1 2007-11-05 Ben Pfaff <blp@gnu.org>
3 Patch #6258. Reviewed by John Darrington.
5 * file-handle-def.c (fh_lock): Add comment that TYPE should be
6 marked with N_() in the caller. Added these markings to each
7 caller too. Should make i18n easier.
8 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
10 2007-11-03 Ben Pfaff <blp@gnu.org>
12 Allow output files to overwrite input files (bug #21280).
14 * file-handle-def.c: Separate locking of files for input and for
15 output, to allow both to take place at once. Also, distinguish a
16 file handle from the identity of the underlying file, because the
17 identity of a file changes over time and the file handle can't
18 represent two different identities.
19 (struct file_handle): Remove `next', `open_cnt', `deleted',
20 `type', `open_mode', `aux', `identity' members. Change `id' from
21 char array to char *. Add `ref_cnt' member.
22 (file_handle_from_ll) New function.
23 (file_handles) Removed.
24 (named_handles) New variable.
25 (fh_init) Initialize named_handles.
26 (fh_done) Remove name from all named_handles.
27 (free_handle) Updated for modified struct file_handle.
28 (unname_handle) New function.
29 (fh_ref) New function.
30 (fh_from_file_name) Removed.
31 (fh_unref) New function.
32 (fh_unname) New function.
33 (fh_from_id) Rewritten.
34 (create_handle) Updated for modified struct file_handle.
35 (fh_inline_file) Increment the handle's ref count.
36 (fh_create_file) Updated for modified struct file_handle.
37 (fh_create_scratch) Ditto.
43 (fh_get_id) Updated for modified struct file_handle.
44 (fh_get_default_handle) Increment the handle's ref count.
45 (fh_set_default_handle) Handle ref counts.
46 (struct fh_lock) New structure.
47 (locks) New static var.
48 (fh_lock) New function.
49 (fh_unlock) New function.
50 (fh_lock_get_aux) New function.
51 (fh_lock_set_aux) New function.
52 (fh_is_locked) New function.
53 (make_key) New function.
54 (free_key) New function.
55 (compare_fh_locks) New function.
56 (hash_fh_lock) New function.
58 * file-handle-def.h (enum fh_access) New enum.
60 * file-name.c: Made file_identity the same in all supported
62 (struct file_identity): New `name' member.
63 (fn_get_identity): For a file that doesn't exist, get the
64 dev/inode of its directory plus its name. If even the directory
65 doesn't exist, just use its name. Merge the Windows
66 implementation into the Unix one.
67 (fn_compare_file_identities): Rewritten. Merge the Windows
68 implementation into the Unix one.
69 (fn_hash_identity): New function.
71 * make-file.c (struct replace_file): New structure.
72 (all_files): New static var.
73 (replace_file_start): New function.
74 (replace_file_commit): New function.
75 (replace_file_abort): New function.
76 (free_replace_file): New function.
77 (unlink_replace_files): New function.
79 * por-file-reader.c (struct pfm_reader): Add `lock' member.
80 (close_reader): Unlock file.
81 (pfm_open_reader): Lock file.
83 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
85 (pfm_open_writer): Lock file and prepare for its replacement.
86 (close_writer): Unlock file.
88 * scratch-handle.h (struct scratch_handle): Add unique_id so that
89 different generations of a scratch file can be distinguished.
91 * scratch-reader.c (scratch_reader_open): Verify that the file is
94 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
95 remove scratch_handle member.
96 (scratch_writer_open): Lock handle. Prepare to replace handle
97 data, instead of doing it immediately.
98 (scratch_writer_casewriter_destroy): Replace handle data and
101 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
102 (sfm_open_reader): Lock file.
103 (close_reader): Unlock file.
105 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
107 (sfm_open_writer): Lock file and prepare for its replacement.
108 (close_writer): Unlock file.
110 2007-11-02 Ben Pfaff <blp@gnu.org>
112 * data-out.c (output_number): Use isfinite (from C99) instead of
113 gsl_isfinite, since we now have universal support for it in
115 (output_infinite): Ditto.
116 [!HAVE_ROUND] (round): Remove definition, since we now have a
117 replacement in gnulib.
119 2007-10-27 John Darrington <john@darrington.wattle.id.au>
121 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
122 wasn't initialised correctly. Closes bugs #21330 and 21397
125 2007-10-08 Ben Pfaff <blp@gnu.org>
127 Fix bug #21280. Thanks to John Darrington for review.
129 * file-name.c (create_stream): New function.
131 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
132 file handle before creating the file, to ensure that we don't
133 truncate a file that we're reading. Make code easier to read by
136 * sys-file-write.c (sfm_open_writer): Ditto.
138 2007-10-01 Ben Pfaff <blp@gnu.org>
140 Fix bug #21192. Thanks to John Darrington for review.
142 * casereader.c (casereader_read): Decrement case_cnt before
143 calling the casereader's "read" member function, so that we
144 interact properly with lazy_casereader.
146 * datasheet.c: Add regression test for above bug fix.
147 (clone_datasheet): New function.
148 (lazy_callback): New function.
149 (check_datasheet_casereader): New function.
150 (check_datasheet): Check datasheet contents are reported correctly
151 through an ordinary casereader and a lazy casereader.
152 (clone_model): Use clone_datasheet.
154 2007-09-24 Ben Pfaff <blp@gnu.org>
156 Patch #6210. Reviewed by John Darrington.
157 * settings.c: Drop "long view", which is not needed any longer.
158 (static var long_view): Removed.
159 (force_long_view): Removed.
160 (get_viewwidth): Removed.
161 (init_viewport): Removed long_view reference.
163 2007-09-19 John Darrington <john@darrington.wattle.id.au>
165 * settings.c settings.h: Changed viewport's length and width to be
166 owned by the user interface which uses the data library. This allows
167 better abstraction, and makes dynamically adjustable dimensions easier.
169 2007-09-18 Ben Pfaff <blp@gnu.org>
171 * procedure.c (proc_extract_active_file_data): New function.
173 * lazy-casereader.h: New file.
175 * lazy-casereader.c: New file.
177 * casereader.c (casereader_dynamic_cast): New function.
179 2007-09-14 Ben Pfaff <blp@gnu.org>
181 * dictionary.c (dict_clone): Copy case indexes from cloned
182 dictionary. Fixes bug #21061. Reviewed by John Darrington.
184 2007-09-13 John Darrington <john@darrington.wattle.id.au>
186 * value.c value.h (create_value): New function.
188 2007-09-12 Ben Pfaff <blp@gnu.org>
190 Make it clear that translator casereader and casewriter translate
191 functions are supposed to destroy their input case.
193 * casereader-translator.c (struct casereader_translator): Change
194 input case parameter of translate member function from const
195 struct ccase * to struct ccase *.
196 (casereader_create_translator): Ditto, for translate parameter.
198 * casewriter-translator.c (struct casewriter_translator): Ditto.
199 (casewriter_create_translator): Ditto.
201 2007-08-27 John Darrington <john@darrington.wattle.id.au>
203 * sys-file-reader.c (read_display_parameters): Force display width
204 to 8, if the sys file says 0 (like SPSS does).
206 2007-08-12 Ben Pfaff <blp@gnu.org>
208 * dictionary.c (dict_dump): New function.
210 2007-08-12 Ben Pfaff <blp@gnu.org>
212 Drop dict_compactor in favor of using the new struct case_map.
214 * dictionary.c (struct copy_map): Removed.
215 (struct dict_compactor): Removed.
216 (dict_make_compactor): Removed.
217 (dict_compactor_compact): Removed.
218 (dict_compactor_destroy): Removed.
220 * procedure.c (struct data_set): Change `compactor' member to be a
222 (proc_open): Use case_map_to_compact_dict instead of
224 (proc_casereader_read): Use case_map_execute instead of
225 dict_compactor_compact.
226 (proc_commit): Use case_map_destroy instead of
227 dict_compactor_destroy.
229 * scratch-writer.c (struct scratch_writer): Change `compactor'
230 member to be a struct case_map *.
231 (scratch_writer_open): Use case_map_to_compact_dict instead of
233 (scratch_writer_casewriter_write): Use case_map_execute instead of
234 dict_compactor_compact.
236 2007-08-12 Ben Pfaff <blp@gnu.org>
238 * automake.mk: Add case-map.c, case-map.h.
240 * case-map.c: New file.
242 * case-map.h: New file.
244 2007-08-12 Ben Pfaff <blp@gnu.org>
246 * dictionary.c (dict_compact_values): Don't delete scratch
247 variables as well as compacting case indexes. Update all callers.
248 (dict_get_compacted_value_cnt): Rename dict_count_values and
249 change interface. Update all callers.
250 (dict_get_compacted_value_cnt): Remove.
251 (dict_compacting_would_shrink): Remove.
252 (dict_compacting_would_change): Remove.
253 (dict_make_compactor): Add new parameter. Update all callers.
255 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
256 instead of dict_count_values, changing an O(N) operation into
259 2007-08-12 Ben Pfaff <blp@gnu.org>
261 * casereader.c (casereader_read): Don't require cases read by a
262 casereader to be exactly the expected size: as long as they're big
265 2007-08-12 Ben Pfaff <blp@gnu.org>
267 Make casewriters keep track of the number of `union value's in
268 each case. This is useful for two reasons: casewriter_write can
269 then check that the case being written is large enough, and later
270 recipients of the casewriter can determine the size of the case.
272 * casewriter-translator.c (casewriter_create_translator): Add
275 * casewriter.c (struct casewriter): Add value_cnt member.
276 (casewriter_write): Check that the case passed in is big enough.
277 (casewriter_get_value_cnt): New function.
278 (casewriter_create): Add value_cnt parameter.
280 2007-08-09 Ben Pfaff <blp@gnu.org>
282 Fix bug reported by Jason Stover.
283 * settings.c: Move get_termcap_viewport and in particular the
284 #include for <curses.h> to the end of the file. curses.h
285 redefines bool on some systems (e.g. OpenBSD), which causes
286 disagreement between uses of bool before its inclusion and
288 Tested by Jason Stover.
290 2007-07-29 Ben Pfaff <blp@gnu.org>
292 Provisional fix for bug #18692 and bug #20161. Reviewed by John
295 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
296 argument (never "rb" or "wb") because SUSv3 says that only those
297 modes are defined, and glibc in fact rejects other modes.
299 Open portable files with fn_open so that they can be read from
300 pipes. Fix missing fh_close call to go along with fh_open.
301 Report an error if the file close reports an error.
302 * por-file-reader.c (close_reader): New function.
303 (por_file_casereader_destroy): Use close_reader.
304 (pfm_open_reader): Open file with fn_open.
306 2007-07-28 Ben Pfaff <blp@gnu.org>
308 Make PSPP able to read all the portable files I could find on the
309 web. Thanks to John Darrington for review. Bug #17620.
310 * por-file-reader.c (struct pfm_reader): New member `line_length'.
311 (error): Print file offset in hexadecimal.
312 (warning): New function.
313 (advance): Treat lines less than 80 bytes long as padded to 80
315 (pfm_open_reader): Call read_documents if we find an "E" record.
316 (convert_format): Convert invalid formats to the default format
317 instead of aborting reading the file.
318 (read_variables): Rename duplicate variable names instead of
319 aborting reading the file.
320 (read_value_label): Allow string variables of different widths to
321 be assigned value labels in the same record. Replace duplicate
322 value labels instead of aborting.
323 (read_documents): New function.
325 * por-file-writer.c (pfm_open_writer): Call write_documents if the
326 dictionary has documents.
327 (write_documents): New function.
329 2007-07-25 Ben Pfaff <blp@gnu.org>
331 Fix bugs related to bug #17213.
333 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
334 since the former is what config.h has. Include the needed ncurses
336 (static var echo) Rename to `do_echo' because the original name is
337 the same as an ncurses identifier.
338 (get_termcap_viewport) Use error instead of msg.
340 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
341 (fn_close): Don't close stdin, stdout, stderr.
343 2007-07-26 John Darrington <john@darrington.wattle.id.au>
345 * procedure.c procedure.h: Added callbacks which get invoked whenever
346 a dataset's transformation chain changes.
348 2007-07-24 Ben Pfaff <blp@gnu.org>
351 * sys-file-writer.c (write_variable_display_parameters): Use new
352 var_default_display_width function to choose display width of
353 segments after the first one in a given variable.
354 * variable.c (var_create): Use var_default_display_width to pick
355 new variable's display width.
356 (var_default_display_width): New function.
357 Reviewed by John Darrington.
359 2007-07-24 Ben Pfaff <blp@gnu.org>
362 * por-file-writer.c (write_variables): Write weight variable.
363 Reviewed by John Darrington.
365 2007-07-23 Ben Pfaff <blp@gnu.org>
367 Improvements to system file reader and writer.
369 First, move all detailed knowledge of very long strings into
370 sys-file-private.[ch], so that this nasty stuff can be isolated.
372 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
373 (EFFECTIVE_VLS_CHUNK): New macro.
374 (min_int): New function.
375 (max_int): New function.
376 (sfm_width_to_bytes): Rewrite.
377 (sfm_width_to_octs): New function.
378 (sfm_segment_alloc_width): New function.
379 (sfm_segment_alloc_bytes): New function.
380 (sfm_segment_used_bytes): New function.
381 (sfm_segment_offset): New function.
382 (sfm_segment_effective_offset): New function.
383 (sfm_dictionary_to_sfm_vars): New function.
385 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
386 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
387 (struct sfm_var): New structure.
389 Next, improvements to the system file reader.
391 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
392 type casenumber. Added `version_major', `version_minor',
395 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
396 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
397 Change `case_cnt' to type casenumber. Removed `has_vls'.
398 (struct sfm_var): Removed.
399 (sfm_open_reader): Don't warn on wrong case size if the file was
400 written by SPSS 13, which tends to get it wrong. Use
401 sfm_dictionary_to_sfm_vars.
402 (read_header): Always output system file info.
403 (read_variable_record): Simplify code for reading missing values.
404 (read_machine_int32_info): Save version numbers from system file
405 into info struct passed as new argument.
406 (read_long_string_map): Restructured to use new sys-file-private
408 (read_value_labels): Use size_overflow_p.
409 (sys_file_casereader_read): Get rid of distinction between fast
410 and slow paths. Use information provided by sys-file-primate's
411 struct sfm_var to simplify code.
412 (skip_whole_strings): New function.
413 (read_int32): Renamed read_int. Changed return value to int.
415 (read_flt64): Renamed read_float. Changed return value to
416 double. Updated all callers.
417 (int32_to_native): Removed. Changed callers to use
419 (flt64_to_double): Removed. Changed callers to use float_convert.
421 Finally, get rid of int32, flt64 terminology and types in system
422 file writer. The former wasn't very useful since a POSIX "int"
423 can hold the whole range of int32 and we generally didn't have a
424 need for it to be exactly-32-bits, just at-least-32-bits. The
425 latter was inconvenient because we had to assume that it could be
426 different from double and thereby convert special values SYSMIS,
427 HIGHEST, LOWEST to and from it in multiple places. Instead, now
428 we just use "int" and "double" in most places, and do conversions,
429 if necessary, very close to where we do I/O. This change meant
430 that the writer code couldn't represent records in the file as C
431 structs any longer, but that's no great loss. The code actually
432 seems to be more readable without them.
434 Simplify the compression buffering code: only buffer as much as
435 necessary, which is no more than eight 8-byte units at any given
438 * sys-file-writer.c (typedef flt64): Removed.
439 (macro second_lowest_flt64): Removed.
440 (struct sysfile_header): Removed.
441 (struct sysfile_variable): Removed.
442 (struct sfm_writer): Removed `needs_translation', `has_vls',
443 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
444 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
445 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
446 compression buffering by `opcodes', `opcode_cnt', `data',
447 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
448 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
449 data writer functions instead of structures.
450 (calc_oct_idx): New function.
451 (write_header): Use simple data writer functions instead of
453 (write_format_spec): Renamed write_format. New argument.
454 (write_variable_continuation_records): New function.
455 (write_variable): Use simple data writer functions instead of
456 structures. Use write_variable_continuation_records. Write
457 entire very long string instead of requiring caller to understand
459 (write_value_labels): Use simple data writer functions instead of
461 (write_documents): Ditto.
462 (write_variable_display_parameters): Use sys-file-private
463 functions to simplify. Use simple data writer functions instead
465 (write_vls_length_table): Use simple data writer functions instead
467 (write_longvar_table): Ditto.
468 (write_rec_7_34): Break into new functions
469 write_integer_info_record, write_float_info_record. Use simple
470 data writer functions instead of structures.
471 (buf_write): Removed.
472 (append_string_max): Removed.
473 (ensure_buf_space): Removed.
474 (sys_file_casewriter_write): Get rid of the distinction between
475 fast and slow paths, which didn't seem to be too useful. Use new
476 functions write_case_uncompressed, write_case_compressed.
477 (put_instruction): Removed.
478 (put_element): Removed.
479 (write_compressed_data): Removed.
480 (close_writer): Use flush_compressed. Only write case count to
481 system file if it will fit in the field.
482 (write_case_compressed): New function.
483 (write_case_uncompressed): New function.
484 (flush_compressed): New function.
485 (put_cmp_opcode): New function.
486 (put_cmp_number): New function.
487 (write_int): New function.
488 (convert_double_to_output_format): New function.
489 (write_float): New function.
490 (write_value): New function.
491 (write_string): New function.
492 (write_bytes): New function.
493 (write_zeros): New function.
494 (write_spaces): New function.
496 Reviewed by John Darrington.
498 2007-07-22 Ben Pfaff <blp@gnu.org>
500 Don't try to write very long strings to portable files. The
501 format does not support it.
503 * por-file-writer.c (MAX_POR_WIDTH): New macro.
504 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
505 (write_format): Add arg to take width to resize format to.
506 (write_value): Limit width of value written to MAX_POR_WIDTH.
507 (write_variables): Limit width of variable and its output formats
509 Reviewed by John Darrington.
511 2007-07-22 Ben Pfaff <blp@gnu.org>
513 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
514 local variable instead of literal 5.
515 Reviewed by John Darrington.
517 2007-07-22 Ben Pfaff <blp@gnu.org>
519 Fix problems with uniqueness of short names in system files with
520 very long string variables. Now a variable may have multiple
523 * automake.mk (src_data_libdata_a_SOURCES): Add new files
524 short-names.c, short-names.h.
526 * dictionary.c (dict_clone): Clone all the short names.
527 (compare_strings): Move into short-names.c.
528 (hash_strings): Ditto.
529 (set_var_short_name_suffix): Ditto.
530 (dict_assign_short_names): Ditto, rename short_names_assign,
531 change to assign all short names.
533 * por-file-writer.c (write_variables): Use short_names_assign
534 instead of dict_assign_short_names.
536 * short-names.c: New file.
538 * short-names.h: New file.
540 * sys-file-private.c (sfm_width_to_segments): New function.
542 * sys-file-reader.c (read_long_var_name_map): Save and restore all
543 the short names, not just the first one.
545 * sys-file-writer.c (cont_var_name): Removed.
546 (sfm_open_writer): Use short_names_assign instead of
547 dict_assign_short_names. Use unique short names assigned by
548 short_names_assign instead of those generated by cont_var_name.
550 * variable.c (struct variable): Remove `short_name' member,
551 replace by `short_names' and `short_name_cnt'.
552 (var_create) Initialize new members.
553 (var_get_short_name_cnt): New function.
554 (var_get_short_name): Now takes an index argument. Changed most
556 (var_set_short_name): Ditto.
557 (var_clear_short_name): Renamed var_clear_short_names, changed to
558 clear all short names.
560 Reviewed by John Darrington.
562 2007-07-22 Ben Pfaff <blp@gnu.org>
564 * variable.c (var_set_width): Use new var_set_width function.
566 * missing-values.c (mv_n_values): Drop assertion, which was not
569 * format.c (fmt_default_for_width): New function.
570 (fmt_resize): New function.
572 Reviewed by John Darrington.
574 2007-07-18 John Darrington <john@darrington.wattle.id.au>
576 * datasheet.c (datasheet_delete_columns): Added assertion to check
577 we're not deleting outside the range of the sheet.
580 * dictionary.c dictionary.h variable.c: Added the ability for string
581 variables to be resized.
583 * vardict.h: Added some prototypes (moved from dictionary.h) as
584 these should only be called by variable.c
587 2007-07-14 John Darrington <john@darrington.wattle.id.au>
589 * sfm-reader.c: Respect case_cnt field in file header.
591 2007-07-01 John Darrington <john@darrington.wattle.id.au>
593 * transformation.c transformation.h (trns_chain_execute): Changed the
594 signature (Patch #6057)
596 2007-06-10 Ben Pfaff <blp@gnu.org>
598 * casereader-filter.c (casereader_filter_destroy): Make sure to
599 write all the remaining excluded cases to the casewriter, if any.
601 * caseinit.c (init_list_destroy): Rewrite.
602 (init_list_clear): Ditto.
604 * casegrouper.c (casegrouper_get_next_group): Always set *reader
605 to null when returning false.
607 2007-06-06 Ben Pfaff <blp@gnu.org>
609 Actually implement the new procedure code and adapt all of its
610 clients to match. Also adapt all of the other case sources and
611 sinks in the tree and their clients to use the
612 casereader/casewriter infrastructure.
614 * automake.mk: Add and remove files.
616 * any-reader.c: Change into a casereader.
617 * por-file-reader.c: Ditto.
618 * scratch-reader.c: Ditto.
619 * sys-file-reader.c: Ditto.
621 * any-writer.c: Change into a casewriter.
622 * por-file-writer.c: Ditto.
623 * scratch-writer.c: Ditto.
624 * sys-file-writer.c: Ditto.
626 * procedure.c: Change to use casereader, casewriter, caseinit, and
627 other new infrastructure.
629 * scratch-handle.c: Adapt to new infrastructure.
631 * case-sink.c: Removed, now dead code.
632 * case-sink.h: Ditto.
633 * case-source.c: Ditto.
634 * case-source.h: Ditto.
635 * casefile-factory.c: Ditto.
636 * casefile-private.h: Ditto.
639 * casefilter.c: Ditto.
640 * casefilter.h: Ditto.
643 * fastfile-factory.c: Ditto.
644 * fastfile-factory.h: Ditto.
645 * storage-stream.c: Ditto.
646 * storage-stream.h: Ditto.
648 2007-06-06 Ben Pfaff <blp@gnu.org>
652 * automake.mk: Add new files.
654 * datasheet.c: New file.
656 * datasheet.h: New file.
658 2007-06-06 Ben Pfaff <blp@gnu.org>
660 Until now, the procedure code has provided a case to the
661 case_source, which has filled in the data values that come from
662 the active file. "Left" data values that don't come from the
663 active file naturally stay the same from case to case, because the
664 procedure code keeps using that same case.
666 One of the compromises that comes with the new procedure code is
667 that the active file allocates and provides its own case, which
668 the procedure code then has to resize to provide room for any
669 other variables that should go in the case and then fill in the
670 values of "left" variables. Then, when we're done with that case,
671 we have to save the values of "left" variables to copy into the
672 next case read from the active file.
674 The caseinit code helps with this.
676 * automake.mk: Add new files.
678 * caseinit.c: New file.
680 * caseinit.h: New file.
682 2007-06-06 Ben Pfaff <blp@gnu.org>
684 * value.h (value_cnt_from_width): New function.
686 * variable.c (var_get_value_cnt): Use new function.
688 2007-06-06 Ben Pfaff <blp@gnu.org>
690 Add casegrouper, to allow cases read from a given casereader to be
691 broken into groups, each of which has its own casereader.
692 Generally cases are grouped based on having equal values for some
695 * automake.mk: Add new files.
697 * casegrouper.c: New file.
699 * casegrouper.h: New file.
701 2007-06-06 Ben Pfaff <blp@gnu.org>
703 Add interface to lexicographical ordering of cases.
705 * automake.mk: Add new files.
707 * case-ordering.c: New file.
709 * case-ordering.h: New file.
711 2007-06-06 Ben Pfaff <blp@gnu.org>
713 Add casereaders and casewriters, the basis of the new data processing
714 implementation. A casereader is a uniform interface to reading cases
715 from a data source; a casewriter is a uniform interface to writing
716 cases to a data sink.
718 * automake.mk: Add new files.
720 * casereader-filter.c: New file.
722 * casereader-provider.h: New file.
724 * casereader-translator.c: New file.
726 * casereader.c: New file.
728 * casereader.h: New file.
730 * casewriter-provider.h: New file.
732 * casewriter-translator.c: New file.
734 * casewriter.c: New file.
736 * casewriter.h: New file.
738 2007-06-06 Ben Pfaff <blp@gnu.org>
740 "casewindow" data structure that extends the deque (from libpspp)
741 of cases with the ability to dump cases to disk if we get too many
744 * automake.mk: Add new files.
746 * casewindow.c: New file.
748 * casewindow.h: New file.
750 2007-06-06 Ben Pfaff <blp@gnu.org>
752 sparse_cases data structure that augments a sparse_array of cases
753 with the ability to dump cases to disk if we get too many cases in
756 * automake.mk: Add new files.
758 * sparse-cases.c: New file.
760 * sparse-cases.h: New file.
762 2007-06-06 Ben Pfaff <blp@gnu.org>
764 Adds a low-level on-disk case array data structure.
766 * automake.mk: Add new files.
768 * case-tmpfile.c: New file.
770 * case-tmpfile.h: New file.
772 2007-06-06 Ben Pfaff <blp@gnu.org>
774 In a couple of places we calculate the maximum number of cases to
775 keep in memory based on the user-defined workspace. Enable
776 centralizing the calculation through a new function.
778 * settings.c (get_workspace_cases): New function.
780 2007-06-06 Ben Pfaff <blp@gnu.org>
782 The casenumber type is defined in transformations.h, but case.h is
783 a more sensible place. Move it.
785 * case.h (CASENUMBER_MAX): New macro.
786 (typedef casenumber): Move here, from transformations.h.
788 2007-06-03 Ben Pfaff <blp@gnu.org>
790 Slightly generalize case_to_values and case_from_values functions.
792 * case.c (case_to_values): Rename case_copy_out, change interface.
793 (case_from_values): Rename case_copy_in, change interface.
795 * fastfile.c (fastfilereader_get_next_case): Update caller.
796 (write_case_to_disk): Ditto.
798 2007-06-02 Ben Pfaff <blp@gnu.org>
800 Clean up after a forgotten part of patch #5829.
802 * casedeque.h: Remove unused file.
804 * automake.mk: Remove casedeque.h from sources.
806 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
808 * category.c: Removed redundant #include
810 2007-05-06 Ben Pfaff <blp@gnu.org>
812 Abstract the documents within a dictionary a little better.
813 Thanks to John Darrington for suggestion, initial version, and
816 * dictionary.c (struct dictionary): Change `documents' member from
817 char * to struct string.
818 (dict_clear): Destroy struct string.
819 (dict_get_documents): Convert struct string to char *.
820 (dict_set_documents): Set struct string. Pad to 80-character
822 (dict_clear_documents): New function.
823 (dict_add_document_line): New function.
824 (dict_get_document_line_cnt): New function.
825 (dict_get_document_line): New function.
827 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
829 * sys-file-reader.c (read_documents): Use new document functions.
831 2007-04-19 John Darrington <john@darrington.wattle.id.au>
833 * sys-file-reader.c: When reading a system file which has no
834 long name table, automatically create one where the long names
835 are the lower case versions of the short names.
837 2007-04-22 Ben Pfaff <blp@gnu.org>
839 * dictionary.c (dict_set_split_vars): dict_destroy expects that
840 dict_clear will free most data related to the dictionary.
841 dict_clear does a decent job, except that dict_set_split_vars on
842 some systems won't actually free the dict's "split" member.
843 Instead, it'll allocate a 1-byte region. Fix this.
845 * value.c (value_copy): New function.
846 (value_set_missing): Ditto.
848 2007-04-22 John Darrington <john@darrington.wattle.id.au>
850 * Deleted existing category.h and moved cat-routines.h into
851 category.h Encapsulated struct cat_vals better.
853 2007-04-19 John Darrington <john@darrington.wattle.id.au>
855 * sys-file-reader.c: When reading a system file which has no
856 long name table, automatically create one where the long names
857 are the lower case versions of the short names.
859 2007-04-16 John Darrington <john@darrington.wattle.id.au>
861 * sys-file-reader.c: Some versions of Other Software seem to
862 produce system files with string variables' measure set to
863 zero. We'll assume these are supposed to be nominal variables.
865 2007-03-30 Ben Pfaff <blp@gnu.org>
867 * procedure.c: Adapt to new deque data structure.
869 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
870 Ben Pfaff <blp@gnu.org>
872 * file-name.c: Mingw compatibility fixes.
873 (fn_search_path): Use ISSLASH instead of comparing against '/'
875 (fn_dir_name): Use dir_name from gnulib.
876 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
877 (fn_get_identity): Use GetFullPathName instead of canonicalize
878 from gnulib, because the latter does not fully support
879 Windows-style path names. Use this implementation based on the
880 detected presence of Windows instead of the absence of Unix, since
881 the new implementation is Windows-specific.
882 (fn_compare_file_identities): In Windows implementation, compare
883 names case-insensitively.
885 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
887 * make-file.c: Don't include mkstemp.h, because gnulib now causes
888 <stdlib.h> to have the same effect.
890 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
892 * por-file-reader.c: Add missing _() around messages.
894 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
896 * make-file.c: Include "mkstemp.h", without which linking on
899 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
901 Reduce platform dependence.
903 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
904 Everywhere we using this, we really should have just depended on
905 the shell to expand tildes.
906 (fn_search_path): Simplify, given that we don't do tilde expansion
908 (fn_normalize): Removed. Caller changed to use the canonicalize
910 (fn_get_cwd): Removed. Only user was fn_normalize.
911 (fn_is_absolute): Really only test for absolute names.
912 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
914 (fn_readlink): Removed, as it was only used fn_normalize.
915 (fn_exists): Assume the stat function is available; gnulib does.
916 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
918 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
920 * dictionary.c (dict_create_vector_assert): New function.
922 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
924 * file-name.c (fn_normalize): Correct name of function
925 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
926 for pointing this out.
928 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
930 We are using a single member in struct file_handle, the "name"
931 field, for more than one purpose. When it begins with '"', it's a
932 file name; otherwise, it's a token that can be used to identify
933 it. When that assertion fires, it's because we searched for the
934 name case-sensitively as a file name (so that there was no match),
935 and then we try to insert it case-insensitively as a token, which
936 fails because duplicates aren't allowed.
938 Solution: break the two purposes into two separate fields. This
939 fixes the problem and likely makes the code easier to read too.
941 Fixes bug #18922. Thanks to John Darrington for bug report and
944 * file-handle-def.c (struct file_handle): New `id' member.
945 (fh_from_name): Rename fh_from_id. Update all callers.
946 (create_handle): New `id' parameter. Update all callers.
947 (fh_create_file): Ditto.
948 (fh_get_id): New function.
950 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
952 * case.c (case_is_null): Change return type to bool.
954 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
958 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
959 reference counting to share data. This makes it easy for
960 valgrind, etc. to find accesses to cases that have been destroyed
961 but have been kept around by another user's ref-count. This often
962 happens when the data set is small enough to find in memory; if a
963 bigger data set that would overflow to disk were used, then data
964 corruption would occur.
966 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
970 * case.c (case_unshare): Make it check internally whether the
971 ref_cnt is greater than 1, so that the callers don't have to.
972 Update callers not to check.
974 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
976 Before, I was thinking that I might want to get rid of reference
977 counting at some point. Now, I'm pretty sure that it's here to
978 stay. Thus, because we have to store the value_cnt anyway for
979 reference-counted cases, we might as well expose it to users.
981 * case.c (case_get_value_cnt): New function.
982 (case_resize): Drop OLD_CNT argument. Update all callers. Only
983 resize case if its size actually changed.
985 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
986 instead of peeking inside struct case directly.
987 (casefile_append): Ditto.
989 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
991 Get rid of the inlines for the case functions, which made the
992 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
993 the inlines didn't speed up "make check" at all, which is not a
994 perfect benchmark but seems indicative.)
996 * case.c: Remove #ifdef DEBUGGING...#endif around many function
997 definitions. Remove some assertions on nonnull pointers that were
998 redundant with a pointer dereference soon after in the function.
1000 (struct case_data): Move definition here from case.h.
1004 (case_data_wr): Ditto.
1006 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1008 * automake.mk: Add casedeque.h to sources.
1010 * casedeque.h: New file.
1012 * procedure.c: (struct dataset) Change lag_count, lag_head,
1013 lag_queue member into single struct casedeque member. Update all
1014 users to use the casedeque instead.
1017 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1019 * procedure.c: Simplify lagged cases interface. Updated all
1020 clients--well, the only client--to use the simplified interface.
1021 (dataset_n_lag) Removed.
1022 (dataset_set_n_lag) Removed.
1023 (dataset_need_lag) New function.
1025 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1027 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1028 whenever appropriate, but only when the dataset/dictionary is in a
1031 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1033 * dictionary.c dictionary.h : Added callbacks for change of filter and
1034 split variables. Refactored some code to ensure that callbacks get
1035 invoked when appropriate.
1037 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1038 replace_dict callback occurs when permanent_dict replaces the current
1041 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1043 * dictionary.c dictionary.h : Added callback for when the weight
1044 variable of a dictionary changes.
1046 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1048 * dictionary.c dictionary.h : Added replace_source and replace_dict
1049 callbacks, and functions to deal with them.
1051 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1053 Simplify missing value handling.
1055 * missing-values.h (enum mv_class): New type.
1056 (enum mv_type): Moved definition into missing-values.c and renamed
1057 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1058 enums. Updated all uses.
1059 (struct missing_values): Changed type of `type' from `enum
1060 mv_type' to `int' because the definition is no longer exposed.
1062 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1063 parameter. Update all callers.
1064 (mv_is_num_missing): Ditto.
1065 (mv_is_str_missing): Ditto.
1066 (mv_is_value_user_missing): Removed. Changed callers to use
1067 mv_is_value_missing.
1068 (mv_is_num_user_missing): Removed. Changed callers to use
1070 (mv_is_str_user_missing): Removed. Changed callers to use
1072 (mv_is_value_system_missing): Removed. Changed callers to use
1073 mv_is_value_missing.
1074 (mv_set_type): Removed. Changed callers to use mv_clear.
1075 (mv_clear): New function.
1077 * variable.c (var_is_value_missing): Add new enum mv_class
1078 parameter. Update all callers.
1079 (var_is_num_missing): Ditto.
1080 (var_is_str_missing): Ditto.
1081 (var_is_value_user_missing): Removed. Changed callers to use
1082 var_is_value_missing.
1083 (var_is_num_user_missing): Removed. Changed callers to use
1085 (var_is_str_user_missing): Removed. Changed callers to use
1087 (var_is_value_system_missing): Removed. Changed callers to use
1088 var_is_value_missing.
1090 * casefilter.c (struct casefilter): Use enum mv_class in place of
1092 (casefilter_variable_missing): Adapt to new member.
1093 (casefilter_create): Change signature to take enum mv_class,
1096 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1098 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1100 * case-sink.c case-sink.h procedure.c procedure.h
1101 storage-stream.c: Now uses the factory.
1103 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1105 Make it possible to pull cases from the active file with a
1106 function call, instead of requiring indirection through a callback
1109 * case-source.h (struct case_source_class): Change ->read function
1110 to return a single case, instead of calling a callback function
1111 for each case. Change ->destroy function to return an error
1114 * case-source.c (free_case_source): Pass along the value returned
1115 by the case_source ->destroy function.
1117 * procedure.c (struct write_case_data): Removed.
1118 (struct dataset): Added some members to track procedure state.
1119 (procedure): Optimize the trivial case at this level.
1120 (internal_procedure): Re-implement in terms of proc_open,
1121 proc_read, proc_close.
1122 (proc_open) New function.
1123 (proc_read) New function.
1124 (proc_close) New function.
1125 (write_case) Moved into proc_read.
1126 (close_active_file) Moved closing of data source into proc_close.
1128 * storage-source.c: Rewrote to conform with modified
1129 case_source_class interface.
1131 * transformations.c (trns_chain_execute): Added argument to allow
1132 starting execution from an arbitrary transformation. Updated
1135 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1137 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1139 * sys-file-reader.c (read_display_parameters): Don't assume that
1140 MEASURE_* and ALIGN_* have the same values found in system files.
1142 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1144 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1145 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1146 the value of n_MEASURES, which was off by 1 (at least from my
1149 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1151 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1152 callbacks which are invoked when the dictionary or its
1153 variables are changed.
1155 * missing-values.c missing-values.h value-labels.c: Tidied up
1156 consistency checks, and made some of them return false
1157 instead of assert-failing.
1159 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1161 * calendar.c (calendar_days_in_month): New function.
1163 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1165 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1166 the value label's width.
1168 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1170 * sfm-private.h: Move contents into sys-file-writer.c, which is
1171 the only remaining user. Removed Borland C++-specific directives.
1173 * sys-file-reader.c: Clean up and rewrite entire file. The
1174 rewritten version is simpler and better abstracted, and should be
1175 easier to maintain and extend. It avoids using structures to read
1176 file data, which is prone to padding variations among compilers.
1177 It should also handle non-IEEE 754 system files, although I
1178 haven't been able to find any. It has been tested against many
1179 .sav files obtained from the Web and found to produce the same
1180 results as the earlier version of the code, or in some cases
1181 improved results. It is more tolerant of format variations found
1184 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1185 member, putting an enum integer_format in its place. New member
1186 `float_format'. Changed `compressed' member to type bool.
1188 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1190 * dictionary.c (dict_delete_consecutive_vars): New function.
1192 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1194 * file-name.c (fn_search_path): Remove prefix arg that was unused
1195 by any caller. Updated all callers.
1197 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1199 * format.c (fmt_dollar_template): Use user's decimal point
1200 character. Add assertion.
1202 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1204 * format.c (fmt_dollar_template): New function, based on
1205 dollar_format_template from var-type-dialog.c.
1207 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1209 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1212 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1214 Finish converting struct variable to an opaque type. In this
1215 phase, we add remaining setter and getter functions, convert the
1216 remaining PSPP code to use them, and do a bunch of cleanup. The
1217 resulting changes are pervasive but mostly trivial, and only the
1218 notable changes are logged.
1220 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1223 * case.c (case_data): Renamed case_data_idx.
1224 (case_num): Renamed case_num_idx.
1225 (case_str): Renamed case_str_idx.
1226 (case_data_rw): Renamed case_data_rw_idx.
1228 * case.h (case_data): New function with old name and an interface
1229 that takes a variable instead of an index, which is easier to
1230 use. Updated all callers to use the new interface, or to use the
1231 new *_idx function (see above).
1234 (case_data_rw): Ditto.
1236 * category.c (cat_stored_values_destroy): Changed interface to
1237 take a struct cat_vals * instead of a struct variable *.
1239 * dictionary.c (dict_clone): Use new vector_clone function.
1240 (dict_clear) Use new var_destroy function.
1241 (add_var) New function.
1242 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1243 (dict_create_var_assert) Rewrite in terms of add_var.
1244 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1245 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1246 (dict_lookup_var) Use new var_create, var_destroy functions.
1247 (dict_contains_var) Rewrite in terms of new vardict functionality.
1248 (set_var_dict_index) New function.
1249 (set_var_case_index) New function.
1250 (reindex_vars) New function.
1251 (dict_delete_var) Rewrite in terms of new vardict functionality.
1252 (dict_reorder_var) Ditto.
1253 (dict_reorder_vars) Ditto.
1254 (rename_var) New function.
1255 (dict_rename_var) Use rename_var.
1256 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1257 (dict_get_compacted_idx_to_fv) Rename
1258 dict_get_compacted_dict_index_to_case_index, update callers.
1259 (dict_create_vector) Use new vector_create function.
1260 (dict_clear_vectors) Use new vector_destroy function.
1261 (set_var_short_name_suffix) Move here from variable.c, renamed
1262 from var_set_short_name_suffix, make static, update caller.
1264 * sys-file-private.c: New file.
1265 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1266 width_to_bytes, update callers.
1268 * sys-file-private.h: New file. Later it will supplant
1269 sfm-private.h; for now it supplements it.
1270 (macro MIN_VERY_LONG_STRING) New macro.
1271 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1273 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1274 MAX_LONG_STRING was used before.
1276 * sys-file-writer.c: Ditto.
1278 * value-labels.c: Change the paradigm here to be that a null
1279 pointer is OK for a struct val_labs * in most cases; it just
1280 represents an empty set of value labels.
1281 (val_labs_copy) A copy of a null set is a null set.
1282 (val_labs_count) A null set has 0 labels.
1283 (val_labs_replace) Change return type to void. Rewrite for
1285 (val_labs_find) A null set does not contain the value.
1286 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1287 transposed argument order, updated all callers.
1289 * value.c: New file.
1290 (value_dup) Moved here from variable.c.
1291 (compare_values) Ditto.
1294 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1295 (macro MAX_LONG_STRING) Removed, because it was only interesting
1296 for system files, not for general code.
1297 (macro MAX_VERY_LONG_STRING) Ditto.
1298 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1299 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1301 * vardict.h: New file, for an interface between variables and
1304 * variable.c: A lot of functions were moved around, for better
1306 (struct variable) Move definition here, from variable.h.
1307 (var_type_adj) Removed--makes i18n hard.
1308 (var_type_noun) Ditto.
1309 (var_create) New function.
1310 (var_clone) New function.
1311 (var_destroy) New function.
1312 (var_set_name) Assert that variable is not in a dictionary.
1313 (compare_var_names) Rename compare_vars_by_name and fix a couple
1314 of callers who thought the args were strings.
1315 (hash_var_name) Rename hash_var_by_name.
1316 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1317 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1318 (var_is_very_long_string) Removed, because it was only interesting
1319 to system file code.
1320 (var_set_missing_values) Allow the argument to be the wrong width,
1321 as long as we can resize it. Simplify callers who were doing the
1322 resizing themselves.
1323 (var_get_value_labels) New function.
1324 (var_has_value_labels) New function.
1325 (var_set_value_labels) New function.
1326 (alloc_value_labels) New function.
1327 (var_add_value_label) New function.
1328 (var_replace_value_label) New function.
1329 (var_clear_value_labels) New function.
1330 (var_lookup_value_label) New function.
1331 (var_get_value_name) Moved here from variable.c, renamed from
1332 var_get_value_name, transposed argument order, updated all
1334 (var_to_string) Moved here, from variable-label.c.
1335 (var_set_leave) New function.
1336 (var_get_leave) New function.
1337 (var_must_leave) New function.
1338 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1339 set_var_short_name_suffix.
1340 (var_get_dict_index) New function.
1341 (var_get_case_index) New function.
1342 (var_get_obs_vals) New function.
1343 (var_set_obs_vals) New function.
1344 (var_has_obs_vals) New function.
1345 (var_get_vardict) New function.
1346 (var_set_vardict) New function.
1347 (var_has_vardict) New function.
1348 (var_clear_vardict) New function.
1349 (value_dup) Moved to value.c.
1350 (compare_values) Ditto.
1353 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1354 (enum ALPHA) Rename VAR_STRING, update all users.
1356 * vector.c: New file.
1357 (struct vector) Moved here, from variable.h.
1358 (check_widths) New function.
1359 (vector_create) New function.
1360 (vector_clone) New function.
1361 (vector_destroy) New function.
1362 (vector_get_name) New function.
1363 (vector_get_var) New function.
1364 (vector_get_var_cnt) New function.
1365 (compare_vector_ptrs_by_name) New function.
1367 * vector.h: New file.
1369 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1371 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1372 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1375 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1377 * value-labels.c (destroy_atoms): New function.
1378 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1380 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1382 Thanks to Jason Stover for pointing out this problem.
1384 * data-out.c (output_number): Use gsl_finite from GSL, which is
1385 portable, instead of isfinite, which is not.
1388 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1390 * variable.c variable.h (value_dup): New function.
1392 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1394 Start converting struct variable to an opaque type. In this
1395 phase, we add a bunch of setter and getter functions and convert
1396 most of the PSPP code to use them. The resulting changes are
1397 pervasive but mostly trivial, and only the notable changes are
1400 * format.c (fmt_equal): New function.
1402 * variable.c (var_type_is_valid): New function.
1403 (measure_is_valid) Moved here, from format.c.
1404 (alignment_is_valid) Moved here, from format.c.
1405 (var_get_name) New function.
1406 (var_set_name) New function.
1407 (width_to_type) New function.
1408 (var_get_type) New function.
1409 (var_get_width) New function.
1410 (var_set_width) New function.
1411 (var_is_numeric) New function.
1412 (var_is_alpha) New function.
1413 (var_is_short_string) New function.
1414 (var_is_long_string) New function.
1415 (var_is_very_long_string) New function.
1416 (var_get_missing_values) New function.
1417 (var_set_missing_values) New function.
1418 (var_clear_missing_values) New function.
1419 (var_has_missing_values) New function.
1420 (var_is_value_missing) New function.
1421 (var_is_num_missing) New function.
1422 (var_is_str_missing) New function.
1423 (var_is_value_user_missing) New function.
1424 (var_is_num_user_missing) New function.
1425 (var_is_str_user_missing) New function.
1426 (var_is_value_system_missing) New function.
1427 (var_get_print_format) New function.
1428 (var_set_print_format) New function.
1429 (var_get_write_format) New function.
1430 (var_set_write_format) New function.
1431 (var_set_both_formats) New function.
1432 (var_get_label) New function.
1433 (var_set_label) New function.
1434 (var_clear_label) New function.
1435 (var_has_label) New function.
1436 (var_get_measure) New function.
1437 (var_set_measure) New function.
1438 (var_get_display_width) New function.
1439 (var_set_display_width) New function.
1440 (var_get_alignment) New function.
1441 (var_set_alignment) New function.
1442 (var_get_value_cnt) New function.
1443 (var_get_leave) New function.
1444 (var_get_short_name) New function.
1446 * variable.h: (struct variable) Removed "type" and "nv" members;
1447 they are now computed from "width" where needed.
1449 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1451 * missing-values.c (mv_resize): Don't write beyond end of the
1452 allocated buffer when resizing a long string.
1454 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1456 Clean up identifier code: don't require identifier enumerations to
1457 be in a particular order; make better use of string library;
1458 expose less of the internals.
1460 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1461 change interface. Updated all callers.
1462 (lex_id_match) Change interface to use struct substring, update
1464 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1465 (global array keywords[]) Make static, change form. Update all
1466 users to use lex_id_name instead.
1467 (lex_is_keyword) New function.
1468 (lex_id_to_token) Change interface to use struct substring, update
1470 (lex_id_name) New function.
1472 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1473 lex_is_keyword instead.
1474 (T_LAST_KEYWORD) Removed.
1475 (T_N_KEYWORDS) Removed.
1477 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1479 * format.c: (fmt_date_template) Distinguish characters for which a
1480 space is output and any date delimiter is allowed on input, from
1481 those for which a space is output and only a space is allowed on
1482 input. The former is represented by X, the latter by a space.
1483 Also, drop distinction between h and H, changing the former to the
1486 * data-in.c: Completely rewrite internals to conform to SPSS input
1487 formats as closely as possible.
1488 (data_in) Changed external interface by replacing the structure
1489 that was used as a single argument by a set of arguments. Updated
1491 (data_in_finite_line) Removed. Converted all callers to use plain
1493 (data_in_get_integer_format) New function.
1494 (data_in_set_integer_format) New function.
1495 (data_in_get_float_format) New function.
1496 (data_in_set_float_format) New function.
1498 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1499 (struct data_in) Removed.
1501 * data-out.c: (output_date) Drop each component from the input as
1502 it is output, to allow us to drop the distinction between h (a
1503 count of hours) and H (the hour of day) template characters.
1504 Also, handle new X template character.
1505 (output_scientific) Follow more rational rule on when to drop
1506 fraction introduced between SPSS 13 and 15. Updated test case to
1509 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1511 Fix buffer overflow reported by John Darrington.
1513 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1514 realize that DIGITS is a count of nibbles, not of bytes.
1516 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1518 * calendar.c (calendar_offset_to_gregorian) Also return the
1519 year-of-day. Change callers to new interface.
1521 * data-out.c: Completely rewrite internals to conform to SPSS
1522 output formats as completely as possible.
1523 (data_out) Change interface to put input parameters before output
1524 parameters, for consistency with the style I now prefer. Update
1526 (data_out_get_integer_format) New public function.
1527 (data_out_set_integer_format) New public function.
1528 (data_out_get_float_format) New public function.
1529 (data_out_set_float_format) New public function.
1531 * data-out.h: New file. Move prototype for data_out here, from
1534 * format.c: (fmt_step_width) Use equality comparison instead of
1535 bitwise and, for clarity.
1536 (fmt_is_string) Ditto.
1537 (fmt_input_to_output) Fix categories that are translated to F
1540 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1542 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1543 casefile-private.h: Added casefilter to assist commands with missing
1546 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1548 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1550 * settings.c: (route_errors_to_terminal) New variable.
1551 (route_errors_to_listing) New variable.
1552 (get_error_routing_to_terminal) New function.
1553 (set_error_routing_to_terminal) New function.
1554 (get_error_routing_to_listing) New function.
1555 (set_error_routing_to_listing) New function.
1557 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1559 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1561 * format.c: Completely rewrite, to achieve better abstraction.
1562 Rewrite all references to formats in other files.
1564 * format.def: Rewrite and reorganize.
1566 * settings.c: Move everything related to custom currency formats
1567 into format.[ch], changing them in form, so as to group related
1568 code and definitions better. Changed all references to use the
1570 (static var decimal) Removed.
1571 (static var grouping) Removed.
1572 (static var cc) Removed.
1573 (get_decimal) Removed.
1574 (set_decimal) Removed.
1575 (get_grouping) Removed.
1576 (set_grouping) Removed.
1580 * settings.h: (macro CC_CNT) Removed.
1581 (macro CC_WIDTH) Removed.
1582 (struct custom_currency) Removed.
1584 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1586 * data-in.c (data_in): Use switch statement instead of table, to
1587 avoid dependence on the order of the FMT_* enums.
1589 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1591 * data-out.c: (num_to_string) Removed, because it was dead code.
1593 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1595 * data-in.c (parse_trailer): Fix error message.
1597 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1599 * format.c (fmt_is_binary): New function.
1601 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1603 * procedure.c procedure.h: Encapsulated the static data into a single
1606 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1608 * casefile.c (casereader_read_xfer): Always initialize the case,
1609 even on an error condition.
1611 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1613 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1614 entire function useless.
1616 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1618 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1619 casereader method casereader_clone.
1621 * procedure.c pransformations.h: Introduced new type casenum_t
1623 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1625 * variable.c: (width_to_bytes) Rephrase code for clarify.
1627 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1629 * format.c: (fmt_type_from_string) New function.
1630 (fmt_to_string) Include decimals in output if the format has
1631 decimals, even if the format type does not. This way, we can
1632 accurately reproduce incorrect formats in user output.
1633 (check_common_specifier) Make the check for a bad format type an
1634 assertion, so we get bug reports if they show up. Fix message.
1635 Check for decimal places with a format type that doesn't allow
1637 (check_input_specifier) Remove check for FMT_X, which has been
1639 (check_output_specifier) Ditto.
1641 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1643 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1644 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1645 (enum fmt_parse_flags) Removed.
1647 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1649 * casefile.c casefile.h: Converted to an abstract base class.
1650 * casefile-private.h fastfile.c fastfile.h: New files.
1651 * automake.mk procedure.c scratch-writer.c storage-stream.c
1653 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1655 * procedure.c (internal_procedure): Create sink_case with only as
1656 many values as the compacted dictionary.
1658 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1660 Remove "debugging" code that caused plenty of false positives and
1663 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1665 * case.c: Remove all references to `this' member.
1667 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1669 Fix link error noted by Jason Stover.
1671 * storage-stream.c: Include <assert.h>.
1673 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1675 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1676 ALL) and additional underlying system file issues.
1678 Thanks to John Darrington for review.
1680 First problem: var_hash points to variables not owned by the
1681 sys-file-reader, which the caller may free or modify. Use an
1682 array of sfm_vars instead, as done earlier (e.g. CVS version
1685 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1686 members and remove all code that references it. Add vars, var_cnt
1687 members. Remove fix_specials member, which was unused.
1688 (struct sfm_var) Remove name member, which was unused.
1689 (sfm_close_reader) Free vars member instead of var_hash.
1690 (compare_var_shortnames) Removed.
1691 (hash_var_shortname) Removed.
1692 (sfm_open_reader) Fill out vars array.
1693 (compare_var_index) Removed.
1694 (sfm_read_case) Use vars instead of var_hash.
1696 Second problem: we're confused about when we actually have very
1697 long strings, causing us to choose incorrectly between slow path
1698 and fast path in sfm_read_case.
1700 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1701 have very long strings, not when we have long variable names,
1702 which is an unrelated feature.
1704 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1706 * variable.h: Move var_set and variable parsing declarations to
1707 new header, src/language/lexer/variable-parser.h. Modified lots
1708 of files to include the new header.
1710 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1712 * value-labels.c (value_to_string): When there's no value label,
1713 format the variable according to its print format, instead of
1714 always effectively using A or F format.
1716 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1718 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1721 * format.c: Removed tortological assertion.
1723 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1725 Reform string library.
1727 * file-name.c (fn_interp_vars): Change interface to take a
1728 substring as input. Updated all users.
1730 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1732 * format.c (measure_is_valid): Really return false when m >=
1735 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1737 Implement random access to casefiles, for use in GUI.
1739 * casefile.c: (struct casereader) Add `random', `file_ofs',
1740 `buffer_ofs' members.
1741 (casefile_get_random_reader) New function.
1742 (read_open_file) Break part into new function
1743 seek_and_fill_buffer().
1744 (fill_buffer) Update buffer_ofs, file_ofs.
1745 (casereader_seek) New function.
1747 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1749 * settings.c: Added call to i18n{done, init}.
1751 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1753 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1756 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1758 * casefile.c: Convert many uses of `int' to `bool'.
1760 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1762 * transformations.c (trns_chain_destroy): Destroy chain's trns
1763 member, to fix memory leak.
1765 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1767 * storage-stream.c (storage_source_decapsulate): Destroy case
1768 source to fix memory leak.
1770 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1772 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1773 case passed as argument instead of initializing the argument as a
1774 case. Fixes memory leak that showed up in
1775 tests/command/aggregate.sh with scratch files.
1777 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1779 * procedure.c (proc_done): Destroy default_dict, to fix memory
1782 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1784 Simplify procedure_with_splits().
1786 * procedure.c (struct split_aux_data): Removed case_count member.
1787 (procedure_with_splits) Don't initialize case_count.
1788 (split_procedure_case_func) Check whether prev_case is null
1789 instead of case_count.
1790 (split_procedure_end_func) Ditto.
1792 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1794 * case.c (case_move): Do nothing if dst and src are the same
1796 (case_try_create) Merge two similar cases.
1797 (case_copy) Unshare only if data must be actually copied.
1799 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1801 * data-in.c data-out.c dictionary.c sys-file-reader.c
1802 sys-file-writer.c variable.c variable.h: Reworked very long string
1803 support for better encapsulation.
1805 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1807 * value-labels.c (val_labs_can_set_width): New function.
1808 (val_labs_set_width) Clear labels if increasing width to long
1810 (val_labs_destroy) Remove unneeded test for null.
1812 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1814 * value-labels.h: Remove unneeded dependency on variable.h.
1816 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1818 Get rid of `char *c' member in union value, for cleanliness.
1820 * value.h: (union value) Remove `c' member.
1822 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1824 Make dictionary compacting functions a little more general.
1826 * sys-file-writer.c (sfm_open_writer): Use
1827 dict_compacting_would_change().
1828 (does_dict_need_translation) Removed.
1830 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1832 Make dictionary compacting functions a little more general.
1834 * dictionary.c (dict_needs_compaction): Rename
1835 dict_compacting_would_shrink(). Update all callers.
1836 (dict_compacting_would_change) New function.
1838 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1840 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1841 inverted return value (!).
1843 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1845 Continue reforming procedure execution.
1847 * procedure.c: Search and replace "vfm" by "proc". Notably:
1848 (static var vfm_source) Rename proc_source. Update all
1850 (static var vfm_sink) Rename proc_sink. Update all references.
1852 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1854 Continue reforming procedure execution. In this phase, remove
1855 PROCESS IF, which was deprecated anyway and can be easily
1856 simulated with TEMPORARY followed by SELECT IF.
1858 * procedure.c: (open_active_file) Don't call
1859 add_process_if_trns().
1860 (discard_variables) Get rid of redundant call to
1861 proc_cancel_all_transformations().
1862 (add_process_if_trns) Removed.
1863 (process_if_trns_proc) Removed.
1864 (process_if_trns_free) Removed.
1866 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1868 Continue reforming procedure execution. In this phase, add
1869 `const' to the case passed to procedure()'s callback.
1871 Updated all users of procedure() as well.
1873 * procedure.c: (struct write_case_data) Add "const" to ccase
1874 parameter for case_func member.
1875 (procedure) Add "const" to ccase parameter for proc_func
1877 (multipass_case_func) Make ccase parameter const.
1878 (internal_procedure) Add "const" to ccase parameter for case_func
1880 (split_procedure_case_func) Make ccase parameter const.
1881 (multipass_split_case_func) Make ccase parameter const.
1883 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
1885 Continue reforming procedure execution. In this phase, get rid of
1886 the output code for SPLIT FILE groups in procedure.c, which really
1887 shouldn't be doing any output. Move it into the individual
1888 procedures instead. This also adds some flexibility.
1890 Updated many users of procedure_with_splits() and
1891 multipass_procedure_with_splits() to call
1892 output_split_file_values() and to deal with increased use of
1895 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
1896 parameter to begin_func member.
1897 (procedure_with_splits) Add "const struct ccase *" parameter to
1898 begin_func parameter. Make ccase parameter const in proc_func
1900 (split_procedure_case_func) Don't dump split file group. Pass
1902 (dump_splits) Moved to language/dictionary/split-file.c as
1903 output_split_file_values().
1904 (struct multipass_split_aux_data) Add "const struct ccase *"
1905 parameter to split_func member.
1906 (multipass_procedure_with_splits) Add "const struct ccase *"
1907 parameter to split_func parameter.
1908 (multipass_split_case_func) Save new SPLIT FILE case before
1910 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
1912 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
1914 Continue reforming procedure execution. Change
1915 internal_procedure() so that it calls open_active_file() and
1916 close_active_file(), which isolates most of the actual procedure
1919 * procedure.c: (struct write_case_data) Rename `proc_func' member
1920 to `case_func' and update all references.
1921 (procedure) Rewrite as one-line wrapper around
1922 internal_procedure().
1923 (struct multipass_aux_data) New.
1924 (multipass_callback) Renamed multipass_case_func(). Use struct
1925 multipass_aux_data as auxiliary data.
1926 (multipass_end_func) New function.
1927 (multipass_procedure) Rewrite as wrapper for internal_procedure()
1928 that uses multipass_case_func, multipass_end_func.
1929 (internal_procedure) Add `end_func' argument. Move optimization
1930 of trivial case in here. Move call to open_active_file() and
1931 close_active_file() in here. Now assert that vfm_source is
1933 (procedure_with_splits_callback) Rename
1934 split_procedure_case_func().
1935 (split_procedure_end_func) New function.
1936 (multipass_split_callback) Rename multipass_split_case_func.
1937 (multipass_split_end_func) New function.
1938 (discard_variables) No need to test for nonnull vfm_source.
1940 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
1942 Continue reforming procedure execution. Get rid of unused member.
1944 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
1946 (write_case) Don't increment cases_analyzed.
1948 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
1950 Continue reforming procedure execution. In this phase, move
1951 procedure.c and procedure.h from src to src/data. Update
1952 makefiles and #includes accordingly.
1954 * procedure.c: Moved here from src/.
1956 * procedure.h: Moved here from src/.
1958 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
1960 Continue reforming procedure execution. In this phase, get rid of
1961 many global variables, consolidating procedure execution in
1962 procedure.c. Encapsulate transformations in new "struct
1963 trns_chain". Also, change implementation of N OF CASES, FILTER,
1964 and PROCESS IF from special cases to transformations.
1966 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
1969 * dictionary.c: (global variable default_dict) Move to
1972 * variable.h: (TRNS_*) Move to transformations.h.
1973 (struct transformation) Move to transformations.c.
1975 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1977 * sys-file-reader.c: Fixed invalid read problems.
1979 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
1981 * storage-stream.c: Add missing function comments.
1983 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
1985 Continue reforming procedure execution. In this phase, add some
1986 new, needed functionality to storage-stream.
1988 * storage-stream.c: (storage_source_decapsulate) New function.
1990 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
1992 * variable.c (width_to_bytes): Declarations must precede
1993 statements for C90 compliance.
1995 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
1997 * data-out.c, data-in.c, variable.c, variable.h: New functions
1998 copy_mangle and copy_demangle for reading/writing cases; emulates the
1999 way SPSS deals with strings > 255 bytes.
2001 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2002 subtype 14 needed for strings longer than 255 bytes.
2004 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2005 instead of literal values. Also fixed some constness issues.
2007 * format.h: Constness
2009 * sfm-private.h: Renamed the case_size identifier, since I discovered
2010 that SPSS's respect for this variable is very nominal.
2012 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2014 Change case limit type from int to size_t.
2016 * dictionary.c: (struct dictionary) Change type of case_limit
2018 (dict_get_case_limit) Change return type.
2019 (dict_set_case_limit) Change parameter type.
2021 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2023 * variable.h: (struct variable) Rename `reinit' member as `leave'
2024 and invert sense. Fix up all references.
2026 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2028 Continue reforming procedure execution. In this phase, break
2029 procedure.c into multiple files.
2031 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2033 * case-sink.c: New file.
2035 * case-sink.h: New file.
2037 * case-source.c: New file.
2039 * case-source.h: New file.
2041 * storage-stream.c: New file.
2043 * storage-stream.h: New file.
2045 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2047 * variable.h: (struct variable) Remove `init' member and all
2048 references to it from other files. It was initialized in several
2049 places, but nothing really ever used it for anything worthwhile.
2050 Thanks to Jason Stover for pointing out how confusing this
2053 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2055 Continue reforming error message support. In this phase, get rid
2056 of message "titles" and put the message text in `struct error'.
2057 Now `struct error' encapsulates a message more properly.
2059 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2060 Format message ourselves.
2062 * data-in.c: (vdls_error) Ditto.
2064 * por-file-reader.c: (error) Ditto.
2066 * sys-file-reader.c: (corrupt_msg) Ditto.
2068 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2070 GNU standards require "file name" instead of "filename" in
2071 documentation. It's nice for our code to follow the convention
2074 * casefile.c: (struct casefile) Rename `filename' member to
2075 `file_name'. Updated all references.
2077 * file-name.c: [!unix] (struct file_identity) Rename
2078 normalized_filename member to normalized_file_name. Updated all
2081 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2083 We don't really support anything but Unix-like environments well,
2084 so we might as well de-obfuscate by writing directory and path
2085 separators explicitly.
2087 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2089 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2090 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2092 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2095 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2097 GNU standards require "file name" instead of "filename" in
2098 documentation. It's nice for our code to follow the convention
2101 * filename.c: Rename to file-name.c.
2103 * filename.h: Rename to file-name.h. Update all inclusions.
2104 Update header guards.
2106 * automake.mk: Update file names.
2108 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2110 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2112 (fn_basename) Removed (dead code).
2113 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2114 (fn_special_p) Renamed fn_is_special(), all references updated.
2115 (fn_exists_p) Renamed fn_exists(), all references updated.
2117 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2119 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2120 Also, now it only tilde-expands file names, not paths.
2121 (fn_search_path) Tilde-expand one directory at a time.
2123 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2125 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2126 get rid of non-Unixlike version of the code, which has probably
2128 (fn_prepend_dir) Removed (dead code).
2130 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2131 (macro PATH_SEPARATOR_STRING) New.
2132 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2134 Continue reforming error message support. In this phase, we get
2135 rid of VM() and the other msg() support for "verbosity", replacing
2136 it by a new function verbose_msg().
2138 * filename.c: (fn_search_path) Use verbose_msg() instead of
2141 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2143 * sfm-private.h: Get rid of #defines after #error, which makes no
2146 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2148 Get rid of our own int32 type in favor of the standard int32_t
2151 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2154 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2156 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2158 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2160 Remove ill-considered file routines that are no longer used.
2162 * filename.c: (fn_open_ext) Removed.
2163 (fn_close_ext) Removed.
2165 * filename.h: (struct file_ext) Removed.
2167 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2169 * variable.c (var_is_valid_name): Move declarations before code
2172 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2174 * filename.ch (fn_interp_vars): Fixed small buglet.
2176 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2178 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2179 so as to modify the string inline. Thus makeing it easier to
2180 destroy the results when no longer needed.
2182 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2184 * category.c (cat_stored_values_destroy): Fixed memory leak.
2186 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2188 Add some missing frees. Thanks to John Darrington for reporting
2191 * any-writer.c (any_writer_close): Free writer.
2193 * any-reader.c (any_reader_close): Free reader.
2195 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2197 * por-file-reader.c: (error) Mark as NO_RETURN.
2199 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2201 * settings.c: Changed default value of scompress to true.
2203 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2205 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2207 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2210 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2211 here from pref.h.orig.
2213 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2215 * sys-file-reader.c: Fixed bug reading compressed files.
2217 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2219 * Numerous renames. See src/ChangeLog for details.
2221 * Moved files from src directory