1 2007-11-08 Ben Pfaff <blp@gnu.org>
3 * data-in.c: Make formatted data parsing locale-independent.
4 (parse_number): Use c_strtod instead of strtod, to avoid
5 locale-specific behavior.
8 2007-11-06 Ben Pfaff <blp@gnu.org>
10 Patch #6256: add support for binary, 360 file formats. Reviewed
13 * data-in.c (struct data_in): Add `encoding' member.
14 (data_in): Add `encoding' parameter, and re-encode the data passed
15 in where appropriate. Update all callers to pass it in.
16 (parse_A): Implement EBCDIC recoding wart described in manual.
17 (parse_AHEX): Implement EBCDIC recoding.
19 * data-out.c (data_out_legacy): New function.
20 (data_out): Make into a wrapper around data_out_legacy.
22 * file-handle-def.c (struct file_handle): New member `encoding'.
23 (fh_create_file): Set encoding.
24 (fh_default_properties): Set default encoding.
25 (fh_get_legacy_encoding): New function.
27 * file-handle-def.h (enum fh_mode): New modes FH_MODE_FIXED
28 (that replaces FH_MODE_BINARY), FH_MODE_VARIABLE,
29 FH_MODE_360_VARIABLE, FH_MODE_360_SPANNED.
30 (struct fh_properties): New member `encoding'.
32 2007-11-05 Ben Pfaff <blp@gnu.org>
34 Patch #6258. Reviewed by John Darrington.
36 * file-handle-def.c (fh_lock): Add comment that TYPE should be
37 marked with N_() in the caller. Added these markings to each
38 caller too. Should make i18n easier.
39 Suggested by Chusslove Illich <caslav.ilic@gmx.net>.
41 2007-11-03 Ben Pfaff <blp@gnu.org>
43 Allow output files to overwrite input files (bug #21280).
45 * file-handle-def.c: Separate locking of files for input and for
46 output, to allow both to take place at once. Also, distinguish a
47 file handle from the identity of the underlying file, because the
48 identity of a file changes over time and the file handle can't
49 represent two different identities.
50 (struct file_handle): Remove `next', `open_cnt', `deleted',
51 `type', `open_mode', `aux', `identity' members. Change `id' from
52 char array to char *. Add `ref_cnt' member.
53 (file_handle_from_ll) New function.
54 (file_handles) Removed.
55 (named_handles) New variable.
56 (fh_init) Initialize named_handles.
57 (fh_done) Remove name from all named_handles.
58 (free_handle) Updated for modified struct file_handle.
59 (unname_handle) New function.
60 (fh_ref) New function.
61 (fh_from_file_name) Removed.
62 (fh_unref) New function.
63 (fh_unname) New function.
64 (fh_from_id) Rewritten.
65 (create_handle) Updated for modified struct file_handle.
66 (fh_inline_file) Increment the handle's ref count.
67 (fh_create_file) Updated for modified struct file_handle.
68 (fh_create_scratch) Ditto.
74 (fh_get_id) Updated for modified struct file_handle.
75 (fh_get_default_handle) Increment the handle's ref count.
76 (fh_set_default_handle) Handle ref counts.
77 (struct fh_lock) New structure.
78 (locks) New static var.
79 (fh_lock) New function.
80 (fh_unlock) New function.
81 (fh_lock_get_aux) New function.
82 (fh_lock_set_aux) New function.
83 (fh_is_locked) New function.
84 (make_key) New function.
85 (free_key) New function.
86 (compare_fh_locks) New function.
87 (hash_fh_lock) New function.
89 * file-handle-def.h (enum fh_access) New enum.
91 * file-name.c: Made file_identity the same in all supported
93 (struct file_identity): New `name' member.
94 (fn_get_identity): For a file that doesn't exist, get the
95 dev/inode of its directory plus its name. If even the directory
96 doesn't exist, just use its name. Merge the Windows
97 implementation into the Unix one.
98 (fn_compare_file_identities): Rewritten. Merge the Windows
99 implementation into the Unix one.
100 (fn_hash_identity): New function.
102 * make-file.c (struct replace_file): New structure.
103 (all_files): New static var.
104 (replace_file_start): New function.
105 (replace_file_commit): New function.
106 (replace_file_abort): New function.
107 (free_replace_file): New function.
108 (unlink_replace_files): New function.
110 * por-file-reader.c (struct pfm_reader): Add `lock' member.
111 (close_reader): Unlock file.
112 (pfm_open_reader): Lock file.
114 * por-file-writer.c (struct pfm_writer): Add fh_lock, replace_file
116 (pfm_open_writer): Lock file and prepare for its replacement.
117 (close_writer): Unlock file.
119 * scratch-handle.h (struct scratch_handle): Add unique_id so that
120 different generations of a scratch file can be distinguished.
122 * scratch-reader.c (scratch_reader_open): Verify that the file is
125 * scratch-writer.c (struct scratch_writer): Add `lock' and `dict',
126 remove scratch_handle member.
127 (scratch_writer_open): Lock handle. Prepare to replace handle
128 data, instead of doing it immediately.
129 (scratch_writer_casewriter_destroy): Replace handle data and
132 * sys-file-reader.c (struct sfm_reader): Add `lock' member.
133 (sfm_open_reader): Lock file.
134 (close_reader): Unlock file.
136 * sys-file-writer.c (struct sfm_writer): Add fh_lock, replace_file
138 (sfm_open_writer): Lock file and prepare for its replacement.
139 (close_writer): Unlock file.
141 2007-11-02 Ben Pfaff <blp@gnu.org>
143 * data-out.c (output_number): Use isfinite (from C99) instead of
144 gsl_isfinite, since we now have universal support for it in
146 (output_infinite): Ditto.
147 [!HAVE_ROUND] (round): Remove definition, since we now have a
148 replacement in gnulib.
150 2007-10-27 John Darrington <john@darrington.wattle.id.au>
152 * dictionary.c: Fixed bug in dict_clone, where the vardict.dict member
153 wasn't initialised correctly. Closes bugs #21330 and 21397
156 2007-10-08 Ben Pfaff <blp@gnu.org>
158 Fix bug #21280. Thanks to John Darrington for review.
160 * file-name.c (create_stream): New function.
162 * por-file-writer.c (pfm_open_writer): Use fh_open to open the
163 file handle before creating the file, to ensure that we don't
164 truncate a file that we're reading. Make code easier to read by
167 * sys-file-write.c (sfm_open_writer): Ditto.
169 2007-10-01 Ben Pfaff <blp@gnu.org>
171 Fix bug #21192. Thanks to John Darrington for review.
173 * casereader.c (casereader_read): Decrement case_cnt before
174 calling the casereader's "read" member function, so that we
175 interact properly with lazy_casereader.
177 * datasheet.c: Add regression test for above bug fix.
178 (clone_datasheet): New function.
179 (lazy_callback): New function.
180 (check_datasheet_casereader): New function.
181 (check_datasheet): Check datasheet contents are reported correctly
182 through an ordinary casereader and a lazy casereader.
183 (clone_model): Use clone_datasheet.
185 2007-09-24 Ben Pfaff <blp@gnu.org>
187 Patch #6210. Reviewed by John Darrington.
188 * settings.c: Drop "long view", which is not needed any longer.
189 (static var long_view): Removed.
190 (force_long_view): Removed.
191 (get_viewwidth): Removed.
192 (init_viewport): Removed long_view reference.
194 2007-09-19 John Darrington <john@darrington.wattle.id.au>
196 * settings.c settings.h: Changed viewport's length and width to be
197 owned by the user interface which uses the data library. This allows
198 better abstraction, and makes dynamically adjustable dimensions easier.
200 2007-09-18 Ben Pfaff <blp@gnu.org>
202 * procedure.c (proc_extract_active_file_data): New function.
204 * lazy-casereader.h: New file.
206 * lazy-casereader.c: New file.
208 * casereader.c (casereader_dynamic_cast): New function.
210 2007-09-14 Ben Pfaff <blp@gnu.org>
212 * dictionary.c (dict_clone): Copy case indexes from cloned
213 dictionary. Fixes bug #21061. Reviewed by John Darrington.
215 2007-09-13 John Darrington <john@darrington.wattle.id.au>
217 * value.c value.h (create_value): New function.
219 2007-09-12 Ben Pfaff <blp@gnu.org>
221 Make it clear that translator casereader and casewriter translate
222 functions are supposed to destroy their input case.
224 * casereader-translator.c (struct casereader_translator): Change
225 input case parameter of translate member function from const
226 struct ccase * to struct ccase *.
227 (casereader_create_translator): Ditto, for translate parameter.
229 * casewriter-translator.c (struct casewriter_translator): Ditto.
230 (casewriter_create_translator): Ditto.
232 2007-08-27 John Darrington <john@darrington.wattle.id.au>
234 * sys-file-reader.c (read_display_parameters): Force display width
235 to 8, if the sys file says 0 (like SPSS does).
237 2007-08-12 Ben Pfaff <blp@gnu.org>
239 * dictionary.c (dict_dump): New function.
241 2007-08-12 Ben Pfaff <blp@gnu.org>
243 Drop dict_compactor in favor of using the new struct case_map.
245 * dictionary.c (struct copy_map): Removed.
246 (struct dict_compactor): Removed.
247 (dict_make_compactor): Removed.
248 (dict_compactor_compact): Removed.
249 (dict_compactor_destroy): Removed.
251 * procedure.c (struct data_set): Change `compactor' member to be a
253 (proc_open): Use case_map_to_compact_dict instead of
255 (proc_casereader_read): Use case_map_execute instead of
256 dict_compactor_compact.
257 (proc_commit): Use case_map_destroy instead of
258 dict_compactor_destroy.
260 * scratch-writer.c (struct scratch_writer): Change `compactor'
261 member to be a struct case_map *.
262 (scratch_writer_open): Use case_map_to_compact_dict instead of
264 (scratch_writer_casewriter_write): Use case_map_execute instead of
265 dict_compactor_compact.
267 2007-08-12 Ben Pfaff <blp@gnu.org>
269 * automake.mk: Add case-map.c, case-map.h.
271 * case-map.c: New file.
273 * case-map.h: New file.
275 2007-08-12 Ben Pfaff <blp@gnu.org>
277 * dictionary.c (dict_compact_values): Don't delete scratch
278 variables as well as compacting case indexes. Update all callers.
279 (dict_get_compacted_value_cnt): Rename dict_count_values and
280 change interface. Update all callers.
281 (dict_get_compacted_value_cnt): Remove.
282 (dict_compacting_would_shrink): Remove.
283 (dict_compacting_would_change): Remove.
284 (dict_make_compactor): Add new parameter. Update all callers.
286 * procedure.c (proc_casereader_read): Use casewriter_get_value_cnt
287 instead of dict_count_values, changing an O(N) operation into
290 2007-08-12 Ben Pfaff <blp@gnu.org>
292 * casereader.c (casereader_read): Don't require cases read by a
293 casereader to be exactly the expected size: as long as they're big
296 2007-08-12 Ben Pfaff <blp@gnu.org>
298 Make casewriters keep track of the number of `union value's in
299 each case. This is useful for two reasons: casewriter_write can
300 then check that the case being written is large enough, and later
301 recipients of the casewriter can determine the size of the case.
303 * casewriter-translator.c (casewriter_create_translator): Add
306 * casewriter.c (struct casewriter): Add value_cnt member.
307 (casewriter_write): Check that the case passed in is big enough.
308 (casewriter_get_value_cnt): New function.
309 (casewriter_create): Add value_cnt parameter.
311 2007-08-09 Ben Pfaff <blp@gnu.org>
313 Fix bug reported by Jason Stover.
314 * settings.c: Move get_termcap_viewport and in particular the
315 #include for <curses.h> to the end of the file. curses.h
316 redefines bool on some systems (e.g. OpenBSD), which causes
317 disagreement between uses of bool before its inclusion and
319 Tested by Jason Stover.
321 2007-07-29 Ben Pfaff <blp@gnu.org>
323 Provisional fix for bug #18692 and bug #20161. Reviewed by John
326 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
327 argument (never "rb" or "wb") because SUSv3 says that only those
328 modes are defined, and glibc in fact rejects other modes.
330 Open portable files with fn_open so that they can be read from
331 pipes. Fix missing fh_close call to go along with fh_open.
332 Report an error if the file close reports an error.
333 * por-file-reader.c (close_reader): New function.
334 (por_file_casereader_destroy): Use close_reader.
335 (pfm_open_reader): Open file with fn_open.
337 2007-07-28 Ben Pfaff <blp@gnu.org>
339 Make PSPP able to read all the portable files I could find on the
340 web. Thanks to John Darrington for review. Bug #17620.
341 * por-file-reader.c (struct pfm_reader): New member `line_length'.
342 (error): Print file offset in hexadecimal.
343 (warning): New function.
344 (advance): Treat lines less than 80 bytes long as padded to 80
346 (pfm_open_reader): Call read_documents if we find an "E" record.
347 (convert_format): Convert invalid formats to the default format
348 instead of aborting reading the file.
349 (read_variables): Rename duplicate variable names instead of
350 aborting reading the file.
351 (read_value_label): Allow string variables of different widths to
352 be assigned value labels in the same record. Replace duplicate
353 value labels instead of aborting.
354 (read_documents): New function.
356 * por-file-writer.c (pfm_open_writer): Call write_documents if the
357 dictionary has documents.
358 (write_documents): New function.
360 2007-07-25 Ben Pfaff <blp@gnu.org>
362 Fix bugs related to bug #17213.
364 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
365 since the former is what config.h has. Include the needed ncurses
367 (static var echo) Rename to `do_echo' because the original name is
368 the same as an ncurses identifier.
369 (get_termcap_viewport) Use error instead of msg.
371 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
372 (fn_close): Don't close stdin, stdout, stderr.
374 2007-07-26 John Darrington <john@darrington.wattle.id.au>
376 * procedure.c procedure.h: Added callbacks which get invoked whenever
377 a dataset's transformation chain changes.
379 2007-07-24 Ben Pfaff <blp@gnu.org>
382 * sys-file-writer.c (write_variable_display_parameters): Use new
383 var_default_display_width function to choose display width of
384 segments after the first one in a given variable.
385 * variable.c (var_create): Use var_default_display_width to pick
386 new variable's display width.
387 (var_default_display_width): New function.
388 Reviewed by John Darrington.
390 2007-07-24 Ben Pfaff <blp@gnu.org>
393 * por-file-writer.c (write_variables): Write weight variable.
394 Reviewed by John Darrington.
396 2007-07-23 Ben Pfaff <blp@gnu.org>
398 Improvements to system file reader and writer.
400 First, move all detailed knowledge of very long strings into
401 sys-file-private.[ch], so that this nasty stuff can be isolated.
403 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
404 (EFFECTIVE_VLS_CHUNK): New macro.
405 (min_int): New function.
406 (max_int): New function.
407 (sfm_width_to_bytes): Rewrite.
408 (sfm_width_to_octs): New function.
409 (sfm_segment_alloc_width): New function.
410 (sfm_segment_alloc_bytes): New function.
411 (sfm_segment_used_bytes): New function.
412 (sfm_segment_offset): New function.
413 (sfm_segment_effective_offset): New function.
414 (sfm_dictionary_to_sfm_vars): New function.
416 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
417 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
418 (struct sfm_var): New structure.
420 Next, improvements to the system file reader.
422 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
423 type casenumber. Added `version_major', `version_minor',
426 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
427 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
428 Change `case_cnt' to type casenumber. Removed `has_vls'.
429 (struct sfm_var): Removed.
430 (sfm_open_reader): Don't warn on wrong case size if the file was
431 written by SPSS 13, which tends to get it wrong. Use
432 sfm_dictionary_to_sfm_vars.
433 (read_header): Always output system file info.
434 (read_variable_record): Simplify code for reading missing values.
435 (read_machine_int32_info): Save version numbers from system file
436 into info struct passed as new argument.
437 (read_long_string_map): Restructured to use new sys-file-private
439 (read_value_labels): Use size_overflow_p.
440 (sys_file_casereader_read): Get rid of distinction between fast
441 and slow paths. Use information provided by sys-file-primate's
442 struct sfm_var to simplify code.
443 (skip_whole_strings): New function.
444 (read_int32): Renamed read_int. Changed return value to int.
446 (read_flt64): Renamed read_float. Changed return value to
447 double. Updated all callers.
448 (int32_to_native): Removed. Changed callers to use
450 (flt64_to_double): Removed. Changed callers to use float_convert.
452 Finally, get rid of int32, flt64 terminology and types in system
453 file writer. The former wasn't very useful since a POSIX "int"
454 can hold the whole range of int32 and we generally didn't have a
455 need for it to be exactly-32-bits, just at-least-32-bits. The
456 latter was inconvenient because we had to assume that it could be
457 different from double and thereby convert special values SYSMIS,
458 HIGHEST, LOWEST to and from it in multiple places. Instead, now
459 we just use "int" and "double" in most places, and do conversions,
460 if necessary, very close to where we do I/O. This change meant
461 that the writer code couldn't represent records in the file as C
462 structs any longer, but that's no great loss. The code actually
463 seems to be more readable without them.
465 Simplify the compression buffering code: only buffer as much as
466 necessary, which is no more than eight 8-byte units at any given
469 * sys-file-writer.c (typedef flt64): Removed.
470 (macro second_lowest_flt64): Removed.
471 (struct sysfile_header): Removed.
472 (struct sysfile_variable): Removed.
473 (struct sfm_writer): Removed `needs_translation', `has_vls',
474 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
475 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
476 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
477 compression buffering by `opcodes', `opcode_cnt', `data',
478 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
479 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
480 data writer functions instead of structures.
481 (calc_oct_idx): New function.
482 (write_header): Use simple data writer functions instead of
484 (write_format_spec): Renamed write_format. New argument.
485 (write_variable_continuation_records): New function.
486 (write_variable): Use simple data writer functions instead of
487 structures. Use write_variable_continuation_records. Write
488 entire very long string instead of requiring caller to understand
490 (write_value_labels): Use simple data writer functions instead of
492 (write_documents): Ditto.
493 (write_variable_display_parameters): Use sys-file-private
494 functions to simplify. Use simple data writer functions instead
496 (write_vls_length_table): Use simple data writer functions instead
498 (write_longvar_table): Ditto.
499 (write_rec_7_34): Break into new functions
500 write_integer_info_record, write_float_info_record. Use simple
501 data writer functions instead of structures.
502 (buf_write): Removed.
503 (append_string_max): Removed.
504 (ensure_buf_space): Removed.
505 (sys_file_casewriter_write): Get rid of the distinction between
506 fast and slow paths, which didn't seem to be too useful. Use new
507 functions write_case_uncompressed, write_case_compressed.
508 (put_instruction): Removed.
509 (put_element): Removed.
510 (write_compressed_data): Removed.
511 (close_writer): Use flush_compressed. Only write case count to
512 system file if it will fit in the field.
513 (write_case_compressed): New function.
514 (write_case_uncompressed): New function.
515 (flush_compressed): New function.
516 (put_cmp_opcode): New function.
517 (put_cmp_number): New function.
518 (write_int): New function.
519 (convert_double_to_output_format): New function.
520 (write_float): New function.
521 (write_value): New function.
522 (write_string): New function.
523 (write_bytes): New function.
524 (write_zeros): New function.
525 (write_spaces): New function.
527 Reviewed by John Darrington.
529 2007-07-22 Ben Pfaff <blp@gnu.org>
531 Don't try to write very long strings to portable files. The
532 format does not support it.
534 * por-file-writer.c (MAX_POR_WIDTH): New macro.
535 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
536 (write_format): Add arg to take width to resize format to.
537 (write_value): Limit width of value written to MAX_POR_WIDTH.
538 (write_variables): Limit width of variable and its output formats
540 Reviewed by John Darrington.
542 2007-07-22 Ben Pfaff <blp@gnu.org>
544 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
545 local variable instead of literal 5.
546 Reviewed by John Darrington.
548 2007-07-22 Ben Pfaff <blp@gnu.org>
550 Fix problems with uniqueness of short names in system files with
551 very long string variables. Now a variable may have multiple
554 * automake.mk (src_data_libdata_a_SOURCES): Add new files
555 short-names.c, short-names.h.
557 * dictionary.c (dict_clone): Clone all the short names.
558 (compare_strings): Move into short-names.c.
559 (hash_strings): Ditto.
560 (set_var_short_name_suffix): Ditto.
561 (dict_assign_short_names): Ditto, rename short_names_assign,
562 change to assign all short names.
564 * por-file-writer.c (write_variables): Use short_names_assign
565 instead of dict_assign_short_names.
567 * short-names.c: New file.
569 * short-names.h: New file.
571 * sys-file-private.c (sfm_width_to_segments): New function.
573 * sys-file-reader.c (read_long_var_name_map): Save and restore all
574 the short names, not just the first one.
576 * sys-file-writer.c (cont_var_name): Removed.
577 (sfm_open_writer): Use short_names_assign instead of
578 dict_assign_short_names. Use unique short names assigned by
579 short_names_assign instead of those generated by cont_var_name.
581 * variable.c (struct variable): Remove `short_name' member,
582 replace by `short_names' and `short_name_cnt'.
583 (var_create) Initialize new members.
584 (var_get_short_name_cnt): New function.
585 (var_get_short_name): Now takes an index argument. Changed most
587 (var_set_short_name): Ditto.
588 (var_clear_short_name): Renamed var_clear_short_names, changed to
589 clear all short names.
591 Reviewed by John Darrington.
593 2007-07-22 Ben Pfaff <blp@gnu.org>
595 * variable.c (var_set_width): Use new var_set_width function.
597 * missing-values.c (mv_n_values): Drop assertion, which was not
600 * format.c (fmt_default_for_width): New function.
601 (fmt_resize): New function.
603 Reviewed by John Darrington.
605 2007-07-18 John Darrington <john@darrington.wattle.id.au>
607 * datasheet.c (datasheet_delete_columns): Added assertion to check
608 we're not deleting outside the range of the sheet.
611 * dictionary.c dictionary.h variable.c: Added the ability for string
612 variables to be resized.
614 * vardict.h: Added some prototypes (moved from dictionary.h) as
615 these should only be called by variable.c
618 2007-07-14 John Darrington <john@darrington.wattle.id.au>
620 * sfm-reader.c: Respect case_cnt field in file header.
622 2007-07-01 John Darrington <john@darrington.wattle.id.au>
624 * transformation.c transformation.h (trns_chain_execute): Changed the
625 signature (Patch #6057)
627 2007-06-10 Ben Pfaff <blp@gnu.org>
629 * casereader-filter.c (casereader_filter_destroy): Make sure to
630 write all the remaining excluded cases to the casewriter, if any.
632 * caseinit.c (init_list_destroy): Rewrite.
633 (init_list_clear): Ditto.
635 * casegrouper.c (casegrouper_get_next_group): Always set *reader
636 to null when returning false.
638 2007-06-06 Ben Pfaff <blp@gnu.org>
640 Actually implement the new procedure code and adapt all of its
641 clients to match. Also adapt all of the other case sources and
642 sinks in the tree and their clients to use the
643 casereader/casewriter infrastructure.
645 * automake.mk: Add and remove files.
647 * any-reader.c: Change into a casereader.
648 * por-file-reader.c: Ditto.
649 * scratch-reader.c: Ditto.
650 * sys-file-reader.c: Ditto.
652 * any-writer.c: Change into a casewriter.
653 * por-file-writer.c: Ditto.
654 * scratch-writer.c: Ditto.
655 * sys-file-writer.c: Ditto.
657 * procedure.c: Change to use casereader, casewriter, caseinit, and
658 other new infrastructure.
660 * scratch-handle.c: Adapt to new infrastructure.
662 * case-sink.c: Removed, now dead code.
663 * case-sink.h: Ditto.
664 * case-source.c: Ditto.
665 * case-source.h: Ditto.
666 * casefile-factory.c: Ditto.
667 * casefile-private.h: Ditto.
670 * casefilter.c: Ditto.
671 * casefilter.h: Ditto.
674 * fastfile-factory.c: Ditto.
675 * fastfile-factory.h: Ditto.
676 * storage-stream.c: Ditto.
677 * storage-stream.h: Ditto.
679 2007-06-06 Ben Pfaff <blp@gnu.org>
683 * automake.mk: Add new files.
685 * datasheet.c: New file.
687 * datasheet.h: New file.
689 2007-06-06 Ben Pfaff <blp@gnu.org>
691 Until now, the procedure code has provided a case to the
692 case_source, which has filled in the data values that come from
693 the active file. "Left" data values that don't come from the
694 active file naturally stay the same from case to case, because the
695 procedure code keeps using that same case.
697 One of the compromises that comes with the new procedure code is
698 that the active file allocates and provides its own case, which
699 the procedure code then has to resize to provide room for any
700 other variables that should go in the case and then fill in the
701 values of "left" variables. Then, when we're done with that case,
702 we have to save the values of "left" variables to copy into the
703 next case read from the active file.
705 The caseinit code helps with this.
707 * automake.mk: Add new files.
709 * caseinit.c: New file.
711 * caseinit.h: New file.
713 2007-06-06 Ben Pfaff <blp@gnu.org>
715 * value.h (value_cnt_from_width): New function.
717 * variable.c (var_get_value_cnt): Use new function.
719 2007-06-06 Ben Pfaff <blp@gnu.org>
721 Add casegrouper, to allow cases read from a given casereader to be
722 broken into groups, each of which has its own casereader.
723 Generally cases are grouped based on having equal values for some
726 * automake.mk: Add new files.
728 * casegrouper.c: New file.
730 * casegrouper.h: New file.
732 2007-06-06 Ben Pfaff <blp@gnu.org>
734 Add interface to lexicographical ordering of cases.
736 * automake.mk: Add new files.
738 * case-ordering.c: New file.
740 * case-ordering.h: New file.
742 2007-06-06 Ben Pfaff <blp@gnu.org>
744 Add casereaders and casewriters, the basis of the new data processing
745 implementation. A casereader is a uniform interface to reading cases
746 from a data source; a casewriter is a uniform interface to writing
747 cases to a data sink.
749 * automake.mk: Add new files.
751 * casereader-filter.c: New file.
753 * casereader-provider.h: New file.
755 * casereader-translator.c: New file.
757 * casereader.c: New file.
759 * casereader.h: New file.
761 * casewriter-provider.h: New file.
763 * casewriter-translator.c: New file.
765 * casewriter.c: New file.
767 * casewriter.h: New file.
769 2007-06-06 Ben Pfaff <blp@gnu.org>
771 "casewindow" data structure that extends the deque (from libpspp)
772 of cases with the ability to dump cases to disk if we get too many
775 * automake.mk: Add new files.
777 * casewindow.c: New file.
779 * casewindow.h: New file.
781 2007-06-06 Ben Pfaff <blp@gnu.org>
783 sparse_cases data structure that augments a sparse_array of cases
784 with the ability to dump cases to disk if we get too many cases in
787 * automake.mk: Add new files.
789 * sparse-cases.c: New file.
791 * sparse-cases.h: New file.
793 2007-06-06 Ben Pfaff <blp@gnu.org>
795 Adds a low-level on-disk case array data structure.
797 * automake.mk: Add new files.
799 * case-tmpfile.c: New file.
801 * case-tmpfile.h: New file.
803 2007-06-06 Ben Pfaff <blp@gnu.org>
805 In a couple of places we calculate the maximum number of cases to
806 keep in memory based on the user-defined workspace. Enable
807 centralizing the calculation through a new function.
809 * settings.c (get_workspace_cases): New function.
811 2007-06-06 Ben Pfaff <blp@gnu.org>
813 The casenumber type is defined in transformations.h, but case.h is
814 a more sensible place. Move it.
816 * case.h (CASENUMBER_MAX): New macro.
817 (typedef casenumber): Move here, from transformations.h.
819 2007-06-03 Ben Pfaff <blp@gnu.org>
821 Slightly generalize case_to_values and case_from_values functions.
823 * case.c (case_to_values): Rename case_copy_out, change interface.
824 (case_from_values): Rename case_copy_in, change interface.
826 * fastfile.c (fastfilereader_get_next_case): Update caller.
827 (write_case_to_disk): Ditto.
829 2007-06-02 Ben Pfaff <blp@gnu.org>
831 Clean up after a forgotten part of patch #5829.
833 * casedeque.h: Remove unused file.
835 * automake.mk: Remove casedeque.h from sources.
837 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
839 * category.c: Removed redundant #include
841 2007-05-06 Ben Pfaff <blp@gnu.org>
843 Abstract the documents within a dictionary a little better.
844 Thanks to John Darrington for suggestion, initial version, and
847 * dictionary.c (struct dictionary): Change `documents' member from
848 char * to struct string.
849 (dict_clear): Destroy struct string.
850 (dict_get_documents): Convert struct string to char *.
851 (dict_set_documents): Set struct string. Pad to 80-character
853 (dict_clear_documents): New function.
854 (dict_add_document_line): New function.
855 (dict_get_document_line_cnt): New function.
856 (dict_get_document_line): New function.
858 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
860 * sys-file-reader.c (read_documents): Use new document functions.
862 2007-04-19 John Darrington <john@darrington.wattle.id.au>
864 * sys-file-reader.c: When reading a system file which has no
865 long name table, automatically create one where the long names
866 are the lower case versions of the short names.
868 2007-04-22 Ben Pfaff <blp@gnu.org>
870 * dictionary.c (dict_set_split_vars): dict_destroy expects that
871 dict_clear will free most data related to the dictionary.
872 dict_clear does a decent job, except that dict_set_split_vars on
873 some systems won't actually free the dict's "split" member.
874 Instead, it'll allocate a 1-byte region. Fix this.
876 * value.c (value_copy): New function.
877 (value_set_missing): Ditto.
879 2007-04-22 John Darrington <john@darrington.wattle.id.au>
881 * Deleted existing category.h and moved cat-routines.h into
882 category.h Encapsulated struct cat_vals better.
884 2007-04-19 John Darrington <john@darrington.wattle.id.au>
886 * sys-file-reader.c: When reading a system file which has no
887 long name table, automatically create one where the long names
888 are the lower case versions of the short names.
890 2007-04-16 John Darrington <john@darrington.wattle.id.au>
892 * sys-file-reader.c: Some versions of Other Software seem to
893 produce system files with string variables' measure set to
894 zero. We'll assume these are supposed to be nominal variables.
896 2007-03-30 Ben Pfaff <blp@gnu.org>
898 * procedure.c: Adapt to new deque data structure.
900 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
901 Ben Pfaff <blp@gnu.org>
903 * file-name.c: Mingw compatibility fixes.
904 (fn_search_path): Use ISSLASH instead of comparing against '/'
906 (fn_dir_name): Use dir_name from gnulib.
907 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
908 (fn_get_identity): Use GetFullPathName instead of canonicalize
909 from gnulib, because the latter does not fully support
910 Windows-style path names. Use this implementation based on the
911 detected presence of Windows instead of the absence of Unix, since
912 the new implementation is Windows-specific.
913 (fn_compare_file_identities): In Windows implementation, compare
914 names case-insensitively.
916 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
918 * make-file.c: Don't include mkstemp.h, because gnulib now causes
919 <stdlib.h> to have the same effect.
921 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
923 * por-file-reader.c: Add missing _() around messages.
925 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
927 * make-file.c: Include "mkstemp.h", without which linking on
930 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
932 Reduce platform dependence.
934 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
935 Everywhere we using this, we really should have just depended on
936 the shell to expand tildes.
937 (fn_search_path): Simplify, given that we don't do tilde expansion
939 (fn_normalize): Removed. Caller changed to use the canonicalize
941 (fn_get_cwd): Removed. Only user was fn_normalize.
942 (fn_is_absolute): Really only test for absolute names.
943 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
945 (fn_readlink): Removed, as it was only used fn_normalize.
946 (fn_exists): Assume the stat function is available; gnulib does.
947 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
949 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
951 * dictionary.c (dict_create_vector_assert): New function.
953 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
955 * file-name.c (fn_normalize): Correct name of function
956 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
957 for pointing this out.
959 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
961 We are using a single member in struct file_handle, the "name"
962 field, for more than one purpose. When it begins with '"', it's a
963 file name; otherwise, it's a token that can be used to identify
964 it. When that assertion fires, it's because we searched for the
965 name case-sensitively as a file name (so that there was no match),
966 and then we try to insert it case-insensitively as a token, which
967 fails because duplicates aren't allowed.
969 Solution: break the two purposes into two separate fields. This
970 fixes the problem and likely makes the code easier to read too.
972 Fixes bug #18922. Thanks to John Darrington for bug report and
975 * file-handle-def.c (struct file_handle): New `id' member.
976 (fh_from_name): Rename fh_from_id. Update all callers.
977 (create_handle): New `id' parameter. Update all callers.
978 (fh_create_file): Ditto.
979 (fh_get_id): New function.
981 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
983 * case.c (case_is_null): Change return type to bool.
985 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
989 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
990 reference counting to share data. This makes it easy for
991 valgrind, etc. to find accesses to cases that have been destroyed
992 but have been kept around by another user's ref-count. This often
993 happens when the data set is small enough to find in memory; if a
994 bigger data set that would overflow to disk were used, then data
995 corruption would occur.
997 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
1001 * case.c (case_unshare): Make it check internally whether the
1002 ref_cnt is greater than 1, so that the callers don't have to.
1003 Update callers not to check.
1005 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
1007 Before, I was thinking that I might want to get rid of reference
1008 counting at some point. Now, I'm pretty sure that it's here to
1009 stay. Thus, because we have to store the value_cnt anyway for
1010 reference-counted cases, we might as well expose it to users.
1012 * case.c (case_get_value_cnt): New function.
1013 (case_resize): Drop OLD_CNT argument. Update all callers. Only
1014 resize case if its size actually changed.
1016 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
1017 instead of peeking inside struct case directly.
1018 (casefile_append): Ditto.
1020 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
1022 Get rid of the inlines for the case functions, which made the
1023 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
1024 the inlines didn't speed up "make check" at all, which is not a
1025 perfect benchmark but seems indicative.)
1027 * case.c: Remove #ifdef DEBUGGING...#endif around many function
1028 definitions. Remove some assertions on nonnull pointers that were
1029 redundant with a pointer dereference soon after in the function.
1031 (struct case_data): Move definition here from case.h.
1035 (case_data_wr): Ditto.
1037 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
1039 * automake.mk: Add casedeque.h to sources.
1041 * casedeque.h: New file.
1043 * procedure.c: (struct dataset) Change lag_count, lag_head,
1044 lag_queue member into single struct casedeque member. Update all
1045 users to use the casedeque instead.
1048 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
1050 * procedure.c: Simplify lagged cases interface. Updated all
1051 clients--well, the only client--to use the simplified interface.
1052 (dataset_n_lag) Removed.
1053 (dataset_set_n_lag) Removed.
1054 (dataset_need_lag) New function.
1056 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
1058 * dictionary.c procedure.c: More changes to ensure that callbacks occur
1059 whenever appropriate, but only when the dataset/dictionary is in a
1062 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
1064 * dictionary.c dictionary.h : Added callbacks for change of filter and
1065 split variables. Refactored some code to ensure that callbacks get
1066 invoked when appropriate.
1068 * procedure.c (proc_cancel_temporary_transformations): Make sure that
1069 replace_dict callback occurs when permanent_dict replaces the current
1072 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
1074 * dictionary.c dictionary.h : Added callback for when the weight
1075 variable of a dictionary changes.
1077 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
1079 * dictionary.c dictionary.h : Added replace_source and replace_dict
1080 callbacks, and functions to deal with them.
1082 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
1084 Simplify missing value handling.
1086 * missing-values.h (enum mv_class): New type.
1087 (enum mv_type): Moved definition into missing-values.c and renamed
1088 each MV_* to MVT_*, to distinguish them from the exposed mv_class
1089 enums. Updated all uses.
1090 (struct missing_values): Changed type of `type' from `enum
1091 mv_type' to `int' because the definition is no longer exposed.
1093 * missing-values.c (mv_is_value_missing): Add new enum mv_class
1094 parameter. Update all callers.
1095 (mv_is_num_missing): Ditto.
1096 (mv_is_str_missing): Ditto.
1097 (mv_is_value_user_missing): Removed. Changed callers to use
1098 mv_is_value_missing.
1099 (mv_is_num_user_missing): Removed. Changed callers to use
1101 (mv_is_str_user_missing): Removed. Changed callers to use
1103 (mv_is_value_system_missing): Removed. Changed callers to use
1104 mv_is_value_missing.
1105 (mv_set_type): Removed. Changed callers to use mv_clear.
1106 (mv_clear): New function.
1108 * variable.c (var_is_value_missing): Add new enum mv_class
1109 parameter. Update all callers.
1110 (var_is_num_missing): Ditto.
1111 (var_is_str_missing): Ditto.
1112 (var_is_value_user_missing): Removed. Changed callers to use
1113 var_is_value_missing.
1114 (var_is_num_user_missing): Removed. Changed callers to use
1116 (var_is_str_user_missing): Removed. Changed callers to use
1118 (var_is_value_system_missing): Removed. Changed callers to use
1119 var_is_value_missing.
1121 * casefilter.c (struct casefilter): Use enum mv_class in place of
1123 (casefilter_variable_missing): Adapt to new member.
1124 (casefilter_create): Change signature to take enum mv_class,
1127 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1129 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
1131 * case-sink.c case-sink.h procedure.c procedure.h
1132 storage-stream.c: Now uses the factory.
1134 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
1136 Make it possible to pull cases from the active file with a
1137 function call, instead of requiring indirection through a callback
1140 * case-source.h (struct case_source_class): Change ->read function
1141 to return a single case, instead of calling a callback function
1142 for each case. Change ->destroy function to return an error
1145 * case-source.c (free_case_source): Pass along the value returned
1146 by the case_source ->destroy function.
1148 * procedure.c (struct write_case_data): Removed.
1149 (struct dataset): Added some members to track procedure state.
1150 (procedure): Optimize the trivial case at this level.
1151 (internal_procedure): Re-implement in terms of proc_open,
1152 proc_read, proc_close.
1153 (proc_open) New function.
1154 (proc_read) New function.
1155 (proc_close) New function.
1156 (write_case) Moved into proc_read.
1157 (close_active_file) Moved closing of data source into proc_close.
1159 * storage-source.c: Rewrote to conform with modified
1160 case_source_class interface.
1162 * transformations.c (trns_chain_execute): Added argument to allow
1163 starting execution from an arbitrary transformation. Updated
1166 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
1168 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
1170 * sys-file-reader.c (read_display_parameters): Don't assume that
1171 MEASURE_* and ALIGN_* have the same values found in system files.
1173 * sys-file-writer.c (write_variable_display_parameters): Ditto.
1175 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
1176 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
1177 the value of n_MEASURES, which was off by 1 (at least from my
1180 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1182 * dictionary.c dictionary.h vardict.h variable.c: Added optional
1183 callbacks which are invoked when the dictionary or its
1184 variables are changed.
1186 * missing-values.c missing-values.h value-labels.c: Tidied up
1187 consistency checks, and made some of them return false
1188 instead of assert-failing.
1190 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
1192 * calendar.c (calendar_days_in_month): New function.
1194 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
1196 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
1197 the value label's width.
1199 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
1201 * sfm-private.h: Move contents into sys-file-writer.c, which is
1202 the only remaining user. Removed Borland C++-specific directives.
1204 * sys-file-reader.c: Clean up and rewrite entire file. The
1205 rewritten version is simpler and better abstracted, and should be
1206 easier to maintain and extend. It avoids using structures to read
1207 file data, which is prone to padding variations among compilers.
1208 It should also handle non-IEEE 754 system files, although I
1209 haven't been able to find any. It has been tested against many
1210 .sav files obtained from the Web and found to produce the same
1211 results as the earlier version of the code, or in some cases
1212 improved results. It is more tolerant of format variations found
1215 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
1216 member, putting an enum integer_format in its place. New member
1217 `float_format'. Changed `compressed' member to type bool.
1219 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
1221 * dictionary.c (dict_delete_consecutive_vars): New function.
1223 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
1225 * file-name.c (fn_search_path): Remove prefix arg that was unused
1226 by any caller. Updated all callers.
1228 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
1230 * format.c (fmt_dollar_template): Use user's decimal point
1231 character. Add assertion.
1233 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
1235 * format.c (fmt_dollar_template): New function, based on
1236 dollar_format_template from var-type-dialog.c.
1238 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
1240 * data-out.c (output_scientific): Fix bad assumption that "buf" is
1243 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
1245 Finish converting struct variable to an opaque type. In this
1246 phase, we add remaining setter and getter functions, convert the
1247 remaining PSPP code to use them, and do a bunch of cleanup. The
1248 resulting changes are pervasive but mostly trivial, and only the
1249 notable changes are logged.
1251 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
1254 * case.c (case_data): Renamed case_data_idx.
1255 (case_num): Renamed case_num_idx.
1256 (case_str): Renamed case_str_idx.
1257 (case_data_rw): Renamed case_data_rw_idx.
1259 * case.h (case_data): New function with old name and an interface
1260 that takes a variable instead of an index, which is easier to
1261 use. Updated all callers to use the new interface, or to use the
1262 new *_idx function (see above).
1265 (case_data_rw): Ditto.
1267 * category.c (cat_stored_values_destroy): Changed interface to
1268 take a struct cat_vals * instead of a struct variable *.
1270 * dictionary.c (dict_clone): Use new vector_clone function.
1271 (dict_clear) Use new var_destroy function.
1272 (add_var) New function.
1273 (dict_create_var) Rewrite in terms of dict_create_var_assert.
1274 (dict_create_var_assert) Rewrite in terms of add_var.
1275 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
1276 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
1277 (dict_lookup_var) Use new var_create, var_destroy functions.
1278 (dict_contains_var) Rewrite in terms of new vardict functionality.
1279 (set_var_dict_index) New function.
1280 (set_var_case_index) New function.
1281 (reindex_vars) New function.
1282 (dict_delete_var) Rewrite in terms of new vardict functionality.
1283 (dict_reorder_var) Ditto.
1284 (dict_reorder_vars) Ditto.
1285 (rename_var) New function.
1286 (dict_rename_var) Use rename_var.
1287 (dict_rename_vars) Use pool to simplify code. Use rename_var.
1288 (dict_get_compacted_idx_to_fv) Rename
1289 dict_get_compacted_dict_index_to_case_index, update callers.
1290 (dict_create_vector) Use new vector_create function.
1291 (dict_clear_vectors) Use new vector_destroy function.
1292 (set_var_short_name_suffix) Move here from variable.c, renamed
1293 from var_set_short_name_suffix, make static, update caller.
1295 * sys-file-private.c: New file.
1296 (sfm_width_to_bytes) Moved here from variable.c, renamed from
1297 width_to_bytes, update callers.
1299 * sys-file-private.h: New file. Later it will supplant
1300 sfm-private.h; for now it supplements it.
1301 (macro MIN_VERY_LONG_STRING) New macro.
1302 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
1304 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
1305 MAX_LONG_STRING was used before.
1307 * sys-file-writer.c: Ditto.
1309 * value-labels.c: Change the paradigm here to be that a null
1310 pointer is OK for a struct val_labs * in most cases; it just
1311 represents an empty set of value labels.
1312 (val_labs_copy) A copy of a null set is a null set.
1313 (val_labs_count) A null set has 0 labels.
1314 (val_labs_replace) Change return type to void. Rewrite for
1316 (val_labs_find) A null set does not contain the value.
1317 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1318 transposed argument order, updated all callers.
1320 * value.c: New file.
1321 (value_dup) Moved here from variable.c.
1322 (compare_values) Ditto.
1325 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1326 (macro MAX_LONG_STRING) Removed, because it was only interesting
1327 for system files, not for general code.
1328 (macro MAX_VERY_LONG_STRING) Ditto.
1329 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1330 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1332 * vardict.h: New file, for an interface between variables and
1335 * variable.c: A lot of functions were moved around, for better
1337 (struct variable) Move definition here, from variable.h.
1338 (var_type_adj) Removed--makes i18n hard.
1339 (var_type_noun) Ditto.
1340 (var_create) New function.
1341 (var_clone) New function.
1342 (var_destroy) New function.
1343 (var_set_name) Assert that variable is not in a dictionary.
1344 (compare_var_names) Rename compare_vars_by_name and fix a couple
1345 of callers who thought the args were strings.
1346 (hash_var_name) Rename hash_var_by_name.
1347 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1348 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1349 (var_is_very_long_string) Removed, because it was only interesting
1350 to system file code.
1351 (var_set_missing_values) Allow the argument to be the wrong width,
1352 as long as we can resize it. Simplify callers who were doing the
1353 resizing themselves.
1354 (var_get_value_labels) New function.
1355 (var_has_value_labels) New function.
1356 (var_set_value_labels) New function.
1357 (alloc_value_labels) New function.
1358 (var_add_value_label) New function.
1359 (var_replace_value_label) New function.
1360 (var_clear_value_labels) New function.
1361 (var_lookup_value_label) New function.
1362 (var_get_value_name) Moved here from variable.c, renamed from
1363 var_get_value_name, transposed argument order, updated all
1365 (var_to_string) Moved here, from variable-label.c.
1366 (var_set_leave) New function.
1367 (var_get_leave) New function.
1368 (var_must_leave) New function.
1369 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1370 set_var_short_name_suffix.
1371 (var_get_dict_index) New function.
1372 (var_get_case_index) New function.
1373 (var_get_obs_vals) New function.
1374 (var_set_obs_vals) New function.
1375 (var_has_obs_vals) New function.
1376 (var_get_vardict) New function.
1377 (var_set_vardict) New function.
1378 (var_has_vardict) New function.
1379 (var_clear_vardict) New function.
1380 (value_dup) Moved to value.c.
1381 (compare_values) Ditto.
1384 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1385 (enum ALPHA) Rename VAR_STRING, update all users.
1387 * vector.c: New file.
1388 (struct vector) Moved here, from variable.h.
1389 (check_widths) New function.
1390 (vector_create) New function.
1391 (vector_clone) New function.
1392 (vector_destroy) New function.
1393 (vector_get_name) New function.
1394 (vector_get_var) New function.
1395 (vector_get_var_cnt) New function.
1396 (compare_vector_ptrs_by_name) New function.
1398 * vector.h: New file.
1400 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1402 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1403 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1406 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1408 * value-labels.c (destroy_atoms): New function.
1409 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1411 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1413 Thanks to Jason Stover for pointing out this problem.
1415 * data-out.c (output_number): Use gsl_finite from GSL, which is
1416 portable, instead of isfinite, which is not.
1419 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1421 * variable.c variable.h (value_dup): New function.
1423 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1425 Start converting struct variable to an opaque type. In this
1426 phase, we add a bunch of setter and getter functions and convert
1427 most of the PSPP code to use them. The resulting changes are
1428 pervasive but mostly trivial, and only the notable changes are
1431 * format.c (fmt_equal): New function.
1433 * variable.c (var_type_is_valid): New function.
1434 (measure_is_valid) Moved here, from format.c.
1435 (alignment_is_valid) Moved here, from format.c.
1436 (var_get_name) New function.
1437 (var_set_name) New function.
1438 (width_to_type) New function.
1439 (var_get_type) New function.
1440 (var_get_width) New function.
1441 (var_set_width) New function.
1442 (var_is_numeric) New function.
1443 (var_is_alpha) New function.
1444 (var_is_short_string) New function.
1445 (var_is_long_string) New function.
1446 (var_is_very_long_string) New function.
1447 (var_get_missing_values) New function.
1448 (var_set_missing_values) New function.
1449 (var_clear_missing_values) New function.
1450 (var_has_missing_values) New function.
1451 (var_is_value_missing) New function.
1452 (var_is_num_missing) New function.
1453 (var_is_str_missing) New function.
1454 (var_is_value_user_missing) New function.
1455 (var_is_num_user_missing) New function.
1456 (var_is_str_user_missing) New function.
1457 (var_is_value_system_missing) New function.
1458 (var_get_print_format) New function.
1459 (var_set_print_format) New function.
1460 (var_get_write_format) New function.
1461 (var_set_write_format) New function.
1462 (var_set_both_formats) New function.
1463 (var_get_label) New function.
1464 (var_set_label) New function.
1465 (var_clear_label) New function.
1466 (var_has_label) New function.
1467 (var_get_measure) New function.
1468 (var_set_measure) New function.
1469 (var_get_display_width) New function.
1470 (var_set_display_width) New function.
1471 (var_get_alignment) New function.
1472 (var_set_alignment) New function.
1473 (var_get_value_cnt) New function.
1474 (var_get_leave) New function.
1475 (var_get_short_name) New function.
1477 * variable.h: (struct variable) Removed "type" and "nv" members;
1478 they are now computed from "width" where needed.
1480 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1482 * missing-values.c (mv_resize): Don't write beyond end of the
1483 allocated buffer when resizing a long string.
1485 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1487 Clean up identifier code: don't require identifier enumerations to
1488 be in a particular order; make better use of string library;
1489 expose less of the internals.
1491 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1492 change interface. Updated all callers.
1493 (lex_id_match) Change interface to use struct substring, update
1495 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1496 (global array keywords[]) Make static, change form. Update all
1497 users to use lex_id_name instead.
1498 (lex_is_keyword) New function.
1499 (lex_id_to_token) Change interface to use struct substring, update
1501 (lex_id_name) New function.
1503 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1504 lex_is_keyword instead.
1505 (T_LAST_KEYWORD) Removed.
1506 (T_N_KEYWORDS) Removed.
1508 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1510 * format.c: (fmt_date_template) Distinguish characters for which a
1511 space is output and any date delimiter is allowed on input, from
1512 those for which a space is output and only a space is allowed on
1513 input. The former is represented by X, the latter by a space.
1514 Also, drop distinction between h and H, changing the former to the
1517 * data-in.c: Completely rewrite internals to conform to SPSS input
1518 formats as closely as possible.
1519 (data_in) Changed external interface by replacing the structure
1520 that was used as a single argument by a set of arguments. Updated
1522 (data_in_finite_line) Removed. Converted all callers to use plain
1524 (data_in_get_integer_format) New function.
1525 (data_in_set_integer_format) New function.
1526 (data_in_get_float_format) New function.
1527 (data_in_set_float_format) New function.
1529 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1530 (struct data_in) Removed.
1532 * data-out.c: (output_date) Drop each component from the input as
1533 it is output, to allow us to drop the distinction between h (a
1534 count of hours) and H (the hour of day) template characters.
1535 Also, handle new X template character.
1536 (output_scientific) Follow more rational rule on when to drop
1537 fraction introduced between SPSS 13 and 15. Updated test case to
1540 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1542 Fix buffer overflow reported by John Darrington.
1544 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1545 realize that DIGITS is a count of nibbles, not of bytes.
1547 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1549 * calendar.c (calendar_offset_to_gregorian) Also return the
1550 year-of-day. Change callers to new interface.
1552 * data-out.c: Completely rewrite internals to conform to SPSS
1553 output formats as completely as possible.
1554 (data_out) Change interface to put input parameters before output
1555 parameters, for consistency with the style I now prefer. Update
1557 (data_out_get_integer_format) New public function.
1558 (data_out_set_integer_format) New public function.
1559 (data_out_get_float_format) New public function.
1560 (data_out_set_float_format) New public function.
1562 * data-out.h: New file. Move prototype for data_out here, from
1565 * format.c: (fmt_step_width) Use equality comparison instead of
1566 bitwise and, for clarity.
1567 (fmt_is_string) Ditto.
1568 (fmt_input_to_output) Fix categories that are translated to F
1571 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1573 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1574 casefile-private.h: Added casefilter to assist commands with missing
1577 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1579 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1581 * settings.c: (route_errors_to_terminal) New variable.
1582 (route_errors_to_listing) New variable.
1583 (get_error_routing_to_terminal) New function.
1584 (set_error_routing_to_terminal) New function.
1585 (get_error_routing_to_listing) New function.
1586 (set_error_routing_to_listing) New function.
1588 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1590 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1592 * format.c: Completely rewrite, to achieve better abstraction.
1593 Rewrite all references to formats in other files.
1595 * format.def: Rewrite and reorganize.
1597 * settings.c: Move everything related to custom currency formats
1598 into format.[ch], changing them in form, so as to group related
1599 code and definitions better. Changed all references to use the
1601 (static var decimal) Removed.
1602 (static var grouping) Removed.
1603 (static var cc) Removed.
1604 (get_decimal) Removed.
1605 (set_decimal) Removed.
1606 (get_grouping) Removed.
1607 (set_grouping) Removed.
1611 * settings.h: (macro CC_CNT) Removed.
1612 (macro CC_WIDTH) Removed.
1613 (struct custom_currency) Removed.
1615 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1617 * data-in.c (data_in): Use switch statement instead of table, to
1618 avoid dependence on the order of the FMT_* enums.
1620 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1622 * data-out.c: (num_to_string) Removed, because it was dead code.
1624 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1626 * data-in.c (parse_trailer): Fix error message.
1628 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1630 * format.c (fmt_is_binary): New function.
1632 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1634 * procedure.c procedure.h: Encapsulated the static data into a single
1637 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1639 * casefile.c (casereader_read_xfer): Always initialize the case,
1640 even on an error condition.
1642 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1644 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1645 entire function useless.
1647 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1649 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1650 casereader method casereader_clone.
1652 * procedure.c pransformations.h: Introduced new type casenum_t
1654 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1656 * variable.c: (width_to_bytes) Rephrase code for clarify.
1658 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1660 * format.c: (fmt_type_from_string) New function.
1661 (fmt_to_string) Include decimals in output if the format has
1662 decimals, even if the format type does not. This way, we can
1663 accurately reproduce incorrect formats in user output.
1664 (check_common_specifier) Make the check for a bad format type an
1665 assertion, so we get bug reports if they show up. Fix message.
1666 Check for decimal places with a format type that doesn't allow
1668 (check_input_specifier) Remove check for FMT_X, which has been
1670 (check_output_specifier) Ditto.
1672 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1674 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1675 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1676 (enum fmt_parse_flags) Removed.
1678 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1680 * casefile.c casefile.h: Converted to an abstract base class.
1681 * casefile-private.h fastfile.c fastfile.h: New files.
1682 * automake.mk procedure.c scratch-writer.c storage-stream.c
1684 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1686 * procedure.c (internal_procedure): Create sink_case with only as
1687 many values as the compacted dictionary.
1689 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1691 Remove "debugging" code that caused plenty of false positives and
1694 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1696 * case.c: Remove all references to `this' member.
1698 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1700 Fix link error noted by Jason Stover.
1702 * storage-stream.c: Include <assert.h>.
1704 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1706 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1707 ALL) and additional underlying system file issues.
1709 Thanks to John Darrington for review.
1711 First problem: var_hash points to variables not owned by the
1712 sys-file-reader, which the caller may free or modify. Use an
1713 array of sfm_vars instead, as done earlier (e.g. CVS version
1716 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1717 members and remove all code that references it. Add vars, var_cnt
1718 members. Remove fix_specials member, which was unused.
1719 (struct sfm_var) Remove name member, which was unused.
1720 (sfm_close_reader) Free vars member instead of var_hash.
1721 (compare_var_shortnames) Removed.
1722 (hash_var_shortname) Removed.
1723 (sfm_open_reader) Fill out vars array.
1724 (compare_var_index) Removed.
1725 (sfm_read_case) Use vars instead of var_hash.
1727 Second problem: we're confused about when we actually have very
1728 long strings, causing us to choose incorrectly between slow path
1729 and fast path in sfm_read_case.
1731 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1732 have very long strings, not when we have long variable names,
1733 which is an unrelated feature.
1735 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1737 * variable.h: Move var_set and variable parsing declarations to
1738 new header, src/language/lexer/variable-parser.h. Modified lots
1739 of files to include the new header.
1741 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1743 * value-labels.c (value_to_string): When there's no value label,
1744 format the variable according to its print format, instead of
1745 always effectively using A or F format.
1747 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1749 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1752 * format.c: Removed tortological assertion.
1754 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1756 Reform string library.
1758 * file-name.c (fn_interp_vars): Change interface to take a
1759 substring as input. Updated all users.
1761 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1763 * format.c (measure_is_valid): Really return false when m >=
1766 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1768 Implement random access to casefiles, for use in GUI.
1770 * casefile.c: (struct casereader) Add `random', `file_ofs',
1771 `buffer_ofs' members.
1772 (casefile_get_random_reader) New function.
1773 (read_open_file) Break part into new function
1774 seek_and_fill_buffer().
1775 (fill_buffer) Update buffer_ofs, file_ofs.
1776 (casereader_seek) New function.
1778 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1780 * settings.c: Added call to i18n{done, init}.
1782 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1784 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1787 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1789 * casefile.c: Convert many uses of `int' to `bool'.
1791 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1793 * transformations.c (trns_chain_destroy): Destroy chain's trns
1794 member, to fix memory leak.
1796 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1798 * storage-stream.c (storage_source_decapsulate): Destroy case
1799 source to fix memory leak.
1801 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1803 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1804 case passed as argument instead of initializing the argument as a
1805 case. Fixes memory leak that showed up in
1806 tests/command/aggregate.sh with scratch files.
1808 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1810 * procedure.c (proc_done): Destroy default_dict, to fix memory
1813 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1815 Simplify procedure_with_splits().
1817 * procedure.c (struct split_aux_data): Removed case_count member.
1818 (procedure_with_splits) Don't initialize case_count.
1819 (split_procedure_case_func) Check whether prev_case is null
1820 instead of case_count.
1821 (split_procedure_end_func) Ditto.
1823 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1825 * case.c (case_move): Do nothing if dst and src are the same
1827 (case_try_create) Merge two similar cases.
1828 (case_copy) Unshare only if data must be actually copied.
1830 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1832 * data-in.c data-out.c dictionary.c sys-file-reader.c
1833 sys-file-writer.c variable.c variable.h: Reworked very long string
1834 support for better encapsulation.
1836 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1838 * value-labels.c (val_labs_can_set_width): New function.
1839 (val_labs_set_width) Clear labels if increasing width to long
1841 (val_labs_destroy) Remove unneeded test for null.
1843 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1845 * value-labels.h: Remove unneeded dependency on variable.h.
1847 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1849 Get rid of `char *c' member in union value, for cleanliness.
1851 * value.h: (union value) Remove `c' member.
1853 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1855 Make dictionary compacting functions a little more general.
1857 * sys-file-writer.c (sfm_open_writer): Use
1858 dict_compacting_would_change().
1859 (does_dict_need_translation) Removed.
1861 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1863 Make dictionary compacting functions a little more general.
1865 * dictionary.c (dict_needs_compaction): Rename
1866 dict_compacting_would_shrink(). Update all callers.
1867 (dict_compacting_would_change) New function.
1869 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1871 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1872 inverted return value (!).
1874 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1876 Continue reforming procedure execution.
1878 * procedure.c: Search and replace "vfm" by "proc". Notably:
1879 (static var vfm_source) Rename proc_source. Update all
1881 (static var vfm_sink) Rename proc_sink. Update all references.
1883 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1885 Continue reforming procedure execution. In this phase, remove
1886 PROCESS IF, which was deprecated anyway and can be easily
1887 simulated with TEMPORARY followed by SELECT IF.
1889 * procedure.c: (open_active_file) Don't call
1890 add_process_if_trns().
1891 (discard_variables) Get rid of redundant call to
1892 proc_cancel_all_transformations().
1893 (add_process_if_trns) Removed.
1894 (process_if_trns_proc) Removed.
1895 (process_if_trns_free) Removed.
1897 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1899 Continue reforming procedure execution. In this phase, add
1900 `const' to the case passed to procedure()'s callback.
1902 Updated all users of procedure() as well.
1904 * procedure.c: (struct write_case_data) Add "const" to ccase
1905 parameter for case_func member.
1906 (procedure) Add "const" to ccase parameter for proc_func
1908 (multipass_case_func) Make ccase parameter const.
1909 (internal_procedure) Add "const" to ccase parameter for case_func
1911 (split_procedure_case_func) Make ccase parameter const.
1912 (multipass_split_case_func) Make ccase parameter const.
1914 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
1916 Continue reforming procedure execution. In this phase, get rid of
1917 the output code for SPLIT FILE groups in procedure.c, which really
1918 shouldn't be doing any output. Move it into the individual
1919 procedures instead. This also adds some flexibility.
1921 Updated many users of procedure_with_splits() and
1922 multipass_procedure_with_splits() to call
1923 output_split_file_values() and to deal with increased use of
1926 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
1927 parameter to begin_func member.
1928 (procedure_with_splits) Add "const struct ccase *" parameter to
1929 begin_func parameter. Make ccase parameter const in proc_func
1931 (split_procedure_case_func) Don't dump split file group. Pass
1933 (dump_splits) Moved to language/dictionary/split-file.c as
1934 output_split_file_values().
1935 (struct multipass_split_aux_data) Add "const struct ccase *"
1936 parameter to split_func member.
1937 (multipass_procedure_with_splits) Add "const struct ccase *"
1938 parameter to split_func parameter.
1939 (multipass_split_case_func) Save new SPLIT FILE case before
1941 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
1943 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
1945 Continue reforming procedure execution. Change
1946 internal_procedure() so that it calls open_active_file() and
1947 close_active_file(), which isolates most of the actual procedure
1950 * procedure.c: (struct write_case_data) Rename `proc_func' member
1951 to `case_func' and update all references.
1952 (procedure) Rewrite as one-line wrapper around
1953 internal_procedure().
1954 (struct multipass_aux_data) New.
1955 (multipass_callback) Renamed multipass_case_func(). Use struct
1956 multipass_aux_data as auxiliary data.
1957 (multipass_end_func) New function.
1958 (multipass_procedure) Rewrite as wrapper for internal_procedure()
1959 that uses multipass_case_func, multipass_end_func.
1960 (internal_procedure) Add `end_func' argument. Move optimization
1961 of trivial case in here. Move call to open_active_file() and
1962 close_active_file() in here. Now assert that vfm_source is
1964 (procedure_with_splits_callback) Rename
1965 split_procedure_case_func().
1966 (split_procedure_end_func) New function.
1967 (multipass_split_callback) Rename multipass_split_case_func.
1968 (multipass_split_end_func) New function.
1969 (discard_variables) No need to test for nonnull vfm_source.
1971 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
1973 Continue reforming procedure execution. Get rid of unused member.
1975 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
1977 (write_case) Don't increment cases_analyzed.
1979 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
1981 Continue reforming procedure execution. In this phase, move
1982 procedure.c and procedure.h from src to src/data. Update
1983 makefiles and #includes accordingly.
1985 * procedure.c: Moved here from src/.
1987 * procedure.h: Moved here from src/.
1989 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
1991 Continue reforming procedure execution. In this phase, get rid of
1992 many global variables, consolidating procedure execution in
1993 procedure.c. Encapsulate transformations in new "struct
1994 trns_chain". Also, change implementation of N OF CASES, FILTER,
1995 and PROCESS IF from special cases to transformations.
1997 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
2000 * dictionary.c: (global variable default_dict) Move to
2003 * variable.h: (TRNS_*) Move to transformations.h.
2004 (struct transformation) Move to transformations.c.
2006 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
2008 * sys-file-reader.c: Fixed invalid read problems.
2010 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
2012 * storage-stream.c: Add missing function comments.
2014 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
2016 Continue reforming procedure execution. In this phase, add some
2017 new, needed functionality to storage-stream.
2019 * storage-stream.c: (storage_source_decapsulate) New function.
2021 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
2023 * variable.c (width_to_bytes): Declarations must precede
2024 statements for C90 compliance.
2026 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
2028 * data-out.c, data-in.c, variable.c, variable.h: New functions
2029 copy_mangle and copy_demangle for reading/writing cases; emulates the
2030 way SPSS deals with strings > 255 bytes.
2032 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
2033 subtype 14 needed for strings longer than 255 bytes.
2035 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
2036 instead of literal values. Also fixed some constness issues.
2038 * format.h: Constness
2040 * sfm-private.h: Renamed the case_size identifier, since I discovered
2041 that SPSS's respect for this variable is very nominal.
2043 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
2045 Change case limit type from int to size_t.
2047 * dictionary.c: (struct dictionary) Change type of case_limit
2049 (dict_get_case_limit) Change return type.
2050 (dict_set_case_limit) Change parameter type.
2052 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
2054 * variable.h: (struct variable) Rename `reinit' member as `leave'
2055 and invert sense. Fix up all references.
2057 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
2059 Continue reforming procedure execution. In this phase, break
2060 procedure.c into multiple files.
2062 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
2064 * case-sink.c: New file.
2066 * case-sink.h: New file.
2068 * case-source.c: New file.
2070 * case-source.h: New file.
2072 * storage-stream.c: New file.
2074 * storage-stream.h: New file.
2076 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
2078 * variable.h: (struct variable) Remove `init' member and all
2079 references to it from other files. It was initialized in several
2080 places, but nothing really ever used it for anything worthwhile.
2081 Thanks to Jason Stover for pointing out how confusing this
2084 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
2086 Continue reforming error message support. In this phase, get rid
2087 of message "titles" and put the message text in `struct error'.
2088 Now `struct error' encapsulates a message more properly.
2090 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
2091 Format message ourselves.
2093 * data-in.c: (vdls_error) Ditto.
2095 * por-file-reader.c: (error) Ditto.
2097 * sys-file-reader.c: (corrupt_msg) Ditto.
2099 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
2101 GNU standards require "file name" instead of "filename" in
2102 documentation. It's nice for our code to follow the convention
2105 * casefile.c: (struct casefile) Rename `filename' member to
2106 `file_name'. Updated all references.
2108 * file-name.c: [!unix] (struct file_identity) Rename
2109 normalized_filename member to normalized_file_name. Updated all
2112 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
2114 We don't really support anything but Unix-like environments well,
2115 so we might as well de-obfuscate by writing directory and path
2116 separators explicitly.
2118 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
2120 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
2121 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
2123 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
2126 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
2128 GNU standards require "file name" instead of "filename" in
2129 documentation. It's nice for our code to follow the convention
2132 * filename.c: Rename to file-name.c.
2134 * filename.h: Rename to file-name.h. Update all inclusions.
2135 Update header guards.
2137 * automake.mk: Update file names.
2139 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
2141 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
2143 (fn_basename) Removed (dead code).
2144 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
2145 (fn_special_p) Renamed fn_is_special(), all references updated.
2146 (fn_exists_p) Renamed fn_exists(), all references updated.
2148 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
2150 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
2151 Also, now it only tilde-expands file names, not paths.
2152 (fn_search_path) Tilde-expand one directory at a time.
2154 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
2156 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
2157 get rid of non-Unixlike version of the code, which has probably
2159 (fn_prepend_dir) Removed (dead code).
2161 * filename.h: (macro DIR_SEPARATOR_STRING) New.
2162 (macro PATH_SEPARATOR_STRING) New.
2163 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
2165 Continue reforming error message support. In this phase, we get
2166 rid of VM() and the other msg() support for "verbosity", replacing
2167 it by a new function verbose_msg().
2169 * filename.c: (fn_search_path) Use verbose_msg() instead of
2172 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
2174 * sfm-private.h: Get rid of #defines after #error, which makes no
2177 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
2179 Get rid of our own int32 type in favor of the standard int32_t
2182 * sfm-private.h: (int32 macro) Don't define this anymore. Do
2185 * sys-file-reader.c: Use int32_t instead of int32 throughout.
2187 * sys-file-writer.c: Use int32_t instead of int32 throughout.
2189 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
2191 Remove ill-considered file routines that are no longer used.
2193 * filename.c: (fn_open_ext) Removed.
2194 (fn_close_ext) Removed.
2196 * filename.h: (struct file_ext) Removed.
2198 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
2200 * variable.c (var_is_valid_name): Move declarations before code
2203 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
2205 * filename.ch (fn_interp_vars): Fixed small buglet.
2207 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
2209 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
2210 so as to modify the string inline. Thus makeing it easier to
2211 destroy the results when no longer needed.
2213 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
2215 * category.c (cat_stored_values_destroy): Fixed memory leak.
2217 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
2219 Add some missing frees. Thanks to John Darrington for reporting
2222 * any-writer.c (any_writer_close): Free writer.
2224 * any-reader.c (any_reader_close): Free reader.
2226 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
2228 * por-file-reader.c: (error) Mark as NO_RETURN.
2230 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
2232 * settings.c: Changed default value of scompress to true.
2234 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
2236 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
2238 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
2241 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
2242 here from pref.h.orig.
2244 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
2246 * sys-file-reader.c: Fixed bug reading compressed files.
2248 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
2250 * Numerous renames. See src/ChangeLog for details.
2252 * Moved files from src directory
projects / pspp-builds.git / blob