1 2007-11-03 Ben Pfaff <blp@gnu.org>
3 Allow output files to overwrite input files (bug #21280).
5 * file-handle-def.c: Separate locking of files for input and for
6 output, to allow both to take place at once. Also, distinguish a
7 file handle from the identity of the underlying file, because the
8 identity of a file changes over time and the file handle can't
9 represent two different identities.
10 (struct file_handle): Remove `next', `open_cnt', `deleted',
11 `type', `open_mode', `aux', `identity' members. Change `id' from
12 char array to char *. Add `ref_cnt' member.
13 (file_handle_from_ll) New function.
14 (file_handles) Removed.
15 (named_handles) New variable.
16 (fh_init) Initialize named_handles.
17 (fh_done) Remove name from all named_handles.
18 (free_handle) Updated for modified struct file_handle.
19 (unname_handle) New function.
20 (fh_ref) New function.
21 (fh_from_file_name) Removed.
22 (fh_unref) New function.
23 (fh_unname) New function.
24 (fh_from_id) Rewritten.
25 (create_handle) Updated for modified struct file_handle.
26 (fh_inline_file) Increment the handle's ref count.
27 (fh_create_file) Updated for modified struct file_handle.
28 (fh_create_scratch) Ditto.
34 (fh_get_id) Updated for modified struct file_handle.
35 (fh_get_default_handle) Increment the handle's ref count.
36 (fh_set_default_handle) Handle ref counts.
37 (struct fh_lock) New structure.
38 (locks) New static var.
39 (fh_lock) New function.
40 (fh_unlock) New function.
41 (fh_lock_get_aux) New function.
42 (fh_lock_set_aux) New function.
43 (fh_is_locked) New function.
44 (make_key) New function.
45 (free_key) New function.
46 (compare_fh_locks) New function.
47 (hash_fh_lock) New function.
49 * file-handle-def.h (enum fh_access) New enum.
51 * file-name.c: Made file_identity the same in all supported
53 (struct file_identity): New `name' member.
54 (fn_get_identity): For a file that doesn't exist, get the
55 dev/inode of its directory plus its name. If even the directory
56 doesn't exist, just use its name. Merge the Windows
57 implementation into the Unix one.
58 (fn_compare_file_identities): Rewritten. Merge the Windows
59 implementation into the Unix one.
60 (fn_hash_identity): New function.
62 * make-file.c (struct replace_file): New structure.
63 (all_files): New static var.
64 (replace_file_start): New function.
65 (replace_file_commit): New function.
66 (replace_file_abort): New function.
67 (free_replace_file): New function.
68 (unlink_replace_files): New function.
70 * por-file-reader.c (struct pfm_reader): Add `lock' member.
71 (close_reader): Unlock file.
72 (pfm_open_reader): Lock file.
74 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
76 (pfm_open_writer): Lock file and prepare for its replacement.
77 (close_writer): Unlock file.
79 * scratch-handle.h (struct scratch_handle): Add unique_id so that
80 different generations of a scratch file can be distinguished.
82 * scratch-reader.c (scratch_reader_open): Verify that the file is
85 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
86 remove scratch_handle member.
87 (scratch_writer_open): Lock handle. Prepare to replace handle
88 data, instead of doing it immediately.
89 (scratch_writer_casewriter_destroy): Replace handle data and
92 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
93 (sfm_open_reader): Lock file.
94 (close_reader): Unlock file.
96 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
98 (sfm_open_writer): Lock file and prepare for its replacement.
99 (close_writer): Unlock file.
101 2007-11-02 Ben Pfaff <blp@gnu.org>
103 * data-out.c (output_number): Use isfinite (from C99) instead of
104 gsl_isfinite, since we now have universal support for it in
106 (output_infinite): Ditto.
107 [!HAVE_ROUND] (round): Remove definition, since we now have a
108 replacement in gnulib.
110 2007-10-27 John Darrington <john@darrington.wattle.id.au>
112 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
113 wasn't initialised correctly. Closes bugs #21330 and 21397
116 2007-10-08 Ben Pfaff <blp@gnu.org>
118 Fix bug #21280. Thanks to John Darrington for review.
120 * file-name.c (create_stream): New function.
122 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
123 file handle before creating the file, to ensure that we don't
124 truncate a file that we're reading. Make code easier to read by
127 * sys-file-write.c (sfm_open_writer): Ditto.
129 2007-10-01 Ben Pfaff <blp@gnu.org>
131 Fix bug #21192. Thanks to John Darrington for review.
133 * casereader.c (casereader_read): Decrement case_cnt before
134 calling the casereader's "read" member function, so that we
135 interact properly with lazy_casereader.
137 * datasheet.c: Add regression test for above bug fix.
138 (clone_datasheet): New function.
139 (lazy_callback): New function.
140 (check_datasheet_casereader): New function.
141 (check_datasheet): Check datasheet contents are reported correctly
142 through an ordinary casereader and a lazy casereader.
143 (clone_model): Use clone_datasheet.
145 2007-09-24 Ben Pfaff <blp@gnu.org>
147 Patch #6210. Reviewed by John Darrington.
148 * settings.c: Drop "long view", which is not needed any longer.
149 (static var long_view): Removed.
150 (force_long_view): Removed.
151 (get_viewwidth): Removed.
152 (init_viewport): Removed long_view reference.
154 2007-09-19 John Darrington <john@darrington.wattle.id.au>
156 * settings.c settings.h: Changed viewport's length and width to be
157 owned by the user interface which uses the data library. This allows
158 better abstraction, and makes dynamically adjustable dimensions easier.
160 2007-09-18 Ben Pfaff <blp@gnu.org>
162 * procedure.c (proc_extract_active_file_data): New function.
164 * lazy-casereader.h: New file.
166 * lazy-casereader.c: New file.
168 * casereader.c (casereader_dynamic_cast): New function.
170 2007-09-14 Ben Pfaff <blp@gnu.org>
172 * dictionary.c (dict_clone): Copy case indexes from cloned
173 dictionary. Fixes bug #21061. Reviewed by John Darrington.
175 2007-09-13 John Darrington <john@darrington.wattle.id.au>
177 * value.c value.h (create_value): New function.
179 2007-09-12 Ben Pfaff <blp@gnu.org>
181 Make it clear that translator casereader and casewriter translate
182 functions are supposed to destroy their input case.
184 * casereader-translator.c (struct casereader_translator): Change
185 input case parameter of translate member function from const
186 struct ccase * to struct ccase *.
187 (casereader_create_translator): Ditto, for translate parameter.
189 * casewriter-translator.c (struct casewriter_translator): Ditto.
190 (casewriter_create_translator): Ditto.
192 2007-08-27 John Darrington <john@darrington.wattle.id.au>
194 * sys-file-reader.c (read_display_parameters): Force display width
195 to 8, if the sys file says 0 (like SPSS does).
197 2007-08-12 Ben Pfaff <blp@gnu.org>
199 * dictionary.c (dict_dump): New function.
201 2007-08-12 Ben Pfaff <blp@gnu.org>
203 Drop dict_compactor in favor of using the new struct case_map.
205 * dictionary.c (struct copy_map): Removed.
206 (struct dict_compactor): Removed.
207 (dict_make_compactor): Removed.
208 (dict_compactor_compact): Removed.
209 (dict_compactor_destroy): Removed.
211 * procedure.c (struct data_set): Change `compactor' member to be a
213 (proc_open): Use case_map_to_compact_dict instead of
215 (proc_casereader_read): Use case_map_execute instead of
216 dict_compactor_compact.
217 (proc_commit): Use case_map_destroy instead of
218 dict_compactor_destroy.
220 * scratch-writer.c (struct scratch_writer): Change `compactor'
221 member to be a struct case_map *.
222 (scratch_writer_open): Use case_map_to_compact_dict instead of
224 (scratch_writer_casewriter_write): Use case_map_execute instead of
225 dict_compactor_compact.
227 2007-08-12 Ben Pfaff <blp@gnu.org>
229 * automake.mk: Add case-map.c, case-map.h.
231 * case-map.c: New file.
233 * case-map.h: New file.
235 2007-08-12 Ben Pfaff <blp@gnu.org>
237 * dictionary.c (dict_compact_values): Don't delete scratch
238 variables as well as compacting case indexes. Update all callers.
239 (dict_get_compacted_value_cnt): Rename dict_count_values and
240 change interface. Update all callers.
241 (dict_get_compacted_value_cnt): Remove.
242 (dict_compacting_would_shrink): Remove.
243 (dict_compacting_would_change): Remove.
244 (dict_make_compactor): Add new parameter. Update all callers.
246 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
247 instead of dict_count_values, changing an O(N) operation into
250 2007-08-12 Ben Pfaff <blp@gnu.org>
252 * casereader.c (casereader_read): Don't require cases read by a
253 casereader to be exactly the expected size: as long as they're big
256 2007-08-12 Ben Pfaff <blp@gnu.org>
258 Make casewriters keep track of the number of `union value's in
259 each case. This is useful for two reasons: casewriter_write can
260 then check that the case being written is large enough, and later
261 recipients of the casewriter can determine the size of the case.
263 * casewriter-translator.c (casewriter_create_translator): Add
266 * casewriter.c (struct casewriter): Add value_cnt member.
267 (casewriter_write): Check that the case passed in is big enough.
268 (casewriter_get_value_cnt): New function.
269 (casewriter_create): Add value_cnt parameter.
271 2007-08-09 Ben Pfaff <blp@gnu.org>
273 Fix bug reported by Jason Stover.
274 * settings.c: Move get_termcap_viewport and in particular the
275 #include for <curses.h> to the end of the file. curses.h
276 redefines bool on some systems (e.g. OpenBSD), which causes
277 disagreement between uses of bool before its inclusion and
279 Tested by Jason Stover.
281 2007-07-29 Ben Pfaff <blp@gnu.org>
283 Provisional fix for bug #18692 and bug #20161. Reviewed by John
286 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
287 argument (never "rb" or "wb") because SUSv3 says that only those
288 modes are defined, and glibc in fact rejects other modes.
290 Open portable files with fn_open so that they can be read from
291 pipes. Fix missing fh_close call to go along with fh_open.
292 Report an error if the file close reports an error.
293 * por-file-reader.c (close_reader): New function.
294 (por_file_casereader_destroy): Use close_reader.
295 (pfm_open_reader): Open file with fn_open.
297 2007-07-28 Ben Pfaff <blp@gnu.org>
299 Make PSPP able to read all the portable files I could find on the
300 web. Thanks to John Darrington for review. Bug #17620.
301 * por-file-reader.c (struct pfm_reader): New member `line_length'.
302 (error): Print file offset in hexadecimal.
303 (warning): New function.
304 (advance): Treat lines less than 80 bytes long as padded to 80
306 (pfm_open_reader): Call read_documents if we find an "E" record.
307 (convert_format): Convert invalid formats to the default format
308 instead of aborting reading the file.
309 (read_variables): Rename duplicate variable names instead of
310 aborting reading the file.
311 (read_value_label): Allow string variables of different widths to
312 be assigned value labels in the same record. Replace duplicate
313 value labels instead of aborting.
314 (read_documents): New function.
316 * por-file-writer.c (pfm_open_writer): Call write_documents if the
317 dictionary has documents.
318 (write_documents): New function.
320 2007-07-25 Ben Pfaff <blp@gnu.org>
322 Fix bugs related to bug #17213.
324 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
325 since the former is what config.h has. Include the needed ncurses
327 (static var echo) Rename to `do_echo' because the original name is
328 the same as an ncurses identifier.
329 (get_termcap_viewport) Use error instead of msg.
331 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
332 (fn_close): Don't close stdin, stdout, stderr.
334 2007-07-26 John Darrington <john@darrington.wattle.id.au>
336 * procedure.c procedure.h: Added callbacks which get invoked whenever
337 a dataset's transformation chain changes.
339 2007-07-24 Ben Pfaff <blp@gnu.org>
342 * sys-file-writer.c (write_variable_display_parameters): Use new
343 var_default_display_width function to choose display width of
344 segments after the first one in a given variable.
345 * variable.c (var_create): Use var_default_display_width to pick
346 new variable's display width.
347 (var_default_display_width): New function.
348 Reviewed by John Darrington.
350 2007-07-24 Ben Pfaff <blp@gnu.org>
353 * por-file-writer.c (write_variables): Write weight variable.
354 Reviewed by John Darrington.
356 2007-07-23 Ben Pfaff <blp@gnu.org>
358 Improvements to system file reader and writer.
360 First, move all detailed knowledge of very long strings into
361 sys-file-private.[ch], so that this nasty stuff can be isolated.
363 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
364 (EFFECTIVE_VLS_CHUNK): New macro.
365 (min_int): New function.
366 (max_int): New function.
367 (sfm_width_to_bytes): Rewrite.
368 (sfm_width_to_octs): New function.
369 (sfm_segment_alloc_width): New function.
370 (sfm_segment_alloc_bytes): New function.
371 (sfm_segment_used_bytes): New function.
372 (sfm_segment_offset): New function.
373 (sfm_segment_effective_offset): New function.
374 (sfm_dictionary_to_sfm_vars): New function.
376 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
377 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
378 (struct sfm_var): New structure.
380 Next, improvements to the system file reader.
382 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
383 type casenumber. Added `version_major', `version_minor',
386 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
387 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
388 Change `case_cnt' to type casenumber. Removed `has_vls'.
389 (struct sfm_var): Removed.
390 (sfm_open_reader): Don't warn on wrong case size if the file was
391 written by SPSS 13, which tends to get it wrong. Use
392 sfm_dictionary_to_sfm_vars.
393 (read_header): Always output system file info.
394 (read_variable_record): Simplify code for reading missing values.
395 (read_machine_int32_info): Save version numbers from system file
396 into info struct passed as new argument.
397 (read_long_string_map): Restructured to use new sys-file-private
399 (read_value_labels): Use size_overflow_p.
400 (sys_file_casereader_read): Get rid of distinction between fast
401 and slow paths. Use information provided by sys-file-primate's
402 struct sfm_var to simplify code.
403 (skip_whole_strings): New function.
404 (read_int32): Renamed read_int. Changed return value to int.
406 (read_flt64): Renamed read_float. Changed return value to
407 double. Updated all callers.
408 (int32_to_native): Removed. Changed callers to use
410 (flt64_to_double): Removed. Changed callers to use float_convert.
412 Finally, get rid of int32, flt64 terminology and types in system
413 file writer. The former wasn't very useful since a POSIX "int"
414 can hold the whole range of int32 and we generally didn't have a
415 need for it to be exactly-32-bits, just at-least-32-bits. The
416 latter was inconvenient because we had to assume that it could be
417 different from double and thereby convert special values SYSMIS,
418 HIGHEST, LOWEST to and from it in multiple places. Instead, now
419 we just use "int" and "double" in most places, and do conversions,
420 if necessary, very close to where we do I/O. This change meant
421 that the writer code couldn't represent records in the file as C
422 structs any longer, but that's no great loss. The code actually
423 seems to be more readable without them.
425 Simplify the compression buffering code: only buffer as much as
426 necessary, which is no more than eight 8-byte units at any given
429 * sys-file-writer.c (typedef flt64): Removed.
430 (macro second_lowest_flt64): Removed.
431 (struct sysfile_header): Removed.
432 (struct sysfile_variable): Removed.
433 (struct sfm_writer): Removed `needs_translation', `has_vls',
434 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
435 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
436 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
437 compression buffering by `opcodes', `opcode_cnt', `data',
438 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
439 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
440 data writer functions instead of structures.
441 (calc_oct_idx): New function.
442 (write_header): Use simple data writer functions instead of
444 (write_format_spec): Renamed write_format. New argument.
445 (write_variable_continuation_records): New function.
446 (write_variable): Use simple data writer functions instead of
447 structures. Use write_variable_continuation_records. Write
448 entire very long string instead of requiring caller to understand
450 (write_value_labels): Use simple data writer functions instead of
452 (write_documents): Ditto.
453 (write_variable_display_parameters): Use sys-file-private
454 functions to simplify. Use simple data writer functions instead
456 (write_vls_length_table): Use simple data writer functions instead
458 (write_longvar_table): Ditto.
459 (write_rec_7_34): Break into new functions
460 write_integer_info_record, write_float_info_record. Use simple
461 data writer functions instead of structures.
462 (buf_write): Removed.
463 (append_string_max): Removed.
464 (ensure_buf_space): Removed.
465 (sys_file_casewriter_write): Get rid of the distinction between
466 fast and slow paths, which didn't seem to be too useful. Use new
467 functions write_case_uncompressed, write_case_compressed.
468 (put_instruction): Removed.
469 (put_element): Removed.
470 (write_compressed_data): Removed.
471 (close_writer): Use flush_compressed. Only write case count to
472 system file if it will fit in the field.
473 (write_case_compressed): New function.
474 (write_case_uncompressed): New function.
475 (flush_compressed): New function.
476 (put_cmp_opcode): New function.
477 (put_cmp_number): New function.
478 (write_int): New function.
479 (convert_double_to_output_format): New function.
480 (write_float): New function.
481 (write_value): New function.
482 (write_string): New function.
483 (write_bytes): New function.
484 (write_zeros): New function.
485 (write_spaces): New function.
487 Reviewed by John Darrington.
489 2007-07-22 Ben Pfaff <blp@gnu.org>
491 Don't try to write very long strings to portable files. The
492 format does not support it.
494 * por-file-writer.c (MAX_POR_WIDTH): New macro.
495 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
496 (write_format): Add arg to take width to resize format to.
497 (write_value): Limit width of value written to MAX_POR_WIDTH.
498 (write_variables): Limit width of variable and its output formats
500 Reviewed by John Darrington.
502 2007-07-22 Ben Pfaff <blp@gnu.org>
504 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
505 local variable instead of literal 5.
506 Reviewed by John Darrington.
508 2007-07-22 Ben Pfaff <blp@gnu.org>
510 Fix problems with uniqueness of short names in system files with
511 very long string variables. Now a variable may have multiple
514 * automake.mk (src_data_libdata_a_SOURCES): Add new files
515 short-names.c, short-names.h.
517 * dictionary.c (dict_clone): Clone all the short names.
518 (compare_strings): Move into short-names.c.
519 (hash_strings): Ditto.
520 (set_var_short_name_suffix): Ditto.
521 (dict_assign_short_names): Ditto, rename short_names_assign,
522 change to assign all short names.
524 * por-file-writer.c (write_variables): Use short_names_assign
525 instead of dict_assign_short_names.
527 * short-names.c: New file.
529 * short-names.h: New file.
531 * sys-file-private.c (sfm_width_to_segments): New function.
533 * sys-file-reader.c (read_long_var_name_map): Save and restore all
534 the short names, not just the first one.
536 * sys-file-writer.c (cont_var_name): Removed.
537 (sfm_open_writer): Use short_names_assign instead of
538 dict_assign_short_names. Use unique short names assigned by
539 short_names_assign instead of those generated by cont_var_name.
541 * variable.c (struct variable): Remove `short_name' member,
542 replace by `short_names' and `short_name_cnt'.
543 (var_create) Initialize new members.
544 (var_get_short_name_cnt): New function.
545 (var_get_short_name): Now takes an index argument. Changed most
547 (var_set_short_name): Ditto.
548 (var_clear_short_name): Renamed var_clear_short_names, changed to
549 clear all short names.
551 Reviewed by John Darrington.
553 2007-07-22 Ben Pfaff <blp@gnu.org>
555 * variable.c (var_set_width): Use new var_set_width function.
557 * missing-values.c (mv_n_values): Drop assertion, which was not
560 * format.c (fmt_default_for_width): New function.
561 (fmt_resize): New function.
563 Reviewed by John Darrington.
565 2007-07-18 John Darrington <john@darrington.wattle.id.au>
567 * datasheet.c (datasheet_delete_columns): Added assertion to check
568 we're not deleting outside the range of the sheet.
571 * dictionary.c dictionary.h variable.c: Added the ability for string
572 variables to be resized.
574 * vardict.h: Added some prototypes (moved from dictionary.h) as
575 these should only be called by variable.c
578 2007-07-14 John Darrington <john@darrington.wattle.id.au>
580 * sfm-reader.c: Respect case_cnt field in file header.
582 2007-07-01 John Darrington <john@darrington.wattle.id.au>
584 * transformation.c transformation.h (trns_chain_execute): Changed the
585 signature (Patch #6057)
587 2007-06-10 Ben Pfaff <blp@gnu.org>
589 * casereader-filter.c (casereader_filter_destroy): Make sure to
590 write all the remaining excluded cases to the casewriter, if any.
592 * caseinit.c (init_list_destroy): Rewrite.
593 (init_list_clear): Ditto.
595 * casegrouper.c (casegrouper_get_next_group): Always set *reader
596 to null when returning false.
598 2007-06-06 Ben Pfaff <blp@gnu.org>
600 Actually implement the new procedure code and adapt all of its
601 clients to match. Also adapt all of the other case sources and
602 sinks in the tree and their clients to use the
603 casereader/casewriter infrastructure.
605 * automake.mk: Add and remove files.
607 * any-reader.c: Change into a casereader.
608 * por-file-reader.c: Ditto.
609 * scratch-reader.c: Ditto.
610 * sys-file-reader.c: Ditto.
612 * any-writer.c: Change into a casewriter.
613 * por-file-writer.c: Ditto.
614 * scratch-writer.c: Ditto.
615 * sys-file-writer.c: Ditto.
617 * procedure.c: Change to use casereader, casewriter, caseinit, and
618 other new infrastructure.
620 * scratch-handle.c: Adapt to new infrastructure.
622 * case-sink.c: Removed, now dead code.
623 * case-sink.h: Ditto.
624 * case-source.c: Ditto.
625 * case-source.h: Ditto.
626 * casefile-factory.c: Ditto.
627 * casefile-private.h: Ditto.
630 * casefilter.c: Ditto.
631 * casefilter.h: Ditto.
634 * fastfile-factory.c: Ditto.
635 * fastfile-factory.h: Ditto.
636 * storage-stream.c: Ditto.
637 * storage-stream.h: Ditto.
639 2007-06-06 Ben Pfaff <blp@gnu.org>
643 * automake.mk: Add new files.
645 * datasheet.c: New file.
647 * datasheet.h: New file.
649 2007-06-06 Ben Pfaff <blp@gnu.org>
651 Until now, the procedure code has provided a case to the
652 case_source, which has filled in the data values that come from
653 the active file. "Left" data values that don't come from the
654 active file naturally stay the same from case to case, because the
655 procedure code keeps using that same case.
657 One of the compromises that comes with the new procedure code is
658 that the active file allocates and provides its own case, which
659 the procedure code then has to resize to provide room for any
660 other variables that should go in the case and then fill in the
661 values of "left" variables. Then, when we're done with that case,
662 we have to save the values of "left" variables to copy into the
663 next case read from the active file.
665 The caseinit code helps with this.
667 * automake.mk: Add new files.
669 * caseinit.c: New file.
671 * caseinit.h: New file.
673 2007-06-06 Ben Pfaff <blp@gnu.org>
675 * value.h (value_cnt_from_width): New function.
677 * variable.c (var_get_value_cnt): Use new function.
679 2007-06-06 Ben Pfaff <blp@gnu.org>
681 Add casegrouper, to allow cases read from a given casereader to be
682 broken into groups, each of which has its own casereader.
683 Generally cases are grouped based on having equal values for some
686 * automake.mk: Add new files.
688 * casegrouper.c: New file.
690 * casegrouper.h: New file.
692 2007-06-06 Ben Pfaff <blp@gnu.org>
694 Add interface to lexicographical ordering of cases.
696 * automake.mk: Add new files.
698 * case-ordering.c: New file.
700 * case-ordering.h: New file.
702 2007-06-06 Ben Pfaff <blp@gnu.org>
704 Add casereaders and casewriters, the basis of the new data processing
705 implementation. A casereader is a uniform interface to reading cases
706 from a data source; a casewriter is a uniform interface to writing
707 cases to a data sink.
709 * automake.mk: Add new files.
711 * casereader-filter.c: New file.
713 * casereader-provider.h: New file.
715 * casereader-translator.c: New file.
717 * casereader.c: New file.
719 * casereader.h: New file.
721 * casewriter-provider.h: New file.
723 * casewriter-translator.c: New file.
725 * casewriter.c: New file.
727 * casewriter.h: New file.
729 2007-06-06 Ben Pfaff <blp@gnu.org>
731 "casewindow" data structure that extends the deque (from libpspp)
732 of cases with the ability to dump cases to disk if we get too many
735 * automake.mk: Add new files.
737 * casewindow.c: New file.
739 * casewindow.h: New file.
741 2007-06-06 Ben Pfaff <blp@gnu.org>
743 sparse_cases data structure that augments a sparse_array of cases
744 with the ability to dump cases to disk if we get too many cases in
747 * automake.mk: Add new files.
749 * sparse-cases.c: New file.
751 * sparse-cases.h: New file.
753 2007-06-06 Ben Pfaff <blp@gnu.org>
755 Adds a low-level on-disk case array data structure.
757 * automake.mk: Add new files.
759 * case-tmpfile.c: New file.
761 * case-tmpfile.h: New file.
763 2007-06-06 Ben Pfaff <blp@gnu.org>
765 In a couple of places we calculate the maximum number of cases to
766 keep in memory based on the user-defined workspace. Enable
767 centralizing the calculation through a new function.
769 * settings.c (get_workspace_cases): New function.
771 2007-06-06 Ben Pfaff <blp@gnu.org>
773 The casenumber type is defined in transformations.h, but case.h is
774 a more sensible place. Move it.
776 * case.h (CASENUMBER_MAX): New macro.
777 (typedef casenumber): Move here, from transformations.h.
779 2007-06-03 Ben Pfaff <blp@gnu.org>
781 Slightly generalize case_to_values and case_from_values functions.
783 * case.c (case_to_values): Rename case_copy_out, change interface.
784 (case_from_values): Rename case_copy_in, change interface.
786 * fastfile.c (fastfilereader_get_next_case): Update caller.
787 (write_case_to_disk): Ditto.
789 2007-06-02 Ben Pfaff <blp@gnu.org>
791 Clean up after a forgotten part of patch #5829.
793 * casedeque.h: Remove unused file.
795 * automake.mk: Remove casedeque.h from sources.
797 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
799 * category.c: Removed redundant #include
801 2007-05-06 Ben Pfaff <blp@gnu.org>
803 Abstract the documents within a dictionary a little better.
804 Thanks to John Darrington for suggestion, initial version, and
807 * dictionary.c (struct dictionary): Change `documents' member from
808 char * to struct string.
809 (dict_clear): Destroy struct string.
810 (dict_get_documents): Convert struct string to char *.
811 (dict_set_documents): Set struct string. Pad to 80-character
813 (dict_clear_documents): New function.
814 (dict_add_document_line): New function.
815 (dict_get_document_line_cnt): New function.
816 (dict_get_document_line): New function.
818 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
820 * sys-file-reader.c (read_documents): Use new document functions.
822 2007-04-19 John Darrington <john@darrington.wattle.id.au>
824 * sys-file-reader.c: When reading a system file which has no
825 long name table, automatically create one where the long names
826 are the lower case versions of the short names.
828 2007-04-22 Ben Pfaff <blp@gnu.org>
830 * dictionary.c (dict_set_split_vars): dict_destroy expects that
831 dict_clear will free most data related to the dictionary.
832 dict_clear does a decent job, except that dict_set_split_vars on
833 some systems won't actually free the dict's "split" member.
834 Instead, it'll allocate a 1-byte region. Fix this.
836 * value.c (value_copy): New function.
837 (value_set_missing): Ditto.
839 2007-04-22 John Darrington <john@darrington.wattle.id.au>
841 * Deleted existing category.h and moved cat-routines.h into
842 category.h Encapsulated struct cat_vals better.
844 2007-04-19 John Darrington <john@darrington.wattle.id.au>
846 * sys-file-reader.c: When reading a system file which has no
847 long name table, automatically create one where the long names
848 are the lower case versions of the short names.
850 2007-04-16 John Darrington <john@darrington.wattle.id.au>
852 * sys-file-reader.c: Some versions of Other Software seem to
853 produce system files with string variables' measure set to
854 zero. We'll assume these are supposed to be nominal variables.
856 2007-03-30 Ben Pfaff <blp@gnu.org>
858 * procedure.c: Adapt to new deque data structure.
860 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
861 Ben Pfaff <blp@gnu.org>
863 * file-name.c: Mingw compatibility fixes.
864 (fn_search_path): Use ISSLASH instead of comparing against '/'
866 (fn_dir_name): Use dir_name from gnulib.
867 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
868 (fn_get_identity): Use GetFullPathName instead of canonicalize
869 from gnulib, because the latter does not fully support
870 Windows-style path names. Use this implementation based on the
871 detected presence of Windows instead of the absence of Unix, since
872 the new implementation is Windows-specific.
873 (fn_compare_file_identities): In Windows implementation, compare
874 names case-insensitively.
876 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
878 * make-file.c: Don't include mkstemp.h, because gnulib now causes
879 <stdlib.h> to have the same effect.
881 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
883 * por-file-reader.c: Add missing _() around messages.
885 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
887 * make-file.c: Include "mkstemp.h", without which linking on
890 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
892 Reduce platform dependence.
894 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
895 Everywhere we using this, we really should have just depended on
896 the shell to expand tildes.
897 (fn_search_path): Simplify, given that we don't do tilde expansion
899 (fn_normalize): Removed. Caller changed to use the canonicalize
901 (fn_get_cwd): Removed. Only user was fn_normalize.
902 (fn_is_absolute): Really only test for absolute names.
903 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
905 (fn_readlink): Removed, as it was only used fn_normalize.
906 (fn_exists): Assume the stat function is available; gnulib does.
907 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
909 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
911 * dictionary.c (dict_create_vector_assert): New function.
913 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
915 * file-name.c (fn_normalize): Correct name of function
916 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
917 for pointing this out.
919 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
921 We are using a single member in struct file_handle, the "name"
922 field, for more than one purpose. When it begins with '"', it's a
923 file name; otherwise, it's a token that can be used to identify
924 it. When that assertion fires, it's because we searched for the
925 name case-sensitively as a file name (so that there was no match),
926 and then we try to insert it case-insensitively as a token, which
927 fails because duplicates aren't allowed.
929 Solution: break the two purposes into two separate fields. This
930 fixes the problem and likely makes the code easier to read too.
932 Fixes bug #18922. Thanks to John Darrington for bug report and
935 * file-handle-def.c (struct file_handle): New `id' member.
936 (fh_from_name): Rename fh_from_id. Update all callers.
937 (create_handle): New `id' parameter. Update all callers.
938 (fh_create_file): Ditto.
939 (fh_get_id): New function.
941 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
943 * case.c (case_is_null): Change return type to bool.
945 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
949 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
950 reference counting to share data. This makes it easy for
951 valgrind, etc. to find accesses to cases that have been destroyed
952 but have been kept around by another user's ref-count. This often
953 happens when the data set is small enough to find in memory; if a
954 bigger data set that would overflow to disk were used, then data
955 corruption would occur.
957 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
961 * case.c (case_unshare): Make it check internally whether the
962 ref_cnt is greater than 1, so that the callers don't have to.
963 Update callers not to check.
965 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
967 Before, I was thinking that I might want to get rid of reference
968 counting at some point. Now, I'm pretty sure that it's here to
969 stay. Thus, because we have to store the value_cnt anyway for
970 reference-counted cases, we might as well expose it to users.
972 * case.c (case_get_value_cnt): New function.
973 (case_resize): Drop OLD_CNT argument. Update all callers. Only
974 resize case if its size actually changed.
976 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
977 instead of peeking inside struct case directly.
978 (casefile_append): Ditto.
980 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
982 Get rid of the inlines for the case functions, which made the
983 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
984 the inlines didn't speed up "make check" at all, which is not a
985 perfect benchmark but seems indicative.)
987 * case.c: Remove #ifdef DEBUGGING...#endif around many function
988 definitions. Remove some assertions on nonnull pointers that were
989 redundant with a pointer dereference soon after in the function.
991 (struct case_data): Move definition here from case.h.
995 (case_data_wr): Ditto.
997 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
999 * automake.mk: Add casedeque.h to sources.
1001 * casedeque.h: New file.
1003 * procedure.c: (struct dataset) Change lag_count, lag_head,
1004 lag_queue member into single struct casedeque member. Update all
1005 users to use the casedeque instead.
1008 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1010 * procedure.c: Simplify lagged cases interface. Updated all
1011 clients--well, the only client--to use the simplified interface.
1012 (dataset_n_lag) Removed.
1013 (dataset_set_n_lag) Removed.
1014 (dataset_need_lag) New function.
1016 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1018 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1019 whenever appropriate, but only when the dataset/dictionary is in a
1022 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1024 * dictionary.c dictionary.h : Added callbacks for change of filter and
1025 split variables. Refactored some code to ensure that callbacks get
1026 invoked when appropriate.
1028 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1029 replace_dict callback occurs when permanent_dict replaces the current
1032 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1034 * dictionary.c dictionary.h : Added callback for when the weight
1035 variable of a dictionary changes.
1037 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1039 * dictionary.c dictionary.h : Added replace_source and replace_dict
1040 callbacks, and functions to deal with them.
1042 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1044 Simplify missing value handling.
1046 * missing-values.h (enum mv_class): New type.
1047 (enum mv_type): Moved definition into missing-values.c and renamed
1048 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1049 enums. Updated all uses.
1050 (struct missing_values): Changed type of `type' from `enum
1051 mv_type' to `int' because the definition is no longer exposed.
1053 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1054 parameter. Update all callers.
1055 (mv_is_num_missing): Ditto.
1056 (mv_is_str_missing): Ditto.
1057 (mv_is_value_user_missing): Removed. Changed callers to use
1058 mv_is_value_missing.
1059 (mv_is_num_user_missing): Removed. Changed callers to use
1061 (mv_is_str_user_missing): Removed. Changed callers to use
1063 (mv_is_value_system_missing): Removed. Changed callers to use
1064 mv_is_value_missing.
1065 (mv_set_type): Removed. Changed callers to use mv_clear.
1066 (mv_clear): New function.
1068 * variable.c (var_is_value_missing): Add new enum mv_class
1069 parameter. Update all callers.
1070 (var_is_num_missing): Ditto.
1071 (var_is_str_missing): Ditto.
1072 (var_is_value_user_missing): Removed. Changed callers to use
1073 var_is_value_missing.
1074 (var_is_num_user_missing): Removed. Changed callers to use
1076 (var_is_str_user_missing): Removed. Changed callers to use
1078 (var_is_value_system_missing): Removed. Changed callers to use
1079 var_is_value_missing.
1081 * casefilter.c (struct casefilter): Use enum mv_class in place of
1083 (casefilter_variable_missing): Adapt to new member.
1084 (casefilter_create): Change signature to take enum mv_class,
1087 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1089 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1091 * case-sink.c case-sink.h procedure.c procedure.h
1092 storage-stream.c: Now uses the factory.
1094 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1096 Make it possible to pull cases from the active file with a
1097 function call, instead of requiring indirection through a callback
1100 * case-source.h (struct case_source_class): Change ->read function
1101 to return a single case, instead of calling a callback function
1102 for each case. Change ->destroy function to return an error
1105 * case-source.c (free_case_source): Pass along the value returned
1106 by the case_source ->destroy function.
1108 * procedure.c (struct write_case_data): Removed.
1109 (struct dataset): Added some members to track procedure state.
1110 (procedure): Optimize the trivial case at this level.
1111 (internal_procedure): Re-implement in terms of proc_open,
1112 proc_read, proc_close.
1113 (proc_open) New function.
1114 (proc_read) New function.
1115 (proc_close) New function.
1116 (write_case) Moved into proc_read.
1117 (close_active_file) Moved closing of data source into proc_close.
1119 * storage-source.c: Rewrote to conform with modified
1120 case_source_class interface.
1122 * transformations.c (trns_chain_execute): Added argument to allow
1123 starting execution from an arbitrary transformation. Updated
1126 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1128 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1130 * sys-file-reader.c (read_display_parameters): Don't assume that
1131 MEASURE_* and ALIGN_* have the same values found in system files.
1133 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1135 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1136 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1137 the value of n_MEASURES, which was off by 1 (at least from my
1140 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1142 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1143 callbacks which are invoked when the dictionary or its
1144 variables are changed.
1146 * missing-values.c missing-values.h value-labels.c: Tidied up
1147 consistency checks, and made some of them return false
1148 instead of assert-failing.
1150 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1152 * calendar.c (calendar_days_in_month): New function.
1154 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1156 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1157 the value label's width.
1159 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1161 * sfm-private.h: Move contents into sys-file-writer.c, which is
1162 the only remaining user. Removed Borland C++-specific directives.
1164 * sys-file-reader.c: Clean up and rewrite entire file. The
1165 rewritten version is simpler and better abstracted, and should be
1166 easier to maintain and extend. It avoids using structures to read
1167 file data, which is prone to padding variations among compilers.
1168 It should also handle non-IEEE 754 system files, although I
1169 haven't been able to find any. It has been tested against many
1170 .sav files obtained from the Web and found to produce the same
1171 results as the earlier version of the code, or in some cases
1172 improved results. It is more tolerant of format variations found
1175 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1176 member, putting an enum integer_format in its place. New member
1177 `float_format'. Changed `compressed' member to type bool.
1179 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1181 * dictionary.c (dict_delete_consecutive_vars): New function.
1183 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1185 * file-name.c (fn_search_path): Remove prefix arg that was unused
1186 by any caller. Updated all callers.
1188 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1190 * format.c (fmt_dollar_template): Use user's decimal point
1191 character. Add assertion.
1193 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1195 * format.c (fmt_dollar_template): New function, based on
1196 dollar_format_template from var-type-dialog.c.
1198 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1200 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1203 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1205 Finish converting struct variable to an opaque type. In this
1206 phase, we add remaining setter and getter functions, convert the
1207 remaining PSPP code to use them, and do a bunch of cleanup. The
1208 resulting changes are pervasive but mostly trivial, and only the
1209 notable changes are logged.
1211 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1214 * case.c (case_data): Renamed case_data_idx.
1215 (case_num): Renamed case_num_idx.
1216 (case_str): Renamed case_str_idx.
1217 (case_data_rw): Renamed case_data_rw_idx.
1219 * case.h (case_data): New function with old name and an interface
1220 that takes a variable instead of an index, which is easier to
1221 use. Updated all callers to use the new interface, or to use the
1222 new *_idx function (see above).
1225 (case_data_rw): Ditto.
1227 * category.c (cat_stored_values_destroy): Changed interface to
1228 take a struct cat_vals * instead of a struct variable *.
1230 * dictionary.c (dict_clone): Use new vector_clone function.
1231 (dict_clear) Use new var_destroy function.
1232 (add_var) New function.
1233 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1234 (dict_create_var_assert) Rewrite in terms of add_var.
1235 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1236 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1237 (dict_lookup_var) Use new var_create, var_destroy functions.
1238 (dict_contains_var) Rewrite in terms of new vardict functionality.
1239 (set_var_dict_index) New function.
1240 (set_var_case_index) New function.
1241 (reindex_vars) New function.
1242 (dict_delete_var) Rewrite in terms of new vardict functionality.
1243 (dict_reorder_var) Ditto.
1244 (dict_reorder_vars) Ditto.
1245 (rename_var) New function.
1246 (dict_rename_var) Use rename_var.
1247 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1248 (dict_get_compacted_idx_to_fv) Rename
1249 dict_get_compacted_dict_index_to_case_index, update callers.
1250 (dict_create_vector) Use new vector_create function.
1251 (dict_clear_vectors) Use new vector_destroy function.
1252 (set_var_short_name_suffix) Move here from variable.c, renamed
1253 from var_set_short_name_suffix, make static, update caller.
1255 * sys-file-private.c: New file.
1256 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1257 width_to_bytes, update callers.
1259 * sys-file-private.h: New file. Later it will supplant
1260 sfm-private.h; for now it supplements it.
1261 (macro MIN_VERY_LONG_STRING) New macro.
1262 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1264 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1265 MAX_LONG_STRING was used before.
1267 * sys-file-writer.c: Ditto.
1269 * value-labels.c: Change the paradigm here to be that a null
1270 pointer is OK for a struct val_labs * in most cases; it just
1271 represents an empty set of value labels.
1272 (val_labs_copy) A copy of a null set is a null set.
1273 (val_labs_count) A null set has 0 labels.
1274 (val_labs_replace) Change return type to void. Rewrite for
1276 (val_labs_find) A null set does not contain the value.
1277 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1278 transposed argument order, updated all callers.
1280 * value.c: New file.
1281 (value_dup) Moved here from variable.c.
1282 (compare_values) Ditto.
1285 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1286 (macro MAX_LONG_STRING) Removed, because it was only interesting
1287 for system files, not for general code.
1288 (macro MAX_VERY_LONG_STRING) Ditto.
1289 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1290 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1292 * vardict.h: New file, for an interface between variables and
1295 * variable.c: A lot of functions were moved around, for better
1297 (struct variable) Move definition here, from variable.h.
1298 (var_type_adj) Removed--makes i18n hard.
1299 (var_type_noun) Ditto.
1300 (var_create) New function.
1301 (var_clone) New function.
1302 (var_destroy) New function.
1303 (var_set_name) Assert that variable is not in a dictionary.
1304 (compare_var_names) Rename compare_vars_by_name and fix a couple
1305 of callers who thought the args were strings.
1306 (hash_var_name) Rename hash_var_by_name.
1307 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1308 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1309 (var_is_very_long_string) Removed, because it was only interesting
1310 to system file code.
1311 (var_set_missing_values) Allow the argument to be the wrong width,
1312 as long as we can resize it. Simplify callers who were doing the
1313 resizing themselves.
1314 (var_get_value_labels) New function.
1315 (var_has_value_labels) New function.
1316 (var_set_value_labels) New function.
1317 (alloc_value_labels) New function.
1318 (var_add_value_label) New function.
1319 (var_replace_value_label) New function.
1320 (var_clear_value_labels) New function.
1321 (var_lookup_value_label) New function.
1322 (var_get_value_name) Moved here from variable.c, renamed from
1323 var_get_value_name, transposed argument order, updated all
1325 (var_to_string) Moved here, from variable-label.c.
1326 (var_set_leave) New function.
1327 (var_get_leave) New function.
1328 (var_must_leave) New function.
1329 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1330 set_var_short_name_suffix.
1331 (var_get_dict_index) New function.
1332 (var_get_case_index) New function.
1333 (var_get_obs_vals) New function.
1334 (var_set_obs_vals) New function.
1335 (var_has_obs_vals) New function.
1336 (var_get_vardict) New function.
1337 (var_set_vardict) New function.
1338 (var_has_vardict) New function.
1339 (var_clear_vardict) New function.
1340 (value_dup) Moved to value.c.
1341 (compare_values) Ditto.
1344 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1345 (enum ALPHA) Rename VAR_STRING, update all users.
1347 * vector.c: New file.
1348 (struct vector) Moved here, from variable.h.
1349 (check_widths) New function.
1350 (vector_create) New function.
1351 (vector_clone) New function.
1352 (vector_destroy) New function.
1353 (vector_get_name) New function.
1354 (vector_get_var) New function.
1355 (vector_get_var_cnt) New function.
1356 (compare_vector_ptrs_by_name) New function.
1358 * vector.h: New file.
1360 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1362 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1363 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1366 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1368 * value-labels.c (destroy_atoms): New function.
1369 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1371 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1373 Thanks to Jason Stover for pointing out this problem.
1375 * data-out.c (output_number): Use gsl_finite from GSL, which is
1376 portable, instead of isfinite, which is not.
1379 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1381 * variable.c variable.h (value_dup): New function.
1383 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1385 Start converting struct variable to an opaque type. In this
1386 phase, we add a bunch of setter and getter functions and convert
1387 most of the PSPP code to use them. The resulting changes are
1388 pervasive but mostly trivial, and only the notable changes are
1391 * format.c (fmt_equal): New function.
1393 * variable.c (var_type_is_valid): New function.
1394 (measure_is_valid) Moved here, from format.c.
1395 (alignment_is_valid) Moved here, from format.c.
1396 (var_get_name) New function.
1397 (var_set_name) New function.
1398 (width_to_type) New function.
1399 (var_get_type) New function.
1400 (var_get_width) New function.
1401 (var_set_width) New function.
1402 (var_is_numeric) New function.
1403 (var_is_alpha) New function.
1404 (var_is_short_string) New function.
1405 (var_is_long_string) New function.
1406 (var_is_very_long_string) New function.
1407 (var_get_missing_values) New function.
1408 (var_set_missing_values) New function.
1409 (var_clear_missing_values) New function.
1410 (var_has_missing_values) New function.
1411 (var_is_value_missing) New function.
1412 (var_is_num_missing) New function.
1413 (var_is_str_missing) New function.
1414 (var_is_value_user_missing) New function.
1415 (var_is_num_user_missing) New function.
1416 (var_is_str_user_missing) New function.
1417 (var_is_value_system_missing) New function.
1418 (var_get_print_format) New function.
1419 (var_set_print_format) New function.
1420 (var_get_write_format) New function.
1421 (var_set_write_format) New function.
1422 (var_set_both_formats) New function.
1423 (var_get_label) New function.
1424 (var_set_label) New function.
1425 (var_clear_label) New function.
1426 (var_has_label) New function.
1427 (var_get_measure) New function.
1428 (var_set_measure) New function.
1429 (var_get_display_width) New function.
1430 (var_set_display_width) New function.
1431 (var_get_alignment) New function.
1432 (var_set_alignment) New function.
1433 (var_get_value_cnt) New function.
1434 (var_get_leave) New function.
1435 (var_get_short_name) New function.
1437 * variable.h: (struct variable) Removed "type" and "nv" members;
1438 they are now computed from "width" where needed.
1440 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1442 * missing-values.c (mv_resize): Don't write beyond end of the
1443 allocated buffer when resizing a long string.
1445 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1447 Clean up identifier code: don't require identifier enumerations to
1448 be in a particular order; make better use of string library;
1449 expose less of the internals.
1451 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1452 change interface. Updated all callers.
1453 (lex_id_match) Change interface to use struct substring, update
1455 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1456 (global array keywords[]) Make static, change form. Update all
1457 users to use lex_id_name instead.
1458 (lex_is_keyword) New function.
1459 (lex_id_to_token) Change interface to use struct substring, update
1461 (lex_id_name) New function.
1463 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1464 lex_is_keyword instead.
1465 (T_LAST_KEYWORD) Removed.
1466 (T_N_KEYWORDS) Removed.
1468 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1470 * format.c: (fmt_date_template) Distinguish characters for which a
1471 space is output and any date delimiter is allowed on input, from
1472 those for which a space is output and only a space is allowed on
1473 input. The former is represented by X, the latter by a space.
1474 Also, drop distinction between h and H, changing the former to the
1477 * data-in.c: Completely rewrite internals to conform to SPSS input
1478 formats as closely as possible.
1479 (data_in) Changed external interface by replacing the structure
1480 that was used as a single argument by a set of arguments. Updated
1482 (data_in_finite_line) Removed. Converted all callers to use plain
1484 (data_in_get_integer_format) New function.
1485 (data_in_set_integer_format) New function.
1486 (data_in_get_float_format) New function.
1487 (data_in_set_float_format) New function.
1489 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1490 (struct data_in) Removed.
1492 * data-out.c: (output_date) Drop each component from the input as
1493 it is output, to allow us to drop the distinction between h (a
1494 count of hours) and H (the hour of day) template characters.
1495 Also, handle new X template character.
1496 (output_scientific) Follow more rational rule on when to drop
1497 fraction introduced between SPSS 13 and 15. Updated test case to
1500 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1502 Fix buffer overflow reported by John Darrington.
1504 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1505 realize that DIGITS is a count of nibbles, not of bytes.
1507 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1509 * calendar.c (calendar_offset_to_gregorian) Also return the
1510 year-of-day. Change callers to new interface.
1512 * data-out.c: Completely rewrite internals to conform to SPSS
1513 output formats as completely as possible.
1514 (data_out) Change interface to put input parameters before output
1515 parameters, for consistency with the style I now prefer. Update
1517 (data_out_get_integer_format) New public function.
1518 (data_out_set_integer_format) New public function.
1519 (data_out_get_float_format) New public function.
1520 (data_out_set_float_format) New public function.
1522 * data-out.h: New file. Move prototype for data_out here, from
1525 * format.c: (fmt_step_width) Use equality comparison instead of
1526 bitwise and, for clarity.
1527 (fmt_is_string) Ditto.
1528 (fmt_input_to_output) Fix categories that are translated to F
1531 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1533 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1534 casefile-private.h: Added casefilter to assist commands with missing
1537 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1539 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1541 * settings.c: (route_errors_to_terminal) New variable.
1542 (route_errors_to_listing) New variable.
1543 (get_error_routing_to_terminal) New function.
1544 (set_error_routing_to_terminal) New function.
1545 (get_error_routing_to_listing) New function.
1546 (set_error_routing_to_listing) New function.
1548 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1550 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1552 * format.c: Completely rewrite, to achieve better abstraction.
1553 Rewrite all references to formats in other files.
1555 * format.def: Rewrite and reorganize.
1557 * settings.c: Move everything related to custom currency formats
1558 into format.[ch], changing them in form, so as to group related
1559 code and definitions better. Changed all references to use the
1561 (static var decimal) Removed.
1562 (static var grouping) Removed.
1563 (static var cc) Removed.
1564 (get_decimal) Removed.
1565 (set_decimal) Removed.
1566 (get_grouping) Removed.
1567 (set_grouping) Removed.
1571 * settings.h: (macro CC_CNT) Removed.
1572 (macro CC_WIDTH) Removed.
1573 (struct custom_currency) Removed.
1575 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1577 * data-in.c (data_in): Use switch statement instead of table, to
1578 avoid dependence on the order of the FMT_* enums.
1580 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1582 * data-out.c: (num_to_string) Removed, because it was dead code.
1584 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1586 * data-in.c (parse_trailer): Fix error message.
1588 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1590 * format.c (fmt_is_binary): New function.
1592 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1594 * procedure.c procedure.h: Encapsulated the static data into a single
1597 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1599 * casefile.c (casereader_read_xfer): Always initialize the case,
1600 even on an error condition.
1602 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1604 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1605 entire function useless.
1607 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1609 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1610 casereader method casereader_clone.
1612 * procedure.c pransformations.h: Introduced new type casenum_t
1614 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1616 * variable.c: (width_to_bytes) Rephrase code for clarify.
1618 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1620 * format.c: (fmt_type_from_string) New function.
1621 (fmt_to_string) Include decimals in output if the format has
1622 decimals, even if the format type does not. This way, we can
1623 accurately reproduce incorrect formats in user output.
1624 (check_common_specifier) Make the check for a bad format type an
1625 assertion, so we get bug reports if they show up. Fix message.
1626 Check for decimal places with a format type that doesn't allow
1628 (check_input_specifier) Remove check for FMT_X, which has been
1630 (check_output_specifier) Ditto.
1632 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1634 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1635 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1636 (enum fmt_parse_flags) Removed.
1638 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1640 * casefile.c casefile.h: Converted to an abstract base class.
1641 * casefile-private.h fastfile.c fastfile.h: New files.
1642 * automake.mk procedure.c scratch-writer.c storage-stream.c
1644 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1646 * procedure.c (internal_procedure): Create sink_case with only as
1647 many values as the compacted dictionary.
1649 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1651 Remove "debugging" code that caused plenty of false positives and
1654 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1656 * case.c: Remove all references to `this' member.
1658 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1660 Fix link error noted by Jason Stover.
1662 * storage-stream.c: Include <assert.h>.
1664 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1666 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1667 ALL) and additional underlying system file issues.
1669 Thanks to John Darrington for review.
1671 First problem: var_hash points to variables not owned by the
1672 sys-file-reader, which the caller may free or modify. Use an
1673 array of sfm_vars instead, as done earlier (e.g. CVS version
1676 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1677 members and remove all code that references it. Add vars, var_cnt
1678 members. Remove fix_specials member, which was unused.
1679 (struct sfm_var) Remove name member, which was unused.
1680 (sfm_close_reader) Free vars member instead of var_hash.
1681 (compare_var_shortnames) Removed.
1682 (hash_var_shortname) Removed.
1683 (sfm_open_reader) Fill out vars array.
1684 (compare_var_index) Removed.
1685 (sfm_read_case) Use vars instead of var_hash.
1687 Second problem: we're confused about when we actually have very
1688 long strings, causing us to choose incorrectly between slow path
1689 and fast path in sfm_read_case.
1691 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1692 have very long strings, not when we have long variable names,
1693 which is an unrelated feature.
1695 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1697 * variable.h: Move var_set and variable parsing declarations to
1698 new header, src/language/lexer/variable-parser.h. Modified lots
1699 of files to include the new header.
1701 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1703 * value-labels.c (value_to_string): When there's no value label,
1704 format the variable according to its print format, instead of
1705 always effectively using A or F format.
1707 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1709 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1712 * format.c: Removed tortological assertion.
1714 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1716 Reform string library.
1718 * file-name.c (fn_interp_vars): Change interface to take a
1719 substring as input. Updated all users.
1721 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1723 * format.c (measure_is_valid): Really return false when m >=
1726 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1728 Implement random access to casefiles, for use in GUI.
1730 * casefile.c: (struct casereader) Add `random', `file_ofs',
1731 `buffer_ofs' members.
1732 (casefile_get_random_reader) New function.
1733 (read_open_file) Break part into new function
1734 seek_and_fill_buffer().
1735 (fill_buffer) Update buffer_ofs, file_ofs.
1736 (casereader_seek) New function.
1738 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1740 * settings.c: Added call to i18n{done, init}.
1742 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1744 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1747 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1749 * casefile.c: Convert many uses of `int' to `bool'.
1751 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1753 * transformations.c (trns_chain_destroy): Destroy chain's trns
1754 member, to fix memory leak.
1756 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1758 * storage-stream.c (storage_source_decapsulate): Destroy case
1759 source to fix memory leak.
1761 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1763 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1764 case passed as argument instead of initializing the argument as a
1765 case. Fixes memory leak that showed up in
1766 tests/command/aggregate.sh with scratch files.
1768 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1770 * procedure.c (proc_done): Destroy default_dict, to fix memory
1773 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1775 Simplify procedure_with_splits().
1777 * procedure.c (struct split_aux_data): Removed case_count member.
1778 (procedure_with_splits) Don't initialize case_count.
1779 (split_procedure_case_func) Check whether prev_case is null
1780 instead of case_count.
1781 (split_procedure_end_func) Ditto.
1783 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1785 * case.c (case_move): Do nothing if dst and src are the same
1787 (case_try_create) Merge two similar cases.
1788 (case_copy) Unshare only if data must be actually copied.
1790 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1792 * data-in.c data-out.c dictionary.c sys-file-reader.c
1793 sys-file-writer.c variable.c variable.h: Reworked very long string
1794 support for better encapsulation.
1796 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1798 * value-labels.c (val_labs_can_set_width): New function.
1799 (val_labs_set_width) Clear labels if increasing width to long
1801 (val_labs_destroy) Remove unneeded test for null.
1803 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1805 * value-labels.h: Remove unneeded dependency on variable.h.
1807 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1809 Get rid of `char *c' member in union value, for cleanliness.
1811 * value.h: (union value) Remove `c' member.
1813 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1815 Make dictionary compacting functions a little more general.
1817 * sys-file-writer.c (sfm_open_writer): Use
1818 dict_compacting_would_change().
1819 (does_dict_need_translation) Removed.
1821 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1823 Make dictionary compacting functions a little more general.
1825 * dictionary.c (dict_needs_compaction): Rename
1826 dict_compacting_would_shrink(). Update all callers.
1827 (dict_compacting_would_change) New function.
1829 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1831 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1832 inverted return value (!).
1834 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1836 Continue reforming procedure execution.
1838 * procedure.c: Search and replace "vfm" by "proc". Notably:
1839 (static var vfm_source) Rename proc_source. Update all
1841 (static var vfm_sink) Rename proc_sink. Update all references.
1843 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1845 Continue reforming procedure execution. In this phase, remove
1846 PROCESS IF, which was deprecated anyway and can be easily
1847 simulated with TEMPORARY followed by SELECT IF.
1849 * procedure.c: (open_active_file) Don't call
1850 add_process_if_trns().
1851 (discard_variables) Get rid of redundant call to
1852 proc_cancel_all_transformations().
1853 (add_process_if_trns) Removed.
1854 (process_if_trns_proc) Removed.
1855 (process_if_trns_free) Removed.
1857 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1859 Continue reforming procedure execution. In this phase, add
1860 `const' to the case passed to procedure()'s callback.
1862 Updated all users of procedure() as well.
1864 * procedure.c: (struct write_case_data) Add "const" to ccase
1865 parameter for case_func member.
1866 (procedure) Add "const" to ccase parameter for proc_func
1868 (multipass_case_func) Make ccase parameter const.
1869 (internal_procedure) Add "const" to ccase parameter for case_func
1871 (split_procedure_case_func) Make ccase parameter const.
1872 (multipass_split_case_func) Make ccase parameter const.
1874 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
1876 Continue reforming procedure execution. In this phase, get rid of
1877 the output code for SPLIT FILE groups in procedure.c, which really
1878 shouldn't be doing any output. Move it into the individual
1879 procedures instead. This also adds some flexibility.
1881 Updated many users of procedure_with_splits() and
1882 multipass_procedure_with_splits() to call
1883 output_split_file_values() and to deal with increased use of
1886 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
1887 parameter to begin_func member.
1888 (procedure_with_splits) Add "const struct ccase *" parameter to
1889 begin_func parameter. Make ccase parameter const in proc_func
1891 (split_procedure_case_func) Don't dump split file group. Pass
1893 (dump_splits) Moved to language/dictionary/split-file.c as
1894 output_split_file_values().
1895 (struct multipass_split_aux_data) Add "const struct ccase *"
1896 parameter to split_func member.
1897 (multipass_procedure_with_splits) Add "const struct ccase *"
1898 parameter to split_func parameter.
1899 (multipass_split_case_func) Save new SPLIT FILE case before
1901 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
1903 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
1905 Continue reforming procedure execution. Change
1906 internal_procedure() so that it calls open_active_file() and
1907 close_active_file(), which isolates most of the actual procedure
1910 * procedure.c: (struct write_case_data) Rename `proc_func' member
1911 to `case_func' and update all references.
1912 (procedure) Rewrite as one-line wrapper around
1913 internal_procedure().
1914 (struct multipass_aux_data) New.
1915 (multipass_callback) Renamed multipass_case_func(). Use struct
1916 multipass_aux_data as auxiliary data.
1917 (multipass_end_func) New function.
1918 (multipass_procedure) Rewrite as wrapper for internal_procedure()
1919 that uses multipass_case_func, multipass_end_func.
1920 (internal_procedure) Add `end_func' argument. Move optimization
1921 of trivial case in here. Move call to open_active_file() and
1922 close_active_file() in here. Now assert that vfm_source is
1924 (procedure_with_splits_callback) Rename
1925 split_procedure_case_func().
1926 (split_procedure_end_func) New function.
1927 (multipass_split_callback) Rename multipass_split_case_func.
1928 (multipass_split_end_func) New function.
1929 (discard_variables) No need to test for nonnull vfm_source.
1931 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
1933 Continue reforming procedure execution. Get rid of unused member.
1935 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
1937 (write_case) Don't increment cases_analyzed.
1939 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
1941 Continue reforming procedure execution. In this phase, move
1942 procedure.c and procedure.h from src to src/data. Update
1943 makefiles and #includes accordingly.
1945 * procedure.c: Moved here from src/.
1947 * procedure.h: Moved here from src/.
1949 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
1951 Continue reforming procedure execution. In this phase, get rid of
1952 many global variables, consolidating procedure execution in
1953 procedure.c. Encapsulate transformations in new "struct
1954 trns_chain". Also, change implementation of N OF CASES, FILTER,
1955 and PROCESS IF from special cases to transformations.
1957 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
1960 * dictionary.c: (global variable default_dict) Move to
1963 * variable.h: (TRNS_*) Move to transformations.h.
1964 (struct transformation) Move to transformations.c.
1966 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1968 * sys-file-reader.c: Fixed invalid read problems.
1970 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
1972 * storage-stream.c: Add missing function comments.
1974 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
1976 Continue reforming procedure execution. In this phase, add some
1977 new, needed functionality to storage-stream.
1979 * storage-stream.c: (storage_source_decapsulate) New function.
1981 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
1983 * variable.c (width_to_bytes): Declarations must precede
1984 statements for C90 compliance.
1986 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
1988 * data-out.c, data-in.c, variable.c, variable.h: New functions
1989 copy_mangle and copy_demangle for reading/writing cases; emulates the
1990 way SPSS deals with strings > 255 bytes.
1992 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
1993 subtype 14 needed for strings longer than 255 bytes.
1995 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
1996 instead of literal values. Also fixed some constness issues.
1998 * format.h: Constness
2000 * sfm-private.h: Renamed the case_size identifier, since I discovered
2001 that SPSS's respect for this variable is very nominal.
2003 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2005 Change case limit type from int to size_t.
2007 * dictionary.c: (struct dictionary) Change type of case_limit
2009 (dict_get_case_limit) Change return type.
2010 (dict_set_case_limit) Change parameter type.
2012 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2014 * variable.h: (struct variable) Rename `reinit' member as `leave'
2015 and invert sense. Fix up all references.
2017 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2019 Continue reforming procedure execution. In this phase, break
2020 procedure.c into multiple files.
2022 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2024 * case-sink.c: New file.
2026 * case-sink.h: New file.
2028 * case-source.c: New file.
2030 * case-source.h: New file.
2032 * storage-stream.c: New file.
2034 * storage-stream.h: New file.
2036 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2038 * variable.h: (struct variable) Remove `init' member and all
2039 references to it from other files. It was initialized in several
2040 places, but nothing really ever used it for anything worthwhile.
2041 Thanks to Jason Stover for pointing out how confusing this
2044 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2046 Continue reforming error message support. In this phase, get rid
2047 of message "titles" and put the message text in `struct error'.
2048 Now `struct error' encapsulates a message more properly.
2050 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2051 Format message ourselves.
2053 * data-in.c: (vdls_error) Ditto.
2055 * por-file-reader.c: (error) Ditto.
2057 * sys-file-reader.c: (corrupt_msg) Ditto.
2059 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2061 GNU standards require "file name" instead of "filename" in
2062 documentation. It's nice for our code to follow the convention
2065 * casefile.c: (struct casefile) Rename `filename' member to
2066 `file_name'. Updated all references.
2068 * file-name.c: [!unix] (struct file_identity) Rename
2069 normalized_filename member to normalized_file_name. Updated all
2072 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2074 We don't really support anything but Unix-like environments well,
2075 so we might as well de-obfuscate by writing directory and path
2076 separators explicitly.
2078 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2080 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2081 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2083 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2086 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2088 GNU standards require "file name" instead of "filename" in
2089 documentation. It's nice for our code to follow the convention
2092 * filename.c: Rename to file-name.c.
2094 * filename.h: Rename to file-name.h. Update all inclusions.
2095 Update header guards.
2097 * automake.mk: Update file names.
2099 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2101 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2103 (fn_basename) Removed (dead code).
2104 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2105 (fn_special_p) Renamed fn_is_special(), all references updated.
2106 (fn_exists_p) Renamed fn_exists(), all references updated.
2108 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2110 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2111 Also, now it only tilde-expands file names, not paths.
2112 (fn_search_path) Tilde-expand one directory at a time.
2114 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2116 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2117 get rid of non-Unixlike version of the code, which has probably
2119 (fn_prepend_dir) Removed (dead code).
2121 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2122 (macro PATH_SEPARATOR_STRING) New.
2123 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2125 Continue reforming error message support. In this phase, we get
2126 rid of VM() and the other msg() support for "verbosity", replacing
2127 it by a new function verbose_msg().
2129 * filename.c: (fn_search_path) Use verbose_msg() instead of
2132 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2134 * sfm-private.h: Get rid of #defines after #error, which makes no
2137 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2139 Get rid of our own int32 type in favor of the standard int32_t
2142 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2145 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2147 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2149 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2151 Remove ill-considered file routines that are no longer used.
2153 * filename.c: (fn_open_ext) Removed.
2154 (fn_close_ext) Removed.
2156 * filename.h: (struct file_ext) Removed.
2158 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2160 * variable.c (var_is_valid_name): Move declarations before code
2163 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2165 * filename.ch (fn_interp_vars): Fixed small buglet.
2167 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2169 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2170 so as to modify the string inline. Thus makeing it easier to
2171 destroy the results when no longer needed.
2173 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2175 * category.c (cat_stored_values_destroy): Fixed memory leak.
2177 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2179 Add some missing frees. Thanks to John Darrington for reporting
2182 * any-writer.c (any_writer_close): Free writer.
2184 * any-reader.c (any_reader_close): Free reader.
2186 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2188 * por-file-reader.c: (error) Mark as NO_RETURN.
2190 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2192 * settings.c: Changed default value of scompress to true.
2194 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2196 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2198 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2201 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2202 here from pref.h.orig.
2204 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2206 * sys-file-reader.c: Fixed bug reading compressed files.
2208 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2210 * Numerous renames. See src/ChangeLog for details.
2212 * Moved files from src directory