1 2008-02-02 Ben Pfaff <blp@gnu.org>
5 * sys-file-reader.c (read_variable_record): Allow missing values
6 to be specified on long string variables, but warn about them
7 because PSPP does not yet support them.
8 (read_extension_record): Ignore extension records 20 and 21, which
9 PSPP does not yet support.
11 2008-02-01 Ben Pfaff <blp@gnu.org>
13 Patch #6386. Thanks to John Darrington for review and for the
14 updates to gnumeric-reader.c.
16 * dictionary.c (make_hinted_name): New function.
17 (make_numeric_name): New function.
18 (dict_make_unique_var_name): New function.
20 * gnumeric-reader.c (devise_name): Removed.
21 (munge_name): Removed.
22 (gnumeric_open_reader): Use new function
23 dict_make_unique_var_name.
25 * short-names.c (set_var_short_name_suffix): Use new function
28 2008-01-19 John Darrington <john@darrington.wattle.id.au>
30 * settings.c settings.h: Moved static variables into a
31 single struct. Renamed functions to have a settings_ prefix.
33 2008-01-14 John Darrington <john@darrington.wattle.id.au>
35 * data-out.c (output_AHEX): Corrected number of bytes in
36 call to output_hex. Closes bug #22011
38 2008-01-02 John Darrington <john@darrington.wattle.id.au>
40 * variable.c variable.h: Replaced var_get_value_name with
41 var_append_value_name which doesn't use any static data.
42 Thanks to Ben for review.
44 2007-12-07 Ben Pfaff <blp@gnu.org>
48 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
49 made from an empty casereader produce a casegrouper with no
50 groups, instead of one with one group that has no cases.
52 * casereader.c (casereader_is_empty): New function.
54 2007-12-06 Ben Pfaff <blp@gnu.org>
58 * sys-file-reader.c (read_display_parameters): Handle variable
59 display parameters record with only 2 data items per variable.
60 Reported by Guido Gay <gay@irer.it>.
62 2007-12-04 Ben Pfaff <blp@gnu.org>
64 * identifier.c (lex_id_match_n): New function.
65 (lex_id_match): Reimplement in terms of lex_id_match_n.
67 2007-11-24 Ben Pfaff <blp@gnu.org>
69 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
72 2007-11-24 Ben Pfaff <blp@gnu.org>
74 Fix warning reported when reading back system files that include
75 very long string variables. Thanks to Guido Gay <gay@irer.it> for
78 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
79 was written where `j' was meant.
81 2007-11-18 Ben Pfaff <blp@gnu.org>
83 Properly write variables that include a range of missing values to
84 system files. Thanks to Guido Gay <gay@irer.it> for reporting
87 * sys-file-writer.c (write_variable): Correctly calculate
88 n_missing_values field when writing variables that include a range
91 2007-11-10 Ben Pfaff <blp@gnu.org>
93 Cleanups and bug fixes devised while writing up documentation.
96 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
98 * dict-class.c: New file.
99 (dict_class_from_id): Move here.
100 (dict_class_to_name): Move here.
102 * dict-class.h: New file.
103 (enum dict_class): Move here. Change from consecutive integers to
104 consecutive bits, to make testing for any of multiple values
105 easier. Add new DC_ALL constant.
107 * dictionary.c (struct dictionary): Change `case_limit' from
108 size_t to casenumber.
109 (dict_get_vars): Make final argument an enum dict_class.
110 (dict_get_vars_mutable): Ditto.
111 (dict_get_case_limit): Change return value to casenumber.
112 (dict_set_case_limit): Change final argument to a casenumber.
113 (dict_unset_split_var): Add assertion. Rephrase slightly.
114 (dict_set_label): Use xstrndup to simplify.
116 * format.c (fmt_step_width): AHEX format also needs 2-byte
118 (fmt_set_style): Simplify assertions.
120 * missing-values.c (mv_add_num_range): Rename mv_add_range.
121 Simplify implementation.
122 (mv_has_value): Simplify implementation.
123 (mv_pop_value): Remove the first value, not the last, to avoid
124 having GET followed by SAVE reverse the order of missing values.
125 (mv_peek_value): Rename mv_get_value. Simplify assertion.
126 (mv_has_range): Simplify implementation.
127 (mv_peek_range): Rename mv_get_range.
128 (can_resize_string): Removed.
129 (mv_is_resizable): Use value_is_resizable.
130 (mv_resize): Use value_resize.
132 * short-names.h (SHORT_NAME_LEN): Move here.
134 * val-type.h: New file, for definitions related to type and width
135 of abstract values. Before, these definitions were mixed among
136 those related to "union value" and those related to variables.
137 (macro SYSMIS): Move here.
138 (macro LOWEST): Move here.
139 (macro HIGHEST): Move here.
140 (macro MAX_STRING); Move here.
141 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
142 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
143 All references updated.
144 (val_type_is_valid): New function. Replaces var_type_is_valid.
145 All references updated.
146 (val_type_from_width): New function. Replaces
147 var_type_from_width. All references updated.
149 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
151 (val_labs_can_set_width): Use value_is_resizable.
152 (val_labs_add): Simply return false if the value labels set is too
153 wide, instead of having undefined behavior.
154 (val_labs_replace): Ditto.
155 (val_labs_replace): Ditto.
156 (val_labs_first): Set iterator to null if iteration is complete.
157 (val_labs_first_sorted): Ditto.
158 (val_labs_done): Become a no-op if the iterator is null.
160 * value.c (value_is_resizable): New function.
161 (value_resize): New function.
163 * variable.c (var_get_dict_class): New function.
165 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
168 2007-11-08 Ben Pfaff <blp@gnu.org>
170 * data-in.c: Make formatted data parsing locale-independent.
171 (parse_number): Use c_strtod instead of strtod, to avoid
172 locale-specific behavior.
175 2007-11-06 Ben Pfaff <blp@gnu.org>
177 Patch #6256: add support for binary, 360 file formats. Reviewed
180 * data-in.c (struct data_in): Add `encoding' member.
181 (data_in): Add `encoding' parameter, and re-encode the data passed
182 in where appropriate. Update all callers to pass it in.
183 (parse_A): Implement EBCDIC recoding wart described in manual.
184 (parse_AHEX): Implement EBCDIC recoding.
186 * data-out.c (data_out_legacy): New function.
187 (data_out): Make into a wrapper around data_out_legacy.
189 * file-handle-def.c (struct file_handle): New member `encoding'.
190 (fh_create_file): Set encoding.
191 (fh_default_properties): Set default encoding.
192 (fh_get_legacy_encoding): New function.
194 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
195 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
196 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
197 (struct fh_properties): New member `encoding'.
199 2007-11-05 Ben Pfaff <blp@gnu.org>
201 Patch #6258. Reviewed by John Darrington.
203 * file-handle-def.c (fh_lock): Add comment that TYPE should be
204 marked with N_() in the caller. Added these markings to each
205 caller too. Should make i18n easier.
206 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
208 2007-11-03 Ben Pfaff <blp@gnu.org>
210 Allow output files to overwrite input files (bug #21280).
212 * file-handle-def.c: Separate locking of files for input and for
213 output, to allow both to take place at once. Also, distinguish a
214 file handle from the identity of the underlying file, because the
215 identity of a file changes over time and the file handle can't
216 represent two different identities.
217 (struct file_handle): Remove `next', `open_cnt', `deleted',
218 `type', `open_mode', `aux', `identity' members. Change `id' from
219 char array to char *. Add `ref_cnt' member.
220 (file_handle_from_ll) New function.
221 (file_handles) Removed.
222 (named_handles) New variable.
223 (fh_init) Initialize named_handles.
224 (fh_done) Remove name from all named_handles.
225 (free_handle) Updated for modified struct file_handle.
226 (unname_handle) New function.
227 (fh_ref) New function.
228 (fh_from_file_name) Removed.
229 (fh_unref) New function.
230 (fh_unname) New function.
231 (fh_from_id) Rewritten.
232 (create_handle) Updated for modified struct file_handle.
233 (fh_inline_file) Increment the handle's ref count.
234 (fh_create_file) Updated for modified struct file_handle.
235 (fh_create_scratch) Ditto.
240 (fh_is_open) Removed.
241 (fh_get_id) Updated for modified struct file_handle.
242 (fh_get_default_handle) Increment the handle's ref count.
243 (fh_set_default_handle) Handle ref counts.
244 (struct fh_lock) New structure.
245 (locks) New static var.
246 (fh_lock) New function.
247 (fh_unlock) New function.
248 (fh_lock_get_aux) New function.
249 (fh_lock_set_aux) New function.
250 (fh_is_locked) New function.
251 (make_key) New function.
252 (free_key) New function.
253 (compare_fh_locks) New function.
254 (hash_fh_lock) New function.
256 * file-handle-def.h (enum fh_access) New enum.
258 * file-name.c: Made file_identity the same in all supported
260 (struct file_identity): New `name' member.
261 (fn_get_identity): For a file that doesn't exist, get the
262 dev/inode of its directory plus its name. If even the directory
263 doesn't exist, just use its name. Merge the Windows
264 implementation into the Unix one.
265 (fn_compare_file_identities): Rewritten. Merge the Windows
266 implementation into the Unix one.
267 (fn_hash_identity): New function.
269 * make-file.c (struct replace_file): New structure.
270 (all_files): New static var.
271 (replace_file_start): New function.
272 (replace_file_commit): New function.
273 (replace_file_abort): New function.
274 (free_replace_file): New function.
275 (unlink_replace_files): New function.
277 * por-file-reader.c (struct pfm_reader): Add `lock' member.
278 (close_reader): Unlock file.
279 (pfm_open_reader): Lock file.
281 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
283 (pfm_open_writer): Lock file and prepare for its replacement.
284 (close_writer): Unlock file.
286 * scratch-handle.h (struct scratch_handle): Add unique_id so that
287 different generations of a scratch file can be distinguished.
289 * scratch-reader.c (scratch_reader_open): Verify that the file is
292 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
293 remove scratch_handle member.
294 (scratch_writer_open): Lock handle. Prepare to replace handle
295 data, instead of doing it immediately.
296 (scratch_writer_casewriter_destroy): Replace handle data and
299 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
300 (sfm_open_reader): Lock file.
301 (close_reader): Unlock file.
303 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
305 (sfm_open_writer): Lock file and prepare for its replacement.
306 (close_writer): Unlock file.
308 2007-11-02 Ben Pfaff <blp@gnu.org>
310 * data-out.c (output_number): Use isfinite (from C99) instead of
311 gsl_isfinite, since we now have universal support for it in
313 (output_infinite): Ditto.
314 [!HAVE_ROUND] (round): Remove definition, since we now have a
315 replacement in gnulib.
317 2007-10-27 John Darrington <john@darrington.wattle.id.au>
319 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
320 wasn't initialised correctly. Closes bugs #21330 and 21397
323 2007-10-08 Ben Pfaff <blp@gnu.org>
325 Fix bug #21280. Thanks to John Darrington for review.
327 * file-name.c (create_stream): New function.
329 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
330 file handle before creating the file, to ensure that we don't
331 truncate a file that we're reading. Make code easier to read by
334 * sys-file-write.c (sfm_open_writer): Ditto.
336 2007-10-01 Ben Pfaff <blp@gnu.org>
338 Fix bug #21192. Thanks to John Darrington for review.
340 * casereader.c (casereader_read): Decrement case_cnt before
341 calling the casereader's "read" member function, so that we
342 interact properly with lazy_casereader.
344 * datasheet.c: Add regression test for above bug fix.
345 (clone_datasheet): New function.
346 (lazy_callback): New function.
347 (check_datasheet_casereader): New function.
348 (check_datasheet): Check datasheet contents are reported correctly
349 through an ordinary casereader and a lazy casereader.
350 (clone_model): Use clone_datasheet.
352 2007-09-24 Ben Pfaff <blp@gnu.org>
354 Patch #6210. Reviewed by John Darrington.
355 * settings.c: Drop "long view", which is not needed any longer.
356 (static var long_view): Removed.
357 (force_long_view): Removed.
358 (get_viewwidth): Removed.
359 (init_viewport): Removed long_view reference.
361 2007-09-19 John Darrington <john@darrington.wattle.id.au>
363 * settings.c settings.h: Changed viewport's length and width to be
364 owned by the user interface which uses the data library. This allows
365 better abstraction, and makes dynamically adjustable dimensions easier.
367 2007-09-18 Ben Pfaff <blp@gnu.org>
369 * procedure.c (proc_extract_active_file_data): New function.
371 * lazy-casereader.h: New file.
373 * lazy-casereader.c: New file.
375 * casereader.c (casereader_dynamic_cast): New function.
377 2007-09-14 Ben Pfaff <blp@gnu.org>
379 * dictionary.c (dict_clone): Copy case indexes from cloned
380 dictionary. Fixes bug #21061. Reviewed by John Darrington.
382 2007-09-13 John Darrington <john@darrington.wattle.id.au>
384 * value.c value.h (create_value): New function.
386 2007-09-12 Ben Pfaff <blp@gnu.org>
388 Make it clear that translator casereader and casewriter translate
389 functions are supposed to destroy their input case.
391 * casereader-translator.c (struct casereader_translator): Change
392 input case parameter of translate member function from const
393 struct ccase * to struct ccase *.
394 (casereader_create_translator): Ditto, for translate parameter.
396 * casewriter-translator.c (struct casewriter_translator): Ditto.
397 (casewriter_create_translator): Ditto.
399 2007-08-27 John Darrington <john@darrington.wattle.id.au>
401 * sys-file-reader.c (read_display_parameters): Force display width
402 to 8, if the sys file says 0 (like SPSS does).
404 2007-08-12 Ben Pfaff <blp@gnu.org>
406 * dictionary.c (dict_dump): New function.
408 2007-08-12 Ben Pfaff <blp@gnu.org>
410 Drop dict_compactor in favor of using the new struct case_map.
412 * dictionary.c (struct copy_map): Removed.
413 (struct dict_compactor): Removed.
414 (dict_make_compactor): Removed.
415 (dict_compactor_compact): Removed.
416 (dict_compactor_destroy): Removed.
418 * procedure.c (struct data_set): Change `compactor' member to be a
420 (proc_open): Use case_map_to_compact_dict instead of
422 (proc_casereader_read): Use case_map_execute instead of
423 dict_compactor_compact.
424 (proc_commit): Use case_map_destroy instead of
425 dict_compactor_destroy.
427 * scratch-writer.c (struct scratch_writer): Change `compactor'
428 member to be a struct case_map *.
429 (scratch_writer_open): Use case_map_to_compact_dict instead of
431 (scratch_writer_casewriter_write): Use case_map_execute instead of
432 dict_compactor_compact.
434 2007-08-12 Ben Pfaff <blp@gnu.org>
436 * automake.mk: Add case-map.c, case-map.h.
438 * case-map.c: New file.
440 * case-map.h: New file.
442 2007-08-12 Ben Pfaff <blp@gnu.org>
444 * dictionary.c (dict_compact_values): Don't delete scratch
445 variables as well as compacting case indexes. Update all callers.
446 (dict_get_compacted_value_cnt): Rename dict_count_values and
447 change interface. Update all callers.
448 (dict_get_compacted_value_cnt): Remove.
449 (dict_compacting_would_shrink): Remove.
450 (dict_compacting_would_change): Remove.
451 (dict_make_compactor): Add new parameter. Update all callers.
453 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
454 instead of dict_count_values, changing an O(N) operation into
457 2007-08-12 Ben Pfaff <blp@gnu.org>
459 * casereader.c (casereader_read): Don't require cases read by a
460 casereader to be exactly the expected size: as long as they're big
463 2007-08-12 Ben Pfaff <blp@gnu.org>
465 Make casewriters keep track of the number of `union value's in
466 each case. This is useful for two reasons: casewriter_write can
467 then check that the case being written is large enough, and later
468 recipients of the casewriter can determine the size of the case.
470 * casewriter-translator.c (casewriter_create_translator): Add
473 * casewriter.c (struct casewriter): Add value_cnt member.
474 (casewriter_write): Check that the case passed in is big enough.
475 (casewriter_get_value_cnt): New function.
476 (casewriter_create): Add value_cnt parameter.
478 2007-08-09 Ben Pfaff <blp@gnu.org>
480 Fix bug reported by Jason Stover.
481 * settings.c: Move get_termcap_viewport and in particular the
482 #include for <curses.h> to the end of the file. curses.h
483 redefines bool on some systems (e.g. OpenBSD), which causes
484 disagreement between uses of bool before its inclusion and
486 Tested by Jason Stover.
488 2007-07-29 Ben Pfaff <blp@gnu.org>
490 Provisional fix for bug #18692 and bug #20161. Reviewed by John
493 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
494 argument (never "rb" or "wb") because SUSv3 says that only those
495 modes are defined, and glibc in fact rejects other modes.
497 Open portable files with fn_open so that they can be read from
498 pipes. Fix missing fh_close call to go along with fh_open.
499 Report an error if the file close reports an error.
500 * por-file-reader.c (close_reader): New function.
501 (por_file_casereader_destroy): Use close_reader.
502 (pfm_open_reader): Open file with fn_open.
504 2007-07-28 Ben Pfaff <blp@gnu.org>
506 Make PSPP able to read all the portable files I could find on the
507 web. Thanks to John Darrington for review. Bug #17620.
508 * por-file-reader.c (struct pfm_reader): New member `line_length'.
509 (error): Print file offset in hexadecimal.
510 (warning): New function.
511 (advance): Treat lines less than 80 bytes long as padded to 80
513 (pfm_open_reader): Call read_documents if we find an "E" record.
514 (convert_format): Convert invalid formats to the default format
515 instead of aborting reading the file.
516 (read_variables): Rename duplicate variable names instead of
517 aborting reading the file.
518 (read_value_label): Allow string variables of different widths to
519 be assigned value labels in the same record. Replace duplicate
520 value labels instead of aborting.
521 (read_documents): New function.
523 * por-file-writer.c (pfm_open_writer): Call write_documents if the
524 dictionary has documents.
525 (write_documents): New function.
527 2007-07-25 Ben Pfaff <blp@gnu.org>
529 Fix bugs related to bug #17213.
531 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
532 since the former is what config.h has. Include the needed ncurses
534 (static var echo) Rename to `do_echo' because the original name is
535 the same as an ncurses identifier.
536 (get_termcap_viewport) Use error instead of msg.
538 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
539 (fn_close): Don't close stdin, stdout, stderr.
541 2007-07-26 John Darrington <john@darrington.wattle.id.au>
543 * procedure.c procedure.h: Added callbacks which get invoked whenever
544 a dataset's transformation chain changes.
546 2007-07-24 Ben Pfaff <blp@gnu.org>
549 * sys-file-writer.c (write_variable_display_parameters): Use new
550 var_default_display_width function to choose display width of
551 segments after the first one in a given variable.
552 * variable.c (var_create): Use var_default_display_width to pick
553 new variable's display width.
554 (var_default_display_width): New function.
555 Reviewed by John Darrington.
557 2007-07-24 Ben Pfaff <blp@gnu.org>
560 * por-file-writer.c (write_variables): Write weight variable.
561 Reviewed by John Darrington.
563 2007-07-23 Ben Pfaff <blp@gnu.org>
565 Improvements to system file reader and writer.
567 First, move all detailed knowledge of very long strings into
568 sys-file-private.[ch], so that this nasty stuff can be isolated.
570 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
571 (EFFECTIVE_VLS_CHUNK): New macro.
572 (min_int): New function.
573 (max_int): New function.
574 (sfm_width_to_bytes): Rewrite.
575 (sfm_width_to_octs): New function.
576 (sfm_segment_alloc_width): New function.
577 (sfm_segment_alloc_bytes): New function.
578 (sfm_segment_used_bytes): New function.
579 (sfm_segment_offset): New function.
580 (sfm_segment_effective_offset): New function.
581 (sfm_dictionary_to_sfm_vars): New function.
583 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
584 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
585 (struct sfm_var): New structure.
587 Next, improvements to the system file reader.
589 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
590 type casenumber. Added `version_major', `version_minor',
593 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
594 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
595 Change `case_cnt' to type casenumber. Removed `has_vls'.
596 (struct sfm_var): Removed.
597 (sfm_open_reader): Don't warn on wrong case size if the file was
598 written by SPSS 13, which tends to get it wrong. Use
599 sfm_dictionary_to_sfm_vars.
600 (read_header): Always output system file info.
601 (read_variable_record): Simplify code for reading missing values.
602 (read_machine_int32_info): Save version numbers from system file
603 into info struct passed as new argument.
604 (read_long_string_map): Restructured to use new sys-file-private
606 (read_value_labels): Use size_overflow_p.
607 (sys_file_casereader_read): Get rid of distinction between fast
608 and slow paths. Use information provided by sys-file-primate's
609 struct sfm_var to simplify code.
610 (skip_whole_strings): New function.
611 (read_int32): Renamed read_int. Changed return value to int.
613 (read_flt64): Renamed read_float. Changed return value to
614 double. Updated all callers.
615 (int32_to_native): Removed. Changed callers to use
617 (flt64_to_double): Removed. Changed callers to use float_convert.
619 Finally, get rid of int32, flt64 terminology and types in system
620 file writer. The former wasn't very useful since a POSIX "int"
621 can hold the whole range of int32 and we generally didn't have a
622 need for it to be exactly-32-bits, just at-least-32-bits. The
623 latter was inconvenient because we had to assume that it could be
624 different from double and thereby convert special values SYSMIS,
625 HIGHEST, LOWEST to and from it in multiple places. Instead, now
626 we just use "int" and "double" in most places, and do conversions,
627 if necessary, very close to where we do I/O. This change meant
628 that the writer code couldn't represent records in the file as C
629 structs any longer, but that's no great loss. The code actually
630 seems to be more readable without them.
632 Simplify the compression buffering code: only buffer as much as
633 necessary, which is no more than eight 8-byte units at any given
636 * sys-file-writer.c (typedef flt64): Removed.
637 (macro second_lowest_flt64): Removed.
638 (struct sysfile_header): Removed.
639 (struct sysfile_variable): Removed.
640 (struct sfm_writer): Removed `needs_translation', `has_vls',
641 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
642 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
643 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
644 compression buffering by `opcodes', `opcode_cnt', `data',
645 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
646 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
647 data writer functions instead of structures.
648 (calc_oct_idx): New function.
649 (write_header): Use simple data writer functions instead of
651 (write_format_spec): Renamed write_format. New argument.
652 (write_variable_continuation_records): New function.
653 (write_variable): Use simple data writer functions instead of
654 structures. Use write_variable_continuation_records. Write
655 entire very long string instead of requiring caller to understand
657 (write_value_labels): Use simple data writer functions instead of
659 (write_documents): Ditto.
660 (write_variable_display_parameters): Use sys-file-private
661 functions to simplify. Use simple data writer functions instead
663 (write_vls_length_table): Use simple data writer functions instead
665 (write_longvar_table): Ditto.
666 (write_rec_7_34): Break into new functions
667 write_integer_info_record, write_float_info_record. Use simple
668 data writer functions instead of structures.
669 (buf_write): Removed.
670 (append_string_max): Removed.
671 (ensure_buf_space): Removed.
672 (sys_file_casewriter_write): Get rid of the distinction between
673 fast and slow paths, which didn't seem to be too useful. Use new
674 functions write_case_uncompressed, write_case_compressed.
675 (put_instruction): Removed.
676 (put_element): Removed.
677 (write_compressed_data): Removed.
678 (close_writer): Use flush_compressed. Only write case count to
679 system file if it will fit in the field.
680 (write_case_compressed): New function.
681 (write_case_uncompressed): New function.
682 (flush_compressed): New function.
683 (put_cmp_opcode): New function.
684 (put_cmp_number): New function.
685 (write_int): New function.
686 (convert_double_to_output_format): New function.
687 (write_float): New function.
688 (write_value): New function.
689 (write_string): New function.
690 (write_bytes): New function.
691 (write_zeros): New function.
692 (write_spaces): New function.
694 Reviewed by John Darrington.
696 2007-07-22 Ben Pfaff <blp@gnu.org>
698 Don't try to write very long strings to portable files. The
699 format does not support it.
701 * por-file-writer.c (MAX_POR_WIDTH): New macro.
702 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
703 (write_format): Add arg to take width to resize format to.
704 (write_value): Limit width of value written to MAX_POR_WIDTH.
705 (write_variables): Limit width of variable and its output formats
707 Reviewed by John Darrington.
709 2007-07-22 Ben Pfaff <blp@gnu.org>
711 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
712 local variable instead of literal 5.
713 Reviewed by John Darrington.
715 2007-07-22 Ben Pfaff <blp@gnu.org>
717 Fix problems with uniqueness of short names in system files with
718 very long string variables. Now a variable may have multiple
721 * automake.mk (src_data_libdata_a_SOURCES): Add new files
722 short-names.c, short-names.h.
724 * dictionary.c (dict_clone): Clone all the short names.
725 (compare_strings): Move into short-names.c.
726 (hash_strings): Ditto.
727 (set_var_short_name_suffix): Ditto.
728 (dict_assign_short_names): Ditto, rename short_names_assign,
729 change to assign all short names.
731 * por-file-writer.c (write_variables): Use short_names_assign
732 instead of dict_assign_short_names.
734 * short-names.c: New file.
736 * short-names.h: New file.
738 * sys-file-private.c (sfm_width_to_segments): New function.
740 * sys-file-reader.c (read_long_var_name_map): Save and restore all
741 the short names, not just the first one.
743 * sys-file-writer.c (cont_var_name): Removed.
744 (sfm_open_writer): Use short_names_assign instead of
745 dict_assign_short_names. Use unique short names assigned by
746 short_names_assign instead of those generated by cont_var_name.
748 * variable.c (struct variable): Remove `short_name' member,
749 replace by `short_names' and `short_name_cnt'.
750 (var_create) Initialize new members.
751 (var_get_short_name_cnt): New function.
752 (var_get_short_name): Now takes an index argument. Changed most
754 (var_set_short_name): Ditto.
755 (var_clear_short_name): Renamed var_clear_short_names, changed to
756 clear all short names.
758 Reviewed by John Darrington.
760 2007-07-22 Ben Pfaff <blp@gnu.org>
762 * variable.c (var_set_width): Use new var_set_width function.
764 * missing-values.c (mv_n_values): Drop assertion, which was not
767 * format.c (fmt_default_for_width): New function.
768 (fmt_resize): New function.
770 Reviewed by John Darrington.
772 2007-07-18 John Darrington <john@darrington.wattle.id.au>
774 * datasheet.c (datasheet_delete_columns): Added assertion to check
775 we're not deleting outside the range of the sheet.
778 * dictionary.c dictionary.h variable.c: Added the ability for string
779 variables to be resized.
781 * vardict.h: Added some prototypes (moved from dictionary.h) as
782 these should only be called by variable.c
785 2007-07-14 John Darrington <john@darrington.wattle.id.au>
787 * sfm-reader.c: Respect case_cnt field in file header.
789 2007-07-01 John Darrington <john@darrington.wattle.id.au>
791 * transformation.c transformation.h (trns_chain_execute): Changed the
792 signature (Patch #6057)
794 2007-06-10 Ben Pfaff <blp@gnu.org>
796 * casereader-filter.c (casereader_filter_destroy): Make sure to
797 write all the remaining excluded cases to the casewriter, if any.
799 * caseinit.c (init_list_destroy): Rewrite.
800 (init_list_clear): Ditto.
802 * casegrouper.c (casegrouper_get_next_group): Always set *reader
803 to null when returning false.
805 2007-06-06 Ben Pfaff <blp@gnu.org>
807 Actually implement the new procedure code and adapt all of its
808 clients to match. Also adapt all of the other case sources and
809 sinks in the tree and their clients to use the
810 casereader/casewriter infrastructure.
812 * automake.mk: Add and remove files.
814 * any-reader.c: Change into a casereader.
815 * por-file-reader.c: Ditto.
816 * scratch-reader.c: Ditto.
817 * sys-file-reader.c: Ditto.
819 * any-writer.c: Change into a casewriter.
820 * por-file-writer.c: Ditto.
821 * scratch-writer.c: Ditto.
822 * sys-file-writer.c: Ditto.
824 * procedure.c: Change to use casereader, casewriter, caseinit, and
825 other new infrastructure.
827 * scratch-handle.c: Adapt to new infrastructure.
829 * case-sink.c: Removed, now dead code.
830 * case-sink.h: Ditto.
831 * case-source.c: Ditto.
832 * case-source.h: Ditto.
833 * casefile-factory.c: Ditto.
834 * casefile-private.h: Ditto.
837 * casefilter.c: Ditto.
838 * casefilter.h: Ditto.
841 * fastfile-factory.c: Ditto.
842 * fastfile-factory.h: Ditto.
843 * storage-stream.c: Ditto.
844 * storage-stream.h: Ditto.
846 2007-06-06 Ben Pfaff <blp@gnu.org>
850 * automake.mk: Add new files.
852 * datasheet.c: New file.
854 * datasheet.h: New file.
856 2007-06-06 Ben Pfaff <blp@gnu.org>
858 Until now, the procedure code has provided a case to the
859 case_source, which has filled in the data values that come from
860 the active file. "Left" data values that don't come from the
861 active file naturally stay the same from case to case, because the
862 procedure code keeps using that same case.
864 One of the compromises that comes with the new procedure code is
865 that the active file allocates and provides its own case, which
866 the procedure code then has to resize to provide room for any
867 other variables that should go in the case and then fill in the
868 values of "left" variables. Then, when we're done with that case,
869 we have to save the values of "left" variables to copy into the
870 next case read from the active file.
872 The caseinit code helps with this.
874 * automake.mk: Add new files.
876 * caseinit.c: New file.
878 * caseinit.h: New file.
880 2007-06-06 Ben Pfaff <blp@gnu.org>
882 * value.h (value_cnt_from_width): New function.
884 * variable.c (var_get_value_cnt): Use new function.
886 2007-06-06 Ben Pfaff <blp@gnu.org>
888 Add casegrouper, to allow cases read from a given casereader to be
889 broken into groups, each of which has its own casereader.
890 Generally cases are grouped based on having equal values for some
893 * automake.mk: Add new files.
895 * casegrouper.c: New file.
897 * casegrouper.h: New file.
899 2007-06-06 Ben Pfaff <blp@gnu.org>
901 Add interface to lexicographical ordering of cases.
903 * automake.mk: Add new files.
905 * case-ordering.c: New file.
907 * case-ordering.h: New file.
909 2007-06-06 Ben Pfaff <blp@gnu.org>
911 Add casereaders and casewriters, the basis of the new data processing
912 implementation. A casereader is a uniform interface to reading cases
913 from a data source; a casewriter is a uniform interface to writing
914 cases to a data sink.
916 * automake.mk: Add new files.
918 * casereader-filter.c: New file.
920 * casereader-provider.h: New file.
922 * casereader-translator.c: New file.
924 * casereader.c: New file.
926 * casereader.h: New file.
928 * casewriter-provider.h: New file.
930 * casewriter-translator.c: New file.
932 * casewriter.c: New file.
934 * casewriter.h: New file.
936 2007-06-06 Ben Pfaff <blp@gnu.org>
938 "casewindow" data structure that extends the deque (from libpspp)
939 of cases with the ability to dump cases to disk if we get too many
942 * automake.mk: Add new files.
944 * casewindow.c: New file.
946 * casewindow.h: New file.
948 2007-06-06 Ben Pfaff <blp@gnu.org>
950 sparse_cases data structure that augments a sparse_array of cases
951 with the ability to dump cases to disk if we get too many cases in
954 * automake.mk: Add new files.
956 * sparse-cases.c: New file.
958 * sparse-cases.h: New file.
960 2007-06-06 Ben Pfaff <blp@gnu.org>
962 Adds a low-level on-disk case array data structure.
964 * automake.mk: Add new files.
966 * case-tmpfile.c: New file.
968 * case-tmpfile.h: New file.
970 2007-06-06 Ben Pfaff <blp@gnu.org>
972 In a couple of places we calculate the maximum number of cases to
973 keep in memory based on the user-defined workspace. Enable
974 centralizing the calculation through a new function.
976 * settings.c (get_workspace_cases): New function.
978 2007-06-06 Ben Pfaff <blp@gnu.org>
980 The casenumber type is defined in transformations.h, but case.h is
981 a more sensible place. Move it.
983 * case.h (CASENUMBER_MAX): New macro.
984 (typedef casenumber): Move here, from transformations.h.
986 2007-06-03 Ben Pfaff <blp@gnu.org>
988 Slightly generalize case_to_values and case_from_values functions.
990 * case.c (case_to_values): Rename case_copy_out, change interface.
991 (case_from_values): Rename case_copy_in, change interface.
993 * fastfile.c (fastfilereader_get_next_case): Update caller.
994 (write_case_to_disk): Ditto.
996 2007-06-02 Ben Pfaff <blp@gnu.org>
998 Clean up after a forgotten part of patch #5829.
1000 * casedeque.h: Remove unused file.
1002 * automake.mk: Remove casedeque.h from sources.
1004 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1006 * category.c: Removed redundant #include
1008 2007-05-06 Ben Pfaff <blp@gnu.org>
1010 Abstract the documents within a dictionary a little better.
1011 Thanks to John Darrington for suggestion, initial version, and
1012 review. Patch #5917.
1014 * dictionary.c (struct dictionary): Change `documents' member from
1015 char * to struct string.
1016 (dict_clear): Destroy struct string.
1017 (dict_get_documents): Convert struct string to char *.
1018 (dict_set_documents): Set struct string. Pad to 80-character
1020 (dict_clear_documents): New function.
1021 (dict_add_document_line): New function.
1022 (dict_get_document_line_cnt): New function.
1023 (dict_get_document_line): New function.
1025 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1027 * sys-file-reader.c (read_documents): Use new document functions.
1029 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1031 * sys-file-reader.c: When reading a system file which has no
1032 long name table, automatically create one where the long names
1033 are the lower case versions of the short names.
1035 2007-04-22 Ben Pfaff <blp@gnu.org>
1037 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1038 dict_clear will free most data related to the dictionary.
1039 dict_clear does a decent job, except that dict_set_split_vars on
1040 some systems won't actually free the dict's "split" member.
1041 Instead, it'll allocate a 1-byte region. Fix this.
1043 * value.c (value_copy): New function.
1044 (value_set_missing): Ditto.
1046 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1048 * Deleted existing category.h and moved cat-routines.h into
1049 category.h Encapsulated struct cat_vals better.
1051 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1053 * sys-file-reader.c: When reading a system file which has no
1054 long name table, automatically create one where the long names
1055 are the lower case versions of the short names.
1057 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1059 * sys-file-reader.c: Some versions of Other Software seem to
1060 produce system files with string variables' measure set to
1061 zero. We'll assume these are supposed to be nominal variables.
1063 2007-03-30 Ben Pfaff <blp@gnu.org>
1065 * procedure.c: Adapt to new deque data structure.
1067 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1068 Ben Pfaff <blp@gnu.org>
1070 * file-name.c: Mingw compatibility fixes.
1071 (fn_search_path): Use ISSLASH instead of comparing against '/'
1073 (fn_dir_name): Use dir_name from gnulib.
1074 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1075 (fn_get_identity): Use GetFullPathName instead of canonicalize
1076 from gnulib, because the latter does not fully support
1077 Windows-style path names. Use this implementation based on the
1078 detected presence of Windows instead of the absence of Unix, since
1079 the new implementation is Windows-specific.
1080 (fn_compare_file_identities): In Windows implementation, compare
1081 names case-insensitively.
1083 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1085 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1086 <stdlib.h> to have the same effect.
1088 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1090 * por-file-reader.c: Add missing _() around messages.
1092 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1094 * make-file.c: Include "mkstemp.h", without which linking on
1097 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1099 Reduce platform dependence.
1101 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1102 Everywhere we using this, we really should have just depended on
1103 the shell to expand tildes.
1104 (fn_search_path): Simplify, given that we don't do tilde expansion
1106 (fn_normalize): Removed. Caller changed to use the canonicalize
1108 (fn_get_cwd): Removed. Only user was fn_normalize.
1109 (fn_is_absolute): Really only test for absolute names.
1110 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1112 (fn_readlink): Removed, as it was only used fn_normalize.
1113 (fn_exists): Assume the stat function is available; gnulib does.
1114 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1116 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1118 * dictionary.c (dict_create_vector_assert): New function.
1120 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1122 * file-name.c (fn_normalize): Correct name of function
1123 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1124 for pointing this out.
1126 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1128 We are using a single member in struct file_handle, the "name"
1129 field, for more than one purpose. When it begins with '"', it's a
1130 file name; otherwise, it's a token that can be used to identify
1131 it. When that assertion fires, it's because we searched for the
1132 name case-sensitively as a file name (so that there was no match),
1133 and then we try to insert it case-insensitively as a token, which
1134 fails because duplicates aren't allowed.
1136 Solution: break the two purposes into two separate fields. This
1137 fixes the problem and likely makes the code easier to read too.
1139 Fixes bug #18922. Thanks to John Darrington for bug report and
1142 * file-handle-def.c (struct file_handle): New `id' member.
1143 (fh_from_name): Rename fh_from_id. Update all callers.
1144 (create_handle): New `id' parameter. Update all callers.
1145 (fh_create_file): Ditto.
1146 (fh_get_id): New function.
1148 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1150 * case.c (case_is_null): Change return type to bool.
1152 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1156 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1157 reference counting to share data. This makes it easy for
1158 valgrind, etc. to find accesses to cases that have been destroyed
1159 but have been kept around by another user's ref-count. This often
1160 happens when the data set is small enough to find in memory; if a
1161 bigger data set that would overflow to disk were used, then data
1162 corruption would occur.
1164 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1168 * case.c (case_unshare): Make it check internally whether the
1169 ref_cnt is greater than 1, so that the callers don't have to.
1170 Update callers not to check.
1172 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1174 Before, I was thinking that I might want to get rid of reference
1175 counting at some point. Now, I'm pretty sure that it's here to
1176 stay. Thus, because we have to store the value_cnt anyway for
1177 reference-counted cases, we might as well expose it to users.
1179 * case.c (case_get_value_cnt): New function.
1180 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1181 resize case if its size actually changed.
1183 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1184 instead of peeking inside struct case directly.
1185 (casefile_append): Ditto.
1187 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1189 Get rid of the inlines for the case functions, which made the
1190 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1191 the inlines didn't speed up "make check" at all, which is not a
1192 perfect benchmark but seems indicative.)
1194 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1195 definitions. Remove some assertions on nonnull pointers that were
1196 redundant with a pointer dereference soon after in the function.
1198 (struct case_data): Move definition here from case.h.
1202 (case_data_wr): Ditto.
1204 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1206 * automake.mk: Add casedeque.h to sources.
1208 * casedeque.h: New file.
1210 * procedure.c: (struct dataset) Change lag_count, lag_head,
1211 lag_queue member into single struct casedeque member. Update all
1212 users to use the casedeque instead.
1215 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1217 * procedure.c: Simplify lagged cases interface. Updated all
1218 clients--well, the only client--to use the simplified interface.
1219 (dataset_n_lag) Removed.
1220 (dataset_set_n_lag) Removed.
1221 (dataset_need_lag) New function.
1223 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1225 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1226 whenever appropriate, but only when the dataset/dictionary is in a
1229 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1231 * dictionary.c dictionary.h : Added callbacks for change of filter and
1232 split variables. Refactored some code to ensure that callbacks get
1233 invoked when appropriate.
1235 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1236 replace_dict callback occurs when permanent_dict replaces the current
1239 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1241 * dictionary.c dictionary.h : Added callback for when the weight
1242 variable of a dictionary changes.
1244 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1246 * dictionary.c dictionary.h : Added replace_source and replace_dict
1247 callbacks, and functions to deal with them.
1249 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1251 Simplify missing value handling.
1253 * missing-values.h (enum mv_class): New type.
1254 (enum mv_type): Moved definition into missing-values.c and renamed
1255 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1256 enums. Updated all uses.
1257 (struct missing_values): Changed type of `type' from `enum
1258 mv_type' to `int' because the definition is no longer exposed.
1260 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1261 parameter. Update all callers.
1262 (mv_is_num_missing): Ditto.
1263 (mv_is_str_missing): Ditto.
1264 (mv_is_value_user_missing): Removed. Changed callers to use
1265 mv_is_value_missing.
1266 (mv_is_num_user_missing): Removed. Changed callers to use
1268 (mv_is_str_user_missing): Removed. Changed callers to use
1270 (mv_is_value_system_missing): Removed. Changed callers to use
1271 mv_is_value_missing.
1272 (mv_set_type): Removed. Changed callers to use mv_clear.
1273 (mv_clear): New function.
1275 * variable.c (var_is_value_missing): Add new enum mv_class
1276 parameter. Update all callers.
1277 (var_is_num_missing): Ditto.
1278 (var_is_str_missing): Ditto.
1279 (var_is_value_user_missing): Removed. Changed callers to use
1280 var_is_value_missing.
1281 (var_is_num_user_missing): Removed. Changed callers to use
1283 (var_is_str_user_missing): Removed. Changed callers to use
1285 (var_is_value_system_missing): Removed. Changed callers to use
1286 var_is_value_missing.
1288 * casefilter.c (struct casefilter): Use enum mv_class in place of
1290 (casefilter_variable_missing): Adapt to new member.
1291 (casefilter_create): Change signature to take enum mv_class,
1294 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1296 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1298 * case-sink.c case-sink.h procedure.c procedure.h
1299 storage-stream.c: Now uses the factory.
1301 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1303 Make it possible to pull cases from the active file with a
1304 function call, instead of requiring indirection through a callback
1307 * case-source.h (struct case_source_class): Change ->read function
1308 to return a single case, instead of calling a callback function
1309 for each case. Change ->destroy function to return an error
1312 * case-source.c (free_case_source): Pass along the value returned
1313 by the case_source ->destroy function.
1315 * procedure.c (struct write_case_data): Removed.
1316 (struct dataset): Added some members to track procedure state.
1317 (procedure): Optimize the trivial case at this level.
1318 (internal_procedure): Re-implement in terms of proc_open,
1319 proc_read, proc_close.
1320 (proc_open) New function.
1321 (proc_read) New function.
1322 (proc_close) New function.
1323 (write_case) Moved into proc_read.
1324 (close_active_file) Moved closing of data source into proc_close.
1326 * storage-source.c: Rewrote to conform with modified
1327 case_source_class interface.
1329 * transformations.c (trns_chain_execute): Added argument to allow
1330 starting execution from an arbitrary transformation. Updated
1333 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1335 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1337 * sys-file-reader.c (read_display_parameters): Don't assume that
1338 MEASURE_* and ALIGN_* have the same values found in system files.
1340 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1342 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1343 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1344 the value of n_MEASURES, which was off by 1 (at least from my
1347 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1349 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1350 callbacks which are invoked when the dictionary or its
1351 variables are changed.
1353 * missing-values.c missing-values.h value-labels.c: Tidied up
1354 consistency checks, and made some of them return false
1355 instead of assert-failing.
1357 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1359 * calendar.c (calendar_days_in_month): New function.
1361 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1363 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1364 the value label's width.
1366 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1368 * sfm-private.h: Move contents into sys-file-writer.c, which is
1369 the only remaining user. Removed Borland C++-specific directives.
1371 * sys-file-reader.c: Clean up and rewrite entire file. The
1372 rewritten version is simpler and better abstracted, and should be
1373 easier to maintain and extend. It avoids using structures to read
1374 file data, which is prone to padding variations among compilers.
1375 It should also handle non-IEEE 754 system files, although I
1376 haven't been able to find any. It has been tested against many
1377 .sav files obtained from the Web and found to produce the same
1378 results as the earlier version of the code, or in some cases
1379 improved results. It is more tolerant of format variations found
1382 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1383 member, putting an enum integer_format in its place. New member
1384 `float_format'. Changed `compressed' member to type bool.
1386 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1388 * dictionary.c (dict_delete_consecutive_vars): New function.
1390 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1392 * file-name.c (fn_search_path): Remove prefix arg that was unused
1393 by any caller. Updated all callers.
1395 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1397 * format.c (fmt_dollar_template): Use user's decimal point
1398 character. Add assertion.
1400 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1402 * format.c (fmt_dollar_template): New function, based on
1403 dollar_format_template from var-type-dialog.c.
1405 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1407 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1410 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1412 Finish converting struct variable to an opaque type. In this
1413 phase, we add remaining setter and getter functions, convert the
1414 remaining PSPP code to use them, and do a bunch of cleanup. The
1415 resulting changes are pervasive but mostly trivial, and only the
1416 notable changes are logged.
1418 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1421 * case.c (case_data): Renamed case_data_idx.
1422 (case_num): Renamed case_num_idx.
1423 (case_str): Renamed case_str_idx.
1424 (case_data_rw): Renamed case_data_rw_idx.
1426 * case.h (case_data): New function with old name and an interface
1427 that takes a variable instead of an index, which is easier to
1428 use. Updated all callers to use the new interface, or to use the
1429 new *_idx function (see above).
1432 (case_data_rw): Ditto.
1434 * category.c (cat_stored_values_destroy): Changed interface to
1435 take a struct cat_vals * instead of a struct variable *.
1437 * dictionary.c (dict_clone): Use new vector_clone function.
1438 (dict_clear) Use new var_destroy function.
1439 (add_var) New function.
1440 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1441 (dict_create_var_assert) Rewrite in terms of add_var.
1442 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1443 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1444 (dict_lookup_var) Use new var_create, var_destroy functions.
1445 (dict_contains_var) Rewrite in terms of new vardict functionality.
1446 (set_var_dict_index) New function.
1447 (set_var_case_index) New function.
1448 (reindex_vars) New function.
1449 (dict_delete_var) Rewrite in terms of new vardict functionality.
1450 (dict_reorder_var) Ditto.
1451 (dict_reorder_vars) Ditto.
1452 (rename_var) New function.
1453 (dict_rename_var) Use rename_var.
1454 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1455 (dict_get_compacted_idx_to_fv) Rename
1456 dict_get_compacted_dict_index_to_case_index, update callers.
1457 (dict_create_vector) Use new vector_create function.
1458 (dict_clear_vectors) Use new vector_destroy function.
1459 (set_var_short_name_suffix) Move here from variable.c, renamed
1460 from var_set_short_name_suffix, make static, update caller.
1462 * sys-file-private.c: New file.
1463 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1464 width_to_bytes, update callers.
1466 * sys-file-private.h: New file. Later it will supplant
1467 sfm-private.h; for now it supplements it.
1468 (macro MIN_VERY_LONG_STRING) New macro.
1469 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1471 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1472 MAX_LONG_STRING was used before.
1474 * sys-file-writer.c: Ditto.
1476 * value-labels.c: Change the paradigm here to be that a null
1477 pointer is OK for a struct val_labs * in most cases; it just
1478 represents an empty set of value labels.
1479 (val_labs_copy) A copy of a null set is a null set.
1480 (val_labs_count) A null set has 0 labels.
1481 (val_labs_replace) Change return type to void. Rewrite for
1483 (val_labs_find) A null set does not contain the value.
1484 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1485 transposed argument order, updated all callers.
1487 * value.c: New file.
1488 (value_dup) Moved here from variable.c.
1489 (compare_values) Ditto.
1492 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1493 (macro MAX_LONG_STRING) Removed, because it was only interesting
1494 for system files, not for general code.
1495 (macro MAX_VERY_LONG_STRING) Ditto.
1496 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1497 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1499 * vardict.h: New file, for an interface between variables and
1502 * variable.c: A lot of functions were moved around, for better
1504 (struct variable) Move definition here, from variable.h.
1505 (var_type_adj) Removed--makes i18n hard.
1506 (var_type_noun) Ditto.
1507 (var_create) New function.
1508 (var_clone) New function.
1509 (var_destroy) New function.
1510 (var_set_name) Assert that variable is not in a dictionary.
1511 (compare_var_names) Rename compare_vars_by_name and fix a couple
1512 of callers who thought the args were strings.
1513 (hash_var_name) Rename hash_var_by_name.
1514 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1515 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1516 (var_is_very_long_string) Removed, because it was only interesting
1517 to system file code.
1518 (var_set_missing_values) Allow the argument to be the wrong width,
1519 as long as we can resize it. Simplify callers who were doing the
1520 resizing themselves.
1521 (var_get_value_labels) New function.
1522 (var_has_value_labels) New function.
1523 (var_set_value_labels) New function.
1524 (alloc_value_labels) New function.
1525 (var_add_value_label) New function.
1526 (var_replace_value_label) New function.
1527 (var_clear_value_labels) New function.
1528 (var_lookup_value_label) New function.
1529 (var_get_value_name) Moved here from variable.c, renamed from
1530 var_get_value_name, transposed argument order, updated all
1532 (var_to_string) Moved here, from variable-label.c.
1533 (var_set_leave) New function.
1534 (var_get_leave) New function.
1535 (var_must_leave) New function.
1536 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1537 set_var_short_name_suffix.
1538 (var_get_dict_index) New function.
1539 (var_get_case_index) New function.
1540 (var_get_obs_vals) New function.
1541 (var_set_obs_vals) New function.
1542 (var_has_obs_vals) New function.
1543 (var_get_vardict) New function.
1544 (var_set_vardict) New function.
1545 (var_has_vardict) New function.
1546 (var_clear_vardict) New function.
1547 (value_dup) Moved to value.c.
1548 (compare_values) Ditto.
1551 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1552 (enum ALPHA) Rename VAR_STRING, update all users.
1554 * vector.c: New file.
1555 (struct vector) Moved here, from variable.h.
1556 (check_widths) New function.
1557 (vector_create) New function.
1558 (vector_clone) New function.
1559 (vector_destroy) New function.
1560 (vector_get_name) New function.
1561 (vector_get_var) New function.
1562 (vector_get_var_cnt) New function.
1563 (compare_vector_ptrs_by_name) New function.
1565 * vector.h: New file.
1567 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1569 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1570 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1573 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1575 * value-labels.c (destroy_atoms): New function.
1576 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1578 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1580 Thanks to Jason Stover for pointing out this problem.
1582 * data-out.c (output_number): Use gsl_finite from GSL, which is
1583 portable, instead of isfinite, which is not.
1586 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1588 * variable.c variable.h (value_dup): New function.
1590 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1592 Start converting struct variable to an opaque type. In this
1593 phase, we add a bunch of setter and getter functions and convert
1594 most of the PSPP code to use them. The resulting changes are
1595 pervasive but mostly trivial, and only the notable changes are
1598 * format.c (fmt_equal): New function.
1600 * variable.c (var_type_is_valid): New function.
1601 (measure_is_valid) Moved here, from format.c.
1602 (alignment_is_valid) Moved here, from format.c.
1603 (var_get_name) New function.
1604 (var_set_name) New function.
1605 (width_to_type) New function.
1606 (var_get_type) New function.
1607 (var_get_width) New function.
1608 (var_set_width) New function.
1609 (var_is_numeric) New function.
1610 (var_is_alpha) New function.
1611 (var_is_short_string) New function.
1612 (var_is_long_string) New function.
1613 (var_is_very_long_string) New function.
1614 (var_get_missing_values) New function.
1615 (var_set_missing_values) New function.
1616 (var_clear_missing_values) New function.
1617 (var_has_missing_values) New function.
1618 (var_is_value_missing) New function.
1619 (var_is_num_missing) New function.
1620 (var_is_str_missing) New function.
1621 (var_is_value_user_missing) New function.
1622 (var_is_num_user_missing) New function.
1623 (var_is_str_user_missing) New function.
1624 (var_is_value_system_missing) New function.
1625 (var_get_print_format) New function.
1626 (var_set_print_format) New function.
1627 (var_get_write_format) New function.
1628 (var_set_write_format) New function.
1629 (var_set_both_formats) New function.
1630 (var_get_label) New function.
1631 (var_set_label) New function.
1632 (var_clear_label) New function.
1633 (var_has_label) New function.
1634 (var_get_measure) New function.
1635 (var_set_measure) New function.
1636 (var_get_display_width) New function.
1637 (var_set_display_width) New function.
1638 (var_get_alignment) New function.
1639 (var_set_alignment) New function.
1640 (var_get_value_cnt) New function.
1641 (var_get_leave) New function.
1642 (var_get_short_name) New function.
1644 * variable.h: (struct variable) Removed "type" and "nv" members;
1645 they are now computed from "width" where needed.
1647 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1649 * missing-values.c (mv_resize): Don't write beyond end of the
1650 allocated buffer when resizing a long string.
1652 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1654 Clean up identifier code: don't require identifier enumerations to
1655 be in a particular order; make better use of string library;
1656 expose less of the internals.
1658 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1659 change interface. Updated all callers.
1660 (lex_id_match) Change interface to use struct substring, update
1662 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1663 (global array keywords[]) Make static, change form. Update all
1664 users to use lex_id_name instead.
1665 (lex_is_keyword) New function.
1666 (lex_id_to_token) Change interface to use struct substring, update
1668 (lex_id_name) New function.
1670 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1671 lex_is_keyword instead.
1672 (T_LAST_KEYWORD) Removed.
1673 (T_N_KEYWORDS) Removed.
1675 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1677 * format.c: (fmt_date_template) Distinguish characters for which a
1678 space is output and any date delimiter is allowed on input, from
1679 those for which a space is output and only a space is allowed on
1680 input. The former is represented by X, the latter by a space.
1681 Also, drop distinction between h and H, changing the former to the
1684 * data-in.c: Completely rewrite internals to conform to SPSS input
1685 formats as closely as possible.
1686 (data_in) Changed external interface by replacing the structure
1687 that was used as a single argument by a set of arguments. Updated
1689 (data_in_finite_line) Removed. Converted all callers to use plain
1691 (data_in_get_integer_format) New function.
1692 (data_in_set_integer_format) New function.
1693 (data_in_get_float_format) New function.
1694 (data_in_set_float_format) New function.
1696 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1697 (struct data_in) Removed.
1699 * data-out.c: (output_date) Drop each component from the input as
1700 it is output, to allow us to drop the distinction between h (a
1701 count of hours) and H (the hour of day) template characters.
1702 Also, handle new X template character.
1703 (output_scientific) Follow more rational rule on when to drop
1704 fraction introduced between SPSS 13 and 15. Updated test case to
1707 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1709 Fix buffer overflow reported by John Darrington.
1711 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1712 realize that DIGITS is a count of nibbles, not of bytes.
1714 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1716 * calendar.c (calendar_offset_to_gregorian) Also return the
1717 year-of-day. Change callers to new interface.
1719 * data-out.c: Completely rewrite internals to conform to SPSS
1720 output formats as completely as possible.
1721 (data_out) Change interface to put input parameters before output
1722 parameters, for consistency with the style I now prefer. Update
1724 (data_out_get_integer_format) New public function.
1725 (data_out_set_integer_format) New public function.
1726 (data_out_get_float_format) New public function.
1727 (data_out_set_float_format) New public function.
1729 * data-out.h: New file. Move prototype for data_out here, from
1732 * format.c: (fmt_step_width) Use equality comparison instead of
1733 bitwise and, for clarity.
1734 (fmt_is_string) Ditto.
1735 (fmt_input_to_output) Fix categories that are translated to F
1738 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1740 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1741 casefile-private.h: Added casefilter to assist commands with missing
1744 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1746 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1748 * settings.c: (route_errors_to_terminal) New variable.
1749 (route_errors_to_listing) New variable.
1750 (get_error_routing_to_terminal) New function.
1751 (set_error_routing_to_terminal) New function.
1752 (get_error_routing_to_listing) New function.
1753 (set_error_routing_to_listing) New function.
1755 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1757 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1759 * format.c: Completely rewrite, to achieve better abstraction.
1760 Rewrite all references to formats in other files.
1762 * format.def: Rewrite and reorganize.
1764 * settings.c: Move everything related to custom currency formats
1765 into format.[ch], changing them in form, so as to group related
1766 code and definitions better. Changed all references to use the
1768 (static var decimal) Removed.
1769 (static var grouping) Removed.
1770 (static var cc) Removed.
1771 (get_decimal) Removed.
1772 (set_decimal) Removed.
1773 (get_grouping) Removed.
1774 (set_grouping) Removed.
1778 * settings.h: (macro CC_CNT) Removed.
1779 (macro CC_WIDTH) Removed.
1780 (struct custom_currency) Removed.
1782 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1784 * data-in.c (data_in): Use switch statement instead of table, to
1785 avoid dependence on the order of the FMT_* enums.
1787 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1789 * data-out.c: (num_to_string) Removed, because it was dead code.
1791 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1793 * data-in.c (parse_trailer): Fix error message.
1795 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1797 * format.c (fmt_is_binary): New function.
1799 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1801 * procedure.c procedure.h: Encapsulated the static data into a single
1804 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1806 * casefile.c (casereader_read_xfer): Always initialize the case,
1807 even on an error condition.
1809 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1811 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1812 entire function useless.
1814 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1816 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1817 casereader method casereader_clone.
1819 * procedure.c pransformations.h: Introduced new type casenum_t
1821 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1823 * variable.c: (width_to_bytes) Rephrase code for clarify.
1825 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1827 * format.c: (fmt_type_from_string) New function.
1828 (fmt_to_string) Include decimals in output if the format has
1829 decimals, even if the format type does not. This way, we can
1830 accurately reproduce incorrect formats in user output.
1831 (check_common_specifier) Make the check for a bad format type an
1832 assertion, so we get bug reports if they show up. Fix message.
1833 Check for decimal places with a format type that doesn't allow
1835 (check_input_specifier) Remove check for FMT_X, which has been
1837 (check_output_specifier) Ditto.
1839 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1841 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1842 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1843 (enum fmt_parse_flags) Removed.
1845 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1847 * casefile.c casefile.h: Converted to an abstract base class.
1848 * casefile-private.h fastfile.c fastfile.h: New files.
1849 * automake.mk procedure.c scratch-writer.c storage-stream.c
1851 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1853 * procedure.c (internal_procedure): Create sink_case with only as
1854 many values as the compacted dictionary.
1856 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1858 Remove "debugging" code that caused plenty of false positives and
1861 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1863 * case.c: Remove all references to `this' member.
1865 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1867 Fix link error noted by Jason Stover.
1869 * storage-stream.c: Include <assert.h>.
1871 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1873 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1874 ALL) and additional underlying system file issues.
1876 Thanks to John Darrington for review.
1878 First problem: var_hash points to variables not owned by the
1879 sys-file-reader, which the caller may free or modify. Use an
1880 array of sfm_vars instead, as done earlier (e.g. CVS version
1883 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1884 members and remove all code that references it. Add vars, var_cnt
1885 members. Remove fix_specials member, which was unused.
1886 (struct sfm_var) Remove name member, which was unused.
1887 (sfm_close_reader) Free vars member instead of var_hash.
1888 (compare_var_shortnames) Removed.
1889 (hash_var_shortname) Removed.
1890 (sfm_open_reader) Fill out vars array.
1891 (compare_var_index) Removed.
1892 (sfm_read_case) Use vars instead of var_hash.
1894 Second problem: we're confused about when we actually have very
1895 long strings, causing us to choose incorrectly between slow path
1896 and fast path in sfm_read_case.
1898 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1899 have very long strings, not when we have long variable names,
1900 which is an unrelated feature.
1902 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1904 * variable.h: Move var_set and variable parsing declarations to
1905 new header, src/language/lexer/variable-parser.h. Modified lots
1906 of files to include the new header.
1908 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1910 * value-labels.c (value_to_string): When there's no value label,
1911 format the variable according to its print format, instead of
1912 always effectively using A or F format.
1914 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1916 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1919 * format.c: Removed tortological assertion.
1921 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1923 Reform string library.
1925 * file-name.c (fn_interp_vars): Change interface to take a
1926 substring as input. Updated all users.
1928 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1930 * format.c (measure_is_valid): Really return false when m >=
1933 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1935 Implement random access to casefiles, for use in GUI.
1937 * casefile.c: (struct casereader) Add `random', `file_ofs',
1938 `buffer_ofs' members.
1939 (casefile_get_random_reader) New function.
1940 (read_open_file) Break part into new function
1941 seek_and_fill_buffer().
1942 (fill_buffer) Update buffer_ofs, file_ofs.
1943 (casereader_seek) New function.
1945 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1947 * settings.c: Added call to i18n{done, init}.
1949 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1951 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1954 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1956 * casefile.c: Convert many uses of `int' to `bool'.
1958 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1960 * transformations.c (trns_chain_destroy): Destroy chain's trns
1961 member, to fix memory leak.
1963 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1965 * storage-stream.c (storage_source_decapsulate): Destroy case
1966 source to fix memory leak.
1968 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1970 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1971 case passed as argument instead of initializing the argument as a
1972 case. Fixes memory leak that showed up in
1973 tests/command/aggregate.sh with scratch files.
1975 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1977 * procedure.c (proc_done): Destroy default_dict, to fix memory
1980 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1982 Simplify procedure_with_splits().
1984 * procedure.c (struct split_aux_data): Removed case_count member.
1985 (procedure_with_splits) Don't initialize case_count.
1986 (split_procedure_case_func) Check whether prev_case is null
1987 instead of case_count.
1988 (split_procedure_end_func) Ditto.
1990 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1992 * case.c (case_move): Do nothing if dst and src are the same
1994 (case_try_create) Merge two similar cases.
1995 (case_copy) Unshare only if data must be actually copied.
1997 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1999 * data-in.c data-out.c dictionary.c sys-file-reader.c
2000 sys-file-writer.c variable.c variable.h: Reworked very long string
2001 support for better encapsulation.
2003 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2005 * value-labels.c (val_labs_can_set_width): New function.
2006 (val_labs_set_width) Clear labels if increasing width to long
2008 (val_labs_destroy) Remove unneeded test for null.
2010 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2012 * value-labels.h: Remove unneeded dependency on variable.h.
2014 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2016 Get rid of `char *c' member in union value, for cleanliness.
2018 * value.h: (union value) Remove `c' member.
2020 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2022 Make dictionary compacting functions a little more general.
2024 * sys-file-writer.c (sfm_open_writer): Use
2025 dict_compacting_would_change().
2026 (does_dict_need_translation) Removed.
2028 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2030 Make dictionary compacting functions a little more general.
2032 * dictionary.c (dict_needs_compaction): Rename
2033 dict_compacting_would_shrink(). Update all callers.
2034 (dict_compacting_would_change) New function.
2036 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2038 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2039 inverted return value (!).
2041 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2043 Continue reforming procedure execution.
2045 * procedure.c: Search and replace "vfm" by "proc". Notably:
2046 (static var vfm_source) Rename proc_source. Update all
2048 (static var vfm_sink) Rename proc_sink. Update all references.
2050 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2052 Continue reforming procedure execution. In this phase, remove
2053 PROCESS IF, which was deprecated anyway and can be easily
2054 simulated with TEMPORARY followed by SELECT IF.
2056 * procedure.c: (open_active_file) Don't call
2057 add_process_if_trns().
2058 (discard_variables) Get rid of redundant call to
2059 proc_cancel_all_transformations().
2060 (add_process_if_trns) Removed.
2061 (process_if_trns_proc) Removed.
2062 (process_if_trns_free) Removed.
2064 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2066 Continue reforming procedure execution. In this phase, add
2067 `const' to the case passed to procedure()'s callback.
2069 Updated all users of procedure() as well.
2071 * procedure.c: (struct write_case_data) Add "const" to ccase
2072 parameter for case_func member.
2073 (procedure) Add "const" to ccase parameter for proc_func
2075 (multipass_case_func) Make ccase parameter const.
2076 (internal_procedure) Add "const" to ccase parameter for case_func
2078 (split_procedure_case_func) Make ccase parameter const.
2079 (multipass_split_case_func) Make ccase parameter const.
2081 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2083 Continue reforming procedure execution. In this phase, get rid of
2084 the output code for SPLIT FILE groups in procedure.c, which really
2085 shouldn't be doing any output. Move it into the individual
2086 procedures instead. This also adds some flexibility.
2088 Updated many users of procedure_with_splits() and
2089 multipass_procedure_with_splits() to call
2090 output_split_file_values() and to deal with increased use of
2093 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2094 parameter to begin_func member.
2095 (procedure_with_splits) Add "const struct ccase *" parameter to
2096 begin_func parameter. Make ccase parameter const in proc_func
2098 (split_procedure_case_func) Don't dump split file group. Pass
2100 (dump_splits) Moved to language/dictionary/split-file.c as
2101 output_split_file_values().
2102 (struct multipass_split_aux_data) Add "const struct ccase *"
2103 parameter to split_func member.
2104 (multipass_procedure_with_splits) Add "const struct ccase *"
2105 parameter to split_func parameter.
2106 (multipass_split_case_func) Save new SPLIT FILE case before
2108 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2110 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2112 Continue reforming procedure execution. Change
2113 internal_procedure() so that it calls open_active_file() and
2114 close_active_file(), which isolates most of the actual procedure
2117 * procedure.c: (struct write_case_data) Rename `proc_func' member
2118 to `case_func' and update all references.
2119 (procedure) Rewrite as one-line wrapper around
2120 internal_procedure().
2121 (struct multipass_aux_data) New.
2122 (multipass_callback) Renamed multipass_case_func(). Use struct
2123 multipass_aux_data as auxiliary data.
2124 (multipass_end_func) New function.
2125 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2126 that uses multipass_case_func, multipass_end_func.
2127 (internal_procedure) Add `end_func' argument. Move optimization
2128 of trivial case in here. Move call to open_active_file() and
2129 close_active_file() in here. Now assert that vfm_source is
2131 (procedure_with_splits_callback) Rename
2132 split_procedure_case_func().
2133 (split_procedure_end_func) New function.
2134 (multipass_split_callback) Rename multipass_split_case_func.
2135 (multipass_split_end_func) New function.
2136 (discard_variables) No need to test for nonnull vfm_source.
2138 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2140 Continue reforming procedure execution. Get rid of unused member.
2142 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2144 (write_case) Don't increment cases_analyzed.
2146 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2148 Continue reforming procedure execution. In this phase, move
2149 procedure.c and procedure.h from src to src/data. Update
2150 makefiles and #includes accordingly.
2152 * procedure.c: Moved here from src/.
2154 * procedure.h: Moved here from src/.
2156 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2158 Continue reforming procedure execution. In this phase, get rid of
2159 many global variables, consolidating procedure execution in
2160 procedure.c. Encapsulate transformations in new "struct
2161 trns_chain". Also, change implementation of N OF CASES, FILTER,
2162 and PROCESS IF from special cases to transformations.
2164 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2167 * dictionary.c: (global variable default_dict) Move to
2170 * variable.h: (TRNS_*) Move to transformations.h.
2171 (struct transformation) Move to transformations.c.
2173 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2175 * sys-file-reader.c: Fixed invalid read problems.
2177 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2179 * storage-stream.c: Add missing function comments.
2181 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2183 Continue reforming procedure execution. In this phase, add some
2184 new, needed functionality to storage-stream.
2186 * storage-stream.c: (storage_source_decapsulate) New function.
2188 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2190 * variable.c (width_to_bytes): Declarations must precede
2191 statements for C90 compliance.
2193 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2195 * data-out.c, data-in.c, variable.c, variable.h: New functions
2196 copy_mangle and copy_demangle for reading/writing cases; emulates the
2197 way SPSS deals with strings > 255 bytes.
2199 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2200 subtype 14 needed for strings longer than 255 bytes.
2202 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2203 instead of literal values. Also fixed some constness issues.
2205 * format.h: Constness
2207 * sfm-private.h: Renamed the case_size identifier, since I discovered
2208 that SPSS's respect for this variable is very nominal.
2210 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2212 Change case limit type from int to size_t.
2214 * dictionary.c: (struct dictionary) Change type of case_limit
2216 (dict_get_case_limit) Change return type.
2217 (dict_set_case_limit) Change parameter type.
2219 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2221 * variable.h: (struct variable) Rename `reinit' member as `leave'
2222 and invert sense. Fix up all references.
2224 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2226 Continue reforming procedure execution. In this phase, break
2227 procedure.c into multiple files.
2229 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2231 * case-sink.c: New file.
2233 * case-sink.h: New file.
2235 * case-source.c: New file.
2237 * case-source.h: New file.
2239 * storage-stream.c: New file.
2241 * storage-stream.h: New file.
2243 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2245 * variable.h: (struct variable) Remove `init' member and all
2246 references to it from other files. It was initialized in several
2247 places, but nothing really ever used it for anything worthwhile.
2248 Thanks to Jason Stover for pointing out how confusing this
2251 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2253 Continue reforming error message support. In this phase, get rid
2254 of message "titles" and put the message text in `struct error'.
2255 Now `struct error' encapsulates a message more properly.
2257 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2258 Format message ourselves.
2260 * data-in.c: (vdls_error) Ditto.
2262 * por-file-reader.c: (error) Ditto.
2264 * sys-file-reader.c: (corrupt_msg) Ditto.
2266 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2268 GNU standards require "file name" instead of "filename" in
2269 documentation. It's nice for our code to follow the convention
2272 * casefile.c: (struct casefile) Rename `filename' member to
2273 `file_name'. Updated all references.
2275 * file-name.c: [!unix] (struct file_identity) Rename
2276 normalized_filename member to normalized_file_name. Updated all
2279 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2281 We don't really support anything but Unix-like environments well,
2282 so we might as well de-obfuscate by writing directory and path
2283 separators explicitly.
2285 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2287 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2288 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2290 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2293 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2295 GNU standards require "file name" instead of "filename" in
2296 documentation. It's nice for our code to follow the convention
2299 * filename.c: Rename to file-name.c.
2301 * filename.h: Rename to file-name.h. Update all inclusions.
2302 Update header guards.
2304 * automake.mk: Update file names.
2306 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2308 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2310 (fn_basename) Removed (dead code).
2311 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2312 (fn_special_p) Renamed fn_is_special(), all references updated.
2313 (fn_exists_p) Renamed fn_exists(), all references updated.
2315 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2317 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2318 Also, now it only tilde-expands file names, not paths.
2319 (fn_search_path) Tilde-expand one directory at a time.
2321 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2323 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2324 get rid of non-Unixlike version of the code, which has probably
2326 (fn_prepend_dir) Removed (dead code).
2328 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2329 (macro PATH_SEPARATOR_STRING) New.
2330 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2332 Continue reforming error message support. In this phase, we get
2333 rid of VM() and the other msg() support for "verbosity", replacing
2334 it by a new function verbose_msg().
2336 * filename.c: (fn_search_path) Use verbose_msg() instead of
2339 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2341 * sfm-private.h: Get rid of #defines after #error, which makes no
2344 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2346 Get rid of our own int32 type in favor of the standard int32_t
2349 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2352 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2354 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2356 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2358 Remove ill-considered file routines that are no longer used.
2360 * filename.c: (fn_open_ext) Removed.
2361 (fn_close_ext) Removed.
2363 * filename.h: (struct file_ext) Removed.
2365 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2367 * variable.c (var_is_valid_name): Move declarations before code
2370 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2372 * filename.ch (fn_interp_vars): Fixed small buglet.
2374 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2376 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2377 so as to modify the string inline. Thus makeing it easier to
2378 destroy the results when no longer needed.
2380 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2382 * category.c (cat_stored_values_destroy): Fixed memory leak.
2384 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2386 Add some missing frees. Thanks to John Darrington for reporting
2389 * any-writer.c (any_writer_close): Free writer.
2391 * any-reader.c (any_reader_close): Free reader.
2393 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2395 * por-file-reader.c: (error) Mark as NO_RETURN.
2397 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2399 * settings.c: Changed default value of scompress to true.
2401 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2403 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2405 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2408 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2409 here from pref.h.orig.
2411 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2413 * sys-file-reader.c: Fixed bug reading compressed files.
2415 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2417 * Numerous renames. See src/ChangeLog for details.
2419 * Moved files from src directory