1 2008-02-02 John Darrington <john@darrington.wattle.id.au>
3 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
4 for reviewing this code.
6 2008-02-02 Ben Pfaff <blp@gnu.org>
10 * sys-file-reader.c (read_variable_record): Allow missing values
11 to be specified on long string variables, but warn about them
12 because PSPP does not yet support them.
13 (read_extension_record): Ignore extension records 20 and 21, which
14 PSPP does not yet support.
15 (read_header): Fix error message when floating-point format cannot
18 2008-02-01 Ben Pfaff <blp@gnu.org>
20 Patch #6386. Thanks to John Darrington for review and for the
21 updates to gnumeric-reader.c.
23 * dictionary.c (make_hinted_name): New function.
24 (make_numeric_name): New function.
25 (dict_make_unique_var_name): New function.
27 * gnumeric-reader.c (devise_name): Removed.
28 (munge_name): Removed.
29 (gnumeric_open_reader): Use new function
30 dict_make_unique_var_name.
32 * short-names.c (set_var_short_name_suffix): Use new function
35 2008-01-19 John Darrington <john@darrington.wattle.id.au>
37 * settings.c settings.h: Moved static variables into a
38 single struct. Renamed functions to have a settings_ prefix.
40 2008-01-14 John Darrington <john@darrington.wattle.id.au>
42 * data-out.c (output_AHEX): Corrected number of bytes in
43 call to output_hex. Closes bug #22011
45 2008-01-02 John Darrington <john@darrington.wattle.id.au>
47 * variable.c variable.h: Replaced var_get_value_name with
48 var_append_value_name which doesn't use any static data.
49 Thanks to Ben for review.
51 2007-12-07 Ben Pfaff <blp@gnu.org>
55 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
56 made from an empty casereader produce a casegrouper with no
57 groups, instead of one with one group that has no cases.
59 * casereader.c (casereader_is_empty): New function.
61 2007-12-06 Ben Pfaff <blp@gnu.org>
65 * sys-file-reader.c (read_display_parameters): Handle variable
66 display parameters record with only 2 data items per variable.
67 Reported by Guido Gay <gay@irer.it>.
69 2007-12-04 Ben Pfaff <blp@gnu.org>
71 * identifier.c (lex_id_match_n): New function.
72 (lex_id_match): Reimplement in terms of lex_id_match_n.
74 2007-11-24 Ben Pfaff <blp@gnu.org>
76 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
79 2007-11-24 Ben Pfaff <blp@gnu.org>
81 Fix warning reported when reading back system files that include
82 very long string variables. Thanks to Guido Gay <gay@irer.it> for
85 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
86 was written where `j' was meant.
88 2007-11-18 Ben Pfaff <blp@gnu.org>
90 Properly write variables that include a range of missing values to
91 system files. Thanks to Guido Gay <gay@irer.it> for reporting
94 * sys-file-writer.c (write_variable): Correctly calculate
95 n_missing_values field when writing variables that include a range
98 2007-11-10 Ben Pfaff <blp@gnu.org>
100 Cleanups and bug fixes devised while writing up documentation.
103 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
105 * dict-class.c: New file.
106 (dict_class_from_id): Move here.
107 (dict_class_to_name): Move here.
109 * dict-class.h: New file.
110 (enum dict_class): Move here. Change from consecutive integers to
111 consecutive bits, to make testing for any of multiple values
112 easier. Add new DC_ALL constant.
114 * dictionary.c (struct dictionary): Change `case_limit' from
115 size_t to casenumber.
116 (dict_get_vars): Make final argument an enum dict_class.
117 (dict_get_vars_mutable): Ditto.
118 (dict_get_case_limit): Change return value to casenumber.
119 (dict_set_case_limit): Change final argument to a casenumber.
120 (dict_unset_split_var): Add assertion. Rephrase slightly.
121 (dict_set_label): Use xstrndup to simplify.
123 * format.c (fmt_step_width): AHEX format also needs 2-byte
125 (fmt_set_style): Simplify assertions.
127 * missing-values.c (mv_add_num_range): Rename mv_add_range.
128 Simplify implementation.
129 (mv_has_value): Simplify implementation.
130 (mv_pop_value): Remove the first value, not the last, to avoid
131 having GET followed by SAVE reverse the order of missing values.
132 (mv_peek_value): Rename mv_get_value. Simplify assertion.
133 (mv_has_range): Simplify implementation.
134 (mv_peek_range): Rename mv_get_range.
135 (can_resize_string): Removed.
136 (mv_is_resizable): Use value_is_resizable.
137 (mv_resize): Use value_resize.
139 * short-names.h (SHORT_NAME_LEN): Move here.
141 * val-type.h: New file, for definitions related to type and width
142 of abstract values. Before, these definitions were mixed among
143 those related to "union value" and those related to variables.
144 (macro SYSMIS): Move here.
145 (macro LOWEST): Move here.
146 (macro HIGHEST): Move here.
147 (macro MAX_STRING); Move here.
148 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
149 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
150 All references updated.
151 (val_type_is_valid): New function. Replaces var_type_is_valid.
152 All references updated.
153 (val_type_from_width): New function. Replaces
154 var_type_from_width. All references updated.
156 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
158 (val_labs_can_set_width): Use value_is_resizable.
159 (val_labs_add): Simply return false if the value labels set is too
160 wide, instead of having undefined behavior.
161 (val_labs_replace): Ditto.
162 (val_labs_replace): Ditto.
163 (val_labs_first): Set iterator to null if iteration is complete.
164 (val_labs_first_sorted): Ditto.
165 (val_labs_done): Become a no-op if the iterator is null.
167 * value.c (value_is_resizable): New function.
168 (value_resize): New function.
170 * variable.c (var_get_dict_class): New function.
172 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
175 2007-11-08 Ben Pfaff <blp@gnu.org>
177 * data-in.c: Make formatted data parsing locale-independent.
178 (parse_number): Use c_strtod instead of strtod, to avoid
179 locale-specific behavior.
182 2007-11-06 Ben Pfaff <blp@gnu.org>
184 Patch #6256: add support for binary, 360 file formats. Reviewed
187 * data-in.c (struct data_in): Add `encoding' member.
188 (data_in): Add `encoding' parameter, and re-encode the data passed
189 in where appropriate. Update all callers to pass it in.
190 (parse_A): Implement EBCDIC recoding wart described in manual.
191 (parse_AHEX): Implement EBCDIC recoding.
193 * data-out.c (data_out_legacy): New function.
194 (data_out): Make into a wrapper around data_out_legacy.
196 * file-handle-def.c (struct file_handle): New member `encoding'.
197 (fh_create_file): Set encoding.
198 (fh_default_properties): Set default encoding.
199 (fh_get_legacy_encoding): New function.
201 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
202 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
203 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
204 (struct fh_properties): New member `encoding'.
206 2007-11-05 Ben Pfaff <blp@gnu.org>
208 Patch #6258. Reviewed by John Darrington.
210 * file-handle-def.c (fh_lock): Add comment that TYPE should be
211 marked with N_() in the caller. Added these markings to each
212 caller too. Should make i18n easier.
213 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
215 2007-11-03 Ben Pfaff <blp@gnu.org>
217 Allow output files to overwrite input files (bug #21280).
219 * file-handle-def.c: Separate locking of files for input and for
220 output, to allow both to take place at once. Also, distinguish a
221 file handle from the identity of the underlying file, because the
222 identity of a file changes over time and the file handle can't
223 represent two different identities.
224 (struct file_handle): Remove `next', `open_cnt', `deleted',
225 `type', `open_mode', `aux', `identity' members. Change `id' from
226 char array to char *. Add `ref_cnt' member.
227 (file_handle_from_ll) New function.
228 (file_handles) Removed.
229 (named_handles) New variable.
230 (fh_init) Initialize named_handles.
231 (fh_done) Remove name from all named_handles.
232 (free_handle) Updated for modified struct file_handle.
233 (unname_handle) New function.
234 (fh_ref) New function.
235 (fh_from_file_name) Removed.
236 (fh_unref) New function.
237 (fh_unname) New function.
238 (fh_from_id) Rewritten.
239 (create_handle) Updated for modified struct file_handle.
240 (fh_inline_file) Increment the handle's ref count.
241 (fh_create_file) Updated for modified struct file_handle.
242 (fh_create_scratch) Ditto.
247 (fh_is_open) Removed.
248 (fh_get_id) Updated for modified struct file_handle.
249 (fh_get_default_handle) Increment the handle's ref count.
250 (fh_set_default_handle) Handle ref counts.
251 (struct fh_lock) New structure.
252 (locks) New static var.
253 (fh_lock) New function.
254 (fh_unlock) New function.
255 (fh_lock_get_aux) New function.
256 (fh_lock_set_aux) New function.
257 (fh_is_locked) New function.
258 (make_key) New function.
259 (free_key) New function.
260 (compare_fh_locks) New function.
261 (hash_fh_lock) New function.
263 * file-handle-def.h (enum fh_access) New enum.
265 * file-name.c: Made file_identity the same in all supported
267 (struct file_identity): New `name' member.
268 (fn_get_identity): For a file that doesn't exist, get the
269 dev/inode of its directory plus its name. If even the directory
270 doesn't exist, just use its name. Merge the Windows
271 implementation into the Unix one.
272 (fn_compare_file_identities): Rewritten. Merge the Windows
273 implementation into the Unix one.
274 (fn_hash_identity): New function.
276 * make-file.c (struct replace_file): New structure.
277 (all_files): New static var.
278 (replace_file_start): New function.
279 (replace_file_commit): New function.
280 (replace_file_abort): New function.
281 (free_replace_file): New function.
282 (unlink_replace_files): New function.
284 * por-file-reader.c (struct pfm_reader): Add `lock' member.
285 (close_reader): Unlock file.
286 (pfm_open_reader): Lock file.
288 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
290 (pfm_open_writer): Lock file and prepare for its replacement.
291 (close_writer): Unlock file.
293 * scratch-handle.h (struct scratch_handle): Add unique_id so that
294 different generations of a scratch file can be distinguished.
296 * scratch-reader.c (scratch_reader_open): Verify that the file is
299 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
300 remove scratch_handle member.
301 (scratch_writer_open): Lock handle. Prepare to replace handle
302 data, instead of doing it immediately.
303 (scratch_writer_casewriter_destroy): Replace handle data and
306 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
307 (sfm_open_reader): Lock file.
308 (close_reader): Unlock file.
310 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
312 (sfm_open_writer): Lock file and prepare for its replacement.
313 (close_writer): Unlock file.
315 2007-11-02 Ben Pfaff <blp@gnu.org>
317 * data-out.c (output_number): Use isfinite (from C99) instead of
318 gsl_isfinite, since we now have universal support for it in
320 (output_infinite): Ditto.
321 [!HAVE_ROUND] (round): Remove definition, since we now have a
322 replacement in gnulib.
324 2007-10-27 John Darrington <john@darrington.wattle.id.au>
326 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
327 wasn't initialised correctly. Closes bugs #21330 and 21397
330 2007-10-08 Ben Pfaff <blp@gnu.org>
332 Fix bug #21280. Thanks to John Darrington for review.
334 * file-name.c (create_stream): New function.
336 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
337 file handle before creating the file, to ensure that we don't
338 truncate a file that we're reading. Make code easier to read by
341 * sys-file-write.c (sfm_open_writer): Ditto.
343 2007-10-01 Ben Pfaff <blp@gnu.org>
345 Fix bug #21192. Thanks to John Darrington for review.
347 * casereader.c (casereader_read): Decrement case_cnt before
348 calling the casereader's "read" member function, so that we
349 interact properly with lazy_casereader.
351 * datasheet.c: Add regression test for above bug fix.
352 (clone_datasheet): New function.
353 (lazy_callback): New function.
354 (check_datasheet_casereader): New function.
355 (check_datasheet): Check datasheet contents are reported correctly
356 through an ordinary casereader and a lazy casereader.
357 (clone_model): Use clone_datasheet.
359 2007-09-24 Ben Pfaff <blp@gnu.org>
361 Patch #6210. Reviewed by John Darrington.
362 * settings.c: Drop "long view", which is not needed any longer.
363 (static var long_view): Removed.
364 (force_long_view): Removed.
365 (get_viewwidth): Removed.
366 (init_viewport): Removed long_view reference.
368 2007-09-19 John Darrington <john@darrington.wattle.id.au>
370 * settings.c settings.h: Changed viewport's length and width to be
371 owned by the user interface which uses the data library. This allows
372 better abstraction, and makes dynamically adjustable dimensions easier.
374 2007-09-18 Ben Pfaff <blp@gnu.org>
376 * procedure.c (proc_extract_active_file_data): New function.
378 * lazy-casereader.h: New file.
380 * lazy-casereader.c: New file.
382 * casereader.c (casereader_dynamic_cast): New function.
384 2007-09-14 Ben Pfaff <blp@gnu.org>
386 * dictionary.c (dict_clone): Copy case indexes from cloned
387 dictionary. Fixes bug #21061. Reviewed by John Darrington.
389 2007-09-13 John Darrington <john@darrington.wattle.id.au>
391 * value.c value.h (create_value): New function.
393 2007-09-12 Ben Pfaff <blp@gnu.org>
395 Make it clear that translator casereader and casewriter translate
396 functions are supposed to destroy their input case.
398 * casereader-translator.c (struct casereader_translator): Change
399 input case parameter of translate member function from const
400 struct ccase * to struct ccase *.
401 (casereader_create_translator): Ditto, for translate parameter.
403 * casewriter-translator.c (struct casewriter_translator): Ditto.
404 (casewriter_create_translator): Ditto.
406 2007-08-27 John Darrington <john@darrington.wattle.id.au>
408 * sys-file-reader.c (read_display_parameters): Force display width
409 to 8, if the sys file says 0 (like SPSS does).
411 2007-08-12 Ben Pfaff <blp@gnu.org>
413 * dictionary.c (dict_dump): New function.
415 2007-08-12 Ben Pfaff <blp@gnu.org>
417 Drop dict_compactor in favor of using the new struct case_map.
419 * dictionary.c (struct copy_map): Removed.
420 (struct dict_compactor): Removed.
421 (dict_make_compactor): Removed.
422 (dict_compactor_compact): Removed.
423 (dict_compactor_destroy): Removed.
425 * procedure.c (struct data_set): Change `compactor' member to be a
427 (proc_open): Use case_map_to_compact_dict instead of
429 (proc_casereader_read): Use case_map_execute instead of
430 dict_compactor_compact.
431 (proc_commit): Use case_map_destroy instead of
432 dict_compactor_destroy.
434 * scratch-writer.c (struct scratch_writer): Change `compactor'
435 member to be a struct case_map *.
436 (scratch_writer_open): Use case_map_to_compact_dict instead of
438 (scratch_writer_casewriter_write): Use case_map_execute instead of
439 dict_compactor_compact.
441 2007-08-12 Ben Pfaff <blp@gnu.org>
443 * automake.mk: Add case-map.c, case-map.h.
445 * case-map.c: New file.
447 * case-map.h: New file.
449 2007-08-12 Ben Pfaff <blp@gnu.org>
451 * dictionary.c (dict_compact_values): Don't delete scratch
452 variables as well as compacting case indexes. Update all callers.
453 (dict_get_compacted_value_cnt): Rename dict_count_values and
454 change interface. Update all callers.
455 (dict_get_compacted_value_cnt): Remove.
456 (dict_compacting_would_shrink): Remove.
457 (dict_compacting_would_change): Remove.
458 (dict_make_compactor): Add new parameter. Update all callers.
460 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
461 instead of dict_count_values, changing an O(N) operation into
464 2007-08-12 Ben Pfaff <blp@gnu.org>
466 * casereader.c (casereader_read): Don't require cases read by a
467 casereader to be exactly the expected size: as long as they're big
470 2007-08-12 Ben Pfaff <blp@gnu.org>
472 Make casewriters keep track of the number of `union value's in
473 each case. This is useful for two reasons: casewriter_write can
474 then check that the case being written is large enough, and later
475 recipients of the casewriter can determine the size of the case.
477 * casewriter-translator.c (casewriter_create_translator): Add
480 * casewriter.c (struct casewriter): Add value_cnt member.
481 (casewriter_write): Check that the case passed in is big enough.
482 (casewriter_get_value_cnt): New function.
483 (casewriter_create): Add value_cnt parameter.
485 2007-08-09 Ben Pfaff <blp@gnu.org>
487 Fix bug reported by Jason Stover.
488 * settings.c: Move get_termcap_viewport and in particular the
489 #include for <curses.h> to the end of the file. curses.h
490 redefines bool on some systems (e.g. OpenBSD), which causes
491 disagreement between uses of bool before its inclusion and
493 Tested by Jason Stover.
495 2007-07-29 Ben Pfaff <blp@gnu.org>
497 Provisional fix for bug #18692 and bug #20161. Reviewed by John
500 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
501 argument (never "rb" or "wb") because SUSv3 says that only those
502 modes are defined, and glibc in fact rejects other modes.
504 Open portable files with fn_open so that they can be read from
505 pipes. Fix missing fh_close call to go along with fh_open.
506 Report an error if the file close reports an error.
507 * por-file-reader.c (close_reader): New function.
508 (por_file_casereader_destroy): Use close_reader.
509 (pfm_open_reader): Open file with fn_open.
511 2007-07-28 Ben Pfaff <blp@gnu.org>
513 Make PSPP able to read all the portable files I could find on the
514 web. Thanks to John Darrington for review. Bug #17620.
515 * por-file-reader.c (struct pfm_reader): New member `line_length'.
516 (error): Print file offset in hexadecimal.
517 (warning): New function.
518 (advance): Treat lines less than 80 bytes long as padded to 80
520 (pfm_open_reader): Call read_documents if we find an "E" record.
521 (convert_format): Convert invalid formats to the default format
522 instead of aborting reading the file.
523 (read_variables): Rename duplicate variable names instead of
524 aborting reading the file.
525 (read_value_label): Allow string variables of different widths to
526 be assigned value labels in the same record. Replace duplicate
527 value labels instead of aborting.
528 (read_documents): New function.
530 * por-file-writer.c (pfm_open_writer): Call write_documents if the
531 dictionary has documents.
532 (write_documents): New function.
534 2007-07-25 Ben Pfaff <blp@gnu.org>
536 Fix bugs related to bug #17213.
538 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
539 since the former is what config.h has. Include the needed ncurses
541 (static var echo) Rename to `do_echo' because the original name is
542 the same as an ncurses identifier.
543 (get_termcap_viewport) Use error instead of msg.
545 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
546 (fn_close): Don't close stdin, stdout, stderr.
548 2007-07-26 John Darrington <john@darrington.wattle.id.au>
550 * procedure.c procedure.h: Added callbacks which get invoked whenever
551 a dataset's transformation chain changes.
553 2007-07-24 Ben Pfaff <blp@gnu.org>
556 * sys-file-writer.c (write_variable_display_parameters): Use new
557 var_default_display_width function to choose display width of
558 segments after the first one in a given variable.
559 * variable.c (var_create): Use var_default_display_width to pick
560 new variable's display width.
561 (var_default_display_width): New function.
562 Reviewed by John Darrington.
564 2007-07-24 Ben Pfaff <blp@gnu.org>
567 * por-file-writer.c (write_variables): Write weight variable.
568 Reviewed by John Darrington.
570 2007-07-23 Ben Pfaff <blp@gnu.org>
572 Improvements to system file reader and writer.
574 First, move all detailed knowledge of very long strings into
575 sys-file-private.[ch], so that this nasty stuff can be isolated.
577 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
578 (EFFECTIVE_VLS_CHUNK): New macro.
579 (min_int): New function.
580 (max_int): New function.
581 (sfm_width_to_bytes): Rewrite.
582 (sfm_width_to_octs): New function.
583 (sfm_segment_alloc_width): New function.
584 (sfm_segment_alloc_bytes): New function.
585 (sfm_segment_used_bytes): New function.
586 (sfm_segment_offset): New function.
587 (sfm_segment_effective_offset): New function.
588 (sfm_dictionary_to_sfm_vars): New function.
590 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
591 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
592 (struct sfm_var): New structure.
594 Next, improvements to the system file reader.
596 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
597 type casenumber. Added `version_major', `version_minor',
600 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
601 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
602 Change `case_cnt' to type casenumber. Removed `has_vls'.
603 (struct sfm_var): Removed.
604 (sfm_open_reader): Don't warn on wrong case size if the file was
605 written by SPSS 13, which tends to get it wrong. Use
606 sfm_dictionary_to_sfm_vars.
607 (read_header): Always output system file info.
608 (read_variable_record): Simplify code for reading missing values.
609 (read_machine_int32_info): Save version numbers from system file
610 into info struct passed as new argument.
611 (read_long_string_map): Restructured to use new sys-file-private
613 (read_value_labels): Use size_overflow_p.
614 (sys_file_casereader_read): Get rid of distinction between fast
615 and slow paths. Use information provided by sys-file-primate's
616 struct sfm_var to simplify code.
617 (skip_whole_strings): New function.
618 (read_int32): Renamed read_int. Changed return value to int.
620 (read_flt64): Renamed read_float. Changed return value to
621 double. Updated all callers.
622 (int32_to_native): Removed. Changed callers to use
624 (flt64_to_double): Removed. Changed callers to use float_convert.
626 Finally, get rid of int32, flt64 terminology and types in system
627 file writer. The former wasn't very useful since a POSIX "int"
628 can hold the whole range of int32 and we generally didn't have a
629 need for it to be exactly-32-bits, just at-least-32-bits. The
630 latter was inconvenient because we had to assume that it could be
631 different from double and thereby convert special values SYSMIS,
632 HIGHEST, LOWEST to and from it in multiple places. Instead, now
633 we just use "int" and "double" in most places, and do conversions,
634 if necessary, very close to where we do I/O. This change meant
635 that the writer code couldn't represent records in the file as C
636 structs any longer, but that's no great loss. The code actually
637 seems to be more readable without them.
639 Simplify the compression buffering code: only buffer as much as
640 necessary, which is no more than eight 8-byte units at any given
643 * sys-file-writer.c (typedef flt64): Removed.
644 (macro second_lowest_flt64): Removed.
645 (struct sysfile_header): Removed.
646 (struct sysfile_variable): Removed.
647 (struct sfm_writer): Removed `needs_translation', `has_vls',
648 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
649 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
650 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
651 compression buffering by `opcodes', `opcode_cnt', `data',
652 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
653 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
654 data writer functions instead of structures.
655 (calc_oct_idx): New function.
656 (write_header): Use simple data writer functions instead of
658 (write_format_spec): Renamed write_format. New argument.
659 (write_variable_continuation_records): New function.
660 (write_variable): Use simple data writer functions instead of
661 structures. Use write_variable_continuation_records. Write
662 entire very long string instead of requiring caller to understand
664 (write_value_labels): Use simple data writer functions instead of
666 (write_documents): Ditto.
667 (write_variable_display_parameters): Use sys-file-private
668 functions to simplify. Use simple data writer functions instead
670 (write_vls_length_table): Use simple data writer functions instead
672 (write_longvar_table): Ditto.
673 (write_rec_7_34): Break into new functions
674 write_integer_info_record, write_float_info_record. Use simple
675 data writer functions instead of structures.
676 (buf_write): Removed.
677 (append_string_max): Removed.
678 (ensure_buf_space): Removed.
679 (sys_file_casewriter_write): Get rid of the distinction between
680 fast and slow paths, which didn't seem to be too useful. Use new
681 functions write_case_uncompressed, write_case_compressed.
682 (put_instruction): Removed.
683 (put_element): Removed.
684 (write_compressed_data): Removed.
685 (close_writer): Use flush_compressed. Only write case count to
686 system file if it will fit in the field.
687 (write_case_compressed): New function.
688 (write_case_uncompressed): New function.
689 (flush_compressed): New function.
690 (put_cmp_opcode): New function.
691 (put_cmp_number): New function.
692 (write_int): New function.
693 (convert_double_to_output_format): New function.
694 (write_float): New function.
695 (write_value): New function.
696 (write_string): New function.
697 (write_bytes): New function.
698 (write_zeros): New function.
699 (write_spaces): New function.
701 Reviewed by John Darrington.
703 2007-07-22 Ben Pfaff <blp@gnu.org>
705 Don't try to write very long strings to portable files. The
706 format does not support it.
708 * por-file-writer.c (MAX_POR_WIDTH): New macro.
709 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
710 (write_format): Add arg to take width to resize format to.
711 (write_value): Limit width of value written to MAX_POR_WIDTH.
712 (write_variables): Limit width of variable and its output formats
714 Reviewed by John Darrington.
716 2007-07-22 Ben Pfaff <blp@gnu.org>
718 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
719 local variable instead of literal 5.
720 Reviewed by John Darrington.
722 2007-07-22 Ben Pfaff <blp@gnu.org>
724 Fix problems with uniqueness of short names in system files with
725 very long string variables. Now a variable may have multiple
728 * automake.mk (src_data_libdata_a_SOURCES): Add new files
729 short-names.c, short-names.h.
731 * dictionary.c (dict_clone): Clone all the short names.
732 (compare_strings): Move into short-names.c.
733 (hash_strings): Ditto.
734 (set_var_short_name_suffix): Ditto.
735 (dict_assign_short_names): Ditto, rename short_names_assign,
736 change to assign all short names.
738 * por-file-writer.c (write_variables): Use short_names_assign
739 instead of dict_assign_short_names.
741 * short-names.c: New file.
743 * short-names.h: New file.
745 * sys-file-private.c (sfm_width_to_segments): New function.
747 * sys-file-reader.c (read_long_var_name_map): Save and restore all
748 the short names, not just the first one.
750 * sys-file-writer.c (cont_var_name): Removed.
751 (sfm_open_writer): Use short_names_assign instead of
752 dict_assign_short_names. Use unique short names assigned by
753 short_names_assign instead of those generated by cont_var_name.
755 * variable.c (struct variable): Remove `short_name' member,
756 replace by `short_names' and `short_name_cnt'.
757 (var_create) Initialize new members.
758 (var_get_short_name_cnt): New function.
759 (var_get_short_name): Now takes an index argument. Changed most
761 (var_set_short_name): Ditto.
762 (var_clear_short_name): Renamed var_clear_short_names, changed to
763 clear all short names.
765 Reviewed by John Darrington.
767 2007-07-22 Ben Pfaff <blp@gnu.org>
769 * variable.c (var_set_width): Use new var_set_width function.
771 * missing-values.c (mv_n_values): Drop assertion, which was not
774 * format.c (fmt_default_for_width): New function.
775 (fmt_resize): New function.
777 Reviewed by John Darrington.
779 2007-07-18 John Darrington <john@darrington.wattle.id.au>
781 * datasheet.c (datasheet_delete_columns): Added assertion to check
782 we're not deleting outside the range of the sheet.
785 * dictionary.c dictionary.h variable.c: Added the ability for string
786 variables to be resized.
788 * vardict.h: Added some prototypes (moved from dictionary.h) as
789 these should only be called by variable.c
792 2007-07-14 John Darrington <john@darrington.wattle.id.au>
794 * sfm-reader.c: Respect case_cnt field in file header.
796 2007-07-01 John Darrington <john@darrington.wattle.id.au>
798 * transformation.c transformation.h (trns_chain_execute): Changed the
799 signature (Patch #6057)
801 2007-06-10 Ben Pfaff <blp@gnu.org>
803 * casereader-filter.c (casereader_filter_destroy): Make sure to
804 write all the remaining excluded cases to the casewriter, if any.
806 * caseinit.c (init_list_destroy): Rewrite.
807 (init_list_clear): Ditto.
809 * casegrouper.c (casegrouper_get_next_group): Always set *reader
810 to null when returning false.
812 2007-06-06 Ben Pfaff <blp@gnu.org>
814 Actually implement the new procedure code and adapt all of its
815 clients to match. Also adapt all of the other case sources and
816 sinks in the tree and their clients to use the
817 casereader/casewriter infrastructure.
819 * automake.mk: Add and remove files.
821 * any-reader.c: Change into a casereader.
822 * por-file-reader.c: Ditto.
823 * scratch-reader.c: Ditto.
824 * sys-file-reader.c: Ditto.
826 * any-writer.c: Change into a casewriter.
827 * por-file-writer.c: Ditto.
828 * scratch-writer.c: Ditto.
829 * sys-file-writer.c: Ditto.
831 * procedure.c: Change to use casereader, casewriter, caseinit, and
832 other new infrastructure.
834 * scratch-handle.c: Adapt to new infrastructure.
836 * case-sink.c: Removed, now dead code.
837 * case-sink.h: Ditto.
838 * case-source.c: Ditto.
839 * case-source.h: Ditto.
840 * casefile-factory.c: Ditto.
841 * casefile-private.h: Ditto.
844 * casefilter.c: Ditto.
845 * casefilter.h: Ditto.
848 * fastfile-factory.c: Ditto.
849 * fastfile-factory.h: Ditto.
850 * storage-stream.c: Ditto.
851 * storage-stream.h: Ditto.
853 2007-06-06 Ben Pfaff <blp@gnu.org>
857 * automake.mk: Add new files.
859 * datasheet.c: New file.
861 * datasheet.h: New file.
863 2007-06-06 Ben Pfaff <blp@gnu.org>
865 Until now, the procedure code has provided a case to the
866 case_source, which has filled in the data values that come from
867 the active file. "Left" data values that don't come from the
868 active file naturally stay the same from case to case, because the
869 procedure code keeps using that same case.
871 One of the compromises that comes with the new procedure code is
872 that the active file allocates and provides its own case, which
873 the procedure code then has to resize to provide room for any
874 other variables that should go in the case and then fill in the
875 values of "left" variables. Then, when we're done with that case,
876 we have to save the values of "left" variables to copy into the
877 next case read from the active file.
879 The caseinit code helps with this.
881 * automake.mk: Add new files.
883 * caseinit.c: New file.
885 * caseinit.h: New file.
887 2007-06-06 Ben Pfaff <blp@gnu.org>
889 * value.h (value_cnt_from_width): New function.
891 * variable.c (var_get_value_cnt): Use new function.
893 2007-06-06 Ben Pfaff <blp@gnu.org>
895 Add casegrouper, to allow cases read from a given casereader to be
896 broken into groups, each of which has its own casereader.
897 Generally cases are grouped based on having equal values for some
900 * automake.mk: Add new files.
902 * casegrouper.c: New file.
904 * casegrouper.h: New file.
906 2007-06-06 Ben Pfaff <blp@gnu.org>
908 Add interface to lexicographical ordering of cases.
910 * automake.mk: Add new files.
912 * case-ordering.c: New file.
914 * case-ordering.h: New file.
916 2007-06-06 Ben Pfaff <blp@gnu.org>
918 Add casereaders and casewriters, the basis of the new data processing
919 implementation. A casereader is a uniform interface to reading cases
920 from a data source; a casewriter is a uniform interface to writing
921 cases to a data sink.
923 * automake.mk: Add new files.
925 * casereader-filter.c: New file.
927 * casereader-provider.h: New file.
929 * casereader-translator.c: New file.
931 * casereader.c: New file.
933 * casereader.h: New file.
935 * casewriter-provider.h: New file.
937 * casewriter-translator.c: New file.
939 * casewriter.c: New file.
941 * casewriter.h: New file.
943 2007-06-06 Ben Pfaff <blp@gnu.org>
945 "casewindow" data structure that extends the deque (from libpspp)
946 of cases with the ability to dump cases to disk if we get too many
949 * automake.mk: Add new files.
951 * casewindow.c: New file.
953 * casewindow.h: New file.
955 2007-06-06 Ben Pfaff <blp@gnu.org>
957 sparse_cases data structure that augments a sparse_array of cases
958 with the ability to dump cases to disk if we get too many cases in
961 * automake.mk: Add new files.
963 * sparse-cases.c: New file.
965 * sparse-cases.h: New file.
967 2007-06-06 Ben Pfaff <blp@gnu.org>
969 Adds a low-level on-disk case array data structure.
971 * automake.mk: Add new files.
973 * case-tmpfile.c: New file.
975 * case-tmpfile.h: New file.
977 2007-06-06 Ben Pfaff <blp@gnu.org>
979 In a couple of places we calculate the maximum number of cases to
980 keep in memory based on the user-defined workspace. Enable
981 centralizing the calculation through a new function.
983 * settings.c (get_workspace_cases): New function.
985 2007-06-06 Ben Pfaff <blp@gnu.org>
987 The casenumber type is defined in transformations.h, but case.h is
988 a more sensible place. Move it.
990 * case.h (CASENUMBER_MAX): New macro.
991 (typedef casenumber): Move here, from transformations.h.
993 2007-06-03 Ben Pfaff <blp@gnu.org>
995 Slightly generalize case_to_values and case_from_values functions.
997 * case.c (case_to_values): Rename case_copy_out, change interface.
998 (case_from_values): Rename case_copy_in, change interface.
1000 * fastfile.c (fastfilereader_get_next_case): Update caller.
1001 (write_case_to_disk): Ditto.
1003 2007-06-02 Ben Pfaff <blp@gnu.org>
1005 Clean up after a forgotten part of patch #5829.
1007 * casedeque.h: Remove unused file.
1009 * automake.mk: Remove casedeque.h from sources.
1011 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1013 * category.c: Removed redundant #include
1015 2007-05-06 Ben Pfaff <blp@gnu.org>
1017 Abstract the documents within a dictionary a little better.
1018 Thanks to John Darrington for suggestion, initial version, and
1019 review. Patch #5917.
1021 * dictionary.c (struct dictionary): Change `documents' member from
1022 char * to struct string.
1023 (dict_clear): Destroy struct string.
1024 (dict_get_documents): Convert struct string to char *.
1025 (dict_set_documents): Set struct string. Pad to 80-character
1027 (dict_clear_documents): New function.
1028 (dict_add_document_line): New function.
1029 (dict_get_document_line_cnt): New function.
1030 (dict_get_document_line): New function.
1032 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1034 * sys-file-reader.c (read_documents): Use new document functions.
1036 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1038 * sys-file-reader.c: When reading a system file which has no
1039 long name table, automatically create one where the long names
1040 are the lower case versions of the short names.
1042 2007-04-22 Ben Pfaff <blp@gnu.org>
1044 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1045 dict_clear will free most data related to the dictionary.
1046 dict_clear does a decent job, except that dict_set_split_vars on
1047 some systems won't actually free the dict's "split" member.
1048 Instead, it'll allocate a 1-byte region. Fix this.
1050 * value.c (value_copy): New function.
1051 (value_set_missing): Ditto.
1053 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1055 * Deleted existing category.h and moved cat-routines.h into
1056 category.h Encapsulated struct cat_vals better.
1058 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1060 * sys-file-reader.c: When reading a system file which has no
1061 long name table, automatically create one where the long names
1062 are the lower case versions of the short names.
1064 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1066 * sys-file-reader.c: Some versions of Other Software seem to
1067 produce system files with string variables' measure set to
1068 zero. We'll assume these are supposed to be nominal variables.
1070 2007-03-30 Ben Pfaff <blp@gnu.org>
1072 * procedure.c: Adapt to new deque data structure.
1074 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1075 Ben Pfaff <blp@gnu.org>
1077 * file-name.c: Mingw compatibility fixes.
1078 (fn_search_path): Use ISSLASH instead of comparing against '/'
1080 (fn_dir_name): Use dir_name from gnulib.
1081 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1082 (fn_get_identity): Use GetFullPathName instead of canonicalize
1083 from gnulib, because the latter does not fully support
1084 Windows-style path names. Use this implementation based on the
1085 detected presence of Windows instead of the absence of Unix, since
1086 the new implementation is Windows-specific.
1087 (fn_compare_file_identities): In Windows implementation, compare
1088 names case-insensitively.
1090 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1092 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1093 <stdlib.h> to have the same effect.
1095 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1097 * por-file-reader.c: Add missing _() around messages.
1099 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1101 * make-file.c: Include "mkstemp.h", without which linking on
1104 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1106 Reduce platform dependence.
1108 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1109 Everywhere we using this, we really should have just depended on
1110 the shell to expand tildes.
1111 (fn_search_path): Simplify, given that we don't do tilde expansion
1113 (fn_normalize): Removed. Caller changed to use the canonicalize
1115 (fn_get_cwd): Removed. Only user was fn_normalize.
1116 (fn_is_absolute): Really only test for absolute names.
1117 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1119 (fn_readlink): Removed, as it was only used fn_normalize.
1120 (fn_exists): Assume the stat function is available; gnulib does.
1121 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1123 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1125 * dictionary.c (dict_create_vector_assert): New function.
1127 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1129 * file-name.c (fn_normalize): Correct name of function
1130 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1131 for pointing this out.
1133 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1135 We are using a single member in struct file_handle, the "name"
1136 field, for more than one purpose. When it begins with '"', it's a
1137 file name; otherwise, it's a token that can be used to identify
1138 it. When that assertion fires, it's because we searched for the
1139 name case-sensitively as a file name (so that there was no match),
1140 and then we try to insert it case-insensitively as a token, which
1141 fails because duplicates aren't allowed.
1143 Solution: break the two purposes into two separate fields. This
1144 fixes the problem and likely makes the code easier to read too.
1146 Fixes bug #18922. Thanks to John Darrington for bug report and
1149 * file-handle-def.c (struct file_handle): New `id' member.
1150 (fh_from_name): Rename fh_from_id. Update all callers.
1151 (create_handle): New `id' parameter. Update all callers.
1152 (fh_create_file): Ditto.
1153 (fh_get_id): New function.
1155 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1157 * case.c (case_is_null): Change return type to bool.
1159 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1163 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1164 reference counting to share data. This makes it easy for
1165 valgrind, etc. to find accesses to cases that have been destroyed
1166 but have been kept around by another user's ref-count. This often
1167 happens when the data set is small enough to find in memory; if a
1168 bigger data set that would overflow to disk were used, then data
1169 corruption would occur.
1171 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1175 * case.c (case_unshare): Make it check internally whether the
1176 ref_cnt is greater than 1, so that the callers don't have to.
1177 Update callers not to check.
1179 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1181 Before, I was thinking that I might want to get rid of reference
1182 counting at some point. Now, I'm pretty sure that it's here to
1183 stay. Thus, because we have to store the value_cnt anyway for
1184 reference-counted cases, we might as well expose it to users.
1186 * case.c (case_get_value_cnt): New function.
1187 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1188 resize case if its size actually changed.
1190 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1191 instead of peeking inside struct case directly.
1192 (casefile_append): Ditto.
1194 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1196 Get rid of the inlines for the case functions, which made the
1197 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1198 the inlines didn't speed up "make check" at all, which is not a
1199 perfect benchmark but seems indicative.)
1201 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1202 definitions. Remove some assertions on nonnull pointers that were
1203 redundant with a pointer dereference soon after in the function.
1205 (struct case_data): Move definition here from case.h.
1209 (case_data_wr): Ditto.
1211 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1213 * automake.mk: Add casedeque.h to sources.
1215 * casedeque.h: New file.
1217 * procedure.c: (struct dataset) Change lag_count, lag_head,
1218 lag_queue member into single struct casedeque member. Update all
1219 users to use the casedeque instead.
1222 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1224 * procedure.c: Simplify lagged cases interface. Updated all
1225 clients--well, the only client--to use the simplified interface.
1226 (dataset_n_lag) Removed.
1227 (dataset_set_n_lag) Removed.
1228 (dataset_need_lag) New function.
1230 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1232 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1233 whenever appropriate, but only when the dataset/dictionary is in a
1236 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1238 * dictionary.c dictionary.h : Added callbacks for change of filter and
1239 split variables. Refactored some code to ensure that callbacks get
1240 invoked when appropriate.
1242 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1243 replace_dict callback occurs when permanent_dict replaces the current
1246 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1248 * dictionary.c dictionary.h : Added callback for when the weight
1249 variable of a dictionary changes.
1251 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1253 * dictionary.c dictionary.h : Added replace_source and replace_dict
1254 callbacks, and functions to deal with them.
1256 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1258 Simplify missing value handling.
1260 * missing-values.h (enum mv_class): New type.
1261 (enum mv_type): Moved definition into missing-values.c and renamed
1262 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1263 enums. Updated all uses.
1264 (struct missing_values): Changed type of `type' from `enum
1265 mv_type' to `int' because the definition is no longer exposed.
1267 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1268 parameter. Update all callers.
1269 (mv_is_num_missing): Ditto.
1270 (mv_is_str_missing): Ditto.
1271 (mv_is_value_user_missing): Removed. Changed callers to use
1272 mv_is_value_missing.
1273 (mv_is_num_user_missing): Removed. Changed callers to use
1275 (mv_is_str_user_missing): Removed. Changed callers to use
1277 (mv_is_value_system_missing): Removed. Changed callers to use
1278 mv_is_value_missing.
1279 (mv_set_type): Removed. Changed callers to use mv_clear.
1280 (mv_clear): New function.
1282 * variable.c (var_is_value_missing): Add new enum mv_class
1283 parameter. Update all callers.
1284 (var_is_num_missing): Ditto.
1285 (var_is_str_missing): Ditto.
1286 (var_is_value_user_missing): Removed. Changed callers to use
1287 var_is_value_missing.
1288 (var_is_num_user_missing): Removed. Changed callers to use
1290 (var_is_str_user_missing): Removed. Changed callers to use
1292 (var_is_value_system_missing): Removed. Changed callers to use
1293 var_is_value_missing.
1295 * casefilter.c (struct casefilter): Use enum mv_class in place of
1297 (casefilter_variable_missing): Adapt to new member.
1298 (casefilter_create): Change signature to take enum mv_class,
1301 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1303 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1305 * case-sink.c case-sink.h procedure.c procedure.h
1306 storage-stream.c: Now uses the factory.
1308 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1310 Make it possible to pull cases from the active file with a
1311 function call, instead of requiring indirection through a callback
1314 * case-source.h (struct case_source_class): Change ->read function
1315 to return a single case, instead of calling a callback function
1316 for each case. Change ->destroy function to return an error
1319 * case-source.c (free_case_source): Pass along the value returned
1320 by the case_source ->destroy function.
1322 * procedure.c (struct write_case_data): Removed.
1323 (struct dataset): Added some members to track procedure state.
1324 (procedure): Optimize the trivial case at this level.
1325 (internal_procedure): Re-implement in terms of proc_open,
1326 proc_read, proc_close.
1327 (proc_open) New function.
1328 (proc_read) New function.
1329 (proc_close) New function.
1330 (write_case) Moved into proc_read.
1331 (close_active_file) Moved closing of data source into proc_close.
1333 * storage-source.c: Rewrote to conform with modified
1334 case_source_class interface.
1336 * transformations.c (trns_chain_execute): Added argument to allow
1337 starting execution from an arbitrary transformation. Updated
1340 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1342 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1344 * sys-file-reader.c (read_display_parameters): Don't assume that
1345 MEASURE_* and ALIGN_* have the same values found in system files.
1347 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1349 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1350 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1351 the value of n_MEASURES, which was off by 1 (at least from my
1354 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1356 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1357 callbacks which are invoked when the dictionary or its
1358 variables are changed.
1360 * missing-values.c missing-values.h value-labels.c: Tidied up
1361 consistency checks, and made some of them return false
1362 instead of assert-failing.
1364 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1366 * calendar.c (calendar_days_in_month): New function.
1368 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1370 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1371 the value label's width.
1373 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1375 * sfm-private.h: Move contents into sys-file-writer.c, which is
1376 the only remaining user. Removed Borland C++-specific directives.
1378 * sys-file-reader.c: Clean up and rewrite entire file. The
1379 rewritten version is simpler and better abstracted, and should be
1380 easier to maintain and extend. It avoids using structures to read
1381 file data, which is prone to padding variations among compilers.
1382 It should also handle non-IEEE 754 system files, although I
1383 haven't been able to find any. It has been tested against many
1384 .sav files obtained from the Web and found to produce the same
1385 results as the earlier version of the code, or in some cases
1386 improved results. It is more tolerant of format variations found
1389 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1390 member, putting an enum integer_format in its place. New member
1391 `float_format'. Changed `compressed' member to type bool.
1393 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1395 * dictionary.c (dict_delete_consecutive_vars): New function.
1397 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1399 * file-name.c (fn_search_path): Remove prefix arg that was unused
1400 by any caller. Updated all callers.
1402 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1404 * format.c (fmt_dollar_template): Use user's decimal point
1405 character. Add assertion.
1407 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1409 * format.c (fmt_dollar_template): New function, based on
1410 dollar_format_template from var-type-dialog.c.
1412 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1414 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1417 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1419 Finish converting struct variable to an opaque type. In this
1420 phase, we add remaining setter and getter functions, convert the
1421 remaining PSPP code to use them, and do a bunch of cleanup. The
1422 resulting changes are pervasive but mostly trivial, and only the
1423 notable changes are logged.
1425 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1428 * case.c (case_data): Renamed case_data_idx.
1429 (case_num): Renamed case_num_idx.
1430 (case_str): Renamed case_str_idx.
1431 (case_data_rw): Renamed case_data_rw_idx.
1433 * case.h (case_data): New function with old name and an interface
1434 that takes a variable instead of an index, which is easier to
1435 use. Updated all callers to use the new interface, or to use the
1436 new *_idx function (see above).
1439 (case_data_rw): Ditto.
1441 * category.c (cat_stored_values_destroy): Changed interface to
1442 take a struct cat_vals * instead of a struct variable *.
1444 * dictionary.c (dict_clone): Use new vector_clone function.
1445 (dict_clear) Use new var_destroy function.
1446 (add_var) New function.
1447 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1448 (dict_create_var_assert) Rewrite in terms of add_var.
1449 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1450 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1451 (dict_lookup_var) Use new var_create, var_destroy functions.
1452 (dict_contains_var) Rewrite in terms of new vardict functionality.
1453 (set_var_dict_index) New function.
1454 (set_var_case_index) New function.
1455 (reindex_vars) New function.
1456 (dict_delete_var) Rewrite in terms of new vardict functionality.
1457 (dict_reorder_var) Ditto.
1458 (dict_reorder_vars) Ditto.
1459 (rename_var) New function.
1460 (dict_rename_var) Use rename_var.
1461 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1462 (dict_get_compacted_idx_to_fv) Rename
1463 dict_get_compacted_dict_index_to_case_index, update callers.
1464 (dict_create_vector) Use new vector_create function.
1465 (dict_clear_vectors) Use new vector_destroy function.
1466 (set_var_short_name_suffix) Move here from variable.c, renamed
1467 from var_set_short_name_suffix, make static, update caller.
1469 * sys-file-private.c: New file.
1470 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1471 width_to_bytes, update callers.
1473 * sys-file-private.h: New file. Later it will supplant
1474 sfm-private.h; for now it supplements it.
1475 (macro MIN_VERY_LONG_STRING) New macro.
1476 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1478 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1479 MAX_LONG_STRING was used before.
1481 * sys-file-writer.c: Ditto.
1483 * value-labels.c: Change the paradigm here to be that a null
1484 pointer is OK for a struct val_labs * in most cases; it just
1485 represents an empty set of value labels.
1486 (val_labs_copy) A copy of a null set is a null set.
1487 (val_labs_count) A null set has 0 labels.
1488 (val_labs_replace) Change return type to void. Rewrite for
1490 (val_labs_find) A null set does not contain the value.
1491 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1492 transposed argument order, updated all callers.
1494 * value.c: New file.
1495 (value_dup) Moved here from variable.c.
1496 (compare_values) Ditto.
1499 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1500 (macro MAX_LONG_STRING) Removed, because it was only interesting
1501 for system files, not for general code.
1502 (macro MAX_VERY_LONG_STRING) Ditto.
1503 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1504 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1506 * vardict.h: New file, for an interface between variables and
1509 * variable.c: A lot of functions were moved around, for better
1511 (struct variable) Move definition here, from variable.h.
1512 (var_type_adj) Removed--makes i18n hard.
1513 (var_type_noun) Ditto.
1514 (var_create) New function.
1515 (var_clone) New function.
1516 (var_destroy) New function.
1517 (var_set_name) Assert that variable is not in a dictionary.
1518 (compare_var_names) Rename compare_vars_by_name and fix a couple
1519 of callers who thought the args were strings.
1520 (hash_var_name) Rename hash_var_by_name.
1521 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1522 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1523 (var_is_very_long_string) Removed, because it was only interesting
1524 to system file code.
1525 (var_set_missing_values) Allow the argument to be the wrong width,
1526 as long as we can resize it. Simplify callers who were doing the
1527 resizing themselves.
1528 (var_get_value_labels) New function.
1529 (var_has_value_labels) New function.
1530 (var_set_value_labels) New function.
1531 (alloc_value_labels) New function.
1532 (var_add_value_label) New function.
1533 (var_replace_value_label) New function.
1534 (var_clear_value_labels) New function.
1535 (var_lookup_value_label) New function.
1536 (var_get_value_name) Moved here from variable.c, renamed from
1537 var_get_value_name, transposed argument order, updated all
1539 (var_to_string) Moved here, from variable-label.c.
1540 (var_set_leave) New function.
1541 (var_get_leave) New function.
1542 (var_must_leave) New function.
1543 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1544 set_var_short_name_suffix.
1545 (var_get_dict_index) New function.
1546 (var_get_case_index) New function.
1547 (var_get_obs_vals) New function.
1548 (var_set_obs_vals) New function.
1549 (var_has_obs_vals) New function.
1550 (var_get_vardict) New function.
1551 (var_set_vardict) New function.
1552 (var_has_vardict) New function.
1553 (var_clear_vardict) New function.
1554 (value_dup) Moved to value.c.
1555 (compare_values) Ditto.
1558 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1559 (enum ALPHA) Rename VAR_STRING, update all users.
1561 * vector.c: New file.
1562 (struct vector) Moved here, from variable.h.
1563 (check_widths) New function.
1564 (vector_create) New function.
1565 (vector_clone) New function.
1566 (vector_destroy) New function.
1567 (vector_get_name) New function.
1568 (vector_get_var) New function.
1569 (vector_get_var_cnt) New function.
1570 (compare_vector_ptrs_by_name) New function.
1572 * vector.h: New file.
1574 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1576 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1577 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1580 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1582 * value-labels.c (destroy_atoms): New function.
1583 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1585 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1587 Thanks to Jason Stover for pointing out this problem.
1589 * data-out.c (output_number): Use gsl_finite from GSL, which is
1590 portable, instead of isfinite, which is not.
1593 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1595 * variable.c variable.h (value_dup): New function.
1597 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1599 Start converting struct variable to an opaque type. In this
1600 phase, we add a bunch of setter and getter functions and convert
1601 most of the PSPP code to use them. The resulting changes are
1602 pervasive but mostly trivial, and only the notable changes are
1605 * format.c (fmt_equal): New function.
1607 * variable.c (var_type_is_valid): New function.
1608 (measure_is_valid) Moved here, from format.c.
1609 (alignment_is_valid) Moved here, from format.c.
1610 (var_get_name) New function.
1611 (var_set_name) New function.
1612 (width_to_type) New function.
1613 (var_get_type) New function.
1614 (var_get_width) New function.
1615 (var_set_width) New function.
1616 (var_is_numeric) New function.
1617 (var_is_alpha) New function.
1618 (var_is_short_string) New function.
1619 (var_is_long_string) New function.
1620 (var_is_very_long_string) New function.
1621 (var_get_missing_values) New function.
1622 (var_set_missing_values) New function.
1623 (var_clear_missing_values) New function.
1624 (var_has_missing_values) New function.
1625 (var_is_value_missing) New function.
1626 (var_is_num_missing) New function.
1627 (var_is_str_missing) New function.
1628 (var_is_value_user_missing) New function.
1629 (var_is_num_user_missing) New function.
1630 (var_is_str_user_missing) New function.
1631 (var_is_value_system_missing) New function.
1632 (var_get_print_format) New function.
1633 (var_set_print_format) New function.
1634 (var_get_write_format) New function.
1635 (var_set_write_format) New function.
1636 (var_set_both_formats) New function.
1637 (var_get_label) New function.
1638 (var_set_label) New function.
1639 (var_clear_label) New function.
1640 (var_has_label) New function.
1641 (var_get_measure) New function.
1642 (var_set_measure) New function.
1643 (var_get_display_width) New function.
1644 (var_set_display_width) New function.
1645 (var_get_alignment) New function.
1646 (var_set_alignment) New function.
1647 (var_get_value_cnt) New function.
1648 (var_get_leave) New function.
1649 (var_get_short_name) New function.
1651 * variable.h: (struct variable) Removed "type" and "nv" members;
1652 they are now computed from "width" where needed.
1654 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1656 * missing-values.c (mv_resize): Don't write beyond end of the
1657 allocated buffer when resizing a long string.
1659 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1661 Clean up identifier code: don't require identifier enumerations to
1662 be in a particular order; make better use of string library;
1663 expose less of the internals.
1665 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1666 change interface. Updated all callers.
1667 (lex_id_match) Change interface to use struct substring, update
1669 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1670 (global array keywords[]) Make static, change form. Update all
1671 users to use lex_id_name instead.
1672 (lex_is_keyword) New function.
1673 (lex_id_to_token) Change interface to use struct substring, update
1675 (lex_id_name) New function.
1677 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1678 lex_is_keyword instead.
1679 (T_LAST_KEYWORD) Removed.
1680 (T_N_KEYWORDS) Removed.
1682 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1684 * format.c: (fmt_date_template) Distinguish characters for which a
1685 space is output and any date delimiter is allowed on input, from
1686 those for which a space is output and only a space is allowed on
1687 input. The former is represented by X, the latter by a space.
1688 Also, drop distinction between h and H, changing the former to the
1691 * data-in.c: Completely rewrite internals to conform to SPSS input
1692 formats as closely as possible.
1693 (data_in) Changed external interface by replacing the structure
1694 that was used as a single argument by a set of arguments. Updated
1696 (data_in_finite_line) Removed. Converted all callers to use plain
1698 (data_in_get_integer_format) New function.
1699 (data_in_set_integer_format) New function.
1700 (data_in_get_float_format) New function.
1701 (data_in_set_float_format) New function.
1703 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1704 (struct data_in) Removed.
1706 * data-out.c: (output_date) Drop each component from the input as
1707 it is output, to allow us to drop the distinction between h (a
1708 count of hours) and H (the hour of day) template characters.
1709 Also, handle new X template character.
1710 (output_scientific) Follow more rational rule on when to drop
1711 fraction introduced between SPSS 13 and 15. Updated test case to
1714 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1716 Fix buffer overflow reported by John Darrington.
1718 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1719 realize that DIGITS is a count of nibbles, not of bytes.
1721 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1723 * calendar.c (calendar_offset_to_gregorian) Also return the
1724 year-of-day. Change callers to new interface.
1726 * data-out.c: Completely rewrite internals to conform to SPSS
1727 output formats as completely as possible.
1728 (data_out) Change interface to put input parameters before output
1729 parameters, for consistency with the style I now prefer. Update
1731 (data_out_get_integer_format) New public function.
1732 (data_out_set_integer_format) New public function.
1733 (data_out_get_float_format) New public function.
1734 (data_out_set_float_format) New public function.
1736 * data-out.h: New file. Move prototype for data_out here, from
1739 * format.c: (fmt_step_width) Use equality comparison instead of
1740 bitwise and, for clarity.
1741 (fmt_is_string) Ditto.
1742 (fmt_input_to_output) Fix categories that are translated to F
1745 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1747 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1748 casefile-private.h: Added casefilter to assist commands with missing
1751 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1753 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1755 * settings.c: (route_errors_to_terminal) New variable.
1756 (route_errors_to_listing) New variable.
1757 (get_error_routing_to_terminal) New function.
1758 (set_error_routing_to_terminal) New function.
1759 (get_error_routing_to_listing) New function.
1760 (set_error_routing_to_listing) New function.
1762 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1764 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1766 * format.c: Completely rewrite, to achieve better abstraction.
1767 Rewrite all references to formats in other files.
1769 * format.def: Rewrite and reorganize.
1771 * settings.c: Move everything related to custom currency formats
1772 into format.[ch], changing them in form, so as to group related
1773 code and definitions better. Changed all references to use the
1775 (static var decimal) Removed.
1776 (static var grouping) Removed.
1777 (static var cc) Removed.
1778 (get_decimal) Removed.
1779 (set_decimal) Removed.
1780 (get_grouping) Removed.
1781 (set_grouping) Removed.
1785 * settings.h: (macro CC_CNT) Removed.
1786 (macro CC_WIDTH) Removed.
1787 (struct custom_currency) Removed.
1789 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1791 * data-in.c (data_in): Use switch statement instead of table, to
1792 avoid dependence on the order of the FMT_* enums.
1794 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1796 * data-out.c: (num_to_string) Removed, because it was dead code.
1798 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1800 * data-in.c (parse_trailer): Fix error message.
1802 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1804 * format.c (fmt_is_binary): New function.
1806 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1808 * procedure.c procedure.h: Encapsulated the static data into a single
1811 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1813 * casefile.c (casereader_read_xfer): Always initialize the case,
1814 even on an error condition.
1816 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1818 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1819 entire function useless.
1821 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1823 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1824 casereader method casereader_clone.
1826 * procedure.c pransformations.h: Introduced new type casenum_t
1828 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1830 * variable.c: (width_to_bytes) Rephrase code for clarify.
1832 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1834 * format.c: (fmt_type_from_string) New function.
1835 (fmt_to_string) Include decimals in output if the format has
1836 decimals, even if the format type does not. This way, we can
1837 accurately reproduce incorrect formats in user output.
1838 (check_common_specifier) Make the check for a bad format type an
1839 assertion, so we get bug reports if they show up. Fix message.
1840 Check for decimal places with a format type that doesn't allow
1842 (check_input_specifier) Remove check for FMT_X, which has been
1844 (check_output_specifier) Ditto.
1846 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1848 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1849 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1850 (enum fmt_parse_flags) Removed.
1852 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1854 * casefile.c casefile.h: Converted to an abstract base class.
1855 * casefile-private.h fastfile.c fastfile.h: New files.
1856 * automake.mk procedure.c scratch-writer.c storage-stream.c
1858 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1860 * procedure.c (internal_procedure): Create sink_case with only as
1861 many values as the compacted dictionary.
1863 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1865 Remove "debugging" code that caused plenty of false positives and
1868 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1870 * case.c: Remove all references to `this' member.
1872 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1874 Fix link error noted by Jason Stover.
1876 * storage-stream.c: Include <assert.h>.
1878 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1880 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1881 ALL) and additional underlying system file issues.
1883 Thanks to John Darrington for review.
1885 First problem: var_hash points to variables not owned by the
1886 sys-file-reader, which the caller may free or modify. Use an
1887 array of sfm_vars instead, as done earlier (e.g. CVS version
1890 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1891 members and remove all code that references it. Add vars, var_cnt
1892 members. Remove fix_specials member, which was unused.
1893 (struct sfm_var) Remove name member, which was unused.
1894 (sfm_close_reader) Free vars member instead of var_hash.
1895 (compare_var_shortnames) Removed.
1896 (hash_var_shortname) Removed.
1897 (sfm_open_reader) Fill out vars array.
1898 (compare_var_index) Removed.
1899 (sfm_read_case) Use vars instead of var_hash.
1901 Second problem: we're confused about when we actually have very
1902 long strings, causing us to choose incorrectly between slow path
1903 and fast path in sfm_read_case.
1905 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1906 have very long strings, not when we have long variable names,
1907 which is an unrelated feature.
1909 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1911 * variable.h: Move var_set and variable parsing declarations to
1912 new header, src/language/lexer/variable-parser.h. Modified lots
1913 of files to include the new header.
1915 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1917 * value-labels.c (value_to_string): When there's no value label,
1918 format the variable according to its print format, instead of
1919 always effectively using A or F format.
1921 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1923 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1926 * format.c: Removed tortological assertion.
1928 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1930 Reform string library.
1932 * file-name.c (fn_interp_vars): Change interface to take a
1933 substring as input. Updated all users.
1935 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1937 * format.c (measure_is_valid): Really return false when m >=
1940 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1942 Implement random access to casefiles, for use in GUI.
1944 * casefile.c: (struct casereader) Add `random', `file_ofs',
1945 `buffer_ofs' members.
1946 (casefile_get_random_reader) New function.
1947 (read_open_file) Break part into new function
1948 seek_and_fill_buffer().
1949 (fill_buffer) Update buffer_ofs, file_ofs.
1950 (casereader_seek) New function.
1952 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1954 * settings.c: Added call to i18n{done, init}.
1956 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1958 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1961 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1963 * casefile.c: Convert many uses of `int' to `bool'.
1965 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1967 * transformations.c (trns_chain_destroy): Destroy chain's trns
1968 member, to fix memory leak.
1970 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1972 * storage-stream.c (storage_source_decapsulate): Destroy case
1973 source to fix memory leak.
1975 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1977 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1978 case passed as argument instead of initializing the argument as a
1979 case. Fixes memory leak that showed up in
1980 tests/command/aggregate.sh with scratch files.
1982 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1984 * procedure.c (proc_done): Destroy default_dict, to fix memory
1987 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1989 Simplify procedure_with_splits().
1991 * procedure.c (struct split_aux_data): Removed case_count member.
1992 (procedure_with_splits) Don't initialize case_count.
1993 (split_procedure_case_func) Check whether prev_case is null
1994 instead of case_count.
1995 (split_procedure_end_func) Ditto.
1997 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1999 * case.c (case_move): Do nothing if dst and src are the same
2001 (case_try_create) Merge two similar cases.
2002 (case_copy) Unshare only if data must be actually copied.
2004 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2006 * data-in.c data-out.c dictionary.c sys-file-reader.c
2007 sys-file-writer.c variable.c variable.h: Reworked very long string
2008 support for better encapsulation.
2010 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2012 * value-labels.c (val_labs_can_set_width): New function.
2013 (val_labs_set_width) Clear labels if increasing width to long
2015 (val_labs_destroy) Remove unneeded test for null.
2017 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2019 * value-labels.h: Remove unneeded dependency on variable.h.
2021 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2023 Get rid of `char *c' member in union value, for cleanliness.
2025 * value.h: (union value) Remove `c' member.
2027 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2029 Make dictionary compacting functions a little more general.
2031 * sys-file-writer.c (sfm_open_writer): Use
2032 dict_compacting_would_change().
2033 (does_dict_need_translation) Removed.
2035 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2037 Make dictionary compacting functions a little more general.
2039 * dictionary.c (dict_needs_compaction): Rename
2040 dict_compacting_would_shrink(). Update all callers.
2041 (dict_compacting_would_change) New function.
2043 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2045 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2046 inverted return value (!).
2048 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2050 Continue reforming procedure execution.
2052 * procedure.c: Search and replace "vfm" by "proc". Notably:
2053 (static var vfm_source) Rename proc_source. Update all
2055 (static var vfm_sink) Rename proc_sink. Update all references.
2057 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2059 Continue reforming procedure execution. In this phase, remove
2060 PROCESS IF, which was deprecated anyway and can be easily
2061 simulated with TEMPORARY followed by SELECT IF.
2063 * procedure.c: (open_active_file) Don't call
2064 add_process_if_trns().
2065 (discard_variables) Get rid of redundant call to
2066 proc_cancel_all_transformations().
2067 (add_process_if_trns) Removed.
2068 (process_if_trns_proc) Removed.
2069 (process_if_trns_free) Removed.
2071 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2073 Continue reforming procedure execution. In this phase, add
2074 `const' to the case passed to procedure()'s callback.
2076 Updated all users of procedure() as well.
2078 * procedure.c: (struct write_case_data) Add "const" to ccase
2079 parameter for case_func member.
2080 (procedure) Add "const" to ccase parameter for proc_func
2082 (multipass_case_func) Make ccase parameter const.
2083 (internal_procedure) Add "const" to ccase parameter for case_func
2085 (split_procedure_case_func) Make ccase parameter const.
2086 (multipass_split_case_func) Make ccase parameter const.
2088 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2090 Continue reforming procedure execution. In this phase, get rid of
2091 the output code for SPLIT FILE groups in procedure.c, which really
2092 shouldn't be doing any output. Move it into the individual
2093 procedures instead. This also adds some flexibility.
2095 Updated many users of procedure_with_splits() and
2096 multipass_procedure_with_splits() to call
2097 output_split_file_values() and to deal with increased use of
2100 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2101 parameter to begin_func member.
2102 (procedure_with_splits) Add "const struct ccase *" parameter to
2103 begin_func parameter. Make ccase parameter const in proc_func
2105 (split_procedure_case_func) Don't dump split file group. Pass
2107 (dump_splits) Moved to language/dictionary/split-file.c as
2108 output_split_file_values().
2109 (struct multipass_split_aux_data) Add "const struct ccase *"
2110 parameter to split_func member.
2111 (multipass_procedure_with_splits) Add "const struct ccase *"
2112 parameter to split_func parameter.
2113 (multipass_split_case_func) Save new SPLIT FILE case before
2115 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2117 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2119 Continue reforming procedure execution. Change
2120 internal_procedure() so that it calls open_active_file() and
2121 close_active_file(), which isolates most of the actual procedure
2124 * procedure.c: (struct write_case_data) Rename `proc_func' member
2125 to `case_func' and update all references.
2126 (procedure) Rewrite as one-line wrapper around
2127 internal_procedure().
2128 (struct multipass_aux_data) New.
2129 (multipass_callback) Renamed multipass_case_func(). Use struct
2130 multipass_aux_data as auxiliary data.
2131 (multipass_end_func) New function.
2132 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2133 that uses multipass_case_func, multipass_end_func.
2134 (internal_procedure) Add `end_func' argument. Move optimization
2135 of trivial case in here. Move call to open_active_file() and
2136 close_active_file() in here. Now assert that vfm_source is
2138 (procedure_with_splits_callback) Rename
2139 split_procedure_case_func().
2140 (split_procedure_end_func) New function.
2141 (multipass_split_callback) Rename multipass_split_case_func.
2142 (multipass_split_end_func) New function.
2143 (discard_variables) No need to test for nonnull vfm_source.
2145 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2147 Continue reforming procedure execution. Get rid of unused member.
2149 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2151 (write_case) Don't increment cases_analyzed.
2153 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2155 Continue reforming procedure execution. In this phase, move
2156 procedure.c and procedure.h from src to src/data. Update
2157 makefiles and #includes accordingly.
2159 * procedure.c: Moved here from src/.
2161 * procedure.h: Moved here from src/.
2163 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2165 Continue reforming procedure execution. In this phase, get rid of
2166 many global variables, consolidating procedure execution in
2167 procedure.c. Encapsulate transformations in new "struct
2168 trns_chain". Also, change implementation of N OF CASES, FILTER,
2169 and PROCESS IF from special cases to transformations.
2171 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2174 * dictionary.c: (global variable default_dict) Move to
2177 * variable.h: (TRNS_*) Move to transformations.h.
2178 (struct transformation) Move to transformations.c.
2180 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2182 * sys-file-reader.c: Fixed invalid read problems.
2184 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2186 * storage-stream.c: Add missing function comments.
2188 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2190 Continue reforming procedure execution. In this phase, add some
2191 new, needed functionality to storage-stream.
2193 * storage-stream.c: (storage_source_decapsulate) New function.
2195 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2197 * variable.c (width_to_bytes): Declarations must precede
2198 statements for C90 compliance.
2200 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2202 * data-out.c, data-in.c, variable.c, variable.h: New functions
2203 copy_mangle and copy_demangle for reading/writing cases; emulates the
2204 way SPSS deals with strings > 255 bytes.
2206 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2207 subtype 14 needed for strings longer than 255 bytes.
2209 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2210 instead of literal values. Also fixed some constness issues.
2212 * format.h: Constness
2214 * sfm-private.h: Renamed the case_size identifier, since I discovered
2215 that SPSS's respect for this variable is very nominal.
2217 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2219 Change case limit type from int to size_t.
2221 * dictionary.c: (struct dictionary) Change type of case_limit
2223 (dict_get_case_limit) Change return type.
2224 (dict_set_case_limit) Change parameter type.
2226 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2228 * variable.h: (struct variable) Rename `reinit' member as `leave'
2229 and invert sense. Fix up all references.
2231 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2233 Continue reforming procedure execution. In this phase, break
2234 procedure.c into multiple files.
2236 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2238 * case-sink.c: New file.
2240 * case-sink.h: New file.
2242 * case-source.c: New file.
2244 * case-source.h: New file.
2246 * storage-stream.c: New file.
2248 * storage-stream.h: New file.
2250 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2252 * variable.h: (struct variable) Remove `init' member and all
2253 references to it from other files. It was initialized in several
2254 places, but nothing really ever used it for anything worthwhile.
2255 Thanks to Jason Stover for pointing out how confusing this
2258 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2260 Continue reforming error message support. In this phase, get rid
2261 of message "titles" and put the message text in `struct error'.
2262 Now `struct error' encapsulates a message more properly.
2264 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2265 Format message ourselves.
2267 * data-in.c: (vdls_error) Ditto.
2269 * por-file-reader.c: (error) Ditto.
2271 * sys-file-reader.c: (corrupt_msg) Ditto.
2273 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2275 GNU standards require "file name" instead of "filename" in
2276 documentation. It's nice for our code to follow the convention
2279 * casefile.c: (struct casefile) Rename `filename' member to
2280 `file_name'. Updated all references.
2282 * file-name.c: [!unix] (struct file_identity) Rename
2283 normalized_filename member to normalized_file_name. Updated all
2286 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2288 We don't really support anything but Unix-like environments well,
2289 so we might as well de-obfuscate by writing directory and path
2290 separators explicitly.
2292 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2294 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2295 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2297 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2300 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2302 GNU standards require "file name" instead of "filename" in
2303 documentation. It's nice for our code to follow the convention
2306 * filename.c: Rename to file-name.c.
2308 * filename.h: Rename to file-name.h. Update all inclusions.
2309 Update header guards.
2311 * automake.mk: Update file names.
2313 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2315 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2317 (fn_basename) Removed (dead code).
2318 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2319 (fn_special_p) Renamed fn_is_special(), all references updated.
2320 (fn_exists_p) Renamed fn_exists(), all references updated.
2322 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2324 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2325 Also, now it only tilde-expands file names, not paths.
2326 (fn_search_path) Tilde-expand one directory at a time.
2328 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2330 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2331 get rid of non-Unixlike version of the code, which has probably
2333 (fn_prepend_dir) Removed (dead code).
2335 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2336 (macro PATH_SEPARATOR_STRING) New.
2337 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2339 Continue reforming error message support. In this phase, we get
2340 rid of VM() and the other msg() support for "verbosity", replacing
2341 it by a new function verbose_msg().
2343 * filename.c: (fn_search_path) Use verbose_msg() instead of
2346 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2348 * sfm-private.h: Get rid of #defines after #error, which makes no
2351 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2353 Get rid of our own int32 type in favor of the standard int32_t
2356 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2359 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2361 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2363 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2365 Remove ill-considered file routines that are no longer used.
2367 * filename.c: (fn_open_ext) Removed.
2368 (fn_close_ext) Removed.
2370 * filename.h: (struct file_ext) Removed.
2372 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2374 * variable.c (var_is_valid_name): Move declarations before code
2377 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2379 * filename.ch (fn_interp_vars): Fixed small buglet.
2381 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2383 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2384 so as to modify the string inline. Thus makeing it easier to
2385 destroy the results when no longer needed.
2387 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2389 * category.c (cat_stored_values_destroy): Fixed memory leak.
2391 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2393 Add some missing frees. Thanks to John Darrington for reporting
2396 * any-writer.c (any_writer_close): Free writer.
2398 * any-reader.c (any_reader_close): Free reader.
2400 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2402 * por-file-reader.c: (error) Mark as NO_RETURN.
2404 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2406 * settings.c: Changed default value of scompress to true.
2408 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2410 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2412 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2415 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2416 here from pref.h.orig.
2418 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2420 * sys-file-reader.c: Fixed bug reading compressed files.
2422 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2424 * Numerous renames. See src/ChangeLog for details.
2426 * Moved files from src directory
projects / pspp-builds.git / blob