1 2008-02-01 Ben Pfaff <blp@gnu.org>
3 Patch #6386. Thanks to John Darrington for review and for the
4 updates to gnumeric-reader.c.
6 * dictionary.c (make_hinted_name): New function.
7 (make_numeric_name): New function.
8 (dict_make_unique_var_name): New function.
10 * gnumeric-reader.c (devise_name): Removed.
11 (munge_name): Removed.
12 (gnumeric_open_reader): Use new function
13 dict_make_unique_var_name.
15 * short-names.c (set_var_short_name_suffix): Use new function
18 2008-01-19 John Darrington <john@darrington.wattle.id.au>
20 * settings.c settings.h: Moved static variables into a
21 single struct. Renamed functions to have a settings_ prefix.
23 2008-01-14 John Darrington <john@darrington.wattle.id.au>
25 * data-out.c (output_AHEX): Corrected number of bytes in
26 call to output_hex. Closes bug #22011
28 2008-01-02 John Darrington <john@darrington.wattle.id.au>
30 * variable.c variable.h: Replaced var_get_value_name with
31 var_append_value_name which doesn't use any static data.
32 Thanks to Ben for review.
34 2007-12-07 Ben Pfaff <blp@gnu.org>
38 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
39 made from an empty casereader produce a casegrouper with no
40 groups, instead of one with one group that has no cases.
42 * casereader.c (casereader_is_empty): New function.
44 2007-12-06 Ben Pfaff <blp@gnu.org>
48 * sys-file-reader.c (read_display_parameters): Handle variable
49 display parameters record with only 2 data items per variable.
50 Reported by Guido Gay <gay@irer.it>.
52 2007-12-04 Ben Pfaff <blp@gnu.org>
54 * identifier.c (lex_id_match_n): New function.
55 (lex_id_match): Reimplement in terms of lex_id_match_n.
57 2007-11-24 Ben Pfaff <blp@gnu.org>
59 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
62 2007-11-24 Ben Pfaff <blp@gnu.org>
64 Fix warning reported when reading back system files that include
65 very long string variables. Thanks to Guido Gay <gay@irer.it> for
68 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
69 was written where `j' was meant.
71 2007-11-18 Ben Pfaff <blp@gnu.org>
73 Properly write variables that include a range of missing values to
74 system files. Thanks to Guido Gay <gay@irer.it> for reporting
77 * sys-file-writer.c (write_variable): Correctly calculate
78 n_missing_values field when writing variables that include a range
81 2007-11-10 Ben Pfaff <blp@gnu.org>
83 Cleanups and bug fixes devised while writing up documentation.
86 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
88 * dict-class.c: New file.
89 (dict_class_from_id): Move here.
90 (dict_class_to_name): Move here.
92 * dict-class.h: New file.
93 (enum dict_class): Move here. Change from consecutive integers to
94 consecutive bits, to make testing for any of multiple values
95 easier. Add new DC_ALL constant.
97 * dictionary.c (struct dictionary): Change `case_limit' from
99 (dict_get_vars): Make final argument an enum dict_class.
100 (dict_get_vars_mutable): Ditto.
101 (dict_get_case_limit): Change return value to casenumber.
102 (dict_set_case_limit): Change final argument to a casenumber.
103 (dict_unset_split_var): Add assertion. Rephrase slightly.
104 (dict_set_label): Use xstrndup to simplify.
106 * format.c (fmt_step_width): AHEX format also needs 2-byte
108 (fmt_set_style): Simplify assertions.
110 * missing-values.c (mv_add_num_range): Rename mv_add_range.
111 Simplify implementation.
112 (mv_has_value): Simplify implementation.
113 (mv_pop_value): Remove the first value, not the last, to avoid
114 having GET followed by SAVE reverse the order of missing values.
115 (mv_peek_value): Rename mv_get_value. Simplify assertion.
116 (mv_has_range): Simplify implementation.
117 (mv_peek_range): Rename mv_get_range.
118 (can_resize_string): Removed.
119 (mv_is_resizable): Use value_is_resizable.
120 (mv_resize): Use value_resize.
122 * short-names.h (SHORT_NAME_LEN): Move here.
124 * val-type.h: New file, for definitions related to type and width
125 of abstract values. Before, these definitions were mixed among
126 those related to "union value" and those related to variables.
127 (macro SYSMIS): Move here.
128 (macro LOWEST): Move here.
129 (macro HIGHEST): Move here.
130 (macro MAX_STRING); Move here.
131 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
132 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
133 All references updated.
134 (val_type_is_valid): New function. Replaces var_type_is_valid.
135 All references updated.
136 (val_type_from_width): New function. Replaces
137 var_type_from_width. All references updated.
139 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
141 (val_labs_can_set_width): Use value_is_resizable.
142 (val_labs_add): Simply return false if the value labels set is too
143 wide, instead of having undefined behavior.
144 (val_labs_replace): Ditto.
145 (val_labs_replace): Ditto.
146 (val_labs_first): Set iterator to null if iteration is complete.
147 (val_labs_first_sorted): Ditto.
148 (val_labs_done): Become a no-op if the iterator is null.
150 * value.c (value_is_resizable): New function.
151 (value_resize): New function.
153 * variable.c (var_get_dict_class): New function.
155 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
158 2007-11-08 Ben Pfaff <blp@gnu.org>
160 * data-in.c: Make formatted data parsing locale-independent.
161 (parse_number): Use c_strtod instead of strtod, to avoid
162 locale-specific behavior.
165 2007-11-06 Ben Pfaff <blp@gnu.org>
167 Patch #6256: add support for binary, 360 file formats. Reviewed
170 * data-in.c (struct data_in): Add `encoding' member.
171 (data_in): Add `encoding' parameter, and re-encode the data passed
172 in where appropriate. Update all callers to pass it in.
173 (parse_A): Implement EBCDIC recoding wart described in manual.
174 (parse_AHEX): Implement EBCDIC recoding.
176 * data-out.c (data_out_legacy): New function.
177 (data_out): Make into a wrapper around data_out_legacy.
179 * file-handle-def.c (struct file_handle): New member `encoding'.
180 (fh_create_file): Set encoding.
181 (fh_default_properties): Set default encoding.
182 (fh_get_legacy_encoding): New function.
184 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
185 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
186 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
187 (struct fh_properties): New member `encoding'.
189 2007-11-05 Ben Pfaff <blp@gnu.org>
191 Patch #6258. Reviewed by John Darrington.
193 * file-handle-def.c (fh_lock): Add comment that TYPE should be
194 marked with N_() in the caller. Added these markings to each
195 caller too. Should make i18n easier.
196 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
198 2007-11-03 Ben Pfaff <blp@gnu.org>
200 Allow output files to overwrite input files (bug #21280).
202 * file-handle-def.c: Separate locking of files for input and for
203 output, to allow both to take place at once. Also, distinguish a
204 file handle from the identity of the underlying file, because the
205 identity of a file changes over time and the file handle can't
206 represent two different identities.
207 (struct file_handle): Remove `next', `open_cnt', `deleted',
208 `type', `open_mode', `aux', `identity' members. Change `id' from
209 char array to char *. Add `ref_cnt' member.
210 (file_handle_from_ll) New function.
211 (file_handles) Removed.
212 (named_handles) New variable.
213 (fh_init) Initialize named_handles.
214 (fh_done) Remove name from all named_handles.
215 (free_handle) Updated for modified struct file_handle.
216 (unname_handle) New function.
217 (fh_ref) New function.
218 (fh_from_file_name) Removed.
219 (fh_unref) New function.
220 (fh_unname) New function.
221 (fh_from_id) Rewritten.
222 (create_handle) Updated for modified struct file_handle.
223 (fh_inline_file) Increment the handle's ref count.
224 (fh_create_file) Updated for modified struct file_handle.
225 (fh_create_scratch) Ditto.
230 (fh_is_open) Removed.
231 (fh_get_id) Updated for modified struct file_handle.
232 (fh_get_default_handle) Increment the handle's ref count.
233 (fh_set_default_handle) Handle ref counts.
234 (struct fh_lock) New structure.
235 (locks) New static var.
236 (fh_lock) New function.
237 (fh_unlock) New function.
238 (fh_lock_get_aux) New function.
239 (fh_lock_set_aux) New function.
240 (fh_is_locked) New function.
241 (make_key) New function.
242 (free_key) New function.
243 (compare_fh_locks) New function.
244 (hash_fh_lock) New function.
246 * file-handle-def.h (enum fh_access) New enum.
248 * file-name.c: Made file_identity the same in all supported
250 (struct file_identity): New `name' member.
251 (fn_get_identity): For a file that doesn't exist, get the
252 dev/inode of its directory plus its name. If even the directory
253 doesn't exist, just use its name. Merge the Windows
254 implementation into the Unix one.
255 (fn_compare_file_identities): Rewritten. Merge the Windows
256 implementation into the Unix one.
257 (fn_hash_identity): New function.
259 * make-file.c (struct replace_file): New structure.
260 (all_files): New static var.
261 (replace_file_start): New function.
262 (replace_file_commit): New function.
263 (replace_file_abort): New function.
264 (free_replace_file): New function.
265 (unlink_replace_files): New function.
267 * por-file-reader.c (struct pfm_reader): Add `lock' member.
268 (close_reader): Unlock file.
269 (pfm_open_reader): Lock file.
271 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
273 (pfm_open_writer): Lock file and prepare for its replacement.
274 (close_writer): Unlock file.
276 * scratch-handle.h (struct scratch_handle): Add unique_id so that
277 different generations of a scratch file can be distinguished.
279 * scratch-reader.c (scratch_reader_open): Verify that the file is
282 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
283 remove scratch_handle member.
284 (scratch_writer_open): Lock handle. Prepare to replace handle
285 data, instead of doing it immediately.
286 (scratch_writer_casewriter_destroy): Replace handle data and
289 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
290 (sfm_open_reader): Lock file.
291 (close_reader): Unlock file.
293 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
295 (sfm_open_writer): Lock file and prepare for its replacement.
296 (close_writer): Unlock file.
298 2007-11-02 Ben Pfaff <blp@gnu.org>
300 * data-out.c (output_number): Use isfinite (from C99) instead of
301 gsl_isfinite, since we now have universal support for it in
303 (output_infinite): Ditto.
304 [!HAVE_ROUND] (round): Remove definition, since we now have a
305 replacement in gnulib.
307 2007-10-27 John Darrington <john@darrington.wattle.id.au>
309 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
310 wasn't initialised correctly. Closes bugs #21330 and 21397
313 2007-10-08 Ben Pfaff <blp@gnu.org>
315 Fix bug #21280. Thanks to John Darrington for review.
317 * file-name.c (create_stream): New function.
319 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
320 file handle before creating the file, to ensure that we don't
321 truncate a file that we're reading. Make code easier to read by
324 * sys-file-write.c (sfm_open_writer): Ditto.
326 2007-10-01 Ben Pfaff <blp@gnu.org>
328 Fix bug #21192. Thanks to John Darrington for review.
330 * casereader.c (casereader_read): Decrement case_cnt before
331 calling the casereader's "read" member function, so that we
332 interact properly with lazy_casereader.
334 * datasheet.c: Add regression test for above bug fix.
335 (clone_datasheet): New function.
336 (lazy_callback): New function.
337 (check_datasheet_casereader): New function.
338 (check_datasheet): Check datasheet contents are reported correctly
339 through an ordinary casereader and a lazy casereader.
340 (clone_model): Use clone_datasheet.
342 2007-09-24 Ben Pfaff <blp@gnu.org>
344 Patch #6210. Reviewed by John Darrington.
345 * settings.c: Drop "long view", which is not needed any longer.
346 (static var long_view): Removed.
347 (force_long_view): Removed.
348 (get_viewwidth): Removed.
349 (init_viewport): Removed long_view reference.
351 2007-09-19 John Darrington <john@darrington.wattle.id.au>
353 * settings.c settings.h: Changed viewport's length and width to be
354 owned by the user interface which uses the data library. This allows
355 better abstraction, and makes dynamically adjustable dimensions easier.
357 2007-09-18 Ben Pfaff <blp@gnu.org>
359 * procedure.c (proc_extract_active_file_data): New function.
361 * lazy-casereader.h: New file.
363 * lazy-casereader.c: New file.
365 * casereader.c (casereader_dynamic_cast): New function.
367 2007-09-14 Ben Pfaff <blp@gnu.org>
369 * dictionary.c (dict_clone): Copy case indexes from cloned
370 dictionary. Fixes bug #21061. Reviewed by John Darrington.
372 2007-09-13 John Darrington <john@darrington.wattle.id.au>
374 * value.c value.h (create_value): New function.
376 2007-09-12 Ben Pfaff <blp@gnu.org>
378 Make it clear that translator casereader and casewriter translate
379 functions are supposed to destroy their input case.
381 * casereader-translator.c (struct casereader_translator): Change
382 input case parameter of translate member function from const
383 struct ccase * to struct ccase *.
384 (casereader_create_translator): Ditto, for translate parameter.
386 * casewriter-translator.c (struct casewriter_translator): Ditto.
387 (casewriter_create_translator): Ditto.
389 2007-08-27 John Darrington <john@darrington.wattle.id.au>
391 * sys-file-reader.c (read_display_parameters): Force display width
392 to 8, if the sys file says 0 (like SPSS does).
394 2007-08-12 Ben Pfaff <blp@gnu.org>
396 * dictionary.c (dict_dump): New function.
398 2007-08-12 Ben Pfaff <blp@gnu.org>
400 Drop dict_compactor in favor of using the new struct case_map.
402 * dictionary.c (struct copy_map): Removed.
403 (struct dict_compactor): Removed.
404 (dict_make_compactor): Removed.
405 (dict_compactor_compact): Removed.
406 (dict_compactor_destroy): Removed.
408 * procedure.c (struct data_set): Change `compactor' member to be a
410 (proc_open): Use case_map_to_compact_dict instead of
412 (proc_casereader_read): Use case_map_execute instead of
413 dict_compactor_compact.
414 (proc_commit): Use case_map_destroy instead of
415 dict_compactor_destroy.
417 * scratch-writer.c (struct scratch_writer): Change `compactor'
418 member to be a struct case_map *.
419 (scratch_writer_open): Use case_map_to_compact_dict instead of
421 (scratch_writer_casewriter_write): Use case_map_execute instead of
422 dict_compactor_compact.
424 2007-08-12 Ben Pfaff <blp@gnu.org>
426 * automake.mk: Add case-map.c, case-map.h.
428 * case-map.c: New file.
430 * case-map.h: New file.
432 2007-08-12 Ben Pfaff <blp@gnu.org>
434 * dictionary.c (dict_compact_values): Don't delete scratch
435 variables as well as compacting case indexes. Update all callers.
436 (dict_get_compacted_value_cnt): Rename dict_count_values and
437 change interface. Update all callers.
438 (dict_get_compacted_value_cnt): Remove.
439 (dict_compacting_would_shrink): Remove.
440 (dict_compacting_would_change): Remove.
441 (dict_make_compactor): Add new parameter. Update all callers.
443 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
444 instead of dict_count_values, changing an O(N) operation into
447 2007-08-12 Ben Pfaff <blp@gnu.org>
449 * casereader.c (casereader_read): Don't require cases read by a
450 casereader to be exactly the expected size: as long as they're big
453 2007-08-12 Ben Pfaff <blp@gnu.org>
455 Make casewriters keep track of the number of `union value's in
456 each case. This is useful for two reasons: casewriter_write can
457 then check that the case being written is large enough, and later
458 recipients of the casewriter can determine the size of the case.
460 * casewriter-translator.c (casewriter_create_translator): Add
463 * casewriter.c (struct casewriter): Add value_cnt member.
464 (casewriter_write): Check that the case passed in is big enough.
465 (casewriter_get_value_cnt): New function.
466 (casewriter_create): Add value_cnt parameter.
468 2007-08-09 Ben Pfaff <blp@gnu.org>
470 Fix bug reported by Jason Stover.
471 * settings.c: Move get_termcap_viewport and in particular the
472 #include for <curses.h> to the end of the file. curses.h
473 redefines bool on some systems (e.g. OpenBSD), which causes
474 disagreement between uses of bool before its inclusion and
476 Tested by Jason Stover.
478 2007-07-29 Ben Pfaff <blp@gnu.org>
480 Provisional fix for bug #18692 and bug #20161. Reviewed by John
483 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
484 argument (never "rb" or "wb") because SUSv3 says that only those
485 modes are defined, and glibc in fact rejects other modes.
487 Open portable files with fn_open so that they can be read from
488 pipes. Fix missing fh_close call to go along with fh_open.
489 Report an error if the file close reports an error.
490 * por-file-reader.c (close_reader): New function.
491 (por_file_casereader_destroy): Use close_reader.
492 (pfm_open_reader): Open file with fn_open.
494 2007-07-28 Ben Pfaff <blp@gnu.org>
496 Make PSPP able to read all the portable files I could find on the
497 web. Thanks to John Darrington for review. Bug #17620.
498 * por-file-reader.c (struct pfm_reader): New member `line_length'.
499 (error): Print file offset in hexadecimal.
500 (warning): New function.
501 (advance): Treat lines less than 80 bytes long as padded to 80
503 (pfm_open_reader): Call read_documents if we find an "E" record.
504 (convert_format): Convert invalid formats to the default format
505 instead of aborting reading the file.
506 (read_variables): Rename duplicate variable names instead of
507 aborting reading the file.
508 (read_value_label): Allow string variables of different widths to
509 be assigned value labels in the same record. Replace duplicate
510 value labels instead of aborting.
511 (read_documents): New function.
513 * por-file-writer.c (pfm_open_writer): Call write_documents if the
514 dictionary has documents.
515 (write_documents): New function.
517 2007-07-25 Ben Pfaff <blp@gnu.org>
519 Fix bugs related to bug #17213.
521 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
522 since the former is what config.h has. Include the needed ncurses
524 (static var echo) Rename to `do_echo' because the original name is
525 the same as an ncurses identifier.
526 (get_termcap_viewport) Use error instead of msg.
528 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
529 (fn_close): Don't close stdin, stdout, stderr.
531 2007-07-26 John Darrington <john@darrington.wattle.id.au>
533 * procedure.c procedure.h: Added callbacks which get invoked whenever
534 a dataset's transformation chain changes.
536 2007-07-24 Ben Pfaff <blp@gnu.org>
539 * sys-file-writer.c (write_variable_display_parameters): Use new
540 var_default_display_width function to choose display width of
541 segments after the first one in a given variable.
542 * variable.c (var_create): Use var_default_display_width to pick
543 new variable's display width.
544 (var_default_display_width): New function.
545 Reviewed by John Darrington.
547 2007-07-24 Ben Pfaff <blp@gnu.org>
550 * por-file-writer.c (write_variables): Write weight variable.
551 Reviewed by John Darrington.
553 2007-07-23 Ben Pfaff <blp@gnu.org>
555 Improvements to system file reader and writer.
557 First, move all detailed knowledge of very long strings into
558 sys-file-private.[ch], so that this nasty stuff can be isolated.
560 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
561 (EFFECTIVE_VLS_CHUNK): New macro.
562 (min_int): New function.
563 (max_int): New function.
564 (sfm_width_to_bytes): Rewrite.
565 (sfm_width_to_octs): New function.
566 (sfm_segment_alloc_width): New function.
567 (sfm_segment_alloc_bytes): New function.
568 (sfm_segment_used_bytes): New function.
569 (sfm_segment_offset): New function.
570 (sfm_segment_effective_offset): New function.
571 (sfm_dictionary_to_sfm_vars): New function.
573 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
574 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
575 (struct sfm_var): New structure.
577 Next, improvements to the system file reader.
579 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
580 type casenumber. Added `version_major', `version_minor',
583 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
584 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
585 Change `case_cnt' to type casenumber. Removed `has_vls'.
586 (struct sfm_var): Removed.
587 (sfm_open_reader): Don't warn on wrong case size if the file was
588 written by SPSS 13, which tends to get it wrong. Use
589 sfm_dictionary_to_sfm_vars.
590 (read_header): Always output system file info.
591 (read_variable_record): Simplify code for reading missing values.
592 (read_machine_int32_info): Save version numbers from system file
593 into info struct passed as new argument.
594 (read_long_string_map): Restructured to use new sys-file-private
596 (read_value_labels): Use size_overflow_p.
597 (sys_file_casereader_read): Get rid of distinction between fast
598 and slow paths. Use information provided by sys-file-primate's
599 struct sfm_var to simplify code.
600 (skip_whole_strings): New function.
601 (read_int32): Renamed read_int. Changed return value to int.
603 (read_flt64): Renamed read_float. Changed return value to
604 double. Updated all callers.
605 (int32_to_native): Removed. Changed callers to use
607 (flt64_to_double): Removed. Changed callers to use float_convert.
609 Finally, get rid of int32, flt64 terminology and types in system
610 file writer. The former wasn't very useful since a POSIX "int"
611 can hold the whole range of int32 and we generally didn't have a
612 need for it to be exactly-32-bits, just at-least-32-bits. The
613 latter was inconvenient because we had to assume that it could be
614 different from double and thereby convert special values SYSMIS,
615 HIGHEST, LOWEST to and from it in multiple places. Instead, now
616 we just use "int" and "double" in most places, and do conversions,
617 if necessary, very close to where we do I/O. This change meant
618 that the writer code couldn't represent records in the file as C
619 structs any longer, but that's no great loss. The code actually
620 seems to be more readable without them.
622 Simplify the compression buffering code: only buffer as much as
623 necessary, which is no more than eight 8-byte units at any given
626 * sys-file-writer.c (typedef flt64): Removed.
627 (macro second_lowest_flt64): Removed.
628 (struct sysfile_header): Removed.
629 (struct sysfile_variable): Removed.
630 (struct sfm_writer): Removed `needs_translation', `has_vls',
631 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
632 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
633 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
634 compression buffering by `opcodes', `opcode_cnt', `data',
635 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
636 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
637 data writer functions instead of structures.
638 (calc_oct_idx): New function.
639 (write_header): Use simple data writer functions instead of
641 (write_format_spec): Renamed write_format. New argument.
642 (write_variable_continuation_records): New function.
643 (write_variable): Use simple data writer functions instead of
644 structures. Use write_variable_continuation_records. Write
645 entire very long string instead of requiring caller to understand
647 (write_value_labels): Use simple data writer functions instead of
649 (write_documents): Ditto.
650 (write_variable_display_parameters): Use sys-file-private
651 functions to simplify. Use simple data writer functions instead
653 (write_vls_length_table): Use simple data writer functions instead
655 (write_longvar_table): Ditto.
656 (write_rec_7_34): Break into new functions
657 write_integer_info_record, write_float_info_record. Use simple
658 data writer functions instead of structures.
659 (buf_write): Removed.
660 (append_string_max): Removed.
661 (ensure_buf_space): Removed.
662 (sys_file_casewriter_write): Get rid of the distinction between
663 fast and slow paths, which didn't seem to be too useful. Use new
664 functions write_case_uncompressed, write_case_compressed.
665 (put_instruction): Removed.
666 (put_element): Removed.
667 (write_compressed_data): Removed.
668 (close_writer): Use flush_compressed. Only write case count to
669 system file if it will fit in the field.
670 (write_case_compressed): New function.
671 (write_case_uncompressed): New function.
672 (flush_compressed): New function.
673 (put_cmp_opcode): New function.
674 (put_cmp_number): New function.
675 (write_int): New function.
676 (convert_double_to_output_format): New function.
677 (write_float): New function.
678 (write_value): New function.
679 (write_string): New function.
680 (write_bytes): New function.
681 (write_zeros): New function.
682 (write_spaces): New function.
684 Reviewed by John Darrington.
686 2007-07-22 Ben Pfaff <blp@gnu.org>
688 Don't try to write very long strings to portable files. The
689 format does not support it.
691 * por-file-writer.c (MAX_POR_WIDTH): New macro.
692 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
693 (write_format): Add arg to take width to resize format to.
694 (write_value): Limit width of value written to MAX_POR_WIDTH.
695 (write_variables): Limit width of variable and its output formats
697 Reviewed by John Darrington.
699 2007-07-22 Ben Pfaff <blp@gnu.org>
701 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
702 local variable instead of literal 5.
703 Reviewed by John Darrington.
705 2007-07-22 Ben Pfaff <blp@gnu.org>
707 Fix problems with uniqueness of short names in system files with
708 very long string variables. Now a variable may have multiple
711 * automake.mk (src_data_libdata_a_SOURCES): Add new files
712 short-names.c, short-names.h.
714 * dictionary.c (dict_clone): Clone all the short names.
715 (compare_strings): Move into short-names.c.
716 (hash_strings): Ditto.
717 (set_var_short_name_suffix): Ditto.
718 (dict_assign_short_names): Ditto, rename short_names_assign,
719 change to assign all short names.
721 * por-file-writer.c (write_variables): Use short_names_assign
722 instead of dict_assign_short_names.
724 * short-names.c: New file.
726 * short-names.h: New file.
728 * sys-file-private.c (sfm_width_to_segments): New function.
730 * sys-file-reader.c (read_long_var_name_map): Save and restore all
731 the short names, not just the first one.
733 * sys-file-writer.c (cont_var_name): Removed.
734 (sfm_open_writer): Use short_names_assign instead of
735 dict_assign_short_names. Use unique short names assigned by
736 short_names_assign instead of those generated by cont_var_name.
738 * variable.c (struct variable): Remove `short_name' member,
739 replace by `short_names' and `short_name_cnt'.
740 (var_create) Initialize new members.
741 (var_get_short_name_cnt): New function.
742 (var_get_short_name): Now takes an index argument. Changed most
744 (var_set_short_name): Ditto.
745 (var_clear_short_name): Renamed var_clear_short_names, changed to
746 clear all short names.
748 Reviewed by John Darrington.
750 2007-07-22 Ben Pfaff <blp@gnu.org>
752 * variable.c (var_set_width): Use new var_set_width function.
754 * missing-values.c (mv_n_values): Drop assertion, which was not
757 * format.c (fmt_default_for_width): New function.
758 (fmt_resize): New function.
760 Reviewed by John Darrington.
762 2007-07-18 John Darrington <john@darrington.wattle.id.au>
764 * datasheet.c (datasheet_delete_columns): Added assertion to check
765 we're not deleting outside the range of the sheet.
768 * dictionary.c dictionary.h variable.c: Added the ability for string
769 variables to be resized.
771 * vardict.h: Added some prototypes (moved from dictionary.h) as
772 these should only be called by variable.c
775 2007-07-14 John Darrington <john@darrington.wattle.id.au>
777 * sfm-reader.c: Respect case_cnt field in file header.
779 2007-07-01 John Darrington <john@darrington.wattle.id.au>
781 * transformation.c transformation.h (trns_chain_execute): Changed the
782 signature (Patch #6057)
784 2007-06-10 Ben Pfaff <blp@gnu.org>
786 * casereader-filter.c (casereader_filter_destroy): Make sure to
787 write all the remaining excluded cases to the casewriter, if any.
789 * caseinit.c (init_list_destroy): Rewrite.
790 (init_list_clear): Ditto.
792 * casegrouper.c (casegrouper_get_next_group): Always set *reader
793 to null when returning false.
795 2007-06-06 Ben Pfaff <blp@gnu.org>
797 Actually implement the new procedure code and adapt all of its
798 clients to match. Also adapt all of the other case sources and
799 sinks in the tree and their clients to use the
800 casereader/casewriter infrastructure.
802 * automake.mk: Add and remove files.
804 * any-reader.c: Change into a casereader.
805 * por-file-reader.c: Ditto.
806 * scratch-reader.c: Ditto.
807 * sys-file-reader.c: Ditto.
809 * any-writer.c: Change into a casewriter.
810 * por-file-writer.c: Ditto.
811 * scratch-writer.c: Ditto.
812 * sys-file-writer.c: Ditto.
814 * procedure.c: Change to use casereader, casewriter, caseinit, and
815 other new infrastructure.
817 * scratch-handle.c: Adapt to new infrastructure.
819 * case-sink.c: Removed, now dead code.
820 * case-sink.h: Ditto.
821 * case-source.c: Ditto.
822 * case-source.h: Ditto.
823 * casefile-factory.c: Ditto.
824 * casefile-private.h: Ditto.
827 * casefilter.c: Ditto.
828 * casefilter.h: Ditto.
831 * fastfile-factory.c: Ditto.
832 * fastfile-factory.h: Ditto.
833 * storage-stream.c: Ditto.
834 * storage-stream.h: Ditto.
836 2007-06-06 Ben Pfaff <blp@gnu.org>
840 * automake.mk: Add new files.
842 * datasheet.c: New file.
844 * datasheet.h: New file.
846 2007-06-06 Ben Pfaff <blp@gnu.org>
848 Until now, the procedure code has provided a case to the
849 case_source, which has filled in the data values that come from
850 the active file. "Left" data values that don't come from the
851 active file naturally stay the same from case to case, because the
852 procedure code keeps using that same case.
854 One of the compromises that comes with the new procedure code is
855 that the active file allocates and provides its own case, which
856 the procedure code then has to resize to provide room for any
857 other variables that should go in the case and then fill in the
858 values of "left" variables. Then, when we're done with that case,
859 we have to save the values of "left" variables to copy into the
860 next case read from the active file.
862 The caseinit code helps with this.
864 * automake.mk: Add new files.
866 * caseinit.c: New file.
868 * caseinit.h: New file.
870 2007-06-06 Ben Pfaff <blp@gnu.org>
872 * value.h (value_cnt_from_width): New function.
874 * variable.c (var_get_value_cnt): Use new function.
876 2007-06-06 Ben Pfaff <blp@gnu.org>
878 Add casegrouper, to allow cases read from a given casereader to be
879 broken into groups, each of which has its own casereader.
880 Generally cases are grouped based on having equal values for some
883 * automake.mk: Add new files.
885 * casegrouper.c: New file.
887 * casegrouper.h: New file.
889 2007-06-06 Ben Pfaff <blp@gnu.org>
891 Add interface to lexicographical ordering of cases.
893 * automake.mk: Add new files.
895 * case-ordering.c: New file.
897 * case-ordering.h: New file.
899 2007-06-06 Ben Pfaff <blp@gnu.org>
901 Add casereaders and casewriters, the basis of the new data processing
902 implementation. A casereader is a uniform interface to reading cases
903 from a data source; a casewriter is a uniform interface to writing
904 cases to a data sink.
906 * automake.mk: Add new files.
908 * casereader-filter.c: New file.
910 * casereader-provider.h: New file.
912 * casereader-translator.c: New file.
914 * casereader.c: New file.
916 * casereader.h: New file.
918 * casewriter-provider.h: New file.
920 * casewriter-translator.c: New file.
922 * casewriter.c: New file.
924 * casewriter.h: New file.
926 2007-06-06 Ben Pfaff <blp@gnu.org>
928 "casewindow" data structure that extends the deque (from libpspp)
929 of cases with the ability to dump cases to disk if we get too many
932 * automake.mk: Add new files.
934 * casewindow.c: New file.
936 * casewindow.h: New file.
938 2007-06-06 Ben Pfaff <blp@gnu.org>
940 sparse_cases data structure that augments a sparse_array of cases
941 with the ability to dump cases to disk if we get too many cases in
944 * automake.mk: Add new files.
946 * sparse-cases.c: New file.
948 * sparse-cases.h: New file.
950 2007-06-06 Ben Pfaff <blp@gnu.org>
952 Adds a low-level on-disk case array data structure.
954 * automake.mk: Add new files.
956 * case-tmpfile.c: New file.
958 * case-tmpfile.h: New file.
960 2007-06-06 Ben Pfaff <blp@gnu.org>
962 In a couple of places we calculate the maximum number of cases to
963 keep in memory based on the user-defined workspace. Enable
964 centralizing the calculation through a new function.
966 * settings.c (get_workspace_cases): New function.
968 2007-06-06 Ben Pfaff <blp@gnu.org>
970 The casenumber type is defined in transformations.h, but case.h is
971 a more sensible place. Move it.
973 * case.h (CASENUMBER_MAX): New macro.
974 (typedef casenumber): Move here, from transformations.h.
976 2007-06-03 Ben Pfaff <blp@gnu.org>
978 Slightly generalize case_to_values and case_from_values functions.
980 * case.c (case_to_values): Rename case_copy_out, change interface.
981 (case_from_values): Rename case_copy_in, change interface.
983 * fastfile.c (fastfilereader_get_next_case): Update caller.
984 (write_case_to_disk): Ditto.
986 2007-06-02 Ben Pfaff <blp@gnu.org>
988 Clean up after a forgotten part of patch #5829.
990 * casedeque.h: Remove unused file.
992 * automake.mk: Remove casedeque.h from sources.
994 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
996 * category.c: Removed redundant #include
998 2007-05-06 Ben Pfaff <blp@gnu.org>
1000 Abstract the documents within a dictionary a little better.
1001 Thanks to John Darrington for suggestion, initial version, and
1002 review. Patch #5917.
1004 * dictionary.c (struct dictionary): Change `documents' member from
1005 char * to struct string.
1006 (dict_clear): Destroy struct string.
1007 (dict_get_documents): Convert struct string to char *.
1008 (dict_set_documents): Set struct string. Pad to 80-character
1010 (dict_clear_documents): New function.
1011 (dict_add_document_line): New function.
1012 (dict_get_document_line_cnt): New function.
1013 (dict_get_document_line): New function.
1015 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1017 * sys-file-reader.c (read_documents): Use new document functions.
1019 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1021 * sys-file-reader.c: When reading a system file which has no
1022 long name table, automatically create one where the long names
1023 are the lower case versions of the short names.
1025 2007-04-22 Ben Pfaff <blp@gnu.org>
1027 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1028 dict_clear will free most data related to the dictionary.
1029 dict_clear does a decent job, except that dict_set_split_vars on
1030 some systems won't actually free the dict's "split" member.
1031 Instead, it'll allocate a 1-byte region. Fix this.
1033 * value.c (value_copy): New function.
1034 (value_set_missing): Ditto.
1036 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1038 * Deleted existing category.h and moved cat-routines.h into
1039 category.h Encapsulated struct cat_vals better.
1041 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1043 * sys-file-reader.c: When reading a system file which has no
1044 long name table, automatically create one where the long names
1045 are the lower case versions of the short names.
1047 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1049 * sys-file-reader.c: Some versions of Other Software seem to
1050 produce system files with string variables' measure set to
1051 zero. We'll assume these are supposed to be nominal variables.
1053 2007-03-30 Ben Pfaff <blp@gnu.org>
1055 * procedure.c: Adapt to new deque data structure.
1057 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1058 Ben Pfaff <blp@gnu.org>
1060 * file-name.c: Mingw compatibility fixes.
1061 (fn_search_path): Use ISSLASH instead of comparing against '/'
1063 (fn_dir_name): Use dir_name from gnulib.
1064 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1065 (fn_get_identity): Use GetFullPathName instead of canonicalize
1066 from gnulib, because the latter does not fully support
1067 Windows-style path names. Use this implementation based on the
1068 detected presence of Windows instead of the absence of Unix, since
1069 the new implementation is Windows-specific.
1070 (fn_compare_file_identities): In Windows implementation, compare
1071 names case-insensitively.
1073 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1075 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1076 <stdlib.h> to have the same effect.
1078 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1080 * por-file-reader.c: Add missing _() around messages.
1082 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1084 * make-file.c: Include "mkstemp.h", without which linking on
1087 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1089 Reduce platform dependence.
1091 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1092 Everywhere we using this, we really should have just depended on
1093 the shell to expand tildes.
1094 (fn_search_path): Simplify, given that we don't do tilde expansion
1096 (fn_normalize): Removed. Caller changed to use the canonicalize
1098 (fn_get_cwd): Removed. Only user was fn_normalize.
1099 (fn_is_absolute): Really only test for absolute names.
1100 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1102 (fn_readlink): Removed, as it was only used fn_normalize.
1103 (fn_exists): Assume the stat function is available; gnulib does.
1104 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1106 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1108 * dictionary.c (dict_create_vector_assert): New function.
1110 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1112 * file-name.c (fn_normalize): Correct name of function
1113 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1114 for pointing this out.
1116 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1118 We are using a single member in struct file_handle, the "name"
1119 field, for more than one purpose. When it begins with '"', it's a
1120 file name; otherwise, it's a token that can be used to identify
1121 it. When that assertion fires, it's because we searched for the
1122 name case-sensitively as a file name (so that there was no match),
1123 and then we try to insert it case-insensitively as a token, which
1124 fails because duplicates aren't allowed.
1126 Solution: break the two purposes into two separate fields. This
1127 fixes the problem and likely makes the code easier to read too.
1129 Fixes bug #18922. Thanks to John Darrington for bug report and
1132 * file-handle-def.c (struct file_handle): New `id' member.
1133 (fh_from_name): Rename fh_from_id. Update all callers.
1134 (create_handle): New `id' parameter. Update all callers.
1135 (fh_create_file): Ditto.
1136 (fh_get_id): New function.
1138 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1140 * case.c (case_is_null): Change return type to bool.
1142 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1146 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1147 reference counting to share data. This makes it easy for
1148 valgrind, etc. to find accesses to cases that have been destroyed
1149 but have been kept around by another user's ref-count. This often
1150 happens when the data set is small enough to find in memory; if a
1151 bigger data set that would overflow to disk were used, then data
1152 corruption would occur.
1154 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1158 * case.c (case_unshare): Make it check internally whether the
1159 ref_cnt is greater than 1, so that the callers don't have to.
1160 Update callers not to check.
1162 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1164 Before, I was thinking that I might want to get rid of reference
1165 counting at some point. Now, I'm pretty sure that it's here to
1166 stay. Thus, because we have to store the value_cnt anyway for
1167 reference-counted cases, we might as well expose it to users.
1169 * case.c (case_get_value_cnt): New function.
1170 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1171 resize case if its size actually changed.
1173 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1174 instead of peeking inside struct case directly.
1175 (casefile_append): Ditto.
1177 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1179 Get rid of the inlines for the case functions, which made the
1180 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1181 the inlines didn't speed up "make check" at all, which is not a
1182 perfect benchmark but seems indicative.)
1184 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1185 definitions. Remove some assertions on nonnull pointers that were
1186 redundant with a pointer dereference soon after in the function.
1188 (struct case_data): Move definition here from case.h.
1192 (case_data_wr): Ditto.
1194 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1196 * automake.mk: Add casedeque.h to sources.
1198 * casedeque.h: New file.
1200 * procedure.c: (struct dataset) Change lag_count, lag_head,
1201 lag_queue member into single struct casedeque member. Update all
1202 users to use the casedeque instead.
1205 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1207 * procedure.c: Simplify lagged cases interface. Updated all
1208 clients--well, the only client--to use the simplified interface.
1209 (dataset_n_lag) Removed.
1210 (dataset_set_n_lag) Removed.
1211 (dataset_need_lag) New function.
1213 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1215 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1216 whenever appropriate, but only when the dataset/dictionary is in a
1219 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1221 * dictionary.c dictionary.h : Added callbacks for change of filter and
1222 split variables. Refactored some code to ensure that callbacks get
1223 invoked when appropriate.
1225 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1226 replace_dict callback occurs when permanent_dict replaces the current
1229 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1231 * dictionary.c dictionary.h : Added callback for when the weight
1232 variable of a dictionary changes.
1234 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1236 * dictionary.c dictionary.h : Added replace_source and replace_dict
1237 callbacks, and functions to deal with them.
1239 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1241 Simplify missing value handling.
1243 * missing-values.h (enum mv_class): New type.
1244 (enum mv_type): Moved definition into missing-values.c and renamed
1245 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1246 enums. Updated all uses.
1247 (struct missing_values): Changed type of `type' from `enum
1248 mv_type' to `int' because the definition is no longer exposed.
1250 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1251 parameter. Update all callers.
1252 (mv_is_num_missing): Ditto.
1253 (mv_is_str_missing): Ditto.
1254 (mv_is_value_user_missing): Removed. Changed callers to use
1255 mv_is_value_missing.
1256 (mv_is_num_user_missing): Removed. Changed callers to use
1258 (mv_is_str_user_missing): Removed. Changed callers to use
1260 (mv_is_value_system_missing): Removed. Changed callers to use
1261 mv_is_value_missing.
1262 (mv_set_type): Removed. Changed callers to use mv_clear.
1263 (mv_clear): New function.
1265 * variable.c (var_is_value_missing): Add new enum mv_class
1266 parameter. Update all callers.
1267 (var_is_num_missing): Ditto.
1268 (var_is_str_missing): Ditto.
1269 (var_is_value_user_missing): Removed. Changed callers to use
1270 var_is_value_missing.
1271 (var_is_num_user_missing): Removed. Changed callers to use
1273 (var_is_str_user_missing): Removed. Changed callers to use
1275 (var_is_value_system_missing): Removed. Changed callers to use
1276 var_is_value_missing.
1278 * casefilter.c (struct casefilter): Use enum mv_class in place of
1280 (casefilter_variable_missing): Adapt to new member.
1281 (casefilter_create): Change signature to take enum mv_class,
1284 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1286 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1288 * case-sink.c case-sink.h procedure.c procedure.h
1289 storage-stream.c: Now uses the factory.
1291 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1293 Make it possible to pull cases from the active file with a
1294 function call, instead of requiring indirection through a callback
1297 * case-source.h (struct case_source_class): Change ->read function
1298 to return a single case, instead of calling a callback function
1299 for each case. Change ->destroy function to return an error
1302 * case-source.c (free_case_source): Pass along the value returned
1303 by the case_source ->destroy function.
1305 * procedure.c (struct write_case_data): Removed.
1306 (struct dataset): Added some members to track procedure state.
1307 (procedure): Optimize the trivial case at this level.
1308 (internal_procedure): Re-implement in terms of proc_open,
1309 proc_read, proc_close.
1310 (proc_open) New function.
1311 (proc_read) New function.
1312 (proc_close) New function.
1313 (write_case) Moved into proc_read.
1314 (close_active_file) Moved closing of data source into proc_close.
1316 * storage-source.c: Rewrote to conform with modified
1317 case_source_class interface.
1319 * transformations.c (trns_chain_execute): Added argument to allow
1320 starting execution from an arbitrary transformation. Updated
1323 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1325 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1327 * sys-file-reader.c (read_display_parameters): Don't assume that
1328 MEASURE_* and ALIGN_* have the same values found in system files.
1330 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1332 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1333 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1334 the value of n_MEASURES, which was off by 1 (at least from my
1337 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1339 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1340 callbacks which are invoked when the dictionary or its
1341 variables are changed.
1343 * missing-values.c missing-values.h value-labels.c: Tidied up
1344 consistency checks, and made some of them return false
1345 instead of assert-failing.
1347 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1349 * calendar.c (calendar_days_in_month): New function.
1351 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1353 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1354 the value label's width.
1356 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1358 * sfm-private.h: Move contents into sys-file-writer.c, which is
1359 the only remaining user. Removed Borland C++-specific directives.
1361 * sys-file-reader.c: Clean up and rewrite entire file. The
1362 rewritten version is simpler and better abstracted, and should be
1363 easier to maintain and extend. It avoids using structures to read
1364 file data, which is prone to padding variations among compilers.
1365 It should also handle non-IEEE 754 system files, although I
1366 haven't been able to find any. It has been tested against many
1367 .sav files obtained from the Web and found to produce the same
1368 results as the earlier version of the code, or in some cases
1369 improved results. It is more tolerant of format variations found
1372 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1373 member, putting an enum integer_format in its place. New member
1374 `float_format'. Changed `compressed' member to type bool.
1376 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1378 * dictionary.c (dict_delete_consecutive_vars): New function.
1380 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1382 * file-name.c (fn_search_path): Remove prefix arg that was unused
1383 by any caller. Updated all callers.
1385 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1387 * format.c (fmt_dollar_template): Use user's decimal point
1388 character. Add assertion.
1390 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1392 * format.c (fmt_dollar_template): New function, based on
1393 dollar_format_template from var-type-dialog.c.
1395 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1397 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1400 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1402 Finish converting struct variable to an opaque type. In this
1403 phase, we add remaining setter and getter functions, convert the
1404 remaining PSPP code to use them, and do a bunch of cleanup. The
1405 resulting changes are pervasive but mostly trivial, and only the
1406 notable changes are logged.
1408 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1411 * case.c (case_data): Renamed case_data_idx.
1412 (case_num): Renamed case_num_idx.
1413 (case_str): Renamed case_str_idx.
1414 (case_data_rw): Renamed case_data_rw_idx.
1416 * case.h (case_data): New function with old name and an interface
1417 that takes a variable instead of an index, which is easier to
1418 use. Updated all callers to use the new interface, or to use the
1419 new *_idx function (see above).
1422 (case_data_rw): Ditto.
1424 * category.c (cat_stored_values_destroy): Changed interface to
1425 take a struct cat_vals * instead of a struct variable *.
1427 * dictionary.c (dict_clone): Use new vector_clone function.
1428 (dict_clear) Use new var_destroy function.
1429 (add_var) New function.
1430 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1431 (dict_create_var_assert) Rewrite in terms of add_var.
1432 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1433 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1434 (dict_lookup_var) Use new var_create, var_destroy functions.
1435 (dict_contains_var) Rewrite in terms of new vardict functionality.
1436 (set_var_dict_index) New function.
1437 (set_var_case_index) New function.
1438 (reindex_vars) New function.
1439 (dict_delete_var) Rewrite in terms of new vardict functionality.
1440 (dict_reorder_var) Ditto.
1441 (dict_reorder_vars) Ditto.
1442 (rename_var) New function.
1443 (dict_rename_var) Use rename_var.
1444 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1445 (dict_get_compacted_idx_to_fv) Rename
1446 dict_get_compacted_dict_index_to_case_index, update callers.
1447 (dict_create_vector) Use new vector_create function.
1448 (dict_clear_vectors) Use new vector_destroy function.
1449 (set_var_short_name_suffix) Move here from variable.c, renamed
1450 from var_set_short_name_suffix, make static, update caller.
1452 * sys-file-private.c: New file.
1453 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1454 width_to_bytes, update callers.
1456 * sys-file-private.h: New file. Later it will supplant
1457 sfm-private.h; for now it supplements it.
1458 (macro MIN_VERY_LONG_STRING) New macro.
1459 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1461 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1462 MAX_LONG_STRING was used before.
1464 * sys-file-writer.c: Ditto.
1466 * value-labels.c: Change the paradigm here to be that a null
1467 pointer is OK for a struct val_labs * in most cases; it just
1468 represents an empty set of value labels.
1469 (val_labs_copy) A copy of a null set is a null set.
1470 (val_labs_count) A null set has 0 labels.
1471 (val_labs_replace) Change return type to void. Rewrite for
1473 (val_labs_find) A null set does not contain the value.
1474 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1475 transposed argument order, updated all callers.
1477 * value.c: New file.
1478 (value_dup) Moved here from variable.c.
1479 (compare_values) Ditto.
1482 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1483 (macro MAX_LONG_STRING) Removed, because it was only interesting
1484 for system files, not for general code.
1485 (macro MAX_VERY_LONG_STRING) Ditto.
1486 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1487 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1489 * vardict.h: New file, for an interface between variables and
1492 * variable.c: A lot of functions were moved around, for better
1494 (struct variable) Move definition here, from variable.h.
1495 (var_type_adj) Removed--makes i18n hard.
1496 (var_type_noun) Ditto.
1497 (var_create) New function.
1498 (var_clone) New function.
1499 (var_destroy) New function.
1500 (var_set_name) Assert that variable is not in a dictionary.
1501 (compare_var_names) Rename compare_vars_by_name and fix a couple
1502 of callers who thought the args were strings.
1503 (hash_var_name) Rename hash_var_by_name.
1504 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1505 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1506 (var_is_very_long_string) Removed, because it was only interesting
1507 to system file code.
1508 (var_set_missing_values) Allow the argument to be the wrong width,
1509 as long as we can resize it. Simplify callers who were doing the
1510 resizing themselves.
1511 (var_get_value_labels) New function.
1512 (var_has_value_labels) New function.
1513 (var_set_value_labels) New function.
1514 (alloc_value_labels) New function.
1515 (var_add_value_label) New function.
1516 (var_replace_value_label) New function.
1517 (var_clear_value_labels) New function.
1518 (var_lookup_value_label) New function.
1519 (var_get_value_name) Moved here from variable.c, renamed from
1520 var_get_value_name, transposed argument order, updated all
1522 (var_to_string) Moved here, from variable-label.c.
1523 (var_set_leave) New function.
1524 (var_get_leave) New function.
1525 (var_must_leave) New function.
1526 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1527 set_var_short_name_suffix.
1528 (var_get_dict_index) New function.
1529 (var_get_case_index) New function.
1530 (var_get_obs_vals) New function.
1531 (var_set_obs_vals) New function.
1532 (var_has_obs_vals) New function.
1533 (var_get_vardict) New function.
1534 (var_set_vardict) New function.
1535 (var_has_vardict) New function.
1536 (var_clear_vardict) New function.
1537 (value_dup) Moved to value.c.
1538 (compare_values) Ditto.
1541 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1542 (enum ALPHA) Rename VAR_STRING, update all users.
1544 * vector.c: New file.
1545 (struct vector) Moved here, from variable.h.
1546 (check_widths) New function.
1547 (vector_create) New function.
1548 (vector_clone) New function.
1549 (vector_destroy) New function.
1550 (vector_get_name) New function.
1551 (vector_get_var) New function.
1552 (vector_get_var_cnt) New function.
1553 (compare_vector_ptrs_by_name) New function.
1555 * vector.h: New file.
1557 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1559 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1560 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1563 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1565 * value-labels.c (destroy_atoms): New function.
1566 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1568 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1570 Thanks to Jason Stover for pointing out this problem.
1572 * data-out.c (output_number): Use gsl_finite from GSL, which is
1573 portable, instead of isfinite, which is not.
1576 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1578 * variable.c variable.h (value_dup): New function.
1580 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1582 Start converting struct variable to an opaque type. In this
1583 phase, we add a bunch of setter and getter functions and convert
1584 most of the PSPP code to use them. The resulting changes are
1585 pervasive but mostly trivial, and only the notable changes are
1588 * format.c (fmt_equal): New function.
1590 * variable.c (var_type_is_valid): New function.
1591 (measure_is_valid) Moved here, from format.c.
1592 (alignment_is_valid) Moved here, from format.c.
1593 (var_get_name) New function.
1594 (var_set_name) New function.
1595 (width_to_type) New function.
1596 (var_get_type) New function.
1597 (var_get_width) New function.
1598 (var_set_width) New function.
1599 (var_is_numeric) New function.
1600 (var_is_alpha) New function.
1601 (var_is_short_string) New function.
1602 (var_is_long_string) New function.
1603 (var_is_very_long_string) New function.
1604 (var_get_missing_values) New function.
1605 (var_set_missing_values) New function.
1606 (var_clear_missing_values) New function.
1607 (var_has_missing_values) New function.
1608 (var_is_value_missing) New function.
1609 (var_is_num_missing) New function.
1610 (var_is_str_missing) New function.
1611 (var_is_value_user_missing) New function.
1612 (var_is_num_user_missing) New function.
1613 (var_is_str_user_missing) New function.
1614 (var_is_value_system_missing) New function.
1615 (var_get_print_format) New function.
1616 (var_set_print_format) New function.
1617 (var_get_write_format) New function.
1618 (var_set_write_format) New function.
1619 (var_set_both_formats) New function.
1620 (var_get_label) New function.
1621 (var_set_label) New function.
1622 (var_clear_label) New function.
1623 (var_has_label) New function.
1624 (var_get_measure) New function.
1625 (var_set_measure) New function.
1626 (var_get_display_width) New function.
1627 (var_set_display_width) New function.
1628 (var_get_alignment) New function.
1629 (var_set_alignment) New function.
1630 (var_get_value_cnt) New function.
1631 (var_get_leave) New function.
1632 (var_get_short_name) New function.
1634 * variable.h: (struct variable) Removed "type" and "nv" members;
1635 they are now computed from "width" where needed.
1637 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1639 * missing-values.c (mv_resize): Don't write beyond end of the
1640 allocated buffer when resizing a long string.
1642 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1644 Clean up identifier code: don't require identifier enumerations to
1645 be in a particular order; make better use of string library;
1646 expose less of the internals.
1648 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1649 change interface. Updated all callers.
1650 (lex_id_match) Change interface to use struct substring, update
1652 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1653 (global array keywords[]) Make static, change form. Update all
1654 users to use lex_id_name instead.
1655 (lex_is_keyword) New function.
1656 (lex_id_to_token) Change interface to use struct substring, update
1658 (lex_id_name) New function.
1660 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1661 lex_is_keyword instead.
1662 (T_LAST_KEYWORD) Removed.
1663 (T_N_KEYWORDS) Removed.
1665 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1667 * format.c: (fmt_date_template) Distinguish characters for which a
1668 space is output and any date delimiter is allowed on input, from
1669 those for which a space is output and only a space is allowed on
1670 input. The former is represented by X, the latter by a space.
1671 Also, drop distinction between h and H, changing the former to the
1674 * data-in.c: Completely rewrite internals to conform to SPSS input
1675 formats as closely as possible.
1676 (data_in) Changed external interface by replacing the structure
1677 that was used as a single argument by a set of arguments. Updated
1679 (data_in_finite_line) Removed. Converted all callers to use plain
1681 (data_in_get_integer_format) New function.
1682 (data_in_set_integer_format) New function.
1683 (data_in_get_float_format) New function.
1684 (data_in_set_float_format) New function.
1686 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1687 (struct data_in) Removed.
1689 * data-out.c: (output_date) Drop each component from the input as
1690 it is output, to allow us to drop the distinction between h (a
1691 count of hours) and H (the hour of day) template characters.
1692 Also, handle new X template character.
1693 (output_scientific) Follow more rational rule on when to drop
1694 fraction introduced between SPSS 13 and 15. Updated test case to
1697 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1699 Fix buffer overflow reported by John Darrington.
1701 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1702 realize that DIGITS is a count of nibbles, not of bytes.
1704 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1706 * calendar.c (calendar_offset_to_gregorian) Also return the
1707 year-of-day. Change callers to new interface.
1709 * data-out.c: Completely rewrite internals to conform to SPSS
1710 output formats as completely as possible.
1711 (data_out) Change interface to put input parameters before output
1712 parameters, for consistency with the style I now prefer. Update
1714 (data_out_get_integer_format) New public function.
1715 (data_out_set_integer_format) New public function.
1716 (data_out_get_float_format) New public function.
1717 (data_out_set_float_format) New public function.
1719 * data-out.h: New file. Move prototype for data_out here, from
1722 * format.c: (fmt_step_width) Use equality comparison instead of
1723 bitwise and, for clarity.
1724 (fmt_is_string) Ditto.
1725 (fmt_input_to_output) Fix categories that are translated to F
1728 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1730 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1731 casefile-private.h: Added casefilter to assist commands with missing
1734 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1736 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1738 * settings.c: (route_errors_to_terminal) New variable.
1739 (route_errors_to_listing) New variable.
1740 (get_error_routing_to_terminal) New function.
1741 (set_error_routing_to_terminal) New function.
1742 (get_error_routing_to_listing) New function.
1743 (set_error_routing_to_listing) New function.
1745 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1747 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1749 * format.c: Completely rewrite, to achieve better abstraction.
1750 Rewrite all references to formats in other files.
1752 * format.def: Rewrite and reorganize.
1754 * settings.c: Move everything related to custom currency formats
1755 into format.[ch], changing them in form, so as to group related
1756 code and definitions better. Changed all references to use the
1758 (static var decimal) Removed.
1759 (static var grouping) Removed.
1760 (static var cc) Removed.
1761 (get_decimal) Removed.
1762 (set_decimal) Removed.
1763 (get_grouping) Removed.
1764 (set_grouping) Removed.
1768 * settings.h: (macro CC_CNT) Removed.
1769 (macro CC_WIDTH) Removed.
1770 (struct custom_currency) Removed.
1772 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1774 * data-in.c (data_in): Use switch statement instead of table, to
1775 avoid dependence on the order of the FMT_* enums.
1777 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1779 * data-out.c: (num_to_string) Removed, because it was dead code.
1781 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1783 * data-in.c (parse_trailer): Fix error message.
1785 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1787 * format.c (fmt_is_binary): New function.
1789 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1791 * procedure.c procedure.h: Encapsulated the static data into a single
1794 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1796 * casefile.c (casereader_read_xfer): Always initialize the case,
1797 even on an error condition.
1799 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1801 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1802 entire function useless.
1804 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1806 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1807 casereader method casereader_clone.
1809 * procedure.c pransformations.h: Introduced new type casenum_t
1811 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1813 * variable.c: (width_to_bytes) Rephrase code for clarify.
1815 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1817 * format.c: (fmt_type_from_string) New function.
1818 (fmt_to_string) Include decimals in output if the format has
1819 decimals, even if the format type does not. This way, we can
1820 accurately reproduce incorrect formats in user output.
1821 (check_common_specifier) Make the check for a bad format type an
1822 assertion, so we get bug reports if they show up. Fix message.
1823 Check for decimal places with a format type that doesn't allow
1825 (check_input_specifier) Remove check for FMT_X, which has been
1827 (check_output_specifier) Ditto.
1829 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1831 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1832 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1833 (enum fmt_parse_flags) Removed.
1835 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1837 * casefile.c casefile.h: Converted to an abstract base class.
1838 * casefile-private.h fastfile.c fastfile.h: New files.
1839 * automake.mk procedure.c scratch-writer.c storage-stream.c
1841 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1843 * procedure.c (internal_procedure): Create sink_case with only as
1844 many values as the compacted dictionary.
1846 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1848 Remove "debugging" code that caused plenty of false positives and
1851 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1853 * case.c: Remove all references to `this' member.
1855 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1857 Fix link error noted by Jason Stover.
1859 * storage-stream.c: Include <assert.h>.
1861 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1863 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1864 ALL) and additional underlying system file issues.
1866 Thanks to John Darrington for review.
1868 First problem: var_hash points to variables not owned by the
1869 sys-file-reader, which the caller may free or modify. Use an
1870 array of sfm_vars instead, as done earlier (e.g. CVS version
1873 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1874 members and remove all code that references it. Add vars, var_cnt
1875 members. Remove fix_specials member, which was unused.
1876 (struct sfm_var) Remove name member, which was unused.
1877 (sfm_close_reader) Free vars member instead of var_hash.
1878 (compare_var_shortnames) Removed.
1879 (hash_var_shortname) Removed.
1880 (sfm_open_reader) Fill out vars array.
1881 (compare_var_index) Removed.
1882 (sfm_read_case) Use vars instead of var_hash.
1884 Second problem: we're confused about when we actually have very
1885 long strings, causing us to choose incorrectly between slow path
1886 and fast path in sfm_read_case.
1888 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1889 have very long strings, not when we have long variable names,
1890 which is an unrelated feature.
1892 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1894 * variable.h: Move var_set and variable parsing declarations to
1895 new header, src/language/lexer/variable-parser.h. Modified lots
1896 of files to include the new header.
1898 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1900 * value-labels.c (value_to_string): When there's no value label,
1901 format the variable according to its print format, instead of
1902 always effectively using A or F format.
1904 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1906 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1909 * format.c: Removed tortological assertion.
1911 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1913 Reform string library.
1915 * file-name.c (fn_interp_vars): Change interface to take a
1916 substring as input. Updated all users.
1918 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1920 * format.c (measure_is_valid): Really return false when m >=
1923 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1925 Implement random access to casefiles, for use in GUI.
1927 * casefile.c: (struct casereader) Add `random', `file_ofs',
1928 `buffer_ofs' members.
1929 (casefile_get_random_reader) New function.
1930 (read_open_file) Break part into new function
1931 seek_and_fill_buffer().
1932 (fill_buffer) Update buffer_ofs, file_ofs.
1933 (casereader_seek) New function.
1935 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1937 * settings.c: Added call to i18n{done, init}.
1939 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1941 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1944 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1946 * casefile.c: Convert many uses of `int' to `bool'.
1948 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1950 * transformations.c (trns_chain_destroy): Destroy chain's trns
1951 member, to fix memory leak.
1953 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1955 * storage-stream.c (storage_source_decapsulate): Destroy case
1956 source to fix memory leak.
1958 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1960 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1961 case passed as argument instead of initializing the argument as a
1962 case. Fixes memory leak that showed up in
1963 tests/command/aggregate.sh with scratch files.
1965 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1967 * procedure.c (proc_done): Destroy default_dict, to fix memory
1970 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1972 Simplify procedure_with_splits().
1974 * procedure.c (struct split_aux_data): Removed case_count member.
1975 (procedure_with_splits) Don't initialize case_count.
1976 (split_procedure_case_func) Check whether prev_case is null
1977 instead of case_count.
1978 (split_procedure_end_func) Ditto.
1980 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1982 * case.c (case_move): Do nothing if dst and src are the same
1984 (case_try_create) Merge two similar cases.
1985 (case_copy) Unshare only if data must be actually copied.
1987 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1989 * data-in.c data-out.c dictionary.c sys-file-reader.c
1990 sys-file-writer.c variable.c variable.h: Reworked very long string
1991 support for better encapsulation.
1993 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1995 * value-labels.c (val_labs_can_set_width): New function.
1996 (val_labs_set_width) Clear labels if increasing width to long
1998 (val_labs_destroy) Remove unneeded test for null.
2000 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2002 * value-labels.h: Remove unneeded dependency on variable.h.
2004 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2006 Get rid of `char *c' member in union value, for cleanliness.
2008 * value.h: (union value) Remove `c' member.
2010 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2012 Make dictionary compacting functions a little more general.
2014 * sys-file-writer.c (sfm_open_writer): Use
2015 dict_compacting_would_change().
2016 (does_dict_need_translation) Removed.
2018 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2020 Make dictionary compacting functions a little more general.
2022 * dictionary.c (dict_needs_compaction): Rename
2023 dict_compacting_would_shrink(). Update all callers.
2024 (dict_compacting_would_change) New function.
2026 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2028 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2029 inverted return value (!).
2031 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2033 Continue reforming procedure execution.
2035 * procedure.c: Search and replace "vfm" by "proc". Notably:
2036 (static var vfm_source) Rename proc_source. Update all
2038 (static var vfm_sink) Rename proc_sink. Update all references.
2040 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2042 Continue reforming procedure execution. In this phase, remove
2043 PROCESS IF, which was deprecated anyway and can be easily
2044 simulated with TEMPORARY followed by SELECT IF.
2046 * procedure.c: (open_active_file) Don't call
2047 add_process_if_trns().
2048 (discard_variables) Get rid of redundant call to
2049 proc_cancel_all_transformations().
2050 (add_process_if_trns) Removed.
2051 (process_if_trns_proc) Removed.
2052 (process_if_trns_free) Removed.
2054 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2056 Continue reforming procedure execution. In this phase, add
2057 `const' to the case passed to procedure()'s callback.
2059 Updated all users of procedure() as well.
2061 * procedure.c: (struct write_case_data) Add "const" to ccase
2062 parameter for case_func member.
2063 (procedure) Add "const" to ccase parameter for proc_func
2065 (multipass_case_func) Make ccase parameter const.
2066 (internal_procedure) Add "const" to ccase parameter for case_func
2068 (split_procedure_case_func) Make ccase parameter const.
2069 (multipass_split_case_func) Make ccase parameter const.
2071 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2073 Continue reforming procedure execution. In this phase, get rid of
2074 the output code for SPLIT FILE groups in procedure.c, which really
2075 shouldn't be doing any output. Move it into the individual
2076 procedures instead. This also adds some flexibility.
2078 Updated many users of procedure_with_splits() and
2079 multipass_procedure_with_splits() to call
2080 output_split_file_values() and to deal with increased use of
2083 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2084 parameter to begin_func member.
2085 (procedure_with_splits) Add "const struct ccase *" parameter to
2086 begin_func parameter. Make ccase parameter const in proc_func
2088 (split_procedure_case_func) Don't dump split file group. Pass
2090 (dump_splits) Moved to language/dictionary/split-file.c as
2091 output_split_file_values().
2092 (struct multipass_split_aux_data) Add "const struct ccase *"
2093 parameter to split_func member.
2094 (multipass_procedure_with_splits) Add "const struct ccase *"
2095 parameter to split_func parameter.
2096 (multipass_split_case_func) Save new SPLIT FILE case before
2098 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2100 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2102 Continue reforming procedure execution. Change
2103 internal_procedure() so that it calls open_active_file() and
2104 close_active_file(), which isolates most of the actual procedure
2107 * procedure.c: (struct write_case_data) Rename `proc_func' member
2108 to `case_func' and update all references.
2109 (procedure) Rewrite as one-line wrapper around
2110 internal_procedure().
2111 (struct multipass_aux_data) New.
2112 (multipass_callback) Renamed multipass_case_func(). Use struct
2113 multipass_aux_data as auxiliary data.
2114 (multipass_end_func) New function.
2115 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2116 that uses multipass_case_func, multipass_end_func.
2117 (internal_procedure) Add `end_func' argument. Move optimization
2118 of trivial case in here. Move call to open_active_file() and
2119 close_active_file() in here. Now assert that vfm_source is
2121 (procedure_with_splits_callback) Rename
2122 split_procedure_case_func().
2123 (split_procedure_end_func) New function.
2124 (multipass_split_callback) Rename multipass_split_case_func.
2125 (multipass_split_end_func) New function.
2126 (discard_variables) No need to test for nonnull vfm_source.
2128 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2130 Continue reforming procedure execution. Get rid of unused member.
2132 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2134 (write_case) Don't increment cases_analyzed.
2136 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2138 Continue reforming procedure execution. In this phase, move
2139 procedure.c and procedure.h from src to src/data. Update
2140 makefiles and #includes accordingly.
2142 * procedure.c: Moved here from src/.
2144 * procedure.h: Moved here from src/.
2146 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2148 Continue reforming procedure execution. In this phase, get rid of
2149 many global variables, consolidating procedure execution in
2150 procedure.c. Encapsulate transformations in new "struct
2151 trns_chain". Also, change implementation of N OF CASES, FILTER,
2152 and PROCESS IF from special cases to transformations.
2154 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2157 * dictionary.c: (global variable default_dict) Move to
2160 * variable.h: (TRNS_*) Move to transformations.h.
2161 (struct transformation) Move to transformations.c.
2163 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2165 * sys-file-reader.c: Fixed invalid read problems.
2167 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2169 * storage-stream.c: Add missing function comments.
2171 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2173 Continue reforming procedure execution. In this phase, add some
2174 new, needed functionality to storage-stream.
2176 * storage-stream.c: (storage_source_decapsulate) New function.
2178 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2180 * variable.c (width_to_bytes): Declarations must precede
2181 statements for C90 compliance.
2183 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2185 * data-out.c, data-in.c, variable.c, variable.h: New functions
2186 copy_mangle and copy_demangle for reading/writing cases; emulates the
2187 way SPSS deals with strings > 255 bytes.
2189 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2190 subtype 14 needed for strings longer than 255 bytes.
2192 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2193 instead of literal values. Also fixed some constness issues.
2195 * format.h: Constness
2197 * sfm-private.h: Renamed the case_size identifier, since I discovered
2198 that SPSS's respect for this variable is very nominal.
2200 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2202 Change case limit type from int to size_t.
2204 * dictionary.c: (struct dictionary) Change type of case_limit
2206 (dict_get_case_limit) Change return type.
2207 (dict_set_case_limit) Change parameter type.
2209 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2211 * variable.h: (struct variable) Rename `reinit' member as `leave'
2212 and invert sense. Fix up all references.
2214 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2216 Continue reforming procedure execution. In this phase, break
2217 procedure.c into multiple files.
2219 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2221 * case-sink.c: New file.
2223 * case-sink.h: New file.
2225 * case-source.c: New file.
2227 * case-source.h: New file.
2229 * storage-stream.c: New file.
2231 * storage-stream.h: New file.
2233 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2235 * variable.h: (struct variable) Remove `init' member and all
2236 references to it from other files. It was initialized in several
2237 places, but nothing really ever used it for anything worthwhile.
2238 Thanks to Jason Stover for pointing out how confusing this
2241 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2243 Continue reforming error message support. In this phase, get rid
2244 of message "titles" and put the message text in `struct error'.
2245 Now `struct error' encapsulates a message more properly.
2247 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2248 Format message ourselves.
2250 * data-in.c: (vdls_error) Ditto.
2252 * por-file-reader.c: (error) Ditto.
2254 * sys-file-reader.c: (corrupt_msg) Ditto.
2256 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2258 GNU standards require "file name" instead of "filename" in
2259 documentation. It's nice for our code to follow the convention
2262 * casefile.c: (struct casefile) Rename `filename' member to
2263 `file_name'. Updated all references.
2265 * file-name.c: [!unix] (struct file_identity) Rename
2266 normalized_filename member to normalized_file_name. Updated all
2269 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2271 We don't really support anything but Unix-like environments well,
2272 so we might as well de-obfuscate by writing directory and path
2273 separators explicitly.
2275 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2277 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2278 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2280 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2283 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2285 GNU standards require "file name" instead of "filename" in
2286 documentation. It's nice for our code to follow the convention
2289 * filename.c: Rename to file-name.c.
2291 * filename.h: Rename to file-name.h. Update all inclusions.
2292 Update header guards.
2294 * automake.mk: Update file names.
2296 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2298 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2300 (fn_basename) Removed (dead code).
2301 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2302 (fn_special_p) Renamed fn_is_special(), all references updated.
2303 (fn_exists_p) Renamed fn_exists(), all references updated.
2305 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2307 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2308 Also, now it only tilde-expands file names, not paths.
2309 (fn_search_path) Tilde-expand one directory at a time.
2311 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2313 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2314 get rid of non-Unixlike version of the code, which has probably
2316 (fn_prepend_dir) Removed (dead code).
2318 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2319 (macro PATH_SEPARATOR_STRING) New.
2320 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2322 Continue reforming error message support. In this phase, we get
2323 rid of VM() and the other msg() support for "verbosity", replacing
2324 it by a new function verbose_msg().
2326 * filename.c: (fn_search_path) Use verbose_msg() instead of
2329 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2331 * sfm-private.h: Get rid of #defines after #error, which makes no
2334 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2336 Get rid of our own int32 type in favor of the standard int32_t
2339 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2342 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2344 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2346 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2348 Remove ill-considered file routines that are no longer used.
2350 * filename.c: (fn_open_ext) Removed.
2351 (fn_close_ext) Removed.
2353 * filename.h: (struct file_ext) Removed.
2355 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2357 * variable.c (var_is_valid_name): Move declarations before code
2360 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2362 * filename.ch (fn_interp_vars): Fixed small buglet.
2364 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2366 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2367 so as to modify the string inline. Thus makeing it easier to
2368 destroy the results when no longer needed.
2370 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2372 * category.c (cat_stored_values_destroy): Fixed memory leak.
2374 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2376 Add some missing frees. Thanks to John Darrington for reporting
2379 * any-writer.c (any_writer_close): Free writer.
2381 * any-reader.c (any_reader_close): Free reader.
2383 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2385 * por-file-reader.c: (error) Mark as NO_RETURN.
2387 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2389 * settings.c: Changed default value of scompress to true.
2391 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2393 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2395 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2398 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2399 here from pref.h.orig.
2401 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2403 * sys-file-reader.c: Fixed bug reading compressed files.
2405 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2407 * Numerous renames. See src/ChangeLog for details.
2409 * Moved files from src directory