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.
10 Reviewed by John Darrington.
12 2007-07-24 Ben Pfaff <blp@gnu.org>
15 * por-file-writer.c (write_variables): Write weight variable.
16 Reviewed by John Darrington.
18 2007-07-23 Ben Pfaff <blp@gnu.org>
20 Improvements to system file reader and writer.
22 First, move all detailed knowledge of very long strings into
23 sys-file-private.[ch], so that this nasty stuff can be isolated.
25 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
26 (EFFECTIVE_VLS_CHUNK): New macro.
27 (min_int): New function.
28 (max_int): New function.
29 (sfm_width_to_bytes): Rewrite.
30 (sfm_width_to_octs): New function.
31 (sfm_segment_alloc_width): New function.
32 (sfm_segment_alloc_bytes): New function.
33 (sfm_segment_used_bytes): New function.
34 (sfm_segment_offset): New function.
35 (sfm_segment_effective_offset): New function.
36 (sfm_dictionary_to_sfm_vars): New function.
38 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
39 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
40 (struct sfm_var): New structure.
42 Next, improvements to the system file reader.
44 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
45 type casenumber. Added `version_major', `version_minor',
48 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
49 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
50 Change `case_cnt' to type casenumber. Removed `has_vls'.
51 (struct sfm_var): Removed.
52 (sfm_open_reader): Don't warn on wrong case size if the file was
53 written by SPSS 13, which tends to get it wrong. Use
54 sfm_dictionary_to_sfm_vars.
55 (read_header): Always output system file info.
56 (read_variable_record): Simplify code for reading missing values.
57 (read_machine_int32_info): Save version numbers from system file
58 into info struct passed as new argument.
59 (read_long_string_map): Restructured to use new sys-file-private
61 (read_value_labels): Use size_overflow_p.
62 (sys_file_casereader_read): Get rid of distinction between fast
63 and slow paths. Use information provided by sys-file-primate's
64 struct sfm_var to simplify code.
65 (skip_whole_strings): New function.
66 (read_int32): Renamed read_int. Changed return value to int.
68 (read_flt64): Renamed read_float. Changed return value to
69 double. Updated all callers.
70 (int32_to_native): Removed. Changed callers to use
72 (flt64_to_double): Removed. Changed callers to use float_convert.
74 Finally, get rid of int32, flt64 terminology and types in system
75 file writer. The former wasn't very useful since a POSIX "int"
76 can hold the whole range of int32 and we generally didn't have a
77 need for it to be exactly-32-bits, just at-least-32-bits. The
78 latter was inconvenient because we had to assume that it could be
79 different from double and thereby convert special values SYSMIS,
80 HIGHEST, LOWEST to and from it in multiple places. Instead, now
81 we just use "int" and "double" in most places, and do conversions,
82 if necessary, very close to where we do I/O. This change meant
83 that the writer code couldn't represent records in the file as C
84 structs any longer, but that's no great loss. The code actually
85 seems to be more readable without them.
87 Simplify the compression buffering code: only buffer as much as
88 necessary, which is no more than eight 8-byte units at any given
91 * sys-file-writer.c (typedef flt64): Removed.
92 (macro second_lowest_flt64): Removed.
93 (struct sysfile_header): Removed.
94 (struct sysfile_variable): Removed.
95 (struct sfm_writer): Removed `needs_translation', `has_vls',
96 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
97 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
98 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
99 compression buffering by `opcodes', `opcode_cnt', `data',
100 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
101 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
102 data writer functions instead of structures.
103 (calc_oct_idx): New function.
104 (write_header): Use simple data writer functions instead of
106 (write_format_spec): Renamed write_format. New argument.
107 (write_variable_continuation_records): New function.
108 (write_variable): Use simple data writer functions instead of
109 structures. Use write_variable_continuation_records. Write
110 entire very long string instead of requiring caller to understand
112 (write_value_labels): Use simple data writer functions instead of
114 (write_documents): Ditto.
115 (write_variable_display_parameters): Use sys-file-private
116 functions to simplify. Use simple data writer functions instead
118 (write_vls_length_table): Use simple data writer functions instead
120 (write_longvar_table): Ditto.
121 (write_rec_7_34): Break into new functions
122 write_integer_info_record, write_float_info_record. Use simple
123 data writer functions instead of structures.
124 (buf_write): Removed.
125 (append_string_max): Removed.
126 (ensure_buf_space): Removed.
127 (sys_file_casewriter_write): Get rid of the distinction between
128 fast and slow paths, which didn't seem to be too useful. Use new
129 functions write_case_uncompressed, write_case_compressed.
130 (put_instruction): Removed.
131 (put_element): Removed.
132 (write_compressed_data): Removed.
133 (close_writer): Use flush_compressed. Only write case count to
134 system file if it will fit in the field.
135 (write_case_compressed): New function.
136 (write_case_uncompressed): New function.
137 (flush_compressed): New function.
138 (put_cmp_opcode): New function.
139 (put_cmp_number): New function.
140 (write_int): New function.
141 (convert_double_to_output_format): New function.
142 (write_float): New function.
143 (write_value): New function.
144 (write_string): New function.
145 (write_bytes): New function.
146 (write_zeros): New function.
147 (write_spaces): New function.
149 Reviewed by John Darrington.
151 2007-07-22 Ben Pfaff <blp@gnu.org>
153 Don't try to write very long strings to portable files. The
154 format does not support it.
156 * por-file-writer.c (MAX_POR_WIDTH): New macro.
157 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
158 (write_format): Add arg to take width to resize format to.
159 (write_value): Limit width of value written to MAX_POR_WIDTH.
160 (write_variables): Limit width of variable and its output formats
162 Reviewed by John Darrington.
164 2007-07-22 Ben Pfaff <blp@gnu.org>
166 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
167 local variable instead of literal 5.
168 Reviewed by John Darrington.
170 2007-07-22 Ben Pfaff <blp@gnu.org>
172 Fix problems with uniqueness of short names in system files with
173 very long string variables. Now a variable may have multiple
176 * automake.mk (src_data_libdata_a_SOURCES): Add new files
177 short-names.c, short-names.h.
179 * dictionary.c (dict_clone): Clone all the short names.
180 (compare_strings): Move into short-names.c.
181 (hash_strings): Ditto.
182 (set_var_short_name_suffix): Ditto.
183 (dict_assign_short_names): Ditto, rename short_names_assign,
184 change to assign all short names.
186 * por-file-writer.c (write_variables): Use short_names_assign
187 instead of dict_assign_short_names.
189 * short-names.c: New file.
191 * short-names.h: New file.
193 * sys-file-private.c (sfm_width_to_segments): New function.
195 * sys-file-reader.c (read_long_var_name_map): Save and restore all
196 the short names, not just the first one.
198 * sys-file-writer.c (cont_var_name): Removed.
199 (sfm_open_writer): Use short_names_assign instead of
200 dict_assign_short_names. Use unique short names assigned by
201 short_names_assign instead of those generated by cont_var_name.
203 * variable.c (struct variable): Remove `short_name' member,
204 replace by `short_names' and `short_name_cnt'.
205 (var_create) Initialize new members.
206 (var_get_short_name_cnt): New function.
207 (var_get_short_name): Now takes an index argument. Changed most
209 (var_set_short_name): Ditto.
210 (var_clear_short_name): Renamed var_clear_short_names, changed to
211 clear all short names.
213 Reviewed by John Darrington.
215 2007-07-22 Ben Pfaff <blp@gnu.org>
217 * variable.c (var_set_width): Use new var_set_width function.
219 * missing-values.c (mv_n_values): Drop assertion, which was not
222 * format.c (fmt_default_for_width): New function.
223 (fmt_resize): New function.
225 Reviewed by John Darrington.
227 2007-07-18 John Darrington <john@darrington.wattle.id.au>
229 * datasheet.c (datasheet_delete_columns): Added assertion to check
230 we're not deleting outside the range of the sheet.
233 * dictionary.c dictionary.h variable.c: Added the ability for string
234 variables to be resized.
236 * vardict.h: Added some prototypes (moved from dictionary.h) as
237 these should only be called by variable.c
240 2007-07-14 John Darrington <john@darrington.wattle.id.au>
242 * sfm-reader.c: Respect case_cnt field in file header.
244 2007-07-01 John Darrington <john@darrington.wattle.id.au>
246 * transformation.c transformation.h (trns_chain_execute): Changed the
247 signature (Patch #6057)
249 2007-06-10 Ben Pfaff <blp@gnu.org>
251 * casereader-filter.c (casereader_filter_destroy): Make sure to
252 write all the remaining excluded cases to the casewriter, if any.
254 * caseinit.c (init_list_destroy): Rewrite.
255 (init_list_clear): Ditto.
257 * casegrouper.c (casegrouper_get_next_group): Always set *reader
258 to null when returning false.
260 2007-06-06 Ben Pfaff <blp@gnu.org>
262 Actually implement the new procedure code and adapt all of its
263 clients to match. Also adapt all of the other case sources and
264 sinks in the tree and their clients to use the
265 casereader/casewriter infrastructure.
267 * automake.mk: Add and remove files.
269 * any-reader.c: Change into a casereader.
270 * por-file-reader.c: Ditto.
271 * scratch-reader.c: Ditto.
272 * sys-file-reader.c: Ditto.
274 * any-writer.c: Change into a casewriter.
275 * por-file-writer.c: Ditto.
276 * scratch-writer.c: Ditto.
277 * sys-file-writer.c: Ditto.
279 * procedure.c: Change to use casereader, casewriter, caseinit, and
280 other new infrastructure.
282 * scratch-handle.c: Adapt to new infrastructure.
284 * case-sink.c: Removed, now dead code.
285 * case-sink.h: Ditto.
286 * case-source.c: Ditto.
287 * case-source.h: Ditto.
288 * casefile-factory.c: Ditto.
289 * casefile-private.h: Ditto.
292 * casefilter.c: Ditto.
293 * casefilter.h: Ditto.
296 * fastfile-factory.c: Ditto.
297 * fastfile-factory.h: Ditto.
298 * storage-stream.c: Ditto.
299 * storage-stream.h: Ditto.
301 2007-06-06 Ben Pfaff <blp@gnu.org>
305 * automake.mk: Add new files.
307 * datasheet.c: New file.
309 * datasheet.h: New file.
311 2007-06-06 Ben Pfaff <blp@gnu.org>
313 Until now, the procedure code has provided a case to the
314 case_source, which has filled in the data values that come from
315 the active file. "Left" data values that don't come from the
316 active file naturally stay the same from case to case, because the
317 procedure code keeps using that same case.
319 One of the compromises that comes with the new procedure code is
320 that the active file allocates and provides its own case, which
321 the procedure code then has to resize to provide room for any
322 other variables that should go in the case and then fill in the
323 values of "left" variables. Then, when we're done with that case,
324 we have to save the values of "left" variables to copy into the
325 next case read from the active file.
327 The caseinit code helps with this.
329 * automake.mk: Add new files.
331 * caseinit.c: New file.
333 * caseinit.h: New file.
335 2007-06-06 Ben Pfaff <blp@gnu.org>
337 * value.h (value_cnt_from_width): New function.
339 * variable.c (var_get_value_cnt): Use new function.
341 2007-06-06 Ben Pfaff <blp@gnu.org>
343 Add casegrouper, to allow cases read from a given casereader to be
344 broken into groups, each of which has its own casereader.
345 Generally cases are grouped based on having equal values for some
348 * automake.mk: Add new files.
350 * casegrouper.c: New file.
352 * casegrouper.h: New file.
354 2007-06-06 Ben Pfaff <blp@gnu.org>
356 Add interface to lexicographical ordering of cases.
358 * automake.mk: Add new files.
360 * case-ordering.c: New file.
362 * case-ordering.h: New file.
364 2007-06-06 Ben Pfaff <blp@gnu.org>
366 Add casereaders and casewriters, the basis of the new data processing
367 implementation. A casereader is a uniform interface to reading cases
368 from a data source; a casewriter is a uniform interface to writing
369 cases to a data sink.
371 * automake.mk: Add new files.
373 * casereader-filter.c: New file.
375 * casereader-provider.h: New file.
377 * casereader-translator.c: New file.
379 * casereader.c: New file.
381 * casereader.h: New file.
383 * casewriter-provider.h: New file.
385 * casewriter-translator.c: New file.
387 * casewriter.c: New file.
389 * casewriter.h: New file.
391 2007-06-06 Ben Pfaff <blp@gnu.org>
393 "casewindow" data structure that extends the deque (from libpspp)
394 of cases with the ability to dump cases to disk if we get too many
397 * automake.mk: Add new files.
399 * casewindow.c: New file.
401 * casewindow.h: New file.
403 2007-06-06 Ben Pfaff <blp@gnu.org>
405 sparse_cases data structure that augments a sparse_array of cases
406 with the ability to dump cases to disk if we get too many cases in
409 * automake.mk: Add new files.
411 * sparse-cases.c: New file.
413 * sparse-cases.h: New file.
415 2007-06-06 Ben Pfaff <blp@gnu.org>
417 Adds a low-level on-disk case array data structure.
419 * automake.mk: Add new files.
421 * case-tmpfile.c: New file.
423 * case-tmpfile.h: New file.
425 2007-06-06 Ben Pfaff <blp@gnu.org>
427 In a couple of places we calculate the maximum number of cases to
428 keep in memory based on the user-defined workspace. Enable
429 centralizing the calculation through a new function.
431 * settings.c (get_workspace_cases): New function.
433 2007-06-06 Ben Pfaff <blp@gnu.org>
435 The casenumber type is defined in transformations.h, but case.h is
436 a more sensible place. Move it.
438 * case.h (CASENUMBER_MAX): New macro.
439 (typedef casenumber): Move here, from transformations.h.
441 2007-06-03 Ben Pfaff <blp@gnu.org>
443 Slightly generalize case_to_values and case_from_values functions.
445 * case.c (case_to_values): Rename case_copy_out, change interface.
446 (case_from_values): Rename case_copy_in, change interface.
448 * fastfile.c (fastfilereader_get_next_case): Update caller.
449 (write_case_to_disk): Ditto.
451 2007-06-02 Ben Pfaff <blp@gnu.org>
453 Clean up after a forgotten part of patch #5829.
455 * casedeque.h: Remove unused file.
457 * automake.mk: Remove casedeque.h from sources.
459 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
461 * category.c: Removed redundant #include
463 2007-05-06 Ben Pfaff <blp@gnu.org>
465 Abstract the documents within a dictionary a little better.
466 Thanks to John Darrington for suggestion, initial version, and
469 * dictionary.c (struct dictionary): Change `documents' member from
470 char * to struct string.
471 (dict_clear): Destroy struct string.
472 (dict_get_documents): Convert struct string to char *.
473 (dict_set_documents): Set struct string. Pad to 80-character
475 (dict_clear_documents): New function.
476 (dict_add_document_line): New function.
477 (dict_get_document_line_cnt): New function.
478 (dict_get_document_line): New function.
480 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
482 * sys-file-reader.c (read_documents): Use new document functions.
484 2007-04-19 John Darrington <john@darrington.wattle.id.au>
486 * sys-file-reader.c: When reading a system file which has no
487 long name table, automatically create one where the long names
488 are the lower case versions of the short names.
490 2007-04-22 Ben Pfaff <blp@gnu.org>
492 * dictionary.c (dict_set_split_vars): dict_destroy expects that
493 dict_clear will free most data related to the dictionary.
494 dict_clear does a decent job, except that dict_set_split_vars on
495 some systems won't actually free the dict's "split" member.
496 Instead, it'll allocate a 1-byte region. Fix this.
498 * value.c (value_copy): New function.
499 (value_set_missing): Ditto.
501 2007-04-22 John Darrington <john@darrington.wattle.id.au>
503 * Deleted existing category.h and moved cat-routines.h into
504 category.h Encapsulated struct cat_vals better.
506 2007-04-19 John Darrington <john@darrington.wattle.id.au>
508 * sys-file-reader.c: When reading a system file which has no
509 long name table, automatically create one where the long names
510 are the lower case versions of the short names.
512 2007-04-16 John Darrington <john@darrington.wattle.id.au>
514 * sys-file-reader.c: Some versions of Other Software seem to
515 produce system files with string variables' measure set to
516 zero. We'll assume these are supposed to be nominal variables.
518 2007-03-30 Ben Pfaff <blp@gnu.org>
520 * procedure.c: Adapt to new deque data structure.
522 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
523 Ben Pfaff <blp@gnu.org>
525 * file-name.c: Mingw compatibility fixes.
526 (fn_search_path): Use ISSLASH instead of comparing against '/'
528 (fn_dir_name): Use dir_name from gnulib.
529 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
530 (fn_get_identity): Use GetFullPathName instead of canonicalize
531 from gnulib, because the latter does not fully support
532 Windows-style path names. Use this implementation based on the
533 detected presence of Windows instead of the absence of Unix, since
534 the new implementation is Windows-specific.
535 (fn_compare_file_identities): In Windows implementation, compare
536 names case-insensitively.
538 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
540 * make-file.c: Don't include mkstemp.h, because gnulib now causes
541 <stdlib.h> to have the same effect.
543 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
545 * por-file-reader.c: Add missing _() around messages.
547 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
549 * make-file.c: Include "mkstemp.h", without which linking on
552 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
554 Reduce platform dependence.
556 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
557 Everywhere we using this, we really should have just depended on
558 the shell to expand tildes.
559 (fn_search_path): Simplify, given that we don't do tilde expansion
561 (fn_normalize): Removed. Caller changed to use the canonicalize
563 (fn_get_cwd): Removed. Only user was fn_normalize.
564 (fn_is_absolute): Really only test for absolute names.
565 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
567 (fn_readlink): Removed, as it was only used fn_normalize.
568 (fn_exists): Assume the stat function is available; gnulib does.
569 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
571 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
573 * dictionary.c (dict_create_vector_assert): New function.
575 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
577 * file-name.c (fn_normalize): Correct name of function
578 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
579 for pointing this out.
581 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
583 We are using a single member in struct file_handle, the "name"
584 field, for more than one purpose. When it begins with '"', it's a
585 file name; otherwise, it's a token that can be used to identify
586 it. When that assertion fires, it's because we searched for the
587 name case-sensitively as a file name (so that there was no match),
588 and then we try to insert it case-insensitively as a token, which
589 fails because duplicates aren't allowed.
591 Solution: break the two purposes into two separate fields. This
592 fixes the problem and likely makes the code easier to read too.
594 Fixes bug #18922. Thanks to John Darrington for bug report and
597 * file-handle-def.c (struct file_handle): New `id' member.
598 (fh_from_name): Rename fh_from_id. Update all callers.
599 (create_handle): New `id' parameter. Update all callers.
600 (fh_create_file): Ditto.
601 (fh_get_id): New function.
603 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
605 * case.c (case_is_null): Change return type to bool.
607 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
611 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
612 reference counting to share data. This makes it easy for
613 valgrind, etc. to find accesses to cases that have been destroyed
614 but have been kept around by another user's ref-count. This often
615 happens when the data set is small enough to find in memory; if a
616 bigger data set that would overflow to disk were used, then data
617 corruption would occur.
619 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
623 * case.c (case_unshare): Make it check internally whether the
624 ref_cnt is greater than 1, so that the callers don't have to.
625 Update callers not to check.
627 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
629 Before, I was thinking that I might want to get rid of reference
630 counting at some point. Now, I'm pretty sure that it's here to
631 stay. Thus, because we have to store the value_cnt anyway for
632 reference-counted cases, we might as well expose it to users.
634 * case.c (case_get_value_cnt): New function.
635 (case_resize): Drop OLD_CNT argument. Update all callers. Only
636 resize case if its size actually changed.
638 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
639 instead of peeking inside struct case directly.
640 (casefile_append): Ditto.
642 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
644 Get rid of the inlines for the case functions, which made the
645 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
646 the inlines didn't speed up "make check" at all, which is not a
647 perfect benchmark but seems indicative.)
649 * case.c: Remove #ifdef DEBUGGING...#endif around many function
650 definitions. Remove some assertions on nonnull pointers that were
651 redundant with a pointer dereference soon after in the function.
653 (struct case_data): Move definition here from case.h.
657 (case_data_wr): Ditto.
659 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
661 * automake.mk: Add casedeque.h to sources.
663 * casedeque.h: New file.
665 * procedure.c: (struct dataset) Change lag_count, lag_head,
666 lag_queue member into single struct casedeque member. Update all
667 users to use the casedeque instead.
670 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
672 * procedure.c: Simplify lagged cases interface. Updated all
673 clients--well, the only client--to use the simplified interface.
674 (dataset_n_lag) Removed.
675 (dataset_set_n_lag) Removed.
676 (dataset_need_lag) New function.
678 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
680 * dictionary.c procedure.c: More changes to ensure that callbacks occur
681 whenever appropriate, but only when the dataset/dictionary is in a
684 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
686 * dictionary.c dictionary.h : Added callbacks for change of filter and
687 split variables. Refactored some code to ensure that callbacks get
688 invoked when appropriate.
690 * procedure.c (proc_cancel_temporary_transformations): Make sure that
691 replace_dict callback occurs when permanent_dict replaces the current
694 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
696 * dictionary.c dictionary.h : Added callback for when the weight
697 variable of a dictionary changes.
699 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
701 * dictionary.c dictionary.h : Added replace_source and replace_dict
702 callbacks, and functions to deal with them.
704 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
706 Simplify missing value handling.
708 * missing-values.h (enum mv_class): New type.
709 (enum mv_type): Moved definition into missing-values.c and renamed
710 each MV_* to MVT_*, to distinguish them from the exposed mv_class
711 enums. Updated all uses.
712 (struct missing_values): Changed type of `type' from `enum
713 mv_type' to `int' because the definition is no longer exposed.
715 * missing-values.c (mv_is_value_missing): Add new enum mv_class
716 parameter. Update all callers.
717 (mv_is_num_missing): Ditto.
718 (mv_is_str_missing): Ditto.
719 (mv_is_value_user_missing): Removed. Changed callers to use
721 (mv_is_num_user_missing): Removed. Changed callers to use
723 (mv_is_str_user_missing): Removed. Changed callers to use
725 (mv_is_value_system_missing): Removed. Changed callers to use
727 (mv_set_type): Removed. Changed callers to use mv_clear.
728 (mv_clear): New function.
730 * variable.c (var_is_value_missing): Add new enum mv_class
731 parameter. Update all callers.
732 (var_is_num_missing): Ditto.
733 (var_is_str_missing): Ditto.
734 (var_is_value_user_missing): Removed. Changed callers to use
735 var_is_value_missing.
736 (var_is_num_user_missing): Removed. Changed callers to use
738 (var_is_str_user_missing): Removed. Changed callers to use
740 (var_is_value_system_missing): Removed. Changed callers to use
741 var_is_value_missing.
743 * casefilter.c (struct casefilter): Use enum mv_class in place of
745 (casefilter_variable_missing): Adapt to new member.
746 (casefilter_create): Change signature to take enum mv_class,
749 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
751 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
753 * case-sink.c case-sink.h procedure.c procedure.h
754 storage-stream.c: Now uses the factory.
756 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
758 Make it possible to pull cases from the active file with a
759 function call, instead of requiring indirection through a callback
762 * case-source.h (struct case_source_class): Change ->read function
763 to return a single case, instead of calling a callback function
764 for each case. Change ->destroy function to return an error
767 * case-source.c (free_case_source): Pass along the value returned
768 by the case_source ->destroy function.
770 * procedure.c (struct write_case_data): Removed.
771 (struct dataset): Added some members to track procedure state.
772 (procedure): Optimize the trivial case at this level.
773 (internal_procedure): Re-implement in terms of proc_open,
774 proc_read, proc_close.
775 (proc_open) New function.
776 (proc_read) New function.
777 (proc_close) New function.
778 (write_case) Moved into proc_read.
779 (close_active_file) Moved closing of data source into proc_close.
781 * storage-source.c: Rewrote to conform with modified
782 case_source_class interface.
784 * transformations.c (trns_chain_execute): Added argument to allow
785 starting execution from an arbitrary transformation. Updated
788 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
790 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
792 * sys-file-reader.c (read_display_parameters): Don't assume that
793 MEASURE_* and ALIGN_* have the same values found in system files.
795 * sys-file-writer.c (write_variable_display_parameters): Ditto.
797 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
798 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
799 the value of n_MEASURES, which was off by 1 (at least from my
802 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
804 * dictionary.c dictionary.h vardict.h variable.c: Added optional
805 callbacks which are invoked when the dictionary or its
806 variables are changed.
808 * missing-values.c missing-values.h value-labels.c: Tidied up
809 consistency checks, and made some of them return false
810 instead of assert-failing.
812 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
814 * calendar.c (calendar_days_in_month): New function.
816 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
818 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
819 the value label's width.
821 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
823 * sfm-private.h: Move contents into sys-file-writer.c, which is
824 the only remaining user. Removed Borland C++-specific directives.
826 * sys-file-reader.c: Clean up and rewrite entire file. The
827 rewritten version is simpler and better abstracted, and should be
828 easier to maintain and extend. It avoids using structures to read
829 file data, which is prone to padding variations among compilers.
830 It should also handle non-IEEE 754 system files, although I
831 haven't been able to find any. It has been tested against many
832 .sav files obtained from the Web and found to produce the same
833 results as the earlier version of the code, or in some cases
834 improved results. It is more tolerant of format variations found
837 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
838 member, putting an enum integer_format in its place. New member
839 `float_format'. Changed `compressed' member to type bool.
841 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
843 * dictionary.c (dict_delete_consecutive_vars): New function.
845 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
847 * file-name.c (fn_search_path): Remove prefix arg that was unused
848 by any caller. Updated all callers.
850 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
852 * format.c (fmt_dollar_template): Use user's decimal point
853 character. Add assertion.
855 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
857 * format.c (fmt_dollar_template): New function, based on
858 dollar_format_template from var-type-dialog.c.
860 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
862 * data-out.c (output_scientific): Fix bad assumption that "buf" is
865 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
867 Finish converting struct variable to an opaque type. In this
868 phase, we add remaining setter and getter functions, convert the
869 remaining PSPP code to use them, and do a bunch of cleanup. The
870 resulting changes are pervasive but mostly trivial, and only the
871 notable changes are logged.
873 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
876 * case.c (case_data): Renamed case_data_idx.
877 (case_num): Renamed case_num_idx.
878 (case_str): Renamed case_str_idx.
879 (case_data_rw): Renamed case_data_rw_idx.
881 * case.h (case_data): New function with old name and an interface
882 that takes a variable instead of an index, which is easier to
883 use. Updated all callers to use the new interface, or to use the
884 new *_idx function (see above).
887 (case_data_rw): Ditto.
889 * category.c (cat_stored_values_destroy): Changed interface to
890 take a struct cat_vals * instead of a struct variable *.
892 * dictionary.c (dict_clone): Use new vector_clone function.
893 (dict_clear) Use new var_destroy function.
894 (add_var) New function.
895 (dict_create_var) Rewrite in terms of dict_create_var_assert.
896 (dict_create_var_assert) Rewrite in terms of add_var.
897 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
898 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
899 (dict_lookup_var) Use new var_create, var_destroy functions.
900 (dict_contains_var) Rewrite in terms of new vardict functionality.
901 (set_var_dict_index) New function.
902 (set_var_case_index) New function.
903 (reindex_vars) New function.
904 (dict_delete_var) Rewrite in terms of new vardict functionality.
905 (dict_reorder_var) Ditto.
906 (dict_reorder_vars) Ditto.
907 (rename_var) New function.
908 (dict_rename_var) Use rename_var.
909 (dict_rename_vars) Use pool to simplify code. Use rename_var.
910 (dict_get_compacted_idx_to_fv) Rename
911 dict_get_compacted_dict_index_to_case_index, update callers.
912 (dict_create_vector) Use new vector_create function.
913 (dict_clear_vectors) Use new vector_destroy function.
914 (set_var_short_name_suffix) Move here from variable.c, renamed
915 from var_set_short_name_suffix, make static, update caller.
917 * sys-file-private.c: New file.
918 (sfm_width_to_bytes) Moved here from variable.c, renamed from
919 width_to_bytes, update callers.
921 * sys-file-private.h: New file. Later it will supplant
922 sfm-private.h; for now it supplements it.
923 (macro MIN_VERY_LONG_STRING) New macro.
924 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
926 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
927 MAX_LONG_STRING was used before.
929 * sys-file-writer.c: Ditto.
931 * value-labels.c: Change the paradigm here to be that a null
932 pointer is OK for a struct val_labs * in most cases; it just
933 represents an empty set of value labels.
934 (val_labs_copy) A copy of a null set is a null set.
935 (val_labs_count) A null set has 0 labels.
936 (val_labs_replace) Change return type to void. Rewrite for
938 (val_labs_find) A null set does not contain the value.
939 (value_to_string) Moved to variable.c, renamed var_get_value_name,
940 transposed argument order, updated all callers.
943 (value_dup) Moved here from variable.c.
944 (compare_values) Ditto.
947 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
948 (macro MAX_LONG_STRING) Removed, because it was only interesting
949 for system files, not for general code.
950 (macro MAX_VERY_LONG_STRING) Ditto.
951 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
952 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
954 * vardict.h: New file, for an interface between variables and
957 * variable.c: A lot of functions were moved around, for better
959 (struct variable) Move definition here, from variable.h.
960 (var_type_adj) Removed--makes i18n hard.
961 (var_type_noun) Ditto.
962 (var_create) New function.
963 (var_clone) New function.
964 (var_destroy) New function.
965 (var_set_name) Assert that variable is not in a dictionary.
966 (compare_var_names) Rename compare_vars_by_name and fix a couple
967 of callers who thought the args were strings.
968 (hash_var_name) Rename hash_var_by_name.
969 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
970 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
971 (var_is_very_long_string) Removed, because it was only interesting
973 (var_set_missing_values) Allow the argument to be the wrong width,
974 as long as we can resize it. Simplify callers who were doing the
976 (var_get_value_labels) New function.
977 (var_has_value_labels) New function.
978 (var_set_value_labels) New function.
979 (alloc_value_labels) New function.
980 (var_add_value_label) New function.
981 (var_replace_value_label) New function.
982 (var_clear_value_labels) New function.
983 (var_lookup_value_label) New function.
984 (var_get_value_name) Moved here from variable.c, renamed from
985 var_get_value_name, transposed argument order, updated all
987 (var_to_string) Moved here, from variable-label.c.
988 (var_set_leave) New function.
989 (var_get_leave) New function.
990 (var_must_leave) New function.
991 (var_set_short_name_suffix) Moved to dictionary.c, renamed
992 set_var_short_name_suffix.
993 (var_get_dict_index) New function.
994 (var_get_case_index) New function.
995 (var_get_obs_vals) New function.
996 (var_set_obs_vals) New function.
997 (var_has_obs_vals) New function.
998 (var_get_vardict) New function.
999 (var_set_vardict) New function.
1000 (var_has_vardict) New function.
1001 (var_clear_vardict) New function.
1002 (value_dup) Moved to value.c.
1003 (compare_values) Ditto.
1006 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1007 (enum ALPHA) Rename VAR_STRING, update all users.
1009 * vector.c: New file.
1010 (struct vector) Moved here, from variable.h.
1011 (check_widths) New function.
1012 (vector_create) New function.
1013 (vector_clone) New function.
1014 (vector_destroy) New function.
1015 (vector_get_name) New function.
1016 (vector_get_var) New function.
1017 (vector_get_var_cnt) New function.
1018 (compare_vector_ptrs_by_name) New function.
1020 * vector.h: New file.
1022 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1024 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1025 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1028 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1030 * value-labels.c (destroy_atoms): New function.
1031 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1033 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1035 Thanks to Jason Stover for pointing out this problem.
1037 * data-out.c (output_number): Use gsl_finite from GSL, which is
1038 portable, instead of isfinite, which is not.
1041 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1043 * variable.c variable.h (value_dup): New function.
1045 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1047 Start converting struct variable to an opaque type. In this
1048 phase, we add a bunch of setter and getter functions and convert
1049 most of the PSPP code to use them. The resulting changes are
1050 pervasive but mostly trivial, and only the notable changes are
1053 * format.c (fmt_equal): New function.
1055 * variable.c (var_type_is_valid): New function.
1056 (measure_is_valid) Moved here, from format.c.
1057 (alignment_is_valid) Moved here, from format.c.
1058 (var_get_name) New function.
1059 (var_set_name) New function.
1060 (width_to_type) New function.
1061 (var_get_type) New function.
1062 (var_get_width) New function.
1063 (var_set_width) New function.
1064 (var_is_numeric) New function.
1065 (var_is_alpha) New function.
1066 (var_is_short_string) New function.
1067 (var_is_long_string) New function.
1068 (var_is_very_long_string) New function.
1069 (var_get_missing_values) New function.
1070 (var_set_missing_values) New function.
1071 (var_clear_missing_values) New function.
1072 (var_has_missing_values) New function.
1073 (var_is_value_missing) New function.
1074 (var_is_num_missing) New function.
1075 (var_is_str_missing) New function.
1076 (var_is_value_user_missing) New function.
1077 (var_is_num_user_missing) New function.
1078 (var_is_str_user_missing) New function.
1079 (var_is_value_system_missing) New function.
1080 (var_get_print_format) New function.
1081 (var_set_print_format) New function.
1082 (var_get_write_format) New function.
1083 (var_set_write_format) New function.
1084 (var_set_both_formats) New function.
1085 (var_get_label) New function.
1086 (var_set_label) New function.
1087 (var_clear_label) New function.
1088 (var_has_label) New function.
1089 (var_get_measure) New function.
1090 (var_set_measure) New function.
1091 (var_get_display_width) New function.
1092 (var_set_display_width) New function.
1093 (var_get_alignment) New function.
1094 (var_set_alignment) New function.
1095 (var_get_value_cnt) New function.
1096 (var_get_leave) New function.
1097 (var_get_short_name) New function.
1099 * variable.h: (struct variable) Removed "type" and "nv" members;
1100 they are now computed from "width" where needed.
1102 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1104 * missing-values.c (mv_resize): Don't write beyond end of the
1105 allocated buffer when resizing a long string.
1107 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1109 Clean up identifier code: don't require identifier enumerations to
1110 be in a particular order; make better use of string library;
1111 expose less of the internals.
1113 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1114 change interface. Updated all callers.
1115 (lex_id_match) Change interface to use struct substring, update
1117 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1118 (global array keywords[]) Make static, change form. Update all
1119 users to use lex_id_name instead.
1120 (lex_is_keyword) New function.
1121 (lex_id_to_token) Change interface to use struct substring, update
1123 (lex_id_name) New function.
1125 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1126 lex_is_keyword instead.
1127 (T_LAST_KEYWORD) Removed.
1128 (T_N_KEYWORDS) Removed.
1130 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1132 * format.c: (fmt_date_template) Distinguish characters for which a
1133 space is output and any date delimiter is allowed on input, from
1134 those for which a space is output and only a space is allowed on
1135 input. The former is represented by X, the latter by a space.
1136 Also, drop distinction between h and H, changing the former to the
1139 * data-in.c: Completely rewrite internals to conform to SPSS input
1140 formats as closely as possible.
1141 (data_in) Changed external interface by replacing the structure
1142 that was used as a single argument by a set of arguments. Updated
1144 (data_in_finite_line) Removed. Converted all callers to use plain
1146 (data_in_get_integer_format) New function.
1147 (data_in_set_integer_format) New function.
1148 (data_in_get_float_format) New function.
1149 (data_in_set_float_format) New function.
1151 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1152 (struct data_in) Removed.
1154 * data-out.c: (output_date) Drop each component from the input as
1155 it is output, to allow us to drop the distinction between h (a
1156 count of hours) and H (the hour of day) template characters.
1157 Also, handle new X template character.
1158 (output_scientific) Follow more rational rule on when to drop
1159 fraction introduced between SPSS 13 and 15. Updated test case to
1162 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1164 Fix buffer overflow reported by John Darrington.
1166 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1167 realize that DIGITS is a count of nibbles, not of bytes.
1169 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1171 * calendar.c (calendar_offset_to_gregorian) Also return the
1172 year-of-day. Change callers to new interface.
1174 * data-out.c: Completely rewrite internals to conform to SPSS
1175 output formats as completely as possible.
1176 (data_out) Change interface to put input parameters before output
1177 parameters, for consistency with the style I now prefer. Update
1179 (data_out_get_integer_format) New public function.
1180 (data_out_set_integer_format) New public function.
1181 (data_out_get_float_format) New public function.
1182 (data_out_set_float_format) New public function.
1184 * data-out.h: New file. Move prototype for data_out here, from
1187 * format.c: (fmt_step_width) Use equality comparison instead of
1188 bitwise and, for clarity.
1189 (fmt_is_string) Ditto.
1190 (fmt_input_to_output) Fix categories that are translated to F
1193 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1195 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1196 casefile-private.h: Added casefilter to assist commands with missing
1199 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1201 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1203 * settings.c: (route_errors_to_terminal) New variable.
1204 (route_errors_to_listing) New variable.
1205 (get_error_routing_to_terminal) New function.
1206 (set_error_routing_to_terminal) New function.
1207 (get_error_routing_to_listing) New function.
1208 (set_error_routing_to_listing) New function.
1210 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1212 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1214 * format.c: Completely rewrite, to achieve better abstraction.
1215 Rewrite all references to formats in other files.
1217 * format.def: Rewrite and reorganize.
1219 * settings.c: Move everything related to custom currency formats
1220 into format.[ch], changing them in form, so as to group related
1221 code and definitions better. Changed all references to use the
1223 (static var decimal) Removed.
1224 (static var grouping) Removed.
1225 (static var cc) Removed.
1226 (get_decimal) Removed.
1227 (set_decimal) Removed.
1228 (get_grouping) Removed.
1229 (set_grouping) Removed.
1233 * settings.h: (macro CC_CNT) Removed.
1234 (macro CC_WIDTH) Removed.
1235 (struct custom_currency) Removed.
1237 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1239 * data-in.c (data_in): Use switch statement instead of table, to
1240 avoid dependence on the order of the FMT_* enums.
1242 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1244 * data-out.c: (num_to_string) Removed, because it was dead code.
1246 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1248 * data-in.c (parse_trailer): Fix error message.
1250 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1252 * format.c (fmt_is_binary): New function.
1254 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1256 * procedure.c procedure.h: Encapsulated the static data into a single
1259 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1261 * casefile.c (casereader_read_xfer): Always initialize the case,
1262 even on an error condition.
1264 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1266 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1267 entire function useless.
1269 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1271 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1272 casereader method casereader_clone.
1274 * procedure.c pransformations.h: Introduced new type casenum_t
1276 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1278 * variable.c: (width_to_bytes) Rephrase code for clarify.
1280 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1282 * format.c: (fmt_type_from_string) New function.
1283 (fmt_to_string) Include decimals in output if the format has
1284 decimals, even if the format type does not. This way, we can
1285 accurately reproduce incorrect formats in user output.
1286 (check_common_specifier) Make the check for a bad format type an
1287 assertion, so we get bug reports if they show up. Fix message.
1288 Check for decimal places with a format type that doesn't allow
1290 (check_input_specifier) Remove check for FMT_X, which has been
1292 (check_output_specifier) Ditto.
1294 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1296 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1297 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1298 (enum fmt_parse_flags) Removed.
1300 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1302 * casefile.c casefile.h: Converted to an abstract base class.
1303 * casefile-private.h fastfile.c fastfile.h: New files.
1304 * automake.mk procedure.c scratch-writer.c storage-stream.c
1306 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1308 * procedure.c (internal_procedure): Create sink_case with only as
1309 many values as the compacted dictionary.
1311 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1313 Remove "debugging" code that caused plenty of false positives and
1316 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1318 * case.c: Remove all references to `this' member.
1320 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1322 Fix link error noted by Jason Stover.
1324 * storage-stream.c: Include <assert.h>.
1326 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1328 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1329 ALL) and additional underlying system file issues.
1331 Thanks to John Darrington for review.
1333 First problem: var_hash points to variables not owned by the
1334 sys-file-reader, which the caller may free or modify. Use an
1335 array of sfm_vars instead, as done earlier (e.g. CVS version
1338 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1339 members and remove all code that references it. Add vars, var_cnt
1340 members. Remove fix_specials member, which was unused.
1341 (struct sfm_var) Remove name member, which was unused.
1342 (sfm_close_reader) Free vars member instead of var_hash.
1343 (compare_var_shortnames) Removed.
1344 (hash_var_shortname) Removed.
1345 (sfm_open_reader) Fill out vars array.
1346 (compare_var_index) Removed.
1347 (sfm_read_case) Use vars instead of var_hash.
1349 Second problem: we're confused about when we actually have very
1350 long strings, causing us to choose incorrectly between slow path
1351 and fast path in sfm_read_case.
1353 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1354 have very long strings, not when we have long variable names,
1355 which is an unrelated feature.
1357 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1359 * variable.h: Move var_set and variable parsing declarations to
1360 new header, src/language/lexer/variable-parser.h. Modified lots
1361 of files to include the new header.
1363 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1365 * value-labels.c (value_to_string): When there's no value label,
1366 format the variable according to its print format, instead of
1367 always effectively using A or F format.
1369 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1371 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1374 * format.c: Removed tortological assertion.
1376 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1378 Reform string library.
1380 * file-name.c (fn_interp_vars): Change interface to take a
1381 substring as input. Updated all users.
1383 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1385 * format.c (measure_is_valid): Really return false when m >=
1388 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1390 Implement random access to casefiles, for use in GUI.
1392 * casefile.c: (struct casereader) Add `random', `file_ofs',
1393 `buffer_ofs' members.
1394 (casefile_get_random_reader) New function.
1395 (read_open_file) Break part into new function
1396 seek_and_fill_buffer().
1397 (fill_buffer) Update buffer_ofs, file_ofs.
1398 (casereader_seek) New function.
1400 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1402 * settings.c: Added call to i18n{done, init}.
1404 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1406 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1409 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1411 * casefile.c: Convert many uses of `int' to `bool'.
1413 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1415 * transformations.c (trns_chain_destroy): Destroy chain's trns
1416 member, to fix memory leak.
1418 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1420 * storage-stream.c (storage_source_decapsulate): Destroy case
1421 source to fix memory leak.
1423 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1425 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1426 case passed as argument instead of initializing the argument as a
1427 case. Fixes memory leak that showed up in
1428 tests/command/aggregate.sh with scratch files.
1430 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1432 * procedure.c (proc_done): Destroy default_dict, to fix memory
1435 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1437 Simplify procedure_with_splits().
1439 * procedure.c (struct split_aux_data): Removed case_count member.
1440 (procedure_with_splits) Don't initialize case_count.
1441 (split_procedure_case_func) Check whether prev_case is null
1442 instead of case_count.
1443 (split_procedure_end_func) Ditto.
1445 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1447 * case.c (case_move): Do nothing if dst and src are the same
1449 (case_try_create) Merge two similar cases.
1450 (case_copy) Unshare only if data must be actually copied.
1452 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1454 * data-in.c data-out.c dictionary.c sys-file-reader.c
1455 sys-file-writer.c variable.c variable.h: Reworked very long string
1456 support for better encapsulation.
1458 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1460 * value-labels.c (val_labs_can_set_width): New function.
1461 (val_labs_set_width) Clear labels if increasing width to long
1463 (val_labs_destroy) Remove unneeded test for null.
1465 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1467 * value-labels.h: Remove unneeded dependency on variable.h.
1469 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1471 Get rid of `char *c' member in union value, for cleanliness.
1473 * value.h: (union value) Remove `c' member.
1475 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1477 Make dictionary compacting functions a little more general.
1479 * sys-file-writer.c (sfm_open_writer): Use
1480 dict_compacting_would_change().
1481 (does_dict_need_translation) Removed.
1483 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1485 Make dictionary compacting functions a little more general.
1487 * dictionary.c (dict_needs_compaction): Rename
1488 dict_compacting_would_shrink(). Update all callers.
1489 (dict_compacting_would_change) New function.
1491 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1493 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1494 inverted return value (!).
1496 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1498 Continue reforming procedure execution.
1500 * procedure.c: Search and replace "vfm" by "proc". Notably:
1501 (static var vfm_source) Rename proc_source. Update all
1503 (static var vfm_sink) Rename proc_sink. Update all references.
1505 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1507 Continue reforming procedure execution. In this phase, remove
1508 PROCESS IF, which was deprecated anyway and can be easily
1509 simulated with TEMPORARY followed by SELECT IF.
1511 * procedure.c: (open_active_file) Don't call
1512 add_process_if_trns().
1513 (discard_variables) Get rid of redundant call to
1514 proc_cancel_all_transformations().
1515 (add_process_if_trns) Removed.
1516 (process_if_trns_proc) Removed.
1517 (process_if_trns_free) Removed.
1519 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1521 Continue reforming procedure execution. In this phase, add
1522 `const' to the case passed to procedure()'s callback.
1524 Updated all users of procedure() as well.
1526 * procedure.c: (struct write_case_data) Add "const" to ccase
1527 parameter for case_func member.
1528 (procedure) Add "const" to ccase parameter for proc_func
1530 (multipass_case_func) Make ccase parameter const.
1531 (internal_procedure) Add "const" to ccase parameter for case_func
1533 (split_procedure_case_func) Make ccase parameter const.
1534 (multipass_split_case_func) Make ccase parameter const.
1536 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
1538 Continue reforming procedure execution. In this phase, get rid of
1539 the output code for SPLIT FILE groups in procedure.c, which really
1540 shouldn't be doing any output. Move it into the individual
1541 procedures instead. This also adds some flexibility.
1543 Updated many users of procedure_with_splits() and
1544 multipass_procedure_with_splits() to call
1545 output_split_file_values() and to deal with increased use of
1548 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
1549 parameter to begin_func member.
1550 (procedure_with_splits) Add "const struct ccase *" parameter to
1551 begin_func parameter. Make ccase parameter const in proc_func
1553 (split_procedure_case_func) Don't dump split file group. Pass
1555 (dump_splits) Moved to language/dictionary/split-file.c as
1556 output_split_file_values().
1557 (struct multipass_split_aux_data) Add "const struct ccase *"
1558 parameter to split_func member.
1559 (multipass_procedure_with_splits) Add "const struct ccase *"
1560 parameter to split_func parameter.
1561 (multipass_split_case_func) Save new SPLIT FILE case before
1563 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
1565 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
1567 Continue reforming procedure execution. Change
1568 internal_procedure() so that it calls open_active_file() and
1569 close_active_file(), which isolates most of the actual procedure
1572 * procedure.c: (struct write_case_data) Rename `proc_func' member
1573 to `case_func' and update all references.
1574 (procedure) Rewrite as one-line wrapper around
1575 internal_procedure().
1576 (struct multipass_aux_data) New.
1577 (multipass_callback) Renamed multipass_case_func(). Use struct
1578 multipass_aux_data as auxiliary data.
1579 (multipass_end_func) New function.
1580 (multipass_procedure) Rewrite as wrapper for internal_procedure()
1581 that uses multipass_case_func, multipass_end_func.
1582 (internal_procedure) Add `end_func' argument. Move optimization
1583 of trivial case in here. Move call to open_active_file() and
1584 close_active_file() in here. Now assert that vfm_source is
1586 (procedure_with_splits_callback) Rename
1587 split_procedure_case_func().
1588 (split_procedure_end_func) New function.
1589 (multipass_split_callback) Rename multipass_split_case_func.
1590 (multipass_split_end_func) New function.
1591 (discard_variables) No need to test for nonnull vfm_source.
1593 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
1595 Continue reforming procedure execution. Get rid of unused member.
1597 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
1599 (write_case) Don't increment cases_analyzed.
1601 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
1603 Continue reforming procedure execution. In this phase, move
1604 procedure.c and procedure.h from src to src/data. Update
1605 makefiles and #includes accordingly.
1607 * procedure.c: Moved here from src/.
1609 * procedure.h: Moved here from src/.
1611 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
1613 Continue reforming procedure execution. In this phase, get rid of
1614 many global variables, consolidating procedure execution in
1615 procedure.c. Encapsulate transformations in new "struct
1616 trns_chain". Also, change implementation of N OF CASES, FILTER,
1617 and PROCESS IF from special cases to transformations.
1619 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
1622 * dictionary.c: (global variable default_dict) Move to
1625 * variable.h: (TRNS_*) Move to transformations.h.
1626 (struct transformation) Move to transformations.c.
1628 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1630 * sys-file-reader.c: Fixed invalid read problems.
1632 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
1634 * storage-stream.c: Add missing function comments.
1636 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
1638 Continue reforming procedure execution. In this phase, add some
1639 new, needed functionality to storage-stream.
1641 * storage-stream.c: (storage_source_decapsulate) New function.
1643 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
1645 * variable.c (width_to_bytes): Declarations must precede
1646 statements for C90 compliance.
1648 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
1650 * data-out.c, data-in.c, variable.c, variable.h: New functions
1651 copy_mangle and copy_demangle for reading/writing cases; emulates the
1652 way SPSS deals with strings > 255 bytes.
1654 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
1655 subtype 14 needed for strings longer than 255 bytes.
1657 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
1658 instead of literal values. Also fixed some constness issues.
1660 * format.h: Constness
1662 * sfm-private.h: Renamed the case_size identifier, since I discovered
1663 that SPSS's respect for this variable is very nominal.
1665 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
1667 Change case limit type from int to size_t.
1669 * dictionary.c: (struct dictionary) Change type of case_limit
1671 (dict_get_case_limit) Change return type.
1672 (dict_set_case_limit) Change parameter type.
1674 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
1676 * variable.h: (struct variable) Rename `reinit' member as `leave'
1677 and invert sense. Fix up all references.
1679 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
1681 Continue reforming procedure execution. In this phase, break
1682 procedure.c into multiple files.
1684 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
1686 * case-sink.c: New file.
1688 * case-sink.h: New file.
1690 * case-source.c: New file.
1692 * case-source.h: New file.
1694 * storage-stream.c: New file.
1696 * storage-stream.h: New file.
1698 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
1700 * variable.h: (struct variable) Remove `init' member and all
1701 references to it from other files. It was initialized in several
1702 places, but nothing really ever used it for anything worthwhile.
1703 Thanks to Jason Stover for pointing out how confusing this
1706 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
1708 Continue reforming error message support. In this phase, get rid
1709 of message "titles" and put the message text in `struct error'.
1710 Now `struct error' encapsulates a message more properly.
1712 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
1713 Format message ourselves.
1715 * data-in.c: (vdls_error) Ditto.
1717 * por-file-reader.c: (error) Ditto.
1719 * sys-file-reader.c: (corrupt_msg) Ditto.
1721 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
1723 GNU standards require "file name" instead of "filename" in
1724 documentation. It's nice for our code to follow the convention
1727 * casefile.c: (struct casefile) Rename `filename' member to
1728 `file_name'. Updated all references.
1730 * file-name.c: [!unix] (struct file_identity) Rename
1731 normalized_filename member to normalized_file_name. Updated all
1734 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
1736 We don't really support anything but Unix-like environments well,
1737 so we might as well de-obfuscate by writing directory and path
1738 separators explicitly.
1740 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
1742 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
1743 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
1745 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
1748 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
1750 GNU standards require "file name" instead of "filename" in
1751 documentation. It's nice for our code to follow the convention
1754 * filename.c: Rename to file-name.c.
1756 * filename.h: Rename to file-name.h. Update all inclusions.
1757 Update header guards.
1759 * automake.mk: Update file names.
1761 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
1763 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
1765 (fn_basename) Removed (dead code).
1766 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
1767 (fn_special_p) Renamed fn_is_special(), all references updated.
1768 (fn_exists_p) Renamed fn_exists(), all references updated.
1770 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
1772 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
1773 Also, now it only tilde-expands file names, not paths.
1774 (fn_search_path) Tilde-expand one directory at a time.
1776 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
1778 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
1779 get rid of non-Unixlike version of the code, which has probably
1781 (fn_prepend_dir) Removed (dead code).
1783 * filename.h: (macro DIR_SEPARATOR_STRING) New.
1784 (macro PATH_SEPARATOR_STRING) New.
1785 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
1787 Continue reforming error message support. In this phase, we get
1788 rid of VM() and the other msg() support for "verbosity", replacing
1789 it by a new function verbose_msg().
1791 * filename.c: (fn_search_path) Use verbose_msg() instead of
1794 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
1796 * sfm-private.h: Get rid of #defines after #error, which makes no
1799 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
1801 Get rid of our own int32 type in favor of the standard int32_t
1804 * sfm-private.h: (int32 macro) Don't define this anymore. Do
1807 * sys-file-reader.c: Use int32_t instead of int32 throughout.
1809 * sys-file-writer.c: Use int32_t instead of int32 throughout.
1811 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
1813 Remove ill-considered file routines that are no longer used.
1815 * filename.c: (fn_open_ext) Removed.
1816 (fn_close_ext) Removed.
1818 * filename.h: (struct file_ext) Removed.
1820 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
1822 * variable.c (var_is_valid_name): Move declarations before code
1825 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
1827 * filename.ch (fn_interp_vars): Fixed small buglet.
1829 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
1831 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
1832 so as to modify the string inline. Thus makeing it easier to
1833 destroy the results when no longer needed.
1835 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
1837 * category.c (cat_stored_values_destroy): Fixed memory leak.
1839 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
1841 Add some missing frees. Thanks to John Darrington for reporting
1844 * any-writer.c (any_writer_close): Free writer.
1846 * any-reader.c (any_reader_close): Free reader.
1848 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
1850 * por-file-reader.c: (error) Mark as NO_RETURN.
1852 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
1854 * settings.c: Changed default value of scompress to true.
1856 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
1858 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
1860 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
1863 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
1864 here from pref.h.orig.
1866 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
1868 * sys-file-reader.c: Fixed bug reading compressed files.
1870 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1872 * Numerous renames. See src/ChangeLog for details.
1874 * Moved files from src directory