1 2008-02-09 Ben Pfaff <blp@gnu.org>
3 Add a couple of extensions to GET DATA TYPE=TXT. Patch #6412.
4 Thanks to John Darrington for review.
6 * data-in.c (data_in): Add new argument to designate the last
7 column of the data field being parsed, for use in error messages.
10 * data-parser (struct data_parser): New member `quote_escape'.
11 (data_parser_create): Initialize quote_escape.
12 (data_parser_set_quotes): New function.
13 (cut_field): Support escaped quotes.
14 (parse_delimited_span): Ditto.
15 (parse_delimited_no_span): Ditto.
17 * get-data.c (parse_get_txt): Support ESCAPE extension subcommand
18 in enhanced mode. Only support multiple quote characters in
21 2008-02-06 John Darrington <john@darrington.wattle.id.au>
23 psql-reader.c psql-reader.h: Read more than one tuple at
24 once. Fix bug reading a query which returns no data. Fix bug
25 when transformation followed a reader.
26 Ask the server for the number of records in the query, for the
29 2008-02-05 John Darrington <john@darrington.wattle.id.au>
31 psql-reader.c: So yesterday they release postgresql 8.3.0
32 which has money represented with 64 bits. They must get
35 2008-02-02 John Darrington <john@darrington.wattle.id.au>
37 psql-reader.c psql-reader.h: New files. Thanks to Ben Pfaff
38 for reviewing this code.
40 2008-02-02 Ben Pfaff <blp@gnu.org>
44 * sys-file-reader.c (read_variable_record): Allow missing values
45 to be specified on long string variables, but warn about them
46 because PSPP does not yet support them.
47 (read_extension_record): Ignore extension records 20 and 21, which
48 PSPP does not yet support.
49 (read_header): Fix error message when floating-point format cannot
52 2008-02-01 Ben Pfaff <blp@gnu.org>
54 Patch #6386. Thanks to John Darrington for review and for the
55 updates to gnumeric-reader.c.
57 * dictionary.c (make_hinted_name): New function.
58 (make_numeric_name): New function.
59 (dict_make_unique_var_name): New function.
61 * gnumeric-reader.c (devise_name): Removed.
62 (munge_name): Removed.
63 (gnumeric_open_reader): Use new function
64 dict_make_unique_var_name.
66 * short-names.c (set_var_short_name_suffix): Use new function
69 2008-01-19 John Darrington <john@darrington.wattle.id.au>
71 * settings.c settings.h: Moved static variables into a
72 single struct. Renamed functions to have a settings_ prefix.
74 2008-01-14 John Darrington <john@darrington.wattle.id.au>
76 * data-out.c (output_AHEX): Corrected number of bytes in
77 call to output_hex. Closes bug #22011
79 2008-01-02 John Darrington <john@darrington.wattle.id.au>
81 * variable.c variable.h: Replaced var_get_value_name with
82 var_append_value_name which doesn't use any static data.
83 Thanks to Ben for review.
85 2007-12-07 Ben Pfaff <blp@gnu.org>
89 * casegrouper.c (casegrouper_get_next_group): Cause a casegrouper
90 made from an empty casereader produce a casegrouper with no
91 groups, instead of one with one group that has no cases.
93 * casereader.c (casereader_is_empty): New function.
95 2007-12-06 Ben Pfaff <blp@gnu.org>
99 * sys-file-reader.c (read_display_parameters): Handle variable
100 display parameters record with only 2 data items per variable.
101 Reported by Guido Gay <gay@irer.it>.
103 2007-12-04 Ben Pfaff <blp@gnu.org>
105 * identifier.c (lex_id_match_n): New function.
106 (lex_id_match): Reimplement in terms of lex_id_match_n.
108 2007-11-24 Ben Pfaff <blp@gnu.org>
110 * automake.mk (src_data_libdata_a_SOURCES): Add val-type.h, to fix
113 2007-11-24 Ben Pfaff <blp@gnu.org>
115 Fix warning reported when reading back system files that include
116 very long string variables. Thanks to Guido Gay <gay@irer.it> for
119 * short-names.c (short_names_assign): Fix dumb typo, in which `i'
120 was written where `j' was meant.
122 2007-11-18 Ben Pfaff <blp@gnu.org>
124 Properly write variables that include a range of missing values to
125 system files. Thanks to Guido Gay <gay@irer.it> for reporting
128 * sys-file-writer.c (write_variable): Correctly calculate
129 n_missing_values field when writing variables that include a range
132 2007-11-10 Ben Pfaff <blp@gnu.org>
134 Cleanups and bug fixes devised while writing up documentation.
137 * automake.mk (src_data_libdata_a_SOURCES): Add new files.
139 * dict-class.c: New file.
140 (dict_class_from_id): Move here.
141 (dict_class_to_name): Move here.
143 * dict-class.h: New file.
144 (enum dict_class): Move here. Change from consecutive integers to
145 consecutive bits, to make testing for any of multiple values
146 easier. Add new DC_ALL constant.
148 * dictionary.c (struct dictionary): Change `case_limit' from
149 size_t to casenumber.
150 (dict_get_vars): Make final argument an enum dict_class.
151 (dict_get_vars_mutable): Ditto.
152 (dict_get_case_limit): Change return value to casenumber.
153 (dict_set_case_limit): Change final argument to a casenumber.
154 (dict_unset_split_var): Add assertion. Rephrase slightly.
155 (dict_set_label): Use xstrndup to simplify.
157 * format.c (fmt_step_width): AHEX format also needs 2-byte
159 (fmt_set_style): Simplify assertions.
161 * missing-values.c (mv_add_num_range): Rename mv_add_range.
162 Simplify implementation.
163 (mv_has_value): Simplify implementation.
164 (mv_pop_value): Remove the first value, not the last, to avoid
165 having GET followed by SAVE reverse the order of missing values.
166 (mv_peek_value): Rename mv_get_value. Simplify assertion.
167 (mv_has_range): Simplify implementation.
168 (mv_peek_range): Rename mv_get_range.
169 (can_resize_string): Removed.
170 (mv_is_resizable): Use value_is_resizable.
171 (mv_resize): Use value_resize.
173 * short-names.h (SHORT_NAME_LEN): Move here.
175 * val-type.h: New file, for definitions related to type and width
176 of abstract values. Before, these definitions were mixed among
177 those related to "union value" and those related to variables.
178 (macro SYSMIS): Move here.
179 (macro LOWEST): Move here.
180 (macro HIGHEST): Move here.
181 (macro MAX_STRING); Move here.
182 (enum val_type): New enum with values VAL_NUMERIC and VAL_STRING.
183 Replaces enum var_type that had values VAR_NUMERIC and VAR_STRING.
184 All references updated.
185 (val_type_is_valid): New function. Replaces var_type_is_valid.
186 All references updated.
187 (val_type_from_width): New function. Replaces
188 var_type_from_width. All references updated.
190 * value-labels.c (val_labs_copy): Renamed val_labs_clone. All
192 (val_labs_can_set_width): Use value_is_resizable.
193 (val_labs_add): Simply return false if the value labels set is too
194 wide, instead of having undefined behavior.
195 (val_labs_replace): Ditto.
196 (val_labs_replace): Ditto.
197 (val_labs_first): Set iterator to null if iteration is complete.
198 (val_labs_first_sorted): Ditto.
199 (val_labs_done): Become a no-op if the iterator is null.
201 * value.c (value_is_resizable): New function.
202 (value_resize): New function.
204 * variable.c (var_get_dict_class): New function.
206 * variable.h (macro LONG_NAME_LEN): Rename VAR_NAME_LEN. Update
209 2007-11-08 Ben Pfaff <blp@gnu.org>
211 * data-in.c: Make formatted data parsing locale-independent.
212 (parse_number): Use c_strtod instead of strtod, to avoid
213 locale-specific behavior.
216 2007-11-06 Ben Pfaff <blp@gnu.org>
218 Patch #6256: add support for binary, 360 file formats. Reviewed
221 * data-in.c (struct data_in): Add `encoding' member.
222 (data_in): Add `encoding' parameter, and re-encode the data passed
223 in where appropriate. Update all callers to pass it in.
224 (parse_A): Implement EBCDIC recoding wart described in manual.
225 (parse_AHEX): Implement EBCDIC recoding.
227 * data-out.c (data_out_legacy): New function.
228 (data_out): Make into a wrapper around data_out_legacy.
230 * file-handle-def.c (struct file_handle): New member `encoding'.
231 (fh_create_file): Set encoding.
232 (fh_default_properties): Set default encoding.
233 (fh_get_legacy_encoding): New function.
235 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
236 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
237 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
238 (struct fh_properties): New member `encoding'.
240 2007-11-05 Ben Pfaff <blp@gnu.org>
242 Patch #6258. Reviewed by John Darrington.
244 * file-handle-def.c (fh_lock): Add comment that TYPE should be
245 marked with N_() in the caller. Added these markings to each
246 caller too. Should make i18n easier.
247 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
249 2007-11-03 Ben Pfaff <blp@gnu.org>
251 Allow output files to overwrite input files (bug #21280).
253 * file-handle-def.c: Separate locking of files for input and for
254 output, to allow both to take place at once. Also, distinguish a
255 file handle from the identity of the underlying file, because the
256 identity of a file changes over time and the file handle can't
257 represent two different identities.
258 (struct file_handle): Remove `next', `open_cnt', `deleted',
259 `type', `open_mode', `aux', `identity' members. Change `id' from
260 char array to char *. Add `ref_cnt' member.
261 (file_handle_from_ll) New function.
262 (file_handles) Removed.
263 (named_handles) New variable.
264 (fh_init) Initialize named_handles.
265 (fh_done) Remove name from all named_handles.
266 (free_handle) Updated for modified struct file_handle.
267 (unname_handle) New function.
268 (fh_ref) New function.
269 (fh_from_file_name) Removed.
270 (fh_unref) New function.
271 (fh_unname) New function.
272 (fh_from_id) Rewritten.
273 (create_handle) Updated for modified struct file_handle.
274 (fh_inline_file) Increment the handle's ref count.
275 (fh_create_file) Updated for modified struct file_handle.
276 (fh_create_scratch) Ditto.
281 (fh_is_open) Removed.
282 (fh_get_id) Updated for modified struct file_handle.
283 (fh_get_default_handle) Increment the handle's ref count.
284 (fh_set_default_handle) Handle ref counts.
285 (struct fh_lock) New structure.
286 (locks) New static var.
287 (fh_lock) New function.
288 (fh_unlock) New function.
289 (fh_lock_get_aux) New function.
290 (fh_lock_set_aux) New function.
291 (fh_is_locked) New function.
292 (make_key) New function.
293 (free_key) New function.
294 (compare_fh_locks) New function.
295 (hash_fh_lock) New function.
297 * file-handle-def.h (enum fh_access) New enum.
299 * file-name.c: Made file_identity the same in all supported
301 (struct file_identity): New `name' member.
302 (fn_get_identity): For a file that doesn't exist, get the
303 dev/inode of its directory plus its name. If even the directory
304 doesn't exist, just use its name. Merge the Windows
305 implementation into the Unix one.
306 (fn_compare_file_identities): Rewritten. Merge the Windows
307 implementation into the Unix one.
308 (fn_hash_identity): New function.
310 * make-file.c (struct replace_file): New structure.
311 (all_files): New static var.
312 (replace_file_start): New function.
313 (replace_file_commit): New function.
314 (replace_file_abort): New function.
315 (free_replace_file): New function.
316 (unlink_replace_files): New function.
318 * por-file-reader.c (struct pfm_reader): Add `lock' member.
319 (close_reader): Unlock file.
320 (pfm_open_reader): Lock file.
322 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
324 (pfm_open_writer): Lock file and prepare for its replacement.
325 (close_writer): Unlock file.
327 * scratch-handle.h (struct scratch_handle): Add unique_id so that
328 different generations of a scratch file can be distinguished.
330 * scratch-reader.c (scratch_reader_open): Verify that the file is
333 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
334 remove scratch_handle member.
335 (scratch_writer_open): Lock handle. Prepare to replace handle
336 data, instead of doing it immediately.
337 (scratch_writer_casewriter_destroy): Replace handle data and
340 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
341 (sfm_open_reader): Lock file.
342 (close_reader): Unlock file.
344 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
346 (sfm_open_writer): Lock file and prepare for its replacement.
347 (close_writer): Unlock file.
349 2007-11-02 Ben Pfaff <blp@gnu.org>
351 * data-out.c (output_number): Use isfinite (from C99) instead of
352 gsl_isfinite, since we now have universal support for it in
354 (output_infinite): Ditto.
355 [!HAVE_ROUND] (round): Remove definition, since we now have a
356 replacement in gnulib.
358 2007-10-27 John Darrington <john@darrington.wattle.id.au>
360 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
361 wasn't initialised correctly. Closes bugs #21330 and 21397
364 2007-10-08 Ben Pfaff <blp@gnu.org>
366 Fix bug #21280. Thanks to John Darrington for review.
368 * file-name.c (create_stream): New function.
370 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
371 file handle before creating the file, to ensure that we don't
372 truncate a file that we're reading. Make code easier to read by
375 * sys-file-write.c (sfm_open_writer): Ditto.
377 2007-10-01 Ben Pfaff <blp@gnu.org>
379 Fix bug #21192. Thanks to John Darrington for review.
381 * casereader.c (casereader_read): Decrement case_cnt before
382 calling the casereader's "read" member function, so that we
383 interact properly with lazy_casereader.
385 * datasheet.c: Add regression test for above bug fix.
386 (clone_datasheet): New function.
387 (lazy_callback): New function.
388 (check_datasheet_casereader): New function.
389 (check_datasheet): Check datasheet contents are reported correctly
390 through an ordinary casereader and a lazy casereader.
391 (clone_model): Use clone_datasheet.
393 2007-09-24 Ben Pfaff <blp@gnu.org>
395 Patch #6210. Reviewed by John Darrington.
396 * settings.c: Drop "long view", which is not needed any longer.
397 (static var long_view): Removed.
398 (force_long_view): Removed.
399 (get_viewwidth): Removed.
400 (init_viewport): Removed long_view reference.
402 2007-09-19 John Darrington <john@darrington.wattle.id.au>
404 * settings.c settings.h: Changed viewport's length and width to be
405 owned by the user interface which uses the data library. This allows
406 better abstraction, and makes dynamically adjustable dimensions easier.
408 2007-09-18 Ben Pfaff <blp@gnu.org>
410 * procedure.c (proc_extract_active_file_data): New function.
412 * lazy-casereader.h: New file.
414 * lazy-casereader.c: New file.
416 * casereader.c (casereader_dynamic_cast): New function.
418 2007-09-14 Ben Pfaff <blp@gnu.org>
420 * dictionary.c (dict_clone): Copy case indexes from cloned
421 dictionary. Fixes bug #21061. Reviewed by John Darrington.
423 2007-09-13 John Darrington <john@darrington.wattle.id.au>
425 * value.c value.h (create_value): New function.
427 2007-09-12 Ben Pfaff <blp@gnu.org>
429 Make it clear that translator casereader and casewriter translate
430 functions are supposed to destroy their input case.
432 * casereader-translator.c (struct casereader_translator): Change
433 input case parameter of translate member function from const
434 struct ccase * to struct ccase *.
435 (casereader_create_translator): Ditto, for translate parameter.
437 * casewriter-translator.c (struct casewriter_translator): Ditto.
438 (casewriter_create_translator): Ditto.
440 2007-08-27 John Darrington <john@darrington.wattle.id.au>
442 * sys-file-reader.c (read_display_parameters): Force display width
443 to 8, if the sys file says 0 (like SPSS does).
445 2007-08-12 Ben Pfaff <blp@gnu.org>
447 * dictionary.c (dict_dump): New function.
449 2007-08-12 Ben Pfaff <blp@gnu.org>
451 Drop dict_compactor in favor of using the new struct case_map.
453 * dictionary.c (struct copy_map): Removed.
454 (struct dict_compactor): Removed.
455 (dict_make_compactor): Removed.
456 (dict_compactor_compact): Removed.
457 (dict_compactor_destroy): Removed.
459 * procedure.c (struct data_set): Change `compactor' member to be a
461 (proc_open): Use case_map_to_compact_dict instead of
463 (proc_casereader_read): Use case_map_execute instead of
464 dict_compactor_compact.
465 (proc_commit): Use case_map_destroy instead of
466 dict_compactor_destroy.
468 * scratch-writer.c (struct scratch_writer): Change `compactor'
469 member to be a struct case_map *.
470 (scratch_writer_open): Use case_map_to_compact_dict instead of
472 (scratch_writer_casewriter_write): Use case_map_execute instead of
473 dict_compactor_compact.
475 2007-08-12 Ben Pfaff <blp@gnu.org>
477 * automake.mk: Add case-map.c, case-map.h.
479 * case-map.c: New file.
481 * case-map.h: New file.
483 2007-08-12 Ben Pfaff <blp@gnu.org>
485 * dictionary.c (dict_compact_values): Don't delete scratch
486 variables as well as compacting case indexes. Update all callers.
487 (dict_get_compacted_value_cnt): Rename dict_count_values and
488 change interface. Update all callers.
489 (dict_get_compacted_value_cnt): Remove.
490 (dict_compacting_would_shrink): Remove.
491 (dict_compacting_would_change): Remove.
492 (dict_make_compactor): Add new parameter. Update all callers.
494 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
495 instead of dict_count_values, changing an O(N) operation into
498 2007-08-12 Ben Pfaff <blp@gnu.org>
500 * casereader.c (casereader_read): Don't require cases read by a
501 casereader to be exactly the expected size: as long as they're big
504 2007-08-12 Ben Pfaff <blp@gnu.org>
506 Make casewriters keep track of the number of `union value's in
507 each case. This is useful for two reasons: casewriter_write can
508 then check that the case being written is large enough, and later
509 recipients of the casewriter can determine the size of the case.
511 * casewriter-translator.c (casewriter_create_translator): Add
514 * casewriter.c (struct casewriter): Add value_cnt member.
515 (casewriter_write): Check that the case passed in is big enough.
516 (casewriter_get_value_cnt): New function.
517 (casewriter_create): Add value_cnt parameter.
519 2007-08-09 Ben Pfaff <blp@gnu.org>
521 Fix bug reported by Jason Stover.
522 * settings.c: Move get_termcap_viewport and in particular the
523 #include for <curses.h> to the end of the file. curses.h
524 redefines bool on some systems (e.g. OpenBSD), which causes
525 disagreement between uses of bool before its inclusion and
527 Tested by Jason Stover.
529 2007-07-29 Ben Pfaff <blp@gnu.org>
531 Provisional fix for bug #18692 and bug #20161. Reviewed by John
534 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
535 argument (never "rb" or "wb") because SUSv3 says that only those
536 modes are defined, and glibc in fact rejects other modes.
538 Open portable files with fn_open so that they can be read from
539 pipes. Fix missing fh_close call to go along with fh_open.
540 Report an error if the file close reports an error.
541 * por-file-reader.c (close_reader): New function.
542 (por_file_casereader_destroy): Use close_reader.
543 (pfm_open_reader): Open file with fn_open.
545 2007-07-28 Ben Pfaff <blp@gnu.org>
547 Make PSPP able to read all the portable files I could find on the
548 web. Thanks to John Darrington for review. Bug #17620.
549 * por-file-reader.c (struct pfm_reader): New member `line_length'.
550 (error): Print file offset in hexadecimal.
551 (warning): New function.
552 (advance): Treat lines less than 80 bytes long as padded to 80
554 (pfm_open_reader): Call read_documents if we find an "E" record.
555 (convert_format): Convert invalid formats to the default format
556 instead of aborting reading the file.
557 (read_variables): Rename duplicate variable names instead of
558 aborting reading the file.
559 (read_value_label): Allow string variables of different widths to
560 be assigned value labels in the same record. Replace duplicate
561 value labels instead of aborting.
562 (read_documents): New function.
564 * por-file-writer.c (pfm_open_writer): Call write_documents if the
565 dictionary has documents.
566 (write_documents): New function.
568 2007-07-25 Ben Pfaff <blp@gnu.org>
570 Fix bugs related to bug #17213.
572 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
573 since the former is what config.h has. Include the needed ncurses
575 (static var echo) Rename to `do_echo' because the original name is
576 the same as an ncurses identifier.
577 (get_termcap_viewport) Use error instead of msg.
579 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
580 (fn_close): Don't close stdin, stdout, stderr.
582 2007-07-26 John Darrington <john@darrington.wattle.id.au>
584 * procedure.c procedure.h: Added callbacks which get invoked whenever
585 a dataset's transformation chain changes.
587 2007-07-24 Ben Pfaff <blp@gnu.org>
590 * sys-file-writer.c (write_variable_display_parameters): Use new
591 var_default_display_width function to choose display width of
592 segments after the first one in a given variable.
593 * variable.c (var_create): Use var_default_display_width to pick
594 new variable's display width.
595 (var_default_display_width): New function.
596 Reviewed by John Darrington.
598 2007-07-24 Ben Pfaff <blp@gnu.org>
601 * por-file-writer.c (write_variables): Write weight variable.
602 Reviewed by John Darrington.
604 2007-07-23 Ben Pfaff <blp@gnu.org>
606 Improvements to system file reader and writer.
608 First, move all detailed knowledge of very long strings into
609 sys-file-private.[ch], so that this nasty stuff can be isolated.
611 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
612 (EFFECTIVE_VLS_CHUNK): New macro.
613 (min_int): New function.
614 (max_int): New function.
615 (sfm_width_to_bytes): Rewrite.
616 (sfm_width_to_octs): New function.
617 (sfm_segment_alloc_width): New function.
618 (sfm_segment_alloc_bytes): New function.
619 (sfm_segment_used_bytes): New function.
620 (sfm_segment_offset): New function.
621 (sfm_segment_effective_offset): New function.
622 (sfm_dictionary_to_sfm_vars): New function.
624 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
625 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
626 (struct sfm_var): New structure.
628 Next, improvements to the system file reader.
630 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
631 type casenumber. Added `version_major', `version_minor',
634 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
635 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
636 Change `case_cnt' to type casenumber. Removed `has_vls'.
637 (struct sfm_var): Removed.
638 (sfm_open_reader): Don't warn on wrong case size if the file was
639 written by SPSS 13, which tends to get it wrong. Use
640 sfm_dictionary_to_sfm_vars.
641 (read_header): Always output system file info.
642 (read_variable_record): Simplify code for reading missing values.
643 (read_machine_int32_info): Save version numbers from system file
644 into info struct passed as new argument.
645 (read_long_string_map): Restructured to use new sys-file-private
647 (read_value_labels): Use size_overflow_p.
648 (sys_file_casereader_read): Get rid of distinction between fast
649 and slow paths. Use information provided by sys-file-primate's
650 struct sfm_var to simplify code.
651 (skip_whole_strings): New function.
652 (read_int32): Renamed read_int. Changed return value to int.
654 (read_flt64): Renamed read_float. Changed return value to
655 double. Updated all callers.
656 (int32_to_native): Removed. Changed callers to use
658 (flt64_to_double): Removed. Changed callers to use float_convert.
660 Finally, get rid of int32, flt64 terminology and types in system
661 file writer. The former wasn't very useful since a POSIX "int"
662 can hold the whole range of int32 and we generally didn't have a
663 need for it to be exactly-32-bits, just at-least-32-bits. The
664 latter was inconvenient because we had to assume that it could be
665 different from double and thereby convert special values SYSMIS,
666 HIGHEST, LOWEST to and from it in multiple places. Instead, now
667 we just use "int" and "double" in most places, and do conversions,
668 if necessary, very close to where we do I/O. This change meant
669 that the writer code couldn't represent records in the file as C
670 structs any longer, but that's no great loss. The code actually
671 seems to be more readable without them.
673 Simplify the compression buffering code: only buffer as much as
674 necessary, which is no more than eight 8-byte units at any given
677 * sys-file-writer.c (typedef flt64): Removed.
678 (macro second_lowest_flt64): Removed.
679 (struct sysfile_header): Removed.
680 (struct sysfile_variable): Removed.
681 (struct sfm_writer): Removed `needs_translation', `has_vls',
682 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
683 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
684 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
685 compression buffering by `opcodes', `opcode_cnt', `data',
686 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
687 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
688 data writer functions instead of structures.
689 (calc_oct_idx): New function.
690 (write_header): Use simple data writer functions instead of
692 (write_format_spec): Renamed write_format. New argument.
693 (write_variable_continuation_records): New function.
694 (write_variable): Use simple data writer functions instead of
695 structures. Use write_variable_continuation_records. Write
696 entire very long string instead of requiring caller to understand
698 (write_value_labels): Use simple data writer functions instead of
700 (write_documents): Ditto.
701 (write_variable_display_parameters): Use sys-file-private
702 functions to simplify. Use simple data writer functions instead
704 (write_vls_length_table): Use simple data writer functions instead
706 (write_longvar_table): Ditto.
707 (write_rec_7_34): Break into new functions
708 write_integer_info_record, write_float_info_record. Use simple
709 data writer functions instead of structures.
710 (buf_write): Removed.
711 (append_string_max): Removed.
712 (ensure_buf_space): Removed.
713 (sys_file_casewriter_write): Get rid of the distinction between
714 fast and slow paths, which didn't seem to be too useful. Use new
715 functions write_case_uncompressed, write_case_compressed.
716 (put_instruction): Removed.
717 (put_element): Removed.
718 (write_compressed_data): Removed.
719 (close_writer): Use flush_compressed. Only write case count to
720 system file if it will fit in the field.
721 (write_case_compressed): New function.
722 (write_case_uncompressed): New function.
723 (flush_compressed): New function.
724 (put_cmp_opcode): New function.
725 (put_cmp_number): New function.
726 (write_int): New function.
727 (convert_double_to_output_format): New function.
728 (write_float): New function.
729 (write_value): New function.
730 (write_string): New function.
731 (write_bytes): New function.
732 (write_zeros): New function.
733 (write_spaces): New function.
735 Reviewed by John Darrington.
737 2007-07-22 Ben Pfaff <blp@gnu.org>
739 Don't try to write very long strings to portable files. The
740 format does not support it.
742 * por-file-writer.c (MAX_POR_WIDTH): New macro.
743 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
744 (write_format): Add arg to take width to resize format to.
745 (write_value): Limit width of value written to MAX_POR_WIDTH.
746 (write_variables): Limit width of variable and its output formats
748 Reviewed by John Darrington.
750 2007-07-22 Ben Pfaff <blp@gnu.org>
752 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
753 local variable instead of literal 5.
754 Reviewed by John Darrington.
756 2007-07-22 Ben Pfaff <blp@gnu.org>
758 Fix problems with uniqueness of short names in system files with
759 very long string variables. Now a variable may have multiple
762 * automake.mk (src_data_libdata_a_SOURCES): Add new files
763 short-names.c, short-names.h.
765 * dictionary.c (dict_clone): Clone all the short names.
766 (compare_strings): Move into short-names.c.
767 (hash_strings): Ditto.
768 (set_var_short_name_suffix): Ditto.
769 (dict_assign_short_names): Ditto, rename short_names_assign,
770 change to assign all short names.
772 * por-file-writer.c (write_variables): Use short_names_assign
773 instead of dict_assign_short_names.
775 * short-names.c: New file.
777 * short-names.h: New file.
779 * sys-file-private.c (sfm_width_to_segments): New function.
781 * sys-file-reader.c (read_long_var_name_map): Save and restore all
782 the short names, not just the first one.
784 * sys-file-writer.c (cont_var_name): Removed.
785 (sfm_open_writer): Use short_names_assign instead of
786 dict_assign_short_names. Use unique short names assigned by
787 short_names_assign instead of those generated by cont_var_name.
789 * variable.c (struct variable): Remove `short_name' member,
790 replace by `short_names' and `short_name_cnt'.
791 (var_create) Initialize new members.
792 (var_get_short_name_cnt): New function.
793 (var_get_short_name): Now takes an index argument. Changed most
795 (var_set_short_name): Ditto.
796 (var_clear_short_name): Renamed var_clear_short_names, changed to
797 clear all short names.
799 Reviewed by John Darrington.
801 2007-07-22 Ben Pfaff <blp@gnu.org>
803 * variable.c (var_set_width): Use new var_set_width function.
805 * missing-values.c (mv_n_values): Drop assertion, which was not
808 * format.c (fmt_default_for_width): New function.
809 (fmt_resize): New function.
811 Reviewed by John Darrington.
813 2007-07-18 John Darrington <john@darrington.wattle.id.au>
815 * datasheet.c (datasheet_delete_columns): Added assertion to check
816 we're not deleting outside the range of the sheet.
819 * dictionary.c dictionary.h variable.c: Added the ability for string
820 variables to be resized.
822 * vardict.h: Added some prototypes (moved from dictionary.h) as
823 these should only be called by variable.c
826 2007-07-14 John Darrington <john@darrington.wattle.id.au>
828 * sfm-reader.c: Respect case_cnt field in file header.
830 2007-07-01 John Darrington <john@darrington.wattle.id.au>
832 * transformation.c transformation.h (trns_chain_execute): Changed the
833 signature (Patch #6057)
835 2007-06-10 Ben Pfaff <blp@gnu.org>
837 * casereader-filter.c (casereader_filter_destroy): Make sure to
838 write all the remaining excluded cases to the casewriter, if any.
840 * caseinit.c (init_list_destroy): Rewrite.
841 (init_list_clear): Ditto.
843 * casegrouper.c (casegrouper_get_next_group): Always set *reader
844 to null when returning false.
846 2007-06-06 Ben Pfaff <blp@gnu.org>
848 Actually implement the new procedure code and adapt all of its
849 clients to match. Also adapt all of the other case sources and
850 sinks in the tree and their clients to use the
851 casereader/casewriter infrastructure.
853 * automake.mk: Add and remove files.
855 * any-reader.c: Change into a casereader.
856 * por-file-reader.c: Ditto.
857 * scratch-reader.c: Ditto.
858 * sys-file-reader.c: Ditto.
860 * any-writer.c: Change into a casewriter.
861 * por-file-writer.c: Ditto.
862 * scratch-writer.c: Ditto.
863 * sys-file-writer.c: Ditto.
865 * procedure.c: Change to use casereader, casewriter, caseinit, and
866 other new infrastructure.
868 * scratch-handle.c: Adapt to new infrastructure.
870 * case-sink.c: Removed, now dead code.
871 * case-sink.h: Ditto.
872 * case-source.c: Ditto.
873 * case-source.h: Ditto.
874 * casefile-factory.c: Ditto.
875 * casefile-private.h: Ditto.
878 * casefilter.c: Ditto.
879 * casefilter.h: Ditto.
882 * fastfile-factory.c: Ditto.
883 * fastfile-factory.h: Ditto.
884 * storage-stream.c: Ditto.
885 * storage-stream.h: Ditto.
887 2007-06-06 Ben Pfaff <blp@gnu.org>
891 * automake.mk: Add new files.
893 * datasheet.c: New file.
895 * datasheet.h: New file.
897 2007-06-06 Ben Pfaff <blp@gnu.org>
899 Until now, the procedure code has provided a case to the
900 case_source, which has filled in the data values that come from
901 the active file. "Left" data values that don't come from the
902 active file naturally stay the same from case to case, because the
903 procedure code keeps using that same case.
905 One of the compromises that comes with the new procedure code is
906 that the active file allocates and provides its own case, which
907 the procedure code then has to resize to provide room for any
908 other variables that should go in the case and then fill in the
909 values of "left" variables. Then, when we're done with that case,
910 we have to save the values of "left" variables to copy into the
911 next case read from the active file.
913 The caseinit code helps with this.
915 * automake.mk: Add new files.
917 * caseinit.c: New file.
919 * caseinit.h: New file.
921 2007-06-06 Ben Pfaff <blp@gnu.org>
923 * value.h (value_cnt_from_width): New function.
925 * variable.c (var_get_value_cnt): Use new function.
927 2007-06-06 Ben Pfaff <blp@gnu.org>
929 Add casegrouper, to allow cases read from a given casereader to be
930 broken into groups, each of which has its own casereader.
931 Generally cases are grouped based on having equal values for some
934 * automake.mk: Add new files.
936 * casegrouper.c: New file.
938 * casegrouper.h: New file.
940 2007-06-06 Ben Pfaff <blp@gnu.org>
942 Add interface to lexicographical ordering of cases.
944 * automake.mk: Add new files.
946 * case-ordering.c: New file.
948 * case-ordering.h: New file.
950 2007-06-06 Ben Pfaff <blp@gnu.org>
952 Add casereaders and casewriters, the basis of the new data processing
953 implementation. A casereader is a uniform interface to reading cases
954 from a data source; a casewriter is a uniform interface to writing
955 cases to a data sink.
957 * automake.mk: Add new files.
959 * casereader-filter.c: New file.
961 * casereader-provider.h: New file.
963 * casereader-translator.c: New file.
965 * casereader.c: New file.
967 * casereader.h: New file.
969 * casewriter-provider.h: New file.
971 * casewriter-translator.c: New file.
973 * casewriter.c: New file.
975 * casewriter.h: New file.
977 2007-06-06 Ben Pfaff <blp@gnu.org>
979 "casewindow" data structure that extends the deque (from libpspp)
980 of cases with the ability to dump cases to disk if we get too many
983 * automake.mk: Add new files.
985 * casewindow.c: New file.
987 * casewindow.h: New file.
989 2007-06-06 Ben Pfaff <blp@gnu.org>
991 sparse_cases data structure that augments a sparse_array of cases
992 with the ability to dump cases to disk if we get too many cases in
995 * automake.mk: Add new files.
997 * sparse-cases.c: New file.
999 * sparse-cases.h: New file.
1001 2007-06-06 Ben Pfaff <blp@gnu.org>
1003 Adds a low-level on-disk case array data structure.
1005 * automake.mk: Add new files.
1007 * case-tmpfile.c: New file.
1009 * case-tmpfile.h: New file.
1011 2007-06-06 Ben Pfaff <blp@gnu.org>
1013 In a couple of places we calculate the maximum number of cases to
1014 keep in memory based on the user-defined workspace. Enable
1015 centralizing the calculation through a new function.
1017 * settings.c (get_workspace_cases): New function.
1019 2007-06-06 Ben Pfaff <blp@gnu.org>
1021 The casenumber type is defined in transformations.h, but case.h is
1022 a more sensible place. Move it.
1024 * case.h (CASENUMBER_MAX): New macro.
1025 (typedef casenumber): Move here, from transformations.h.
1027 2007-06-03 Ben Pfaff <blp@gnu.org>
1029 Slightly generalize case_to_values and case_from_values functions.
1031 * case.c (case_to_values): Rename case_copy_out, change interface.
1032 (case_from_values): Rename case_copy_in, change interface.
1034 * fastfile.c (fastfilereader_get_next_case): Update caller.
1035 (write_case_to_disk): Ditto.
1037 2007-06-02 Ben Pfaff <blp@gnu.org>
1039 Clean up after a forgotten part of patch #5829.
1041 * casedeque.h: Remove unused file.
1043 * automake.mk: Remove casedeque.h from sources.
1045 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
1047 * category.c: Removed redundant #include
1049 2007-05-06 Ben Pfaff <blp@gnu.org>
1051 Abstract the documents within a dictionary a little better.
1052 Thanks to John Darrington for suggestion, initial version, and
1053 review. Patch #5917.
1055 * dictionary.c (struct dictionary): Change `documents' member from
1056 char * to struct string.
1057 (dict_clear): Destroy struct string.
1058 (dict_get_documents): Convert struct string to char *.
1059 (dict_set_documents): Set struct string. Pad to 80-character
1061 (dict_clear_documents): New function.
1062 (dict_add_document_line): New function.
1063 (dict_get_document_line_cnt): New function.
1064 (dict_get_document_line): New function.
1066 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
1068 * sys-file-reader.c (read_documents): Use new document functions.
1070 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1072 * sys-file-reader.c: When reading a system file which has no
1073 long name table, automatically create one where the long names
1074 are the lower case versions of the short names.
1076 2007-04-22 Ben Pfaff <blp@gnu.org>
1078 * dictionary.c (dict_set_split_vars): dict_destroy expects that
1079 dict_clear will free most data related to the dictionary.
1080 dict_clear does a decent job, except that dict_set_split_vars on
1081 some systems won't actually free the dict's "split" member.
1082 Instead, it'll allocate a 1-byte region. Fix this.
1084 * value.c (value_copy): New function.
1085 (value_set_missing): Ditto.
1087 2007-04-22 John Darrington <john@darrington.wattle.id.au>
1089 * Deleted existing category.h and moved cat-routines.h into
1090 category.h Encapsulated struct cat_vals better.
1092 2007-04-19 John Darrington <john@darrington.wattle.id.au>
1094 * sys-file-reader.c: When reading a system file which has no
1095 long name table, automatically create one where the long names
1096 are the lower case versions of the short names.
1098 2007-04-16 John Darrington <john@darrington.wattle.id.au>
1100 * sys-file-reader.c: Some versions of Other Software seem to
1101 produce system files with string variables' measure set to
1102 zero. We'll assume these are supposed to be nominal variables.
1104 2007-03-30 Ben Pfaff <blp@gnu.org>
1106 * procedure.c: Adapt to new deque data structure.
1108 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
1109 Ben Pfaff <blp@gnu.org>
1111 * file-name.c: Mingw compatibility fixes.
1112 (fn_search_path): Use ISSLASH instead of comparing against '/'
1114 (fn_dir_name): Use dir_name from gnulib.
1115 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
1116 (fn_get_identity): Use GetFullPathName instead of canonicalize
1117 from gnulib, because the latter does not fully support
1118 Windows-style path names. Use this implementation based on the
1119 detected presence of Windows instead of the absence of Unix, since
1120 the new implementation is Windows-specific.
1121 (fn_compare_file_identities): In Windows implementation, compare
1122 names case-insensitively.
1124 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
1126 * make-file.c: Don't include mkstemp.h, because gnulib now causes
1127 <stdlib.h> to have the same effect.
1129 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
1131 * por-file-reader.c: Add missing _() around messages.
1133 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
1135 * make-file.c: Include "mkstemp.h", without which linking on
1138 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
1140 Reduce platform dependence.
1142 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
1143 Everywhere we using this, we really should have just depended on
1144 the shell to expand tildes.
1145 (fn_search_path): Simplify, given that we don't do tilde expansion
1147 (fn_normalize): Removed. Caller changed to use the canonicalize
1149 (fn_get_cwd): Removed. Only user was fn_normalize.
1150 (fn_is_absolute): Really only test for absolute names.
1151 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
1153 (fn_readlink): Removed, as it was only used fn_normalize.
1154 (fn_exists): Assume the stat function is available; gnulib does.
1155 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
1157 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
1159 * dictionary.c (dict_create_vector_assert): New function.
1161 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
1163 * file-name.c (fn_normalize): Correct name of function
1164 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
1165 for pointing this out.
1167 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
1169 We are using a single member in struct file_handle, the "name"
1170 field, for more than one purpose. When it begins with '"', it's a
1171 file name; otherwise, it's a token that can be used to identify
1172 it. When that assertion fires, it's because we searched for the
1173 name case-sensitively as a file name (so that there was no match),
1174 and then we try to insert it case-insensitively as a token, which
1175 fails because duplicates aren't allowed.
1177 Solution: break the two purposes into two separate fields. This
1178 fixes the problem and likely makes the code easier to read too.
1180 Fixes bug #18922. Thanks to John Darrington for bug report and
1183 * file-handle-def.c (struct file_handle): New `id' member.
1184 (fh_from_name): Rename fh_from_id. Update all callers.
1185 (create_handle): New `id' parameter. Update all callers.
1186 (fh_create_file): Ditto.
1187 (fh_get_id): New function.
1189 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
1191 * case.c (case_is_null): Change return type to bool.
1193 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
1197 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
1198 reference counting to share data. This makes it easy for
1199 valgrind, etc. to find accesses to cases that have been destroyed
1200 but have been kept around by another user's ref-count. This often
1201 happens when the data set is small enough to find in memory; if a
1202 bigger data set that would overflow to disk were used, then data
1203 corruption would occur.
1205 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1209 * case.c (case_unshare): Make it check internally whether the
1210 ref_cnt is greater than 1, so that the callers don't have to.
1211 Update callers not to check.
1213 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1215 Before, I was thinking that I might want to get rid of reference
1216 counting at some point. Now, I'm pretty sure that it's here to
1217 stay. Thus, because we have to store the value_cnt anyway for
1218 reference-counted cases, we might as well expose it to users.
1220 * case.c (case_get_value_cnt): New function.
1221 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1222 resize case if its size actually changed.
1224 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1225 instead of peeking inside struct case directly.
1226 (casefile_append): Ditto.
1228 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1230 Get rid of the inlines for the case functions, which made the
1231 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1232 the inlines didn't speed up "make check" at all, which is not a
1233 perfect benchmark but seems indicative.)
1235 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1236 definitions. Remove some assertions on nonnull pointers that were
1237 redundant with a pointer dereference soon after in the function.
1239 (struct case_data): Move definition here from case.h.
1243 (case_data_wr): Ditto.
1245 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1247 * automake.mk: Add casedeque.h to sources.
1249 * casedeque.h: New file.
1251 * procedure.c: (struct dataset) Change lag_count, lag_head,
1252 lag_queue member into single struct casedeque member. Update all
1253 users to use the casedeque instead.
1256 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1258 * procedure.c: Simplify lagged cases interface. Updated all
1259 clients--well, the only client--to use the simplified interface.
1260 (dataset_n_lag) Removed.
1261 (dataset_set_n_lag) Removed.
1262 (dataset_need_lag) New function.
1264 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1266 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1267 whenever appropriate, but only when the dataset/dictionary is in a
1270 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1272 * dictionary.c dictionary.h : Added callbacks for change of filter and
1273 split variables. Refactored some code to ensure that callbacks get
1274 invoked when appropriate.
1276 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1277 replace_dict callback occurs when permanent_dict replaces the current
1280 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1282 * dictionary.c dictionary.h : Added callback for when the weight
1283 variable of a dictionary changes.
1285 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1287 * dictionary.c dictionary.h : Added replace_source and replace_dict
1288 callbacks, and functions to deal with them.
1290 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1292 Simplify missing value handling.
1294 * missing-values.h (enum mv_class): New type.
1295 (enum mv_type): Moved definition into missing-values.c and renamed
1296 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1297 enums. Updated all uses.
1298 (struct missing_values): Changed type of `type' from `enum
1299 mv_type' to `int' because the definition is no longer exposed.
1301 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1302 parameter. Update all callers.
1303 (mv_is_num_missing): Ditto.
1304 (mv_is_str_missing): Ditto.
1305 (mv_is_value_user_missing): Removed. Changed callers to use
1306 mv_is_value_missing.
1307 (mv_is_num_user_missing): Removed. Changed callers to use
1309 (mv_is_str_user_missing): Removed. Changed callers to use
1311 (mv_is_value_system_missing): Removed. Changed callers to use
1312 mv_is_value_missing.
1313 (mv_set_type): Removed. Changed callers to use mv_clear.
1314 (mv_clear): New function.
1316 * variable.c (var_is_value_missing): Add new enum mv_class
1317 parameter. Update all callers.
1318 (var_is_num_missing): Ditto.
1319 (var_is_str_missing): Ditto.
1320 (var_is_value_user_missing): Removed. Changed callers to use
1321 var_is_value_missing.
1322 (var_is_num_user_missing): Removed. Changed callers to use
1324 (var_is_str_user_missing): Removed. Changed callers to use
1326 (var_is_value_system_missing): Removed. Changed callers to use
1327 var_is_value_missing.
1329 * casefilter.c (struct casefilter): Use enum mv_class in place of
1331 (casefilter_variable_missing): Adapt to new member.
1332 (casefilter_create): Change signature to take enum mv_class,
1335 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1337 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1339 * case-sink.c case-sink.h procedure.c procedure.h
1340 storage-stream.c: Now uses the factory.
1342 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1344 Make it possible to pull cases from the active file with a
1345 function call, instead of requiring indirection through a callback
1348 * case-source.h (struct case_source_class): Change ->read function
1349 to return a single case, instead of calling a callback function
1350 for each case. Change ->destroy function to return an error
1353 * case-source.c (free_case_source): Pass along the value returned
1354 by the case_source ->destroy function.
1356 * procedure.c (struct write_case_data): Removed.
1357 (struct dataset): Added some members to track procedure state.
1358 (procedure): Optimize the trivial case at this level.
1359 (internal_procedure): Re-implement in terms of proc_open,
1360 proc_read, proc_close.
1361 (proc_open) New function.
1362 (proc_read) New function.
1363 (proc_close) New function.
1364 (write_case) Moved into proc_read.
1365 (close_active_file) Moved closing of data source into proc_close.
1367 * storage-source.c: Rewrote to conform with modified
1368 case_source_class interface.
1370 * transformations.c (trns_chain_execute): Added argument to allow
1371 starting execution from an arbitrary transformation. Updated
1374 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1376 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1378 * sys-file-reader.c (read_display_parameters): Don't assume that
1379 MEASURE_* and ALIGN_* have the same values found in system files.
1381 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1383 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1384 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1385 the value of n_MEASURES, which was off by 1 (at least from my
1388 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1390 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1391 callbacks which are invoked when the dictionary or its
1392 variables are changed.
1394 * missing-values.c missing-values.h value-labels.c: Tidied up
1395 consistency checks, and made some of them return false
1396 instead of assert-failing.
1398 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1400 * calendar.c (calendar_days_in_month): New function.
1402 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1404 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1405 the value label's width.
1407 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1409 * sfm-private.h: Move contents into sys-file-writer.c, which is
1410 the only remaining user. Removed Borland C++-specific directives.
1412 * sys-file-reader.c: Clean up and rewrite entire file. The
1413 rewritten version is simpler and better abstracted, and should be
1414 easier to maintain and extend. It avoids using structures to read
1415 file data, which is prone to padding variations among compilers.
1416 It should also handle non-IEEE 754 system files, although I
1417 haven't been able to find any. It has been tested against many
1418 .sav files obtained from the Web and found to produce the same
1419 results as the earlier version of the code, or in some cases
1420 improved results. It is more tolerant of format variations found
1423 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1424 member, putting an enum integer_format in its place. New member
1425 `float_format'. Changed `compressed' member to type bool.
1427 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1429 * dictionary.c (dict_delete_consecutive_vars): New function.
1431 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1433 * file-name.c (fn_search_path): Remove prefix arg that was unused
1434 by any caller. Updated all callers.
1436 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1438 * format.c (fmt_dollar_template): Use user's decimal point
1439 character. Add assertion.
1441 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1443 * format.c (fmt_dollar_template): New function, based on
1444 dollar_format_template from var-type-dialog.c.
1446 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1448 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1451 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1453 Finish converting struct variable to an opaque type. In this
1454 phase, we add remaining setter and getter functions, convert the
1455 remaining PSPP code to use them, and do a bunch of cleanup. The
1456 resulting changes are pervasive but mostly trivial, and only the
1457 notable changes are logged.
1459 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1462 * case.c (case_data): Renamed case_data_idx.
1463 (case_num): Renamed case_num_idx.
1464 (case_str): Renamed case_str_idx.
1465 (case_data_rw): Renamed case_data_rw_idx.
1467 * case.h (case_data): New function with old name and an interface
1468 that takes a variable instead of an index, which is easier to
1469 use. Updated all callers to use the new interface, or to use the
1470 new *_idx function (see above).
1473 (case_data_rw): Ditto.
1475 * category.c (cat_stored_values_destroy): Changed interface to
1476 take a struct cat_vals * instead of a struct variable *.
1478 * dictionary.c (dict_clone): Use new vector_clone function.
1479 (dict_clear) Use new var_destroy function.
1480 (add_var) New function.
1481 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1482 (dict_create_var_assert) Rewrite in terms of add_var.
1483 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1484 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1485 (dict_lookup_var) Use new var_create, var_destroy functions.
1486 (dict_contains_var) Rewrite in terms of new vardict functionality.
1487 (set_var_dict_index) New function.
1488 (set_var_case_index) New function.
1489 (reindex_vars) New function.
1490 (dict_delete_var) Rewrite in terms of new vardict functionality.
1491 (dict_reorder_var) Ditto.
1492 (dict_reorder_vars) Ditto.
1493 (rename_var) New function.
1494 (dict_rename_var) Use rename_var.
1495 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1496 (dict_get_compacted_idx_to_fv) Rename
1497 dict_get_compacted_dict_index_to_case_index, update callers.
1498 (dict_create_vector) Use new vector_create function.
1499 (dict_clear_vectors) Use new vector_destroy function.
1500 (set_var_short_name_suffix) Move here from variable.c, renamed
1501 from var_set_short_name_suffix, make static, update caller.
1503 * sys-file-private.c: New file.
1504 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1505 width_to_bytes, update callers.
1507 * sys-file-private.h: New file. Later it will supplant
1508 sfm-private.h; for now it supplements it.
1509 (macro MIN_VERY_LONG_STRING) New macro.
1510 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1512 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1513 MAX_LONG_STRING was used before.
1515 * sys-file-writer.c: Ditto.
1517 * value-labels.c: Change the paradigm here to be that a null
1518 pointer is OK for a struct val_labs * in most cases; it just
1519 represents an empty set of value labels.
1520 (val_labs_copy) A copy of a null set is a null set.
1521 (val_labs_count) A null set has 0 labels.
1522 (val_labs_replace) Change return type to void. Rewrite for
1524 (val_labs_find) A null set does not contain the value.
1525 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1526 transposed argument order, updated all callers.
1528 * value.c: New file.
1529 (value_dup) Moved here from variable.c.
1530 (compare_values) Ditto.
1533 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1534 (macro MAX_LONG_STRING) Removed, because it was only interesting
1535 for system files, not for general code.
1536 (macro MAX_VERY_LONG_STRING) Ditto.
1537 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1538 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1540 * vardict.h: New file, for an interface between variables and
1543 * variable.c: A lot of functions were moved around, for better
1545 (struct variable) Move definition here, from variable.h.
1546 (var_type_adj) Removed--makes i18n hard.
1547 (var_type_noun) Ditto.
1548 (var_create) New function.
1549 (var_clone) New function.
1550 (var_destroy) New function.
1551 (var_set_name) Assert that variable is not in a dictionary.
1552 (compare_var_names) Rename compare_vars_by_name and fix a couple
1553 of callers who thought the args were strings.
1554 (hash_var_name) Rename hash_var_by_name.
1555 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1556 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1557 (var_is_very_long_string) Removed, because it was only interesting
1558 to system file code.
1559 (var_set_missing_values) Allow the argument to be the wrong width,
1560 as long as we can resize it. Simplify callers who were doing the
1561 resizing themselves.
1562 (var_get_value_labels) New function.
1563 (var_has_value_labels) New function.
1564 (var_set_value_labels) New function.
1565 (alloc_value_labels) New function.
1566 (var_add_value_label) New function.
1567 (var_replace_value_label) New function.
1568 (var_clear_value_labels) New function.
1569 (var_lookup_value_label) New function.
1570 (var_get_value_name) Moved here from variable.c, renamed from
1571 var_get_value_name, transposed argument order, updated all
1573 (var_to_string) Moved here, from variable-label.c.
1574 (var_set_leave) New function.
1575 (var_get_leave) New function.
1576 (var_must_leave) New function.
1577 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1578 set_var_short_name_suffix.
1579 (var_get_dict_index) New function.
1580 (var_get_case_index) New function.
1581 (var_get_obs_vals) New function.
1582 (var_set_obs_vals) New function.
1583 (var_has_obs_vals) New function.
1584 (var_get_vardict) New function.
1585 (var_set_vardict) New function.
1586 (var_has_vardict) New function.
1587 (var_clear_vardict) New function.
1588 (value_dup) Moved to value.c.
1589 (compare_values) Ditto.
1592 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1593 (enum ALPHA) Rename VAR_STRING, update all users.
1595 * vector.c: New file.
1596 (struct vector) Moved here, from variable.h.
1597 (check_widths) New function.
1598 (vector_create) New function.
1599 (vector_clone) New function.
1600 (vector_destroy) New function.
1601 (vector_get_name) New function.
1602 (vector_get_var) New function.
1603 (vector_get_var_cnt) New function.
1604 (compare_vector_ptrs_by_name) New function.
1606 * vector.h: New file.
1608 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1610 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1611 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1614 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1616 * value-labels.c (destroy_atoms): New function.
1617 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1619 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1621 Thanks to Jason Stover for pointing out this problem.
1623 * data-out.c (output_number): Use gsl_finite from GSL, which is
1624 portable, instead of isfinite, which is not.
1627 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1629 * variable.c variable.h (value_dup): New function.
1631 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1633 Start converting struct variable to an opaque type. In this
1634 phase, we add a bunch of setter and getter functions and convert
1635 most of the PSPP code to use them. The resulting changes are
1636 pervasive but mostly trivial, and only the notable changes are
1639 * format.c (fmt_equal): New function.
1641 * variable.c (var_type_is_valid): New function.
1642 (measure_is_valid) Moved here, from format.c.
1643 (alignment_is_valid) Moved here, from format.c.
1644 (var_get_name) New function.
1645 (var_set_name) New function.
1646 (width_to_type) New function.
1647 (var_get_type) New function.
1648 (var_get_width) New function.
1649 (var_set_width) New function.
1650 (var_is_numeric) New function.
1651 (var_is_alpha) New function.
1652 (var_is_short_string) New function.
1653 (var_is_long_string) New function.
1654 (var_is_very_long_string) New function.
1655 (var_get_missing_values) New function.
1656 (var_set_missing_values) New function.
1657 (var_clear_missing_values) New function.
1658 (var_has_missing_values) New function.
1659 (var_is_value_missing) New function.
1660 (var_is_num_missing) New function.
1661 (var_is_str_missing) New function.
1662 (var_is_value_user_missing) New function.
1663 (var_is_num_user_missing) New function.
1664 (var_is_str_user_missing) New function.
1665 (var_is_value_system_missing) New function.
1666 (var_get_print_format) New function.
1667 (var_set_print_format) New function.
1668 (var_get_write_format) New function.
1669 (var_set_write_format) New function.
1670 (var_set_both_formats) New function.
1671 (var_get_label) New function.
1672 (var_set_label) New function.
1673 (var_clear_label) New function.
1674 (var_has_label) New function.
1675 (var_get_measure) New function.
1676 (var_set_measure) New function.
1677 (var_get_display_width) New function.
1678 (var_set_display_width) New function.
1679 (var_get_alignment) New function.
1680 (var_set_alignment) New function.
1681 (var_get_value_cnt) New function.
1682 (var_get_leave) New function.
1683 (var_get_short_name) New function.
1685 * variable.h: (struct variable) Removed "type" and "nv" members;
1686 they are now computed from "width" where needed.
1688 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1690 * missing-values.c (mv_resize): Don't write beyond end of the
1691 allocated buffer when resizing a long string.
1693 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1695 Clean up identifier code: don't require identifier enumerations to
1696 be in a particular order; make better use of string library;
1697 expose less of the internals.
1699 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1700 change interface. Updated all callers.
1701 (lex_id_match) Change interface to use struct substring, update
1703 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1704 (global array keywords[]) Make static, change form. Update all
1705 users to use lex_id_name instead.
1706 (lex_is_keyword) New function.
1707 (lex_id_to_token) Change interface to use struct substring, update
1709 (lex_id_name) New function.
1711 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1712 lex_is_keyword instead.
1713 (T_LAST_KEYWORD) Removed.
1714 (T_N_KEYWORDS) Removed.
1716 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1718 * format.c: (fmt_date_template) Distinguish characters for which a
1719 space is output and any date delimiter is allowed on input, from
1720 those for which a space is output and only a space is allowed on
1721 input. The former is represented by X, the latter by a space.
1722 Also, drop distinction between h and H, changing the former to the
1725 * data-in.c: Completely rewrite internals to conform to SPSS input
1726 formats as closely as possible.
1727 (data_in) Changed external interface by replacing the structure
1728 that was used as a single argument by a set of arguments. Updated
1730 (data_in_finite_line) Removed. Converted all callers to use plain
1732 (data_in_get_integer_format) New function.
1733 (data_in_set_integer_format) New function.
1734 (data_in_get_float_format) New function.
1735 (data_in_set_float_format) New function.
1737 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1738 (struct data_in) Removed.
1740 * data-out.c: (output_date) Drop each component from the input as
1741 it is output, to allow us to drop the distinction between h (a
1742 count of hours) and H (the hour of day) template characters.
1743 Also, handle new X template character.
1744 (output_scientific) Follow more rational rule on when to drop
1745 fraction introduced between SPSS 13 and 15. Updated test case to
1748 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1750 Fix buffer overflow reported by John Darrington.
1752 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1753 realize that DIGITS is a count of nibbles, not of bytes.
1755 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1757 * calendar.c (calendar_offset_to_gregorian) Also return the
1758 year-of-day. Change callers to new interface.
1760 * data-out.c: Completely rewrite internals to conform to SPSS
1761 output formats as completely as possible.
1762 (data_out) Change interface to put input parameters before output
1763 parameters, for consistency with the style I now prefer. Update
1765 (data_out_get_integer_format) New public function.
1766 (data_out_set_integer_format) New public function.
1767 (data_out_get_float_format) New public function.
1768 (data_out_set_float_format) New public function.
1770 * data-out.h: New file. Move prototype for data_out here, from
1773 * format.c: (fmt_step_width) Use equality comparison instead of
1774 bitwise and, for clarity.
1775 (fmt_is_string) Ditto.
1776 (fmt_input_to_output) Fix categories that are translated to F
1779 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1781 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1782 casefile-private.h: Added casefilter to assist commands with missing
1785 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1787 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1789 * settings.c: (route_errors_to_terminal) New variable.
1790 (route_errors_to_listing) New variable.
1791 (get_error_routing_to_terminal) New function.
1792 (set_error_routing_to_terminal) New function.
1793 (get_error_routing_to_listing) New function.
1794 (set_error_routing_to_listing) New function.
1796 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1798 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1800 * format.c: Completely rewrite, to achieve better abstraction.
1801 Rewrite all references to formats in other files.
1803 * format.def: Rewrite and reorganize.
1805 * settings.c: Move everything related to custom currency formats
1806 into format.[ch], changing them in form, so as to group related
1807 code and definitions better. Changed all references to use the
1809 (static var decimal) Removed.
1810 (static var grouping) Removed.
1811 (static var cc) Removed.
1812 (get_decimal) Removed.
1813 (set_decimal) Removed.
1814 (get_grouping) Removed.
1815 (set_grouping) Removed.
1819 * settings.h: (macro CC_CNT) Removed.
1820 (macro CC_WIDTH) Removed.
1821 (struct custom_currency) Removed.
1823 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1825 * data-in.c (data_in): Use switch statement instead of table, to
1826 avoid dependence on the order of the FMT_* enums.
1828 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1830 * data-out.c: (num_to_string) Removed, because it was dead code.
1832 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1834 * data-in.c (parse_trailer): Fix error message.
1836 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1838 * format.c (fmt_is_binary): New function.
1840 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1842 * procedure.c procedure.h: Encapsulated the static data into a single
1845 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1847 * casefile.c (casereader_read_xfer): Always initialize the case,
1848 even on an error condition.
1850 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1852 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1853 entire function useless.
1855 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1857 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1858 casereader method casereader_clone.
1860 * procedure.c pransformations.h: Introduced new type casenum_t
1862 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1864 * variable.c: (width_to_bytes) Rephrase code for clarify.
1866 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1868 * format.c: (fmt_type_from_string) New function.
1869 (fmt_to_string) Include decimals in output if the format has
1870 decimals, even if the format type does not. This way, we can
1871 accurately reproduce incorrect formats in user output.
1872 (check_common_specifier) Make the check for a bad format type an
1873 assertion, so we get bug reports if they show up. Fix message.
1874 Check for decimal places with a format type that doesn't allow
1876 (check_input_specifier) Remove check for FMT_X, which has been
1878 (check_output_specifier) Ditto.
1880 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1882 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1883 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1884 (enum fmt_parse_flags) Removed.
1886 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1888 * casefile.c casefile.h: Converted to an abstract base class.
1889 * casefile-private.h fastfile.c fastfile.h: New files.
1890 * automake.mk procedure.c scratch-writer.c storage-stream.c
1892 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1894 * procedure.c (internal_procedure): Create sink_case with only as
1895 many values as the compacted dictionary.
1897 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1899 Remove "debugging" code that caused plenty of false positives and
1902 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1904 * case.c: Remove all references to `this' member.
1906 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1908 Fix link error noted by Jason Stover.
1910 * storage-stream.c: Include <assert.h>.
1912 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1914 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1915 ALL) and additional underlying system file issues.
1917 Thanks to John Darrington for review.
1919 First problem: var_hash points to variables not owned by the
1920 sys-file-reader, which the caller may free or modify. Use an
1921 array of sfm_vars instead, as done earlier (e.g. CVS version
1924 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1925 members and remove all code that references it. Add vars, var_cnt
1926 members. Remove fix_specials member, which was unused.
1927 (struct sfm_var) Remove name member, which was unused.
1928 (sfm_close_reader) Free vars member instead of var_hash.
1929 (compare_var_shortnames) Removed.
1930 (hash_var_shortname) Removed.
1931 (sfm_open_reader) Fill out vars array.
1932 (compare_var_index) Removed.
1933 (sfm_read_case) Use vars instead of var_hash.
1935 Second problem: we're confused about when we actually have very
1936 long strings, causing us to choose incorrectly between slow path
1937 and fast path in sfm_read_case.
1939 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1940 have very long strings, not when we have long variable names,
1941 which is an unrelated feature.
1943 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1945 * variable.h: Move var_set and variable parsing declarations to
1946 new header, src/language/lexer/variable-parser.h. Modified lots
1947 of files to include the new header.
1949 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1951 * value-labels.c (value_to_string): When there's no value label,
1952 format the variable according to its print format, instead of
1953 always effectively using A or F format.
1955 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1957 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1960 * format.c: Removed tortological assertion.
1962 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1964 Reform string library.
1966 * file-name.c (fn_interp_vars): Change interface to take a
1967 substring as input. Updated all users.
1969 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1971 * format.c (measure_is_valid): Really return false when m >=
1974 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1976 Implement random access to casefiles, for use in GUI.
1978 * casefile.c: (struct casereader) Add `random', `file_ofs',
1979 `buffer_ofs' members.
1980 (casefile_get_random_reader) New function.
1981 (read_open_file) Break part into new function
1982 seek_and_fill_buffer().
1983 (fill_buffer) Update buffer_ofs, file_ofs.
1984 (casereader_seek) New function.
1986 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1988 * settings.c: Added call to i18n{done, init}.
1990 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1992 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1995 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1997 * casefile.c: Convert many uses of `int' to `bool'.
1999 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
2001 * transformations.c (trns_chain_destroy): Destroy chain's trns
2002 member, to fix memory leak.
2004 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
2006 * storage-stream.c (storage_source_decapsulate): Destroy case
2007 source to fix memory leak.
2009 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
2011 * scratch-reader.c (scratch_reader_read_case): Copy into existing
2012 case passed as argument instead of initializing the argument as a
2013 case. Fixes memory leak that showed up in
2014 tests/command/aggregate.sh with scratch files.
2016 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
2018 * procedure.c (proc_done): Destroy default_dict, to fix memory
2021 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
2023 Simplify procedure_with_splits().
2025 * procedure.c (struct split_aux_data): Removed case_count member.
2026 (procedure_with_splits) Don't initialize case_count.
2027 (split_procedure_case_func) Check whether prev_case is null
2028 instead of case_count.
2029 (split_procedure_end_func) Ditto.
2031 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
2033 * case.c (case_move): Do nothing if dst and src are the same
2035 (case_try_create) Merge two similar cases.
2036 (case_copy) Unshare only if data must be actually copied.
2038 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2040 * data-in.c data-out.c dictionary.c sys-file-reader.c
2041 sys-file-writer.c variable.c variable.h: Reworked very long string
2042 support for better encapsulation.
2044 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
2046 * value-labels.c (val_labs_can_set_width): New function.
2047 (val_labs_set_width) Clear labels if increasing width to long
2049 (val_labs_destroy) Remove unneeded test for null.
2051 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
2053 * value-labels.h: Remove unneeded dependency on variable.h.
2055 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
2057 Get rid of `char *c' member in union value, for cleanliness.
2059 * value.h: (union value) Remove `c' member.
2061 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
2063 Make dictionary compacting functions a little more general.
2065 * sys-file-writer.c (sfm_open_writer): Use
2066 dict_compacting_would_change().
2067 (does_dict_need_translation) Removed.
2069 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
2071 Make dictionary compacting functions a little more general.
2073 * dictionary.c (dict_needs_compaction): Rename
2074 dict_compacting_would_shrink(). Update all callers.
2075 (dict_compacting_would_change) New function.
2077 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
2079 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
2080 inverted return value (!).
2082 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
2084 Continue reforming procedure execution.
2086 * procedure.c: Search and replace "vfm" by "proc". Notably:
2087 (static var vfm_source) Rename proc_source. Update all
2089 (static var vfm_sink) Rename proc_sink. Update all references.
2091 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
2093 Continue reforming procedure execution. In this phase, remove
2094 PROCESS IF, which was deprecated anyway and can be easily
2095 simulated with TEMPORARY followed by SELECT IF.
2097 * procedure.c: (open_active_file) Don't call
2098 add_process_if_trns().
2099 (discard_variables) Get rid of redundant call to
2100 proc_cancel_all_transformations().
2101 (add_process_if_trns) Removed.
2102 (process_if_trns_proc) Removed.
2103 (process_if_trns_free) Removed.
2105 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
2107 Continue reforming procedure execution. In this phase, add
2108 `const' to the case passed to procedure()'s callback.
2110 Updated all users of procedure() as well.
2112 * procedure.c: (struct write_case_data) Add "const" to ccase
2113 parameter for case_func member.
2114 (procedure) Add "const" to ccase parameter for proc_func
2116 (multipass_case_func) Make ccase parameter const.
2117 (internal_procedure) Add "const" to ccase parameter for case_func
2119 (split_procedure_case_func) Make ccase parameter const.
2120 (multipass_split_case_func) Make ccase parameter const.
2122 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
2124 Continue reforming procedure execution. In this phase, get rid of
2125 the output code for SPLIT FILE groups in procedure.c, which really
2126 shouldn't be doing any output. Move it into the individual
2127 procedures instead. This also adds some flexibility.
2129 Updated many users of procedure_with_splits() and
2130 multipass_procedure_with_splits() to call
2131 output_split_file_values() and to deal with increased use of
2134 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
2135 parameter to begin_func member.
2136 (procedure_with_splits) Add "const struct ccase *" parameter to
2137 begin_func parameter. Make ccase parameter const in proc_func
2139 (split_procedure_case_func) Don't dump split file group. Pass
2141 (dump_splits) Moved to language/dictionary/split-file.c as
2142 output_split_file_values().
2143 (struct multipass_split_aux_data) Add "const struct ccase *"
2144 parameter to split_func member.
2145 (multipass_procedure_with_splits) Add "const struct ccase *"
2146 parameter to split_func parameter.
2147 (multipass_split_case_func) Save new SPLIT FILE case before
2149 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
2151 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
2153 Continue reforming procedure execution. Change
2154 internal_procedure() so that it calls open_active_file() and
2155 close_active_file(), which isolates most of the actual procedure
2158 * procedure.c: (struct write_case_data) Rename `proc_func' member
2159 to `case_func' and update all references.
2160 (procedure) Rewrite as one-line wrapper around
2161 internal_procedure().
2162 (struct multipass_aux_data) New.
2163 (multipass_callback) Renamed multipass_case_func(). Use struct
2164 multipass_aux_data as auxiliary data.
2165 (multipass_end_func) New function.
2166 (multipass_procedure) Rewrite as wrapper for internal_procedure()
2167 that uses multipass_case_func, multipass_end_func.
2168 (internal_procedure) Add `end_func' argument. Move optimization
2169 of trivial case in here. Move call to open_active_file() and
2170 close_active_file() in here. Now assert that vfm_source is
2172 (procedure_with_splits_callback) Rename
2173 split_procedure_case_func().
2174 (split_procedure_end_func) New function.
2175 (multipass_split_callback) Rename multipass_split_case_func.
2176 (multipass_split_end_func) New function.
2177 (discard_variables) No need to test for nonnull vfm_source.
2179 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
2181 Continue reforming procedure execution. Get rid of unused member.
2183 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
2185 (write_case) Don't increment cases_analyzed.
2187 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
2189 Continue reforming procedure execution. In this phase, move
2190 procedure.c and procedure.h from src to src/data. Update
2191 makefiles and #includes accordingly.
2193 * procedure.c: Moved here from src/.
2195 * procedure.h: Moved here from src/.
2197 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
2199 Continue reforming procedure execution. In this phase, get rid of
2200 many global variables, consolidating procedure execution in
2201 procedure.c. Encapsulate transformations in new "struct
2202 trns_chain". Also, change implementation of N OF CASES, FILTER,
2203 and PROCESS IF from special cases to transformations.
2205 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2208 * dictionary.c: (global variable default_dict) Move to
2211 * variable.h: (TRNS_*) Move to transformations.h.
2212 (struct transformation) Move to transformations.c.
2214 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2216 * sys-file-reader.c: Fixed invalid read problems.
2218 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2220 * storage-stream.c: Add missing function comments.
2222 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2224 Continue reforming procedure execution. In this phase, add some
2225 new, needed functionality to storage-stream.
2227 * storage-stream.c: (storage_source_decapsulate) New function.
2229 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2231 * variable.c (width_to_bytes): Declarations must precede
2232 statements for C90 compliance.
2234 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2236 * data-out.c, data-in.c, variable.c, variable.h: New functions
2237 copy_mangle and copy_demangle for reading/writing cases; emulates the
2238 way SPSS deals with strings > 255 bytes.
2240 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2241 subtype 14 needed for strings longer than 255 bytes.
2243 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2244 instead of literal values. Also fixed some constness issues.
2246 * format.h: Constness
2248 * sfm-private.h: Renamed the case_size identifier, since I discovered
2249 that SPSS's respect for this variable is very nominal.
2251 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2253 Change case limit type from int to size_t.
2255 * dictionary.c: (struct dictionary) Change type of case_limit
2257 (dict_get_case_limit) Change return type.
2258 (dict_set_case_limit) Change parameter type.
2260 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2262 * variable.h: (struct variable) Rename `reinit' member as `leave'
2263 and invert sense. Fix up all references.
2265 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2267 Continue reforming procedure execution. In this phase, break
2268 procedure.c into multiple files.
2270 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2272 * case-sink.c: New file.
2274 * case-sink.h: New file.
2276 * case-source.c: New file.
2278 * case-source.h: New file.
2280 * storage-stream.c: New file.
2282 * storage-stream.h: New file.
2284 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2286 * variable.h: (struct variable) Remove `init' member and all
2287 references to it from other files. It was initialized in several
2288 places, but nothing really ever used it for anything worthwhile.
2289 Thanks to Jason Stover for pointing out how confusing this
2292 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2294 Continue reforming error message support. In this phase, get rid
2295 of message "titles" and put the message text in `struct error'.
2296 Now `struct error' encapsulates a message more properly.
2298 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2299 Format message ourselves.
2301 * data-in.c: (vdls_error) Ditto.
2303 * por-file-reader.c: (error) Ditto.
2305 * sys-file-reader.c: (corrupt_msg) Ditto.
2307 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2309 GNU standards require "file name" instead of "filename" in
2310 documentation. It's nice for our code to follow the convention
2313 * casefile.c: (struct casefile) Rename `filename' member to
2314 `file_name'. Updated all references.
2316 * file-name.c: [!unix] (struct file_identity) Rename
2317 normalized_filename member to normalized_file_name. Updated all
2320 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2322 We don't really support anything but Unix-like environments well,
2323 so we might as well de-obfuscate by writing directory and path
2324 separators explicitly.
2326 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2328 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2329 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2331 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2334 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2336 GNU standards require "file name" instead of "filename" in
2337 documentation. It's nice for our code to follow the convention
2340 * filename.c: Rename to file-name.c.
2342 * filename.h: Rename to file-name.h. Update all inclusions.
2343 Update header guards.
2345 * automake.mk: Update file names.
2347 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2349 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2351 (fn_basename) Removed (dead code).
2352 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2353 (fn_special_p) Renamed fn_is_special(), all references updated.
2354 (fn_exists_p) Renamed fn_exists(), all references updated.
2356 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2358 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2359 Also, now it only tilde-expands file names, not paths.
2360 (fn_search_path) Tilde-expand one directory at a time.
2362 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2364 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2365 get rid of non-Unixlike version of the code, which has probably
2367 (fn_prepend_dir) Removed (dead code).
2369 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2370 (macro PATH_SEPARATOR_STRING) New.
2371 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2373 Continue reforming error message support. In this phase, we get
2374 rid of VM() and the other msg() support for "verbosity", replacing
2375 it by a new function verbose_msg().
2377 * filename.c: (fn_search_path) Use verbose_msg() instead of
2380 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2382 * sfm-private.h: Get rid of #defines after #error, which makes no
2385 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2387 Get rid of our own int32 type in favor of the standard int32_t
2390 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2393 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2395 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2397 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2399 Remove ill-considered file routines that are no longer used.
2401 * filename.c: (fn_open_ext) Removed.
2402 (fn_close_ext) Removed.
2404 * filename.h: (struct file_ext) Removed.
2406 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2408 * variable.c (var_is_valid_name): Move declarations before code
2411 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2413 * filename.ch (fn_interp_vars): Fixed small buglet.
2415 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2417 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2418 so as to modify the string inline. Thus makeing it easier to
2419 destroy the results when no longer needed.
2421 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2423 * category.c (cat_stored_values_destroy): Fixed memory leak.
2425 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2427 Add some missing frees. Thanks to John Darrington for reporting
2430 * any-writer.c (any_writer_close): Free writer.
2432 * any-reader.c (any_reader_close): Free reader.
2434 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2436 * por-file-reader.c: (error) Mark as NO_RETURN.
2438 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2440 * settings.c: Changed default value of scompress to true.
2442 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2444 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2446 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2449 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2450 here from pref.h.orig.
2452 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2454 * sys-file-reader.c: Fixed bug reading compressed files.
2456 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2458 * Numerous renames. See src/ChangeLog for details.
2460 * Moved files from src directory