1 2007-07-24 Ben Pfaff <blp@gnu.org>
4 * sys-file-writer.c (write_variable_display_parameters): Use new
5 var_default_display_width function to choose display width of
6 segments after the first one in a given variable.
7 * variable.c (var_create): Use var_default_display_width to pick
8 new variable's display width.
9 (var_default_display_width): New function.
11 2007-07-24 Ben Pfaff <blp@gnu.org>
14 * por-file-writer.c (write_variables): Write weight variable.
15 Reviewed by John Darrington.
17 2007-07-23 Ben Pfaff <blp@gnu.org>
19 Improvements to system file reader and writer.
21 First, move all detailed knowledge of very long strings into
22 sys-file-private.[ch], so that this nasty stuff can be isolated.
24 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
25 (EFFECTIVE_VLS_CHUNK): New macro.
26 (min_int): New function.
27 (max_int): New function.
28 (sfm_width_to_bytes): Rewrite.
29 (sfm_width_to_octs): New function.
30 (sfm_segment_alloc_width): New function.
31 (sfm_segment_alloc_bytes): New function.
32 (sfm_segment_used_bytes): New function.
33 (sfm_segment_offset): New function.
34 (sfm_segment_effective_offset): New function.
35 (sfm_dictionary_to_sfm_vars): New function.
37 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
38 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
39 (struct sfm_var): New structure.
41 Next, improvements to the system file reader.
43 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
44 type casenumber. Added `version_major', `version_minor',
47 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
48 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
49 Change `case_cnt' to type casenumber. Removed `has_vls'.
50 (struct sfm_var): Removed.
51 (sfm_open_reader): Don't warn on wrong case size if the file was
52 written by SPSS 13, which tends to get it wrong. Use
53 sfm_dictionary_to_sfm_vars.
54 (read_header): Always output system file info.
55 (read_variable_record): Simplify code for reading missing values.
56 (read_machine_int32_info): Save version numbers from system file
57 into info struct passed as new argument.
58 (read_long_string_map): Restructured to use new sys-file-private
60 (read_value_labels): Use size_overflow_p.
61 (sys_file_casereader_read): Get rid of distinction between fast
62 and slow paths. Use information provided by sys-file-primate's
63 struct sfm_var to simplify code.
64 (skip_whole_strings): New function.
65 (read_int32): Renamed read_int. Changed return value to int.
67 (read_flt64): Renamed read_float. Changed return value to
68 double. Updated all callers.
69 (int32_to_native): Removed. Changed callers to use
71 (flt64_to_double): Removed. Changed callers to use float_convert.
73 Finally, get rid of int32, flt64 terminology and types in system
74 file writer. The former wasn't very useful since a POSIX "int"
75 can hold the whole range of int32 and we generally didn't have a
76 need for it to be exactly-32-bits, just at-least-32-bits. The
77 latter was inconvenient because we had to assume that it could be
78 different from double and thereby convert special values SYSMIS,
79 HIGHEST, LOWEST to and from it in multiple places. Instead, now
80 we just use "int" and "double" in most places, and do conversions,
81 if necessary, very close to where we do I/O. This change meant
82 that the writer code couldn't represent records in the file as C
83 structs any longer, but that's no great loss. The code actually
84 seems to be more readable without them.
86 Simplify the compression buffering code: only buffer as much as
87 necessary, which is no more than eight 8-byte units at any given
90 * sys-file-writer.c (typedef flt64): Removed.
91 (macro second_lowest_flt64): Removed.
92 (struct sysfile_header): Removed.
93 (struct sysfile_variable): Removed.
94 (struct sfm_writer): Removed `needs_translation', `has_vls',
95 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
96 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
97 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
98 compression buffering by `opcodes', `opcode_cnt', `data',
99 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
100 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
101 data writer functions instead of structures.
102 (calc_oct_idx): New function.
103 (write_header): Use simple data writer functions instead of
105 (write_format_spec): Renamed write_format. New argument.
106 (write_variable_continuation_records): New function.
107 (write_variable): Use simple data writer functions instead of
108 structures. Use write_variable_continuation_records. Write
109 entire very long string instead of requiring caller to understand
111 (write_value_labels): Use simple data writer functions instead of
113 (write_documents): Ditto.
114 (write_variable_display_parameters): Use sys-file-private
115 functions to simplify. Use simple data writer functions instead
117 (write_vls_length_table): Use simple data writer functions instead
119 (write_longvar_table): Ditto.
120 (write_rec_7_34): Break into new functions
121 write_integer_info_record, write_float_info_record. Use simple
122 data writer functions instead of structures.
123 (buf_write): Removed.
124 (append_string_max): Removed.
125 (ensure_buf_space): Removed.
126 (sys_file_casewriter_write): Get rid of the distinction between
127 fast and slow paths, which didn't seem to be too useful. Use new
128 functions write_case_uncompressed, write_case_compressed.
129 (put_instruction): Removed.
130 (put_element): Removed.
131 (write_compressed_data): Removed.
132 (close_writer): Use flush_compressed. Only write case count to
133 system file if it will fit in the field.
134 (write_case_compressed): New function.
135 (write_case_uncompressed): New function.
136 (flush_compressed): New function.
137 (put_cmp_opcode): New function.
138 (put_cmp_number): New function.
139 (write_int): New function.
140 (convert_double_to_output_format): New function.
141 (write_float): New function.
142 (write_value): New function.
143 (write_string): New function.
144 (write_bytes): New function.
145 (write_zeros): New function.
146 (write_spaces): New function.
148 Reviewed by John Darrington.
150 2007-07-22 Ben Pfaff <blp@gnu.org>
152 Don't try to write very long strings to portable files. The
153 format does not support it.
155 * por-file-writer.c (MAX_POR_WIDTH): New macro.
156 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
157 (write_format): Add arg to take width to resize format to.
158 (write_value): Limit width of value written to MAX_POR_WIDTH.
159 (write_variables): Limit width of variable and its output formats
161 Reviewed by John Darrington.
163 2007-07-22 Ben Pfaff <blp@gnu.org>
165 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
166 local variable instead of literal 5.
167 Reviewed by John Darrington.
169 2007-07-22 Ben Pfaff <blp@gnu.org>
171 Fix problems with uniqueness of short names in system files with
172 very long string variables. Now a variable may have multiple
175 * automake.mk (src_data_libdata_a_SOURCES): Add new files
176 short-names.c, short-names.h.
178 * dictionary.c (dict_clone): Clone all the short names.
179 (compare_strings): Move into short-names.c.
180 (hash_strings): Ditto.
181 (set_var_short_name_suffix): Ditto.
182 (dict_assign_short_names): Ditto, rename short_names_assign,
183 change to assign all short names.
185 * por-file-writer.c (write_variables): Use short_names_assign
186 instead of dict_assign_short_names.
188 * short-names.c: New file.
190 * short-names.h: New file.
192 * sys-file-private.c (sfm_width_to_segments): New function.
194 * sys-file-reader.c (read_long_var_name_map): Save and restore all
195 the short names, not just the first one.
197 * sys-file-writer.c (cont_var_name): Removed.
198 (sfm_open_writer): Use short_names_assign instead of
199 dict_assign_short_names. Use unique short names assigned by
200 short_names_assign instead of those generated by cont_var_name.
202 * variable.c (struct variable): Remove `short_name' member,
203 replace by `short_names' and `short_name_cnt'.
204 (var_create) Initialize new members.
205 (var_get_short_name_cnt): New function.
206 (var_get_short_name): Now takes an index argument. Changed most
208 (var_set_short_name): Ditto.
209 (var_clear_short_name): Renamed var_clear_short_names, changed to
210 clear all short names.
212 Reviewed by John Darrington.
214 2007-07-22 Ben Pfaff <blp@gnu.org>
216 * variable.c (var_set_width): Use new var_set_width function.
218 * missing-values.c (mv_n_values): Drop assertion, which was not
221 * format.c (fmt_default_for_width): New function.
222 (fmt_resize): New function.
224 Reviewed by John Darrington.
226 2007-07-18 John Darrington <john@darrington.wattle.id.au>
228 * datasheet.c (datasheet_delete_columns): Added assertion to check
229 we're not deleting outside the range of the sheet.
232 * dictionary.c dictionary.h variable.c: Added the ability for string
233 variables to be resized.
235 * vardict.h: Added some prototypes (moved from dictionary.h) as
236 these should only be called by variable.c
239 2007-07-14 John Darrington <john@darrington.wattle.id.au>
241 * sfm-reader.c: Respect case_cnt field in file header.
243 2007-07-01 John Darrington <john@darrington.wattle.id.au>
245 * transformation.c transformation.h (trns_chain_execute): Changed the
246 signature (Patch #6057)
248 2007-06-10 Ben Pfaff <blp@gnu.org>
250 * casereader-filter.c (casereader_filter_destroy): Make sure to
251 write all the remaining excluded cases to the casewriter, if any.
253 * caseinit.c (init_list_destroy): Rewrite.
254 (init_list_clear): Ditto.
256 * casegrouper.c (casegrouper_get_next_group): Always set *reader
257 to null when returning false.
259 2007-06-06 Ben Pfaff <blp@gnu.org>
261 Actually implement the new procedure code and adapt all of its
262 clients to match. Also adapt all of the other case sources and
263 sinks in the tree and their clients to use the
264 casereader/casewriter infrastructure.
266 * automake.mk: Add and remove files.
268 * any-reader.c: Change into a casereader.
269 * por-file-reader.c: Ditto.
270 * scratch-reader.c: Ditto.
271 * sys-file-reader.c: Ditto.
273 * any-writer.c: Change into a casewriter.
274 * por-file-writer.c: Ditto.
275 * scratch-writer.c: Ditto.
276 * sys-file-writer.c: Ditto.
278 * procedure.c: Change to use casereader, casewriter, caseinit, and
279 other new infrastructure.
281 * scratch-handle.c: Adapt to new infrastructure.
283 * case-sink.c: Removed, now dead code.
284 * case-sink.h: Ditto.
285 * case-source.c: Ditto.
286 * case-source.h: Ditto.
287 * casefile-factory.c: Ditto.
288 * casefile-private.h: Ditto.
291 * casefilter.c: Ditto.
292 * casefilter.h: Ditto.
295 * fastfile-factory.c: Ditto.
296 * fastfile-factory.h: Ditto.
297 * storage-stream.c: Ditto.
298 * storage-stream.h: Ditto.
300 2007-06-06 Ben Pfaff <blp@gnu.org>
304 * automake.mk: Add new files.
306 * datasheet.c: New file.
308 * datasheet.h: New file.
310 2007-06-06 Ben Pfaff <blp@gnu.org>
312 Until now, the procedure code has provided a case to the
313 case_source, which has filled in the data values that come from
314 the active file. "Left" data values that don't come from the
315 active file naturally stay the same from case to case, because the
316 procedure code keeps using that same case.
318 One of the compromises that comes with the new procedure code is
319 that the active file allocates and provides its own case, which
320 the procedure code then has to resize to provide room for any
321 other variables that should go in the case and then fill in the
322 values of "left" variables. Then, when we're done with that case,
323 we have to save the values of "left" variables to copy into the
324 next case read from the active file.
326 The caseinit code helps with this.
328 * automake.mk: Add new files.
330 * caseinit.c: New file.
332 * caseinit.h: New file.
334 2007-06-06 Ben Pfaff <blp@gnu.org>
336 * value.h (value_cnt_from_width): New function.
338 * variable.c (var_get_value_cnt): Use new function.
340 2007-06-06 Ben Pfaff <blp@gnu.org>
342 Add casegrouper, to allow cases read from a given casereader to be
343 broken into groups, each of which has its own casereader.
344 Generally cases are grouped based on having equal values for some
347 * automake.mk: Add new files.
349 * casegrouper.c: New file.
351 * casegrouper.h: New file.
353 2007-06-06 Ben Pfaff <blp@gnu.org>
355 Add interface to lexicographical ordering of cases.
357 * automake.mk: Add new files.
359 * case-ordering.c: New file.
361 * case-ordering.h: New file.
363 2007-06-06 Ben Pfaff <blp@gnu.org>
365 Add casereaders and casewriters, the basis of the new data processing
366 implementation. A casereader is a uniform interface to reading cases
367 from a data source; a casewriter is a uniform interface to writing
368 cases to a data sink.
370 * automake.mk: Add new files.
372 * casereader-filter.c: New file.
374 * casereader-provider.h: New file.
376 * casereader-translator.c: New file.
378 * casereader.c: New file.
380 * casereader.h: New file.
382 * casewriter-provider.h: New file.
384 * casewriter-translator.c: New file.
386 * casewriter.c: New file.
388 * casewriter.h: New file.
390 2007-06-06 Ben Pfaff <blp@gnu.org>
392 "casewindow" data structure that extends the deque (from libpspp)
393 of cases with the ability to dump cases to disk if we get too many
396 * automake.mk: Add new files.
398 * casewindow.c: New file.
400 * casewindow.h: New file.
402 2007-06-06 Ben Pfaff <blp@gnu.org>
404 sparse_cases data structure that augments a sparse_array of cases
405 with the ability to dump cases to disk if we get too many cases in
408 * automake.mk: Add new files.
410 * sparse-cases.c: New file.
412 * sparse-cases.h: New file.
414 2007-06-06 Ben Pfaff <blp@gnu.org>
416 Adds a low-level on-disk case array data structure.
418 * automake.mk: Add new files.
420 * case-tmpfile.c: New file.
422 * case-tmpfile.h: New file.
424 2007-06-06 Ben Pfaff <blp@gnu.org>
426 In a couple of places we calculate the maximum number of cases to
427 keep in memory based on the user-defined workspace. Enable
428 centralizing the calculation through a new function.
430 * settings.c (get_workspace_cases): New function.
432 2007-06-06 Ben Pfaff <blp@gnu.org>
434 The casenumber type is defined in transformations.h, but case.h is
435 a more sensible place. Move it.
437 * case.h (CASENUMBER_MAX): New macro.
438 (typedef casenumber): Move here, from transformations.h.
440 2007-06-03 Ben Pfaff <blp@gnu.org>
442 Slightly generalize case_to_values and case_from_values functions.
444 * case.c (case_to_values): Rename case_copy_out, change interface.
445 (case_from_values): Rename case_copy_in, change interface.
447 * fastfile.c (fastfilereader_get_next_case): Update caller.
448 (write_case_to_disk): Ditto.
450 2007-06-02 Ben Pfaff <blp@gnu.org>
452 Clean up after a forgotten part of patch #5829.
454 * casedeque.h: Remove unused file.
456 * automake.mk: Remove casedeque.h from sources.
458 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
460 * category.c: Removed redundant #include
462 2007-05-06 Ben Pfaff <blp@gnu.org>
464 Abstract the documents within a dictionary a little better.
465 Thanks to John Darrington for suggestion, initial version, and
468 * dictionary.c (struct dictionary): Change `documents' member from
469 char * to struct string.
470 (dict_clear): Destroy struct string.
471 (dict_get_documents): Convert struct string to char *.
472 (dict_set_documents): Set struct string. Pad to 80-character
474 (dict_clear_documents): New function.
475 (dict_add_document_line): New function.
476 (dict_get_document_line_cnt): New function.
477 (dict_get_document_line): New function.
479 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
481 * sys-file-reader.c (read_documents): Use new document functions.
483 2007-04-19 John Darrington <john@darrington.wattle.id.au>
485 * sys-file-reader.c: When reading a system file which has no
486 long name table, automatically create one where the long names
487 are the lower case versions of the short names.
489 2007-04-22 Ben Pfaff <blp@gnu.org>
491 * dictionary.c (dict_set_split_vars): dict_destroy expects that
492 dict_clear will free most data related to the dictionary.
493 dict_clear does a decent job, except that dict_set_split_vars on
494 some systems won't actually free the dict's "split" member.
495 Instead, it'll allocate a 1-byte region. Fix this.
497 * value.c (value_copy): New function.
498 (value_set_missing): Ditto.
500 2007-04-22 John Darrington <john@darrington.wattle.id.au>
502 * Deleted existing category.h and moved cat-routines.h into
503 category.h Encapsulated struct cat_vals better.
505 2007-04-19 John Darrington <john@darrington.wattle.id.au>
507 * sys-file-reader.c: When reading a system file which has no
508 long name table, automatically create one where the long names
509 are the lower case versions of the short names.
511 2007-04-16 John Darrington <john@darrington.wattle.id.au>
513 * sys-file-reader.c: Some versions of Other Software seem to
514 produce system files with string variables' measure set to
515 zero. We'll assume these are supposed to be nominal variables.
517 2007-03-30 Ben Pfaff <blp@gnu.org>
519 * procedure.c: Adapt to new deque data structure.
521 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
522 Ben Pfaff <blp@gnu.org>
524 * file-name.c: Mingw compatibility fixes.
525 (fn_search_path): Use ISSLASH instead of comparing against '/'
527 (fn_dir_name): Use dir_name from gnulib.
528 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
529 (fn_get_identity): Use GetFullPathName instead of canonicalize
530 from gnulib, because the latter does not fully support
531 Windows-style path names. Use this implementation based on the
532 detected presence of Windows instead of the absence of Unix, since
533 the new implementation is Windows-specific.
534 (fn_compare_file_identities): In Windows implementation, compare
535 names case-insensitively.
537 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
539 * make-file.c: Don't include mkstemp.h, because gnulib now causes
540 <stdlib.h> to have the same effect.
542 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
544 * por-file-reader.c: Add missing _() around messages.
546 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
548 * make-file.c: Include "mkstemp.h", without which linking on
551 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
553 Reduce platform dependence.
555 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
556 Everywhere we using this, we really should have just depended on
557 the shell to expand tildes.
558 (fn_search_path): Simplify, given that we don't do tilde expansion
560 (fn_normalize): Removed. Caller changed to use the canonicalize
562 (fn_get_cwd): Removed. Only user was fn_normalize.
563 (fn_is_absolute): Really only test for absolute names.
564 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
566 (fn_readlink): Removed, as it was only used fn_normalize.
567 (fn_exists): Assume the stat function is available; gnulib does.
568 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
570 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
572 * dictionary.c (dict_create_vector_assert): New function.
574 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
576 * file-name.c (fn_normalize): Correct name of function
577 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
578 for pointing this out.
580 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
582 We are using a single member in struct file_handle, the "name"
583 field, for more than one purpose. When it begins with '"', it's a
584 file name; otherwise, it's a token that can be used to identify
585 it. When that assertion fires, it's because we searched for the
586 name case-sensitively as a file name (so that there was no match),
587 and then we try to insert it case-insensitively as a token, which
588 fails because duplicates aren't allowed.
590 Solution: break the two purposes into two separate fields. This
591 fixes the problem and likely makes the code easier to read too.
593 Fixes bug #18922. Thanks to John Darrington for bug report and
596 * file-handle-def.c (struct file_handle): New `id' member.
597 (fh_from_name): Rename fh_from_id. Update all callers.
598 (create_handle): New `id' parameter. Update all callers.
599 (fh_create_file): Ditto.
600 (fh_get_id): New function.
602 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
604 * case.c (case_is_null): Change return type to bool.
606 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
610 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
611 reference counting to share data. This makes it easy for
612 valgrind, etc. to find accesses to cases that have been destroyed
613 but have been kept around by another user's ref-count. This often
614 happens when the data set is small enough to find in memory; if a
615 bigger data set that would overflow to disk were used, then data
616 corruption would occur.
618 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
622 * case.c (case_unshare): Make it check internally whether the
623 ref_cnt is greater than 1, so that the callers don't have to.
624 Update callers not to check.
626 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
628 Before, I was thinking that I might want to get rid of reference
629 counting at some point. Now, I'm pretty sure that it's here to
630 stay. Thus, because we have to store the value_cnt anyway for
631 reference-counted cases, we might as well expose it to users.
633 * case.c (case_get_value_cnt): New function.
634 (case_resize): Drop OLD_CNT argument. Update all callers. Only
635 resize case if its size actually changed.
637 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
638 instead of peeking inside struct case directly.
639 (casefile_append): Ditto.
641 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
643 Get rid of the inlines for the case functions, which made the
644 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
645 the inlines didn't speed up "make check" at all, which is not a
646 perfect benchmark but seems indicative.)
648 * case.c: Remove #ifdef DEBUGGING...#endif around many function
649 definitions. Remove some assertions on nonnull pointers that were
650 redundant with a pointer dereference soon after in the function.
652 (struct case_data): Move definition here from case.h.
656 (case_data_wr): Ditto.
658 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
660 * automake.mk: Add casedeque.h to sources.
662 * casedeque.h: New file.
664 * procedure.c: (struct dataset) Change lag_count, lag_head,
665 lag_queue member into single struct casedeque member. Update all
666 users to use the casedeque instead.
669 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
671 * procedure.c: Simplify lagged cases interface. Updated all
672 clients--well, the only client--to use the simplified interface.
673 (dataset_n_lag) Removed.
674 (dataset_set_n_lag) Removed.
675 (dataset_need_lag) New function.
677 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
679 * dictionary.c procedure.c: More changes to ensure that callbacks occur
680 whenever appropriate, but only when the dataset/dictionary is in a
683 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
685 * dictionary.c dictionary.h : Added callbacks for change of filter and
686 split variables. Refactored some code to ensure that callbacks get
687 invoked when appropriate.
689 * procedure.c (proc_cancel_temporary_transformations): Make sure that
690 replace_dict callback occurs when permanent_dict replaces the current
693 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
695 * dictionary.c dictionary.h : Added callback for when the weight
696 variable of a dictionary changes.
698 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
700 * dictionary.c dictionary.h : Added replace_source and replace_dict
701 callbacks, and functions to deal with them.
703 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
705 Simplify missing value handling.
707 * missing-values.h (enum mv_class): New type.
708 (enum mv_type): Moved definition into missing-values.c and renamed
709 each MV_* to MVT_*, to distinguish them from the exposed mv_class
710 enums. Updated all uses.
711 (struct missing_values): Changed type of `type' from `enum
712 mv_type' to `int' because the definition is no longer exposed.
714 * missing-values.c (mv_is_value_missing): Add new enum mv_class
715 parameter. Update all callers.
716 (mv_is_num_missing): Ditto.
717 (mv_is_str_missing): Ditto.
718 (mv_is_value_user_missing): Removed. Changed callers to use
720 (mv_is_num_user_missing): Removed. Changed callers to use
722 (mv_is_str_user_missing): Removed. Changed callers to use
724 (mv_is_value_system_missing): Removed. Changed callers to use
726 (mv_set_type): Removed. Changed callers to use mv_clear.
727 (mv_clear): New function.
729 * variable.c (var_is_value_missing): Add new enum mv_class
730 parameter. Update all callers.
731 (var_is_num_missing): Ditto.
732 (var_is_str_missing): Ditto.
733 (var_is_value_user_missing): Removed. Changed callers to use
734 var_is_value_missing.
735 (var_is_num_user_missing): Removed. Changed callers to use
737 (var_is_str_user_missing): Removed. Changed callers to use
739 (var_is_value_system_missing): Removed. Changed callers to use
740 var_is_value_missing.
742 * casefilter.c (struct casefilter): Use enum mv_class in place of
744 (casefilter_variable_missing): Adapt to new member.
745 (casefilter_create): Change signature to take enum mv_class,
748 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
750 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
752 * case-sink.c case-sink.h procedure.c procedure.h
753 storage-stream.c: Now uses the factory.
755 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
757 Make it possible to pull cases from the active file with a
758 function call, instead of requiring indirection through a callback
761 * case-source.h (struct case_source_class): Change ->read function
762 to return a single case, instead of calling a callback function
763 for each case. Change ->destroy function to return an error
766 * case-source.c (free_case_source): Pass along the value returned
767 by the case_source ->destroy function.
769 * procedure.c (struct write_case_data): Removed.
770 (struct dataset): Added some members to track procedure state.
771 (procedure): Optimize the trivial case at this level.
772 (internal_procedure): Re-implement in terms of proc_open,
773 proc_read, proc_close.
774 (proc_open) New function.
775 (proc_read) New function.
776 (proc_close) New function.
777 (write_case) Moved into proc_read.
778 (close_active_file) Moved closing of data source into proc_close.
780 * storage-source.c: Rewrote to conform with modified
781 case_source_class interface.
783 * transformations.c (trns_chain_execute): Added argument to allow
784 starting execution from an arbitrary transformation. Updated
787 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
789 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
791 * sys-file-reader.c (read_display_parameters): Don't assume that
792 MEASURE_* and ALIGN_* have the same values found in system files.
794 * sys-file-writer.c (write_variable_display_parameters): Ditto.
796 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
797 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
798 the value of n_MEASURES, which was off by 1 (at least from my
801 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
803 * dictionary.c dictionary.h vardict.h variable.c: Added optional
804 callbacks which are invoked when the dictionary or its
805 variables are changed.
807 * missing-values.c missing-values.h value-labels.c: Tidied up
808 consistency checks, and made some of them return false
809 instead of assert-failing.
811 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
813 * calendar.c (calendar_days_in_month): New function.
815 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
817 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
818 the value label's width.
820 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
822 * sfm-private.h: Move contents into sys-file-writer.c, which is
823 the only remaining user. Removed Borland C++-specific directives.
825 * sys-file-reader.c: Clean up and rewrite entire file. The
826 rewritten version is simpler and better abstracted, and should be
827 easier to maintain and extend. It avoids using structures to read
828 file data, which is prone to padding variations among compilers.
829 It should also handle non-IEEE 754 system files, although I
830 haven't been able to find any. It has been tested against many
831 .sav files obtained from the Web and found to produce the same
832 results as the earlier version of the code, or in some cases
833 improved results. It is more tolerant of format variations found
836 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
837 member, putting an enum integer_format in its place. New member
838 `float_format'. Changed `compressed' member to type bool.
840 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
842 * dictionary.c (dict_delete_consecutive_vars): New function.
844 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
846 * file-name.c (fn_search_path): Remove prefix arg that was unused
847 by any caller. Updated all callers.
849 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
851 * format.c (fmt_dollar_template): Use user's decimal point
852 character. Add assertion.
854 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
856 * format.c (fmt_dollar_template): New function, based on
857 dollar_format_template from var-type-dialog.c.
859 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
861 * data-out.c (output_scientific): Fix bad assumption that "buf" is
864 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
866 Finish converting struct variable to an opaque type. In this
867 phase, we add remaining setter and getter functions, convert the
868 remaining PSPP code to use them, and do a bunch of cleanup. The
869 resulting changes are pervasive but mostly trivial, and only the
870 notable changes are logged.
872 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
875 * case.c (case_data): Renamed case_data_idx.
876 (case_num): Renamed case_num_idx.
877 (case_str): Renamed case_str_idx.
878 (case_data_rw): Renamed case_data_rw_idx.
880 * case.h (case_data): New function with old name and an interface
881 that takes a variable instead of an index, which is easier to
882 use. Updated all callers to use the new interface, or to use the
883 new *_idx function (see above).
886 (case_data_rw): Ditto.
888 * category.c (cat_stored_values_destroy): Changed interface to
889 take a struct cat_vals * instead of a struct variable *.
891 * dictionary.c (dict_clone): Use new vector_clone function.
892 (dict_clear) Use new var_destroy function.
893 (add_var) New function.
894 (dict_create_var) Rewrite in terms of dict_create_var_assert.
895 (dict_create_var_assert) Rewrite in terms of add_var.
896 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
897 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
898 (dict_lookup_var) Use new var_create, var_destroy functions.
899 (dict_contains_var) Rewrite in terms of new vardict functionality.
900 (set_var_dict_index) New function.
901 (set_var_case_index) New function.
902 (reindex_vars) New function.
903 (dict_delete_var) Rewrite in terms of new vardict functionality.
904 (dict_reorder_var) Ditto.
905 (dict_reorder_vars) Ditto.
906 (rename_var) New function.
907 (dict_rename_var) Use rename_var.
908 (dict_rename_vars) Use pool to simplify code. Use rename_var.
909 (dict_get_compacted_idx_to_fv) Rename
910 dict_get_compacted_dict_index_to_case_index, update callers.
911 (dict_create_vector) Use new vector_create function.
912 (dict_clear_vectors) Use new vector_destroy function.
913 (set_var_short_name_suffix) Move here from variable.c, renamed
914 from var_set_short_name_suffix, make static, update caller.
916 * sys-file-private.c: New file.
917 (sfm_width_to_bytes) Moved here from variable.c, renamed from
918 width_to_bytes, update callers.
920 * sys-file-private.h: New file. Later it will supplant
921 sfm-private.h; for now it supplements it.
922 (macro MIN_VERY_LONG_STRING) New macro.
923 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
925 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
926 MAX_LONG_STRING was used before.
928 * sys-file-writer.c: Ditto.
930 * value-labels.c: Change the paradigm here to be that a null
931 pointer is OK for a struct val_labs * in most cases; it just
932 represents an empty set of value labels.
933 (val_labs_copy) A copy of a null set is a null set.
934 (val_labs_count) A null set has 0 labels.
935 (val_labs_replace) Change return type to void. Rewrite for
937 (val_labs_find) A null set does not contain the value.
938 (value_to_string) Moved to variable.c, renamed var_get_value_name,
939 transposed argument order, updated all callers.
942 (value_dup) Moved here from variable.c.
943 (compare_values) Ditto.
946 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
947 (macro MAX_LONG_STRING) Removed, because it was only interesting
948 for system files, not for general code.
949 (macro MAX_VERY_LONG_STRING) Ditto.
950 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
951 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
953 * vardict.h: New file, for an interface between variables and
956 * variable.c: A lot of functions were moved around, for better
958 (struct variable) Move definition here, from variable.h.
959 (var_type_adj) Removed--makes i18n hard.
960 (var_type_noun) Ditto.
961 (var_create) New function.
962 (var_clone) New function.
963 (var_destroy) New function.
964 (var_set_name) Assert that variable is not in a dictionary.
965 (compare_var_names) Rename compare_vars_by_name and fix a couple
966 of callers who thought the args were strings.
967 (hash_var_name) Rename hash_var_by_name.
968 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
969 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
970 (var_is_very_long_string) Removed, because it was only interesting
972 (var_set_missing_values) Allow the argument to be the wrong width,
973 as long as we can resize it. Simplify callers who were doing the
975 (var_get_value_labels) New function.
976 (var_has_value_labels) New function.
977 (var_set_value_labels) New function.
978 (alloc_value_labels) New function.
979 (var_add_value_label) New function.
980 (var_replace_value_label) New function.
981 (var_clear_value_labels) New function.
982 (var_lookup_value_label) New function.
983 (var_get_value_name) Moved here from variable.c, renamed from
984 var_get_value_name, transposed argument order, updated all
986 (var_to_string) Moved here, from variable-label.c.
987 (var_set_leave) New function.
988 (var_get_leave) New function.
989 (var_must_leave) New function.
990 (var_set_short_name_suffix) Moved to dictionary.c, renamed
991 set_var_short_name_suffix.
992 (var_get_dict_index) New function.
993 (var_get_case_index) New function.
994 (var_get_obs_vals) New function.
995 (var_set_obs_vals) New function.
996 (var_has_obs_vals) New function.
997 (var_get_vardict) New function.
998 (var_set_vardict) New function.
999 (var_has_vardict) New function.
1000 (var_clear_vardict) New function.
1001 (value_dup) Moved to value.c.
1002 (compare_values) Ditto.
1005 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1006 (enum ALPHA) Rename VAR_STRING, update all users.
1008 * vector.c: New file.
1009 (struct vector) Moved here, from variable.h.
1010 (check_widths) New function.
1011 (vector_create) New function.
1012 (vector_clone) New function.
1013 (vector_destroy) New function.
1014 (vector_get_name) New function.
1015 (vector_get_var) New function.
1016 (vector_get_var_cnt) New function.
1017 (compare_vector_ptrs_by_name) New function.
1019 * vector.h: New file.
1021 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1023 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1024 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1027 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1029 * value-labels.c (destroy_atoms): New function.
1030 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1032 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1034 Thanks to Jason Stover for pointing out this problem.
1036 * data-out.c (output_number): Use gsl_finite from GSL, which is
1037 portable, instead of isfinite, which is not.
1040 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1042 * variable.c variable.h (value_dup): New function.
1044 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1046 Start converting struct variable to an opaque type. In this
1047 phase, we add a bunch of setter and getter functions and convert
1048 most of the PSPP code to use them. The resulting changes are
1049 pervasive but mostly trivial, and only the notable changes are
1052 * format.c (fmt_equal): New function.
1054 * variable.c (var_type_is_valid): New function.
1055 (measure_is_valid) Moved here, from format.c.
1056 (alignment_is_valid) Moved here, from format.c.
1057 (var_get_name) New function.
1058 (var_set_name) New function.
1059 (width_to_type) New function.
1060 (var_get_type) New function.
1061 (var_get_width) New function.
1062 (var_set_width) New function.
1063 (var_is_numeric) New function.
1064 (var_is_alpha) New function.
1065 (var_is_short_string) New function.
1066 (var_is_long_string) New function.
1067 (var_is_very_long_string) New function.
1068 (var_get_missing_values) New function.
1069 (var_set_missing_values) New function.
1070 (var_clear_missing_values) New function.
1071 (var_has_missing_values) New function.
1072 (var_is_value_missing) New function.
1073 (var_is_num_missing) New function.
1074 (var_is_str_missing) New function.
1075 (var_is_value_user_missing) New function.
1076 (var_is_num_user_missing) New function.
1077 (var_is_str_user_missing) New function.
1078 (var_is_value_system_missing) New function.
1079 (var_get_print_format) New function.
1080 (var_set_print_format) New function.
1081 (var_get_write_format) New function.
1082 (var_set_write_format) New function.
1083 (var_set_both_formats) New function.
1084 (var_get_label) New function.
1085 (var_set_label) New function.
1086 (var_clear_label) New function.
1087 (var_has_label) New function.
1088 (var_get_measure) New function.
1089 (var_set_measure) New function.
1090 (var_get_display_width) New function.
1091 (var_set_display_width) New function.
1092 (var_get_alignment) New function.
1093 (var_set_alignment) New function.
1094 (var_get_value_cnt) New function.
1095 (var_get_leave) New function.
1096 (var_get_short_name) New function.
1098 * variable.h: (struct variable) Removed "type" and "nv" members;
1099 they are now computed from "width" where needed.
1101 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1103 * missing-values.c (mv_resize): Don't write beyond end of the
1104 allocated buffer when resizing a long string.
1106 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1108 Clean up identifier code: don't require identifier enumerations to
1109 be in a particular order; make better use of string library;
1110 expose less of the internals.
1112 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1113 change interface. Updated all callers.
1114 (lex_id_match) Change interface to use struct substring, update
1116 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1117 (global array keywords[]) Make static, change form. Update all
1118 users to use lex_id_name instead.
1119 (lex_is_keyword) New function.
1120 (lex_id_to_token) Change interface to use struct substring, update
1122 (lex_id_name) New function.
1124 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1125 lex_is_keyword instead.
1126 (T_LAST_KEYWORD) Removed.
1127 (T_N_KEYWORDS) Removed.
1129 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1131 * format.c: (fmt_date_template) Distinguish characters for which a
1132 space is output and any date delimiter is allowed on input, from
1133 those for which a space is output and only a space is allowed on
1134 input. The former is represented by X, the latter by a space.
1135 Also, drop distinction between h and H, changing the former to the
1138 * data-in.c: Completely rewrite internals to conform to SPSS input
1139 formats as closely as possible.
1140 (data_in) Changed external interface by replacing the structure
1141 that was used as a single argument by a set of arguments. Updated
1143 (data_in_finite_line) Removed. Converted all callers to use plain
1145 (data_in_get_integer_format) New function.
1146 (data_in_set_integer_format) New function.
1147 (data_in_get_float_format) New function.
1148 (data_in_set_float_format) New function.
1150 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1151 (struct data_in) Removed.
1153 * data-out.c: (output_date) Drop each component from the input as
1154 it is output, to allow us to drop the distinction between h (a
1155 count of hours) and H (the hour of day) template characters.
1156 Also, handle new X template character.
1157 (output_scientific) Follow more rational rule on when to drop
1158 fraction introduced between SPSS 13 and 15. Updated test case to
1161 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1163 Fix buffer overflow reported by John Darrington.
1165 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1166 realize that DIGITS is a count of nibbles, not of bytes.
1168 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1170 * calendar.c (calendar_offset_to_gregorian) Also return the
1171 year-of-day. Change callers to new interface.
1173 * data-out.c: Completely rewrite internals to conform to SPSS
1174 output formats as completely as possible.
1175 (data_out) Change interface to put input parameters before output
1176 parameters, for consistency with the style I now prefer. Update
1178 (data_out_get_integer_format) New public function.
1179 (data_out_set_integer_format) New public function.
1180 (data_out_get_float_format) New public function.
1181 (data_out_set_float_format) New public function.
1183 * data-out.h: New file. Move prototype for data_out here, from
1186 * format.c: (fmt_step_width) Use equality comparison instead of
1187 bitwise and, for clarity.
1188 (fmt_is_string) Ditto.
1189 (fmt_input_to_output) Fix categories that are translated to F
1192 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1194 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1195 casefile-private.h: Added casefilter to assist commands with missing
1198 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1200 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1202 * settings.c: (route_errors_to_terminal) New variable.
1203 (route_errors_to_listing) New variable.
1204 (get_error_routing_to_terminal) New function.
1205 (set_error_routing_to_terminal) New function.
1206 (get_error_routing_to_listing) New function.
1207 (set_error_routing_to_listing) New function.
1209 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1211 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1213 * format.c: Completely rewrite, to achieve better abstraction.
1214 Rewrite all references to formats in other files.
1216 * format.def: Rewrite and reorganize.
1218 * settings.c: Move everything related to custom currency formats
1219 into format.[ch], changing them in form, so as to group related
1220 code and definitions better. Changed all references to use the
1222 (static var decimal) Removed.
1223 (static var grouping) Removed.
1224 (static var cc) Removed.
1225 (get_decimal) Removed.
1226 (set_decimal) Removed.
1227 (get_grouping) Removed.
1228 (set_grouping) Removed.
1232 * settings.h: (macro CC_CNT) Removed.
1233 (macro CC_WIDTH) Removed.
1234 (struct custom_currency) Removed.
1236 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1238 * data-in.c (data_in): Use switch statement instead of table, to
1239 avoid dependence on the order of the FMT_* enums.
1241 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1243 * data-out.c: (num_to_string) Removed, because it was dead code.
1245 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1247 * data-in.c (parse_trailer): Fix error message.
1249 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1251 * format.c (fmt_is_binary): New function.
1253 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1255 * procedure.c procedure.h: Encapsulated the static data into a single
1258 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1260 * casefile.c (casereader_read_xfer): Always initialize the case,
1261 even on an error condition.
1263 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1265 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1266 entire function useless.
1268 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1270 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1271 casereader method casereader_clone.
1273 * procedure.c pransformations.h: Introduced new type casenum_t
1275 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1277 * variable.c: (width_to_bytes) Rephrase code for clarify.
1279 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1281 * format.c: (fmt_type_from_string) New function.
1282 (fmt_to_string) Include decimals in output if the format has
1283 decimals, even if the format type does not. This way, we can
1284 accurately reproduce incorrect formats in user output.
1285 (check_common_specifier) Make the check for a bad format type an
1286 assertion, so we get bug reports if they show up. Fix message.
1287 Check for decimal places with a format type that doesn't allow
1289 (check_input_specifier) Remove check for FMT_X, which has been
1291 (check_output_specifier) Ditto.
1293 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1295 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1296 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1297 (enum fmt_parse_flags) Removed.
1299 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1301 * casefile.c casefile.h: Converted to an abstract base class.
1302 * casefile-private.h fastfile.c fastfile.h: New files.
1303 * automake.mk procedure.c scratch-writer.c storage-stream.c
1305 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1307 * procedure.c (internal_procedure): Create sink_case with only as
1308 many values as the compacted dictionary.
1310 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1312 Remove "debugging" code that caused plenty of false positives and
1315 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1317 * case.c: Remove all references to `this' member.
1319 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1321 Fix link error noted by Jason Stover.
1323 * storage-stream.c: Include <assert.h>.
1325 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1327 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1328 ALL) and additional underlying system file issues.
1330 Thanks to John Darrington for review.
1332 First problem: var_hash points to variables not owned by the
1333 sys-file-reader, which the caller may free or modify. Use an
1334 array of sfm_vars instead, as done earlier (e.g. CVS version
1337 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1338 members and remove all code that references it. Add vars, var_cnt
1339 members. Remove fix_specials member, which was unused.
1340 (struct sfm_var) Remove name member, which was unused.
1341 (sfm_close_reader) Free vars member instead of var_hash.
1342 (compare_var_shortnames) Removed.
1343 (hash_var_shortname) Removed.
1344 (sfm_open_reader) Fill out vars array.
1345 (compare_var_index) Removed.
1346 (sfm_read_case) Use vars instead of var_hash.
1348 Second problem: we're confused about when we actually have very
1349 long strings, causing us to choose incorrectly between slow path
1350 and fast path in sfm_read_case.
1352 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1353 have very long strings, not when we have long variable names,
1354 which is an unrelated feature.
1356 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1358 * variable.h: Move var_set and variable parsing declarations to
1359 new header, src/language/lexer/variable-parser.h. Modified lots
1360 of files to include the new header.
1362 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1364 * value-labels.c (value_to_string): When there's no value label,
1365 format the variable according to its print format, instead of
1366 always effectively using A or F format.
1368 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1370 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1373 * format.c: Removed tortological assertion.
1375 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1377 Reform string library.
1379 * file-name.c (fn_interp_vars): Change interface to take a
1380 substring as input. Updated all users.
1382 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1384 * format.c (measure_is_valid): Really return false when m >=
1387 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1389 Implement random access to casefiles, for use in GUI.
1391 * casefile.c: (struct casereader) Add `random', `file_ofs',
1392 `buffer_ofs' members.
1393 (casefile_get_random_reader) New function.
1394 (read_open_file) Break part into new function
1395 seek_and_fill_buffer().
1396 (fill_buffer) Update buffer_ofs, file_ofs.
1397 (casereader_seek) New function.
1399 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1401 * settings.c: Added call to i18n{done, init}.
1403 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1405 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1408 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1410 * casefile.c: Convert many uses of `int' to `bool'.
1412 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1414 * transformations.c (trns_chain_destroy): Destroy chain's trns
1415 member, to fix memory leak.
1417 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1419 * storage-stream.c (storage_source_decapsulate): Destroy case
1420 source to fix memory leak.
1422 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1424 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1425 case passed as argument instead of initializing the argument as a
1426 case. Fixes memory leak that showed up in
1427 tests/command/aggregate.sh with scratch files.
1429 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1431 * procedure.c (proc_done): Destroy default_dict, to fix memory
1434 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1436 Simplify procedure_with_splits().
1438 * procedure.c (struct split_aux_data): Removed case_count member.
1439 (procedure_with_splits) Don't initialize case_count.
1440 (split_procedure_case_func) Check whether prev_case is null
1441 instead of case_count.
1442 (split_procedure_end_func) Ditto.
1444 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1446 * case.c (case_move): Do nothing if dst and src are the same
1448 (case_try_create) Merge two similar cases.
1449 (case_copy) Unshare only if data must be actually copied.
1451 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1453 * data-in.c data-out.c dictionary.c sys-file-reader.c
1454 sys-file-writer.c variable.c variable.h: Reworked very long string
1455 support for better encapsulation.
1457 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1459 * value-labels.c (val_labs_can_set_width): New function.
1460 (val_labs_set_width) Clear labels if increasing width to long
1462 (val_labs_destroy) Remove unneeded test for null.
1464 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1466 * value-labels.h: Remove unneeded dependency on variable.h.
1468 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1470 Get rid of `char *c' member in union value, for cleanliness.
1472 * value.h: (union value) Remove `c' member.
1474 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1476 Make dictionary compacting functions a little more general.
1478 * sys-file-writer.c (sfm_open_writer): Use
1479 dict_compacting_would_change().
1480 (does_dict_need_translation) Removed.
1482 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1484 Make dictionary compacting functions a little more general.
1486 * dictionary.c (dict_needs_compaction): Rename
1487 dict_compacting_would_shrink(). Update all callers.
1488 (dict_compacting_would_change) New function.
1490 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1492 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1493 inverted return value (!).
1495 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1497 Continue reforming procedure execution.
1499 * procedure.c: Search and replace "vfm" by "proc". Notably:
1500 (static var vfm_source) Rename proc_source. Update all
1502 (static var vfm_sink) Rename proc_sink. Update all references.
1504 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1506 Continue reforming procedure execution. In this phase, remove
1507 PROCESS IF, which was deprecated anyway and can be easily
1508 simulated with TEMPORARY followed by SELECT IF.
1510 * procedure.c: (open_active_file) Don't call
1511 add_process_if_trns().
1512 (discard_variables) Get rid of redundant call to
1513 proc_cancel_all_transformations().
1514 (add_process_if_trns) Removed.
1515 (process_if_trns_proc) Removed.
1516 (process_if_trns_free) Removed.
1518 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1520 Continue reforming procedure execution. In this phase, add
1521 `const' to the case passed to procedure()'s callback.
1523 Updated all users of procedure() as well.
1525 * procedure.c: (struct write_case_data) Add "const" to ccase
1526 parameter for case_func member.
1527 (procedure) Add "const" to ccase parameter for proc_func
1529 (multipass_case_func) Make ccase parameter const.
1530 (internal_procedure) Add "const" to ccase parameter for case_func
1532 (split_procedure_case_func) Make ccase parameter const.
1533 (multipass_split_case_func) Make ccase parameter const.
1535 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
1537 Continue reforming procedure execution. In this phase, get rid of
1538 the output code for SPLIT FILE groups in procedure.c, which really
1539 shouldn't be doing any output. Move it into the individual
1540 procedures instead. This also adds some flexibility.
1542 Updated many users of procedure_with_splits() and
1543 multipass_procedure_with_splits() to call
1544 output_split_file_values() and to deal with increased use of
1547 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
1548 parameter to begin_func member.
1549 (procedure_with_splits) Add "const struct ccase *" parameter to
1550 begin_func parameter. Make ccase parameter const in proc_func
1552 (split_procedure_case_func) Don't dump split file group. Pass
1554 (dump_splits) Moved to language/dictionary/split-file.c as
1555 output_split_file_values().
1556 (struct multipass_split_aux_data) Add "const struct ccase *"
1557 parameter to split_func member.
1558 (multipass_procedure_with_splits) Add "const struct ccase *"
1559 parameter to split_func parameter.
1560 (multipass_split_case_func) Save new SPLIT FILE case before
1562 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
1564 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
1566 Continue reforming procedure execution. Change
1567 internal_procedure() so that it calls open_active_file() and
1568 close_active_file(), which isolates most of the actual procedure
1571 * procedure.c: (struct write_case_data) Rename `proc_func' member
1572 to `case_func' and update all references.
1573 (procedure) Rewrite as one-line wrapper around
1574 internal_procedure().
1575 (struct multipass_aux_data) New.
1576 (multipass_callback) Renamed multipass_case_func(). Use struct
1577 multipass_aux_data as auxiliary data.
1578 (multipass_end_func) New function.
1579 (multipass_procedure) Rewrite as wrapper for internal_procedure()
1580 that uses multipass_case_func, multipass_end_func.
1581 (internal_procedure) Add `end_func' argument. Move optimization
1582 of trivial case in here. Move call to open_active_file() and
1583 close_active_file() in here. Now assert that vfm_source is
1585 (procedure_with_splits_callback) Rename
1586 split_procedure_case_func().
1587 (split_procedure_end_func) New function.
1588 (multipass_split_callback) Rename multipass_split_case_func.
1589 (multipass_split_end_func) New function.
1590 (discard_variables) No need to test for nonnull vfm_source.
1592 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
1594 Continue reforming procedure execution. Get rid of unused member.
1596 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
1598 (write_case) Don't increment cases_analyzed.
1600 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
1602 Continue reforming procedure execution. In this phase, move
1603 procedure.c and procedure.h from src to src/data. Update
1604 makefiles and #includes accordingly.
1606 * procedure.c: Moved here from src/.
1608 * procedure.h: Moved here from src/.
1610 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
1612 Continue reforming procedure execution. In this phase, get rid of
1613 many global variables, consolidating procedure execution in
1614 procedure.c. Encapsulate transformations in new "struct
1615 trns_chain". Also, change implementation of N OF CASES, FILTER,
1616 and PROCESS IF from special cases to transformations.
1618 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
1621 * dictionary.c: (global variable default_dict) Move to
1624 * variable.h: (TRNS_*) Move to transformations.h.
1625 (struct transformation) Move to transformations.c.
1627 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1629 * sys-file-reader.c: Fixed invalid read problems.
1631 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
1633 * storage-stream.c: Add missing function comments.
1635 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
1637 Continue reforming procedure execution. In this phase, add some
1638 new, needed functionality to storage-stream.
1640 * storage-stream.c: (storage_source_decapsulate) New function.
1642 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
1644 * variable.c (width_to_bytes): Declarations must precede
1645 statements for C90 compliance.
1647 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
1649 * data-out.c, data-in.c, variable.c, variable.h: New functions
1650 copy_mangle and copy_demangle for reading/writing cases; emulates the
1651 way SPSS deals with strings > 255 bytes.
1653 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
1654 subtype 14 needed for strings longer than 255 bytes.
1656 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
1657 instead of literal values. Also fixed some constness issues.
1659 * format.h: Constness
1661 * sfm-private.h: Renamed the case_size identifier, since I discovered
1662 that SPSS's respect for this variable is very nominal.
1664 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
1666 Change case limit type from int to size_t.
1668 * dictionary.c: (struct dictionary) Change type of case_limit
1670 (dict_get_case_limit) Change return type.
1671 (dict_set_case_limit) Change parameter type.
1673 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
1675 * variable.h: (struct variable) Rename `reinit' member as `leave'
1676 and invert sense. Fix up all references.
1678 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
1680 Continue reforming procedure execution. In this phase, break
1681 procedure.c into multiple files.
1683 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
1685 * case-sink.c: New file.
1687 * case-sink.h: New file.
1689 * case-source.c: New file.
1691 * case-source.h: New file.
1693 * storage-stream.c: New file.
1695 * storage-stream.h: New file.
1697 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
1699 * variable.h: (struct variable) Remove `init' member and all
1700 references to it from other files. It was initialized in several
1701 places, but nothing really ever used it for anything worthwhile.
1702 Thanks to Jason Stover for pointing out how confusing this
1705 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
1707 Continue reforming error message support. In this phase, get rid
1708 of message "titles" and put the message text in `struct error'.
1709 Now `struct error' encapsulates a message more properly.
1711 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
1712 Format message ourselves.
1714 * data-in.c: (vdls_error) Ditto.
1716 * por-file-reader.c: (error) Ditto.
1718 * sys-file-reader.c: (corrupt_msg) Ditto.
1720 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
1722 GNU standards require "file name" instead of "filename" in
1723 documentation. It's nice for our code to follow the convention
1726 * casefile.c: (struct casefile) Rename `filename' member to
1727 `file_name'. Updated all references.
1729 * file-name.c: [!unix] (struct file_identity) Rename
1730 normalized_filename member to normalized_file_name. Updated all
1733 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
1735 We don't really support anything but Unix-like environments well,
1736 so we might as well de-obfuscate by writing directory and path
1737 separators explicitly.
1739 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
1741 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
1742 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
1744 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
1747 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
1749 GNU standards require "file name" instead of "filename" in
1750 documentation. It's nice for our code to follow the convention
1753 * filename.c: Rename to file-name.c.
1755 * filename.h: Rename to file-name.h. Update all inclusions.
1756 Update header guards.
1758 * automake.mk: Update file names.
1760 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
1762 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
1764 (fn_basename) Removed (dead code).
1765 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
1766 (fn_special_p) Renamed fn_is_special(), all references updated.
1767 (fn_exists_p) Renamed fn_exists(), all references updated.
1769 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
1771 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
1772 Also, now it only tilde-expands file names, not paths.
1773 (fn_search_path) Tilde-expand one directory at a time.
1775 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
1777 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
1778 get rid of non-Unixlike version of the code, which has probably
1780 (fn_prepend_dir) Removed (dead code).
1782 * filename.h: (macro DIR_SEPARATOR_STRING) New.
1783 (macro PATH_SEPARATOR_STRING) New.
1784 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
1786 Continue reforming error message support. In this phase, we get
1787 rid of VM() and the other msg() support for "verbosity", replacing
1788 it by a new function verbose_msg().
1790 * filename.c: (fn_search_path) Use verbose_msg() instead of
1793 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
1795 * sfm-private.h: Get rid of #defines after #error, which makes no
1798 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
1800 Get rid of our own int32 type in favor of the standard int32_t
1803 * sfm-private.h: (int32 macro) Don't define this anymore. Do
1806 * sys-file-reader.c: Use int32_t instead of int32 throughout.
1808 * sys-file-writer.c: Use int32_t instead of int32 throughout.
1810 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
1812 Remove ill-considered file routines that are no longer used.
1814 * filename.c: (fn_open_ext) Removed.
1815 (fn_close_ext) Removed.
1817 * filename.h: (struct file_ext) Removed.
1819 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
1821 * variable.c (var_is_valid_name): Move declarations before code
1824 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
1826 * filename.ch (fn_interp_vars): Fixed small buglet.
1828 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
1830 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
1831 so as to modify the string inline. Thus makeing it easier to
1832 destroy the results when no longer needed.
1834 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
1836 * category.c (cat_stored_values_destroy): Fixed memory leak.
1838 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
1840 Add some missing frees. Thanks to John Darrington for reporting
1843 * any-writer.c (any_writer_close): Free writer.
1845 * any-reader.c (any_reader_close): Free reader.
1847 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
1849 * por-file-reader.c: (error) Mark as NO_RETURN.
1851 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
1853 * settings.c: Changed default value of scompress to true.
1855 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
1857 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
1859 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
1862 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
1863 here from pref.h.orig.
1865 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
1867 * sys-file-reader.c: Fixed bug reading compressed files.
1869 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1871 * Numerous renames. See src/ChangeLog for details.
1873 * Moved files from src directory