constness
2117 * sfm-private.h: Renamed the case_size identifier, since I discovered
2118 that SPSS's respect for this variable is very nominal.
2120 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2122 Change case limit type from int to size_t.
2124 * dictionary.c: (struct dictionary) Change type of case_limit
2126 (dict_get_case_limit) Change return type.
2127 (dict_set_case_limit) Change parameter type.
2129 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2131 * variable.h: (struct variable) Rename `reinit' member as `leave'
2132 and invert sense. Fix up all references.
2134 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2136 Continue reforming procedure execution. In this phase, break
2137 procedure.c into multiple files.
2139 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2141 * case-sink.c: New file.
2143 * case-sink.h: New file.
2145 * case-source.c: New file.
2147 * case-source.h: New file.
2149 * storage-stream.c: New file.
2151 * storage-stream.h: New file.
2153 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2155 * variable.h: (struct variable) Remove `init' member and all
2156 references to it from other files. It was initialized in several
2157 places, but nothing really ever used it for anything worthwhile.
2158 Thanks to Jason Stover for pointing out how confusing this
2161 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2163 Continue reforming error message support. In this phase, get rid
2164 of message "titles" and put the message text in `struct error'.
2165 Now `struct error' encapsulates a message more properly.
2167 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2168 Format message ourselves.
2170 * data-in.c: (vdls_error) Ditto.
2172 * por-file-reader.c: (error) Ditto.
2174 * sys-file-reader.c: (corrupt_msg) Ditto.
2176 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2178 GNU standards require "file name" instead of "filename" in
2179 documentation. It's nice for our code to follow the convention
2182 * casefile.c: (struct casefile) Rename `filename' member to
2183 `file_name'. Updated all references.
2185 * file-name.c: [!unix] (struct file_identity) Rename
2186 normalized_filename member to normalized_file_name. Updated all
2189 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2191 We don't really support anything but Unix-like environments well,
2192 so we might as well de-obfuscate by writing directory and path
2193 separators explicitly.
2195 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2197 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2198 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2200 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2203 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2205 GNU standards require "file name" instead of "filename" in
2206 documentation. It's nice for our code to follow the convention
2209 * filename.c: Rename to file-name.c.
2211 * filename.h: Rename to file-name.h. Update all inclusions.
2212 Update header guards.
2214 * automake.mk: Update file names.
2216 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2218 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2220 (fn_basename) Removed (dead code).
2221 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2222 (fn_special_p) Renamed fn_is_special(), all references updated.
2223 (fn_exists_p) Renamed fn_exists(), all references updated.
2225 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2227 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2228 Also, now it only tilde-expands file names, not paths.
2229 (fn_search_path) Tilde-expand one directory at a time.
2231 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2233 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2234 get rid of non-Unixlike version of the code, which has probably
2236 (fn_prepend_dir) Removed (dead code).
2238 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2239 (macro PATH_SEPARATOR_STRING) New.
2240 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2242 Continue reforming error message support. In this phase, we get
2243 rid of VM() and the other msg() support for "verbosity", replacing
2244 it by a new function verbose_msg().
2246 * filename.c: (fn_search_path) Use verbose_msg() instead of
2249 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2251 * sfm-private.h: Get rid of #defines after #error, which makes no
2254 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2256 Get rid of our own int32 type in favor of the standard int32_t
2259 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2262 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2264 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2266 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2268 Remove ill-considered file routines that are no longer used.
2270 * filename.c: (fn_open_ext) Removed.
2271 (fn_close_ext) Removed.
2273 * filename.h: (struct file_ext) Removed.
2275 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2277 * variable.c (var_is_valid_name): Move declarations before code
2280 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2282 * filename.ch (fn_interp_vars): Fixed small buglet.
2284 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2286 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2287 so as to modify the string inline. Thus makeing it easier to
2288 destroy the results when no longer needed.
2290 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2292 * category.c (cat_stored_values_destroy): Fixed memory leak.
2294 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2296 Add some missing frees. Thanks to John Darrington for reporting
2299 * any-writer.c (any_writer_close): Free writer.
2301 * any-reader.c (any_reader_close): Free reader.
2303 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2305 * por-file-reader.c: (error) Mark as NO_RETURN.
2307 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2309 * settings.c: Changed default value of scompress to true.
2311 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2313 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2315 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2318 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2319 here from pref.h.orig.
2321 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2323 * sys-file-reader.c: Fixed bug reading compressed files.
2325 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2327 * Numerous renames. See src/ChangeLog for details.
2329 * Moved files from src directory