1 2007-07-25 Ben Pfaff <blp@gnu.org>
3 Fix bugs related to bug #17213.
5 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
6 since the former is what config.h has. Include the needed ncurses
8 (static var echo) Rename to `do_echo' because the original name is
9 the same as an ncurses identifier.
10 (get_termcap_viewport) Use error instead of msg.
12 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
13 (fn_close): Don't close stdin, stdout, stderr.
15 2007-07-26 John Darrington <john@darrington.wattle.id.au>
17 * procedure.c procedure.h: Added callbacks which get invoked whenever
18 a dataset's transformation chain changes.
20 2007-07-24 Ben Pfaff <blp@gnu.org>
23 * sys-file-writer.c (write_variable_display_parameters): Use new
24 var_default_display_width function to choose display width of
25 segments after the first one in a given variable.
26 * variable.c (var_create): Use var_default_display_width to pick
27 new variable's display width.
28 (var_default_display_width): New function.
29 Reviewed by John Darrington.
31 2007-07-24 Ben Pfaff <blp@gnu.org>
34 * por-file-writer.c (write_variables): Write weight variable.
35 Reviewed by John Darrington.
37 2007-07-23 Ben Pfaff <blp@gnu.org>
39 Improvements to system file reader and writer.
41 First, move all detailed knowledge of very long strings into
42 sys-file-private.[ch], so that this nasty stuff can be isolated.
44 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
45 (EFFECTIVE_VLS_CHUNK): New macro.
46 (min_int): New function.
47 (max_int): New function.
48 (sfm_width_to_bytes): Rewrite.
49 (sfm_width_to_octs): New function.
50 (sfm_segment_alloc_width): New function.
51 (sfm_segment_alloc_bytes): New function.
52 (sfm_segment_used_bytes): New function.
53 (sfm_segment_offset): New function.
54 (sfm_segment_effective_offset): New function.
55 (sfm_dictionary_to_sfm_vars): New function.
57 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
58 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
59 (struct sfm_var): New structure.
61 Next, improvements to the system file reader.
63 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
64 type casenumber. Added `version_major', `version_minor',
67 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
68 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
69 Change `case_cnt' to type casenumber. Removed `has_vls'.
70 (struct sfm_var): Removed.
71 (sfm_open_reader): Don't warn on wrong case size if the file was
72 written by SPSS 13, which tends to get it wrong. Use
73 sfm_dictionary_to_sfm_vars.
74 (read_header): Always output system file info.
75 (read_variable_record): Simplify code for reading missing values.
76 (read_machine_int32_info): Save version numbers from system file
77 into info struct passed as new argument.
78 (read_long_string_map): Restructured to use new sys-file-private
80 (read_value_labels): Use size_overflow_p.
81 (sys_file_casereader_read): Get rid of distinction between fast
82 and slow paths. Use information provided by sys-file-primate's
83 struct sfm_var to simplify code.
84 (skip_whole_strings): New function.
85 (read_int32): Renamed read_int. Changed return value to int.
87 (read_flt64): Renamed read_float. Changed return value to
88 double. Updated all callers.
89 (int32_to_native): Removed. Changed callers to use
91 (flt64_to_double): Removed. Changed callers to use float_convert.
93 Finally, get rid of int32, flt64 terminology and types in system
94 file writer. The former wasn't very useful since a POSIX "int"
95 can hold the whole range of int32 and we generally didn't have a
96 need for it to be exactly-32-bits, just at-least-32-bits. The
97 latter was inconvenient because we had to assume that it could be
98 different from double and thereby convert special values SYSMIS,
99 HIGHEST, LOWEST to and from it in multiple places. Instead, now
100 we just use "int" and "double" in most places, and do conversions,
101 if necessary, very close to where we do I/O. This change meant
102 that the writer code couldn't represent records in the file as C
103 structs any longer, but that's no great loss. The code actually
104 seems to be more readable without them.
106 Simplify the compression buffering code: only buffer as much as
107 necessary, which is no more than eight 8-byte units at any given
110 * sys-file-writer.c (typedef flt64): Removed.
111 (macro second_lowest_flt64): Removed.
112 (struct sysfile_header): Removed.
113 (struct sysfile_variable): Removed.
114 (struct sfm_writer): Removed `needs_translation', `has_vls',
115 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
116 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
117 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
118 compression buffering by `opcodes', `opcode_cnt', `data',
119 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
120 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
121 data writer functions instead of structures.
122 (calc_oct_idx): New function.
123 (write_header): Use simple data writer functions instead of
125 (write_format_spec): Renamed write_format. New argument.
126 (write_variable_continuation_records): New function.
127 (write_variable): Use simple data writer functions instead of
128 structures. Use write_variable_continuation_records. Write
129 entire very long string instead of requiring caller to understand
131 (write_value_labels): Use simple data writer functions instead of
133 (write_documents): Ditto.
134 (write_variable_display_parameters): Use sys-file-private
135 functions to simplify. Use simple data writer functions instead
137 (write_vls_length_table): Use simple data writer functions instead
139 (write_longvar_table): Ditto.
140 (write_rec_7_34): Break into new functions
141 write_integer_info_record, write_float_info_record. Use simple
142 data writer functions instead of structures.
143 (buf_write): Removed.
144 (append_string_max): Removed.
145 (ensure_buf_space): Removed.
146 (sys_file_casewriter_write): Get rid of the distinction between
147 fast and slow paths, which didn't seem to be too useful. Use new
148 functions write_case_uncompressed, write_case_compressed.
149 (put_instruction): Removed.
150 (put_element): Removed.
151 (write_compressed_data): Removed.
152 (close_writer): Use flush_compressed. Only write case count to
153 system file if it will fit in the field.
154 (write_case_compressed): New function.
155 (write_case_uncompressed): New function.
156 (flush_compressed): New function.
157 (put_cmp_opcode): New function.
158 (put_cmp_number): New function.
159 (write_int): New function.
160 (convert_double_to_output_format): New function.
161 (write_float): New function.
162 (write_value): New function.
163 (write_string): New function.
164 (write_bytes): New function.
165 (write_zeros): New function.
166 (write_spaces): New function.
168 Reviewed by John Darrington.
170 2007-07-22 Ben Pfaff <blp@gnu.org>
172 Don't try to write very long strings to portable files. The
173 format does not support it.
175 * por-file-writer.c (MAX_POR_WIDTH): New macro.
176 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
177 (write_format): Add arg to take width to resize format to.
178 (write_value): Limit width of value written to MAX_POR_WIDTH.
179 (write_variables): Limit width of variable and its output formats
181 Reviewed by John Darrington.
183 2007-07-22 Ben Pfaff <blp@gnu.org>
185 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
186 local variable instead of literal 5.
187 Reviewed by John Darrington.
189 2007-07-22 Ben Pfaff <blp@gnu.org>
191 Fix problems with uniqueness of short names in system files with
192 very long string variables. Now a variable may have multiple
195 * automake.mk (src_data_libdata_a_SOURCES): Add new files
196 short-names.c, short-names.h.
198 * dictionary.c (dict_clone): Clone all the short names.
199 (compare_strings): Move into short-names.c.
200 (hash_strings): Ditto.
201 (set_var_short_name_suffix): Ditto.
202 (dict_assign_short_names): Ditto, rename short_names_assign,
203 change to assign all short names.
205 * por-file-writer.c (write_variables): Use short_names_assign
206 instead of dict_assign_short_names.
208 * short-names.c: New file.
210 * short-names.h: New file.
212 * sys-file-private.c (sfm_width_to_segments): New function.
214 * sys-file-reader.c (read_long_var_name_map): Save and restore all
215 the short names, not just the first one.
217 * sys-file-writer.c (cont_var_name): Removed.
218 (sfm_open_writer): Use short_names_assign instead of
219 dict_assign_short_names. Use unique short names assigned by
220 short_names_assign instead of those generated by cont_var_name.
222 * variable.c (struct variable): Remove `short_name' member,
223 replace by `short_names' and `short_name_cnt'.
224 (var_create) Initialize new members.
225 (var_get_short_name_cnt): New function.
226 (var_get_short_name): Now takes an index argument. Changed most
228 (var_set_short_name): Ditto.
229 (var_clear_short_name): Renamed var_clear_short_names, changed to
230 clear all short names.
232 Reviewed by John Darrington.
234 2007-07-22 Ben Pfaff <blp@gnu.org>
236 * variable.c (var_set_width): Use new var_set_width function.
238 * missing-values.c (mv_n_values): Drop assertion, which was not
241 * format.c (fmt_default_for_width): New function.
242 (fmt_resize): New function.
244 Reviewed by John Darrington.
246 2007-07-18 John Darrington <john@darrington.wattle.id.au>
248 * datasheet.c (datasheet_delete_columns): Added assertion to check
249 we're not deleting outside the range of the sheet.
252 * dictionary.c dictionary.h variable.c: Added the ability for string
253 variables to be resized.
255 * vardict.h: Added some prototypes (moved from dictionary.h) as
256 these should only be called by variable.c
259 2007-07-14 John Darrington <john@darrington.wattle.id.au>
261 * sfm-reader.c: Respect case_cnt field in file header.
263 2007-07-01 John Darrington <john@darrington.wattle.id.au>
265 * transformation.c transformation.h (trns_chain_execute): Changed the
266 signature (Patch #6057)
268 2007-06-10 Ben Pfaff <blp@gnu.org>
270 * casereader-filter.c (casereader_filter_destroy): Make sure to
271 write all the remaining excluded cases to the casewriter, if any.
273 * caseinit.c (init_list_destroy): Rewrite.
274 (init_list_clear): Ditto.
276 * casegrouper.c (casegrouper_get_next_group): Always set *reader
277 to null when returning false.
279 2007-06-06 Ben Pfaff <blp@gnu.org>
281 Actually implement the new procedure code and adapt all of its
282 clients to match. Also adapt all of the other case sources and
283 sinks in the tree and their clients to use the
284 casereader/casewriter infrastructure.
286 * automake.mk: Add and remove files.
288 * any-reader.c: Change into a casereader.
289 * por-file-reader.c: Ditto.
290 * scratch-reader.c: Ditto.
291 * sys-file-reader.c: Ditto.
293 * any-writer.c: Change into a casewriter.
294 * por-file-writer.c: Ditto.
295 * scratch-writer.c: Ditto.
296 * sys-file-writer.c: Ditto.
298 * procedure.c: Change to use casereader, casewriter, caseinit, and
299 other new infrastructure.
301 * scratch-handle.c: Adapt to new infrastructure.
303 * case-sink.c: Removed, now dead code.
304 * case-sink.h: Ditto.
305 * case-source.c: Ditto.
306 * case-source.h: Ditto.
307 * casefile-factory.c: Ditto.
308 * casefile-private.h: Ditto.
311 * casefilter.c: Ditto.
312 * casefilter.h: Ditto.
315 * fastfile-factory.c: Ditto.
316 * fastfile-factory.h: Ditto.
317 * storage-stream.c: Ditto.
318 * storage-stream.h: Ditto.
320 2007-06-06 Ben Pfaff <blp@gnu.org>
324 * automake.mk: Add new files.
326 * datasheet.c: New file.
328 * datasheet.h: New file.
330 2007-06-06 Ben Pfaff <blp@gnu.org>
332 Until now, the procedure code has provided a case to the
333 case_source, which has filled in the data values that come from
334 the active file. "Left" data values that don't come from the
335 active file naturally stay the same from case to case, because the
336 procedure code keeps using that same case.
338 One of the compromises that comes with the new procedure code is
339 that the active file allocates and provides its own case, which
340 the procedure code then has to resize to provide room for any
341 other variables that should go in the case and then fill in the
342 values of "left" variables. Then, when we're done with that case,
343 we have to save the values of "left" variables to copy into the
344 next case read from the active file.
346 The caseinit code helps with this.
348 * automake.mk: Add new files.
350 * caseinit.c: New file.
352 * caseinit.h: New file.
354 2007-06-06 Ben Pfaff <blp@gnu.org>
356 * value.h (value_cnt_from_width): New function.
358 * variable.c (var_get_value_cnt): Use new function.
360 2007-06-06 Ben Pfaff <blp@gnu.org>
362 Add casegrouper, to allow cases read from a given casereader to be
363 broken into groups, each of which has its own casereader.
364 Generally cases are grouped based on having equal values for some
367 * automake.mk: Add new files.
369 * casegrouper.c: New file.
371 * casegrouper.h: New file.
373 2007-06-06 Ben Pfaff <blp@gnu.org>
375 Add interface to lexicographical ordering of cases.
377 * automake.mk: Add new files.
379 * case-ordering.c: New file.
381 * case-ordering.h: New file.
383 2007-06-06 Ben Pfaff <blp@gnu.org>
385 Add casereaders and casewriters, the basis of the new data processing
386 implementation. A casereader is a uniform interface to reading cases
387 from a data source; a casewriter is a uniform interface to writing
388 cases to a data sink.
390 * automake.mk: Add new files.
392 * casereader-filter.c: New file.
394 * casereader-provider.h: New file.
396 * casereader-translator.c: New file.
398 * casereader.c: New file.
400 * casereader.h: New file.
402 * casewriter-provider.h: New file.
404 * casewriter-translator.c: New file.
406 * casewriter.c: New file.
408 * casewriter.h: New file.
410 2007-06-06 Ben Pfaff <blp@gnu.org>
412 "casewindow" data structure that extends the deque (from libpspp)
413 of cases with the ability to dump cases to disk if we get too many
416 * automake.mk: Add new files.
418 * casewindow.c: New file.
420 * casewindow.h: New file.
422 2007-06-06 Ben Pfaff <blp@gnu.org>
424 sparse_cases data structure that augments a sparse_array of cases
425 with the ability to dump cases to disk if we get too many cases in
428 * automake.mk: Add new files.
430 * sparse-cases.c: New file.
432 * sparse-cases.h: New file.
434 2007-06-06 Ben Pfaff <blp@gnu.org>
436 Adds a low-level on-disk case array data structure.
438 * automake.mk: Add new files.
440 * case-tmpfile.c: New file.
442 * case-tmpfile.h: New file.
444 2007-06-06 Ben Pfaff <blp@gnu.org>
446 In a couple of places we calculate the maximum number of cases to
447 keep in memory based on the user-defined workspace. Enable
448 centralizing the calculation through a new function.
450 * settings.c (get_workspace_cases): New function.
452 2007-06-06 Ben Pfaff <blp@gnu.org>
454 The casenumber type is defined in transformations.h, but case.h is
455 a more sensible place. Move it.
457 * case.h (CASENUMBER_MAX): New macro.
458 (typedef casenumber): Move here, from transformations.h.
460 2007-06-03 Ben Pfaff <blp@gnu.org>
462 Slightly generalize case_to_values and case_from_values functions.
464 * case.c (case_to_values): Rename case_copy_out, change interface.
465 (case_from_values): Rename case_copy_in, change interface.
467 * fastfile.c (fastfilereader_get_next_case): Update caller.
468 (write_case_to_disk): Ditto.
470 2007-06-02 Ben Pfaff <blp@gnu.org>
472 Clean up after a forgotten part of patch #5829.
474 * casedeque.h: Remove unused file.
476 * automake.mk: Remove casedeque.h from sources.
478 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
480 * category.c: Removed redundant #include
482 2007-05-06 Ben Pfaff <blp@gnu.org>
484 Abstract the documents within a dictionary a little better.
485 Thanks to John Darrington for suggestion, initial version, and
488 * dictionary.c (struct dictionary): Change `documents' member from
489 char * to struct string.
490 (dict_clear): Destroy struct string.
491 (dict_get_documents): Convert struct string to char *.
492 (dict_set_documents): Set struct string. Pad to 80-character
494 (dict_clear_documents): New function.
495 (dict_add_document_line): New function.
496 (dict_get_document_line_cnt): New function.
497 (dict_get_document_line): New function.
499 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
501 * sys-file-reader.c (read_documents): Use new document functions.
503 2007-04-19 John Darrington <john@darrington.wattle.id.au>
505 * sys-file-reader.c: When reading a system file which has no
506 long name table, automatically create one where the long names
507 are the lower case versions of the short names.
509 2007-04-22 Ben Pfaff <blp@gnu.org>
511 * dictionary.c (dict_set_split_vars): dict_destroy expects that
512 dict_clear will free most data related to the dictionary.
513 dict_clear does a decent job, except that dict_set_split_vars on
514 some systems won't actually free the dict's "split" member.
515 Instead, it'll allocate a 1-byte region. Fix this.
517 * value.c (value_copy): New function.
518 (value_set_missing): Ditto.
520 2007-04-22 John Darrington <john@darrington.wattle.id.au>
522 * Deleted existing category.h and moved cat-routines.h into
523 category.h Encapsulated struct cat_vals better.
525 2007-04-19 John Darrington <john@darrington.wattle.id.au>
527 * sys-file-reader.c: When reading a system file which has no
528 long name table, automatically create one where the long names
529 are the lower case versions of the short names.
531 2007-04-16 John Darrington <john@darrington.wattle.id.au>
533 * sys-file-reader.c: Some versions of Other Software seem to
534 produce system files with string variables' measure set to
535 zero. We'll assume these are supposed to be nominal variables.
537 2007-03-30 Ben Pfaff <blp@gnu.org>
539 * procedure.c: Adapt to new deque data structure.
541 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
542 Ben Pfaff <blp@gnu.org>
544 * file-name.c: Mingw compatibility fixes.
545 (fn_search_path): Use ISSLASH instead of comparing against '/'
547 (fn_dir_name): Use dir_name from gnulib.
548 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
549 (fn_get_identity): Use GetFullPathName instead of canonicalize
550 from gnulib, because the latter does not fully support
551 Windows-style path names. Use this implementation based on the
552 detected presence of Windows instead of the absence of Unix, since
553 the new implementation is Windows-specific.
554 (fn_compare_file_identities): In Windows implementation, compare
555 names case-insensitively.
557 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
559 * make-file.c: Don't include mkstemp.h, because gnulib now causes
560 <stdlib.h> to have the same effect.
562 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
564 * por-file-reader.c: Add missing _() around messages.
566 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
568 * make-file.c: Include "mkstemp.h", without which linking on
571 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
573 Reduce platform dependence.
575 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
576 Everywhere we using this, we really should have just depended on
577 the shell to expand tildes.
578 (fn_search_path): Simplify, given that we don't do tilde expansion
580 (fn_normalize): Removed. Caller changed to use the canonicalize
582 (fn_get_cwd): Removed. Only user was fn_normalize.
583 (fn_is_absolute): Really only test for absolute names.
584 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
586 (fn_readlink): Removed, as it was only used fn_normalize.
587 (fn_exists): Assume the stat function is available; gnulib does.
588 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
590 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
592 * dictionary.c (dict_create_vector_assert): New function.
594 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
596 * file-name.c (fn_normalize): Correct name of function
597 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
598 for pointing this out.
600 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
602 We are using a single member in struct file_handle, the "name"
603 field, for more than one purpose. When it begins with '"', it's a
604 file name; otherwise, it's a token that can be used to identify
605 it. When that assertion fires, it's because we searched for the
606 name case-sensitively as a file name (so that there was no match),
607 and then we try to insert it case-insensitively as a token, which
608 fails because duplicates aren't allowed.
610 Solution: break the two purposes into two separate fields. This
611 fixes the problem and likely makes the code easier to read too.
613 Fixes bug #18922. Thanks to John Darrington for bug report and
616 * file-handle-def.c (struct file_handle): New `id' member.
617 (fh_from_name): Rename fh_from_id. Update all callers.
618 (create_handle): New `id' parameter. Update all callers.
619 (fh_create_file): Ditto.
620 (fh_get_id): New function.
622 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
624 * case.c (case_is_null): Change return type to bool.
626 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
630 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
631 reference counting to share data. This makes it easy for
632 valgrind, etc. to find accesses to cases that have been destroyed
633 but have been kept around by another user's ref-count. This often
634 happens when the data set is small enough to find in memory; if a
635 bigger data set that would overflow to disk were used, then data
636 corruption would occur.
638 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
642 * case.c (case_unshare): Make it check internally whether the
643 ref_cnt is greater than 1, so that the callers don't have to.
644 Update callers not to check.
646 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
648 Before, I was thinking that I might want to get rid of reference
649 counting at some point. Now, I'm pretty sure that it's here to
650 stay. Thus, because we have to store the value_cnt anyway for
651 reference-counted cases, we might as well expose it to users.
653 * case.c (case_get_value_cnt): New function.
654 (case_resize): Drop OLD_CNT argument. Update all callers. Only
655 resize case if its size actually changed.
657 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
658 instead of peeking inside struct case directly.
659 (casefile_append): Ditto.
661 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
663 Get rid of the inlines for the case functions, which made the
664 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
665 the inlines didn't speed up "make check" at all, which is not a
666 perfect benchmark but seems indicative.)
668 * case.c: Remove #ifdef DEBUGGING...#endif around many function
669 definitions. Remove some assertions on nonnull pointers that were
670 redundant with a pointer dereference soon after in the function.
672 (struct case_data): Move definition here from case.h.
676 (case_data_wr): Ditto.
678 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
680 * automake.mk: Add casedeque.h to sources.
682 * casedeque.h: New file.
684 * procedure.c: (struct dataset) Change lag_count, lag_head,
685 lag_queue member into single struct casedeque member. Update all
686 users to use the casedeque instead.
689 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
691 * procedure.c: Simplify lagged cases interface. Updated all
692 clients--well, the only client--to use the simplified interface.
693 (dataset_n_lag) Removed.
694 (dataset_set_n_lag) Removed.
695 (dataset_need_lag) New function.
697 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
699 * dictionary.c procedure.c: More changes to ensure that callbacks occur
700 whenever appropriate, but only when the dataset/dictionary is in a
703 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
705 * dictionary.c dictionary.h : Added callbacks for change of filter and
706 split variables. Refactored some code to ensure that callbacks get
707 invoked when appropriate.
709 * procedure.c (proc_cancel_temporary_transformations): Make sure that
710 replace_dict callback occurs when permanent_dict replaces the current
713 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
715 * dictionary.c dictionary.h : Added callback for when the weight
716 variable of a dictionary changes.
718 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
720 * dictionary.c dictionary.h : Added replace_source and replace_dict
721 callbacks, and functions to deal with them.
723 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
725 Simplify missing value handling.
727 * missing-values.h (enum mv_class): New type.
728 (enum mv_type): Moved definition into missing-values.c and renamed
729 each MV_* to MVT_*, to distinguish them from the exposed mv_class
730 enums. Updated all uses.
731 (struct missing_values): Changed type of `type' from `enum
732 mv_type' to `int' because the definition is no longer exposed.
734 * missing-values.c (mv_is_value_missing): Add new enum mv_class
735 parameter. Update all callers.
736 (mv_is_num_missing): Ditto.
737 (mv_is_str_missing): Ditto.
738 (mv_is_value_user_missing): Removed. Changed callers to use
740 (mv_is_num_user_missing): Removed. Changed callers to use
742 (mv_is_str_user_missing): Removed. Changed callers to use
744 (mv_is_value_system_missing): Removed. Changed callers to use
746 (mv_set_type): Removed. Changed callers to use mv_clear.
747 (mv_clear): New function.
749 * variable.c (var_is_value_missing): Add new enum mv_class
750 parameter. Update all callers.
751 (var_is_num_missing): Ditto.
752 (var_is_str_missing): Ditto.
753 (var_is_value_user_missing): Removed. Changed callers to use
754 var_is_value_missing.
755 (var_is_num_user_missing): Removed. Changed callers to use
757 (var_is_str_user_missing): Removed. Changed callers to use
759 (var_is_value_system_missing): Removed. Changed callers to use
760 var_is_value_missing.
762 * casefilter.c (struct casefilter): Use enum mv_class in place of
764 (casefilter_variable_missing): Adapt to new member.
765 (casefilter_create): Change signature to take enum mv_class,
768 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
770 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
772 * case-sink.c case-sink.h procedure.c procedure.h
773 storage-stream.c: Now uses the factory.
775 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
777 Make it possible to pull cases from the active file with a
778 function call, instead of requiring indirection through a callback
781 * case-source.h (struct case_source_class): Change ->read function
782 to return a single case, instead of calling a callback function
783 for each case. Change ->destroy function to return an error
786 * case-source.c (free_case_source): Pass along the value returned
787 by the case_source ->destroy function.
789 * procedure.c (struct write_case_data): Removed.
790 (struct dataset): Added some members to track procedure state.
791 (procedure): Optimize the trivial case at this level.
792 (internal_procedure): Re-implement in terms of proc_open,
793 proc_read, proc_close.
794 (proc_open) New function.
795 (proc_read) New function.
796 (proc_close) New function.
797 (write_case) Moved into proc_read.
798 (close_active_file) Moved closing of data source into proc_close.
800 * storage-source.c: Rewrote to conform with modified
801 case_source_class interface.
803 * transformations.c (trns_chain_execute): Added argument to allow
804 starting execution from an arbitrary transformation. Updated
807 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
809 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
811 * sys-file-reader.c (read_display_parameters): Don't assume that
812 MEASURE_* and ALIGN_* have the same values found in system files.
814 * sys-file-writer.c (write_variable_display_parameters): Ditto.
816 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
817 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
818 the value of n_MEASURES, which was off by 1 (at least from my
821 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
823 * dictionary.c dictionary.h vardict.h variable.c: Added optional
824 callbacks which are invoked when the dictionary or its
825 variables are changed.
827 * missing-values.c missing-values.h value-labels.c: Tidied up
828 consistency checks, and made some of them return false
829 instead of assert-failing.
831 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
833 * calendar.c (calendar_days_in_month): New function.
835 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
837 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
838 the value label's width.
840 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
842 * sfm-private.h: Move contents into sys-file-writer.c, which is
843 the only remaining user. Removed Borland C++-specific directives.
845 * sys-file-reader.c: Clean up and rewrite entire file. The
846 rewritten version is simpler and better abstracted, and should be
847 easier to maintain and extend. It avoids using structures to read
848 file data, which is prone to padding variations among compilers.
849 It should also handle non-IEEE 754 system files, although I
850 haven't been able to find any. It has been tested against many
851 .sav files obtained from the Web and found to produce the same
852 results as the earlier version of the code, or in some cases
853 improved results. It is more tolerant of format variations found
856 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
857 member, putting an enum integer_format in its place. New member
858 `float_format'. Changed `compressed' member to type bool.
860 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
862 * dictionary.c (dict_delete_consecutive_vars): New function.
864 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
866 * file-name.c (fn_search_path): Remove prefix arg that was unused
867 by any caller. Updated all callers.
869 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
871 * format.c (fmt_dollar_template): Use user's decimal point
872 character. Add assertion.
874 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
876 * format.c (fmt_dollar_template): New function, based on
877 dollar_format_template from var-type-dialog.c.
879 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
881 * data-out.c (output_scientific): Fix bad assumption that "buf" is
884 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
886 Finish converting struct variable to an opaque type. In this
887 phase, we add remaining setter and getter functions, convert the
888 remaining PSPP code to use them, and do a bunch of cleanup. The
889 resulting changes are pervasive but mostly trivial, and only the
890 notable changes are logged.
892 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
895 * case.c (case_data): Renamed case_data_idx.
896 (case_num): Renamed case_num_idx.
897 (case_str): Renamed case_str_idx.
898 (case_data_rw): Renamed case_data_rw_idx.
900 * case.h (case_data): New function with old name and an interface
901 that takes a variable instead of an index, which is easier to
902 use. Updated all callers to use the new interface, or to use the
903 new *_idx function (see above).
906 (case_data_rw): Ditto.
908 * category.c (cat_stored_values_destroy): Changed interface to
909 take a struct cat_vals * instead of a struct variable *.
911 * dictionary.c (dict_clone): Use new vector_clone function.
912 (dict_clear) Use new var_destroy function.
913 (add_var) New function.
914 (dict_create_var) Rewrite in terms of dict_create_var_assert.
915 (dict_create_var_assert) Rewrite in terms of add_var.
916 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
917 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
918 (dict_lookup_var) Use new var_create, var_destroy functions.
919 (dict_contains_var) Rewrite in terms of new vardict functionality.
920 (set_var_dict_index) New function.
921 (set_var_case_index) New function.
922 (reindex_vars) New function.
923 (dict_delete_var) Rewrite in terms of new vardict functionality.
924 (dict_reorder_var) Ditto.
925 (dict_reorder_vars) Ditto.
926 (rename_var) New function.
927 (dict_rename_var) Use rename_var.
928 (dict_rename_vars) Use pool to simplify code. Use rename_var.
929 (dict_get_compacted_idx_to_fv) Rename
930 dict_get_compacted_dict_index_to_case_index, update callers.
931 (dict_create_vector) Use new vector_create function.
932 (dict_clear_vectors) Use new vector_destroy function.
933 (set_var_short_name_suffix) Move here from variable.c, renamed
934 from var_set_short_name_suffix, make static, update caller.
936 * sys-file-private.c: New file.
937 (sfm_width_to_bytes) Moved here from variable.c, renamed from
938 width_to_bytes, update callers.
940 * sys-file-private.h: New file. Later it will supplant
941 sfm-private.h; for now it supplements it.
942 (macro MIN_VERY_LONG_STRING) New macro.
943 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
945 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
946 MAX_LONG_STRING was used before.
948 * sys-file-writer.c: Ditto.
950 * value-labels.c: Change the paradigm here to be that a null
951 pointer is OK for a struct val_labs * in most cases; it just
952 represents an empty set of value labels.
953 (val_labs_copy) A copy of a null set is a null set.
954 (val_labs_count) A null set has 0 labels.
955 (val_labs_replace) Change return type to void. Rewrite for
957 (val_labs_find) A null set does not contain the value.
958 (value_to_string) Moved to variable.c, renamed var_get_value_name,
959 transposed argument order, updated all callers.
962 (value_dup) Moved here from variable.c.
963 (compare_values) Ditto.
966 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
967 (macro MAX_LONG_STRING) Removed, because it was only interesting
968 for system files, not for general code.
969 (macro MAX_VERY_LONG_STRING) Ditto.
970 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
971 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
973 * vardict.h: New file, for an interface between variables and
976 * variable.c: A lot of functions were moved around, for better
978 (struct variable) Move definition here, from variable.h.
979 (var_type_adj) Removed--makes i18n hard.
980 (var_type_noun) Ditto.
981 (var_create) New function.
982 (var_clone) New function.
983 (var_destroy) New function.
984 (var_set_name) Assert that variable is not in a dictionary.
985 (compare_var_names) Rename compare_vars_by_name and fix a couple
986 of callers who thought the args were strings.
987 (hash_var_name) Rename hash_var_by_name.
988 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
989 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
990 (var_is_very_long_string) Removed, because it was only interesting
992 (var_set_missing_values) Allow the argument to be the wrong width,
993 as long as we can resize it. Simplify callers who were doing the
995 (var_get_value_labels) New function.
996 (var_has_value_labels) New function.
997 (var_set_value_labels) New function.
998 (alloc_value_labels) New function.
999 (var_add_value_label) New function.
1000 (var_replace_value_label) New function.
1001 (var_clear_value_labels) New function.
1002 (var_lookup_value_label) New function.
1003 (var_get_value_name) Moved here from variable.c, renamed from
1004 var_get_value_name, transposed argument order, updated all
1006 (var_to_string) Moved here, from variable-label.c.
1007 (var_set_leave) New function.
1008 (var_get_leave) New function.
1009 (var_must_leave) New function.
1010 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1011 set_var_short_name_suffix.
1012 (var_get_dict_index) New function.
1013 (var_get_case_index) New function.
1014 (var_get_obs_vals) New function.
1015 (var_set_obs_vals) New function.
1016 (var_has_obs_vals) New function.
1017 (var_get_vardict) New function.
1018 (var_set_vardict) New function.
1019 (var_has_vardict) New function.
1020 (var_clear_vardict) New function.
1021 (value_dup) Moved to value.c.
1022 (compare_values) Ditto.
1025 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1026 (enum ALPHA) Rename VAR_STRING, update all users.
1028 * vector.c: New file.
1029 (struct vector) Moved here, from variable.h.
1030 (check_widths) New function.
1031 (vector_create) New function.
1032 (vector_clone) New function.
1033 (vector_destroy) New function.
1034 (vector_get_name) New function.
1035 (vector_get_var) New function.
1036 (vector_get_var_cnt) New function.
1037 (compare_vector_ptrs_by_name) New function.
1039 * vector.h: New file.
1041 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1043 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1044 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1047 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1049 * value-labels.c (destroy_atoms): New function.
1050 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1052 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1054 Thanks to Jason Stover for pointing out this problem.
1056 * data-out.c (output_number): Use gsl_finite from GSL, which is
1057 portable, instead of isfinite, which is not.
1060 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1062 * variable.c variable.h (value_dup): New function.
1064 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1066 Start converting struct variable to an opaque type. In this
1067 phase, we add a bunch of setter and getter functions and convert
1068 most of the PSPP code to use them. The resulting changes are
1069 pervasive but mostly trivial, and only the notable changes are
1072 * format.c (fmt_equal): New function.
1074 * variable.c (var_type_is_valid): New function.
1075 (measure_is_valid) Moved here, from format.c.
1076 (alignment_is_valid) Moved here, from format.c.
1077 (var_get_name) New function.
1078 (var_set_name) New function.
1079 (width_to_type) New function.
1080 (var_get_type) New function.
1081 (var_get_width) New function.
1082 (var_set_width) New function.
1083 (var_is_numeric) New function.
1084 (var_is_alpha) New function.
1085 (var_is_short_string) New function.
1086 (var_is_long_string) New function.
1087 (var_is_very_long_string) New function.
1088 (var_get_missing_values) New function.
1089 (var_set_missing_values) New function.
1090 (var_clear_missing_values) New function.
1091 (var_has_missing_values) New function.
1092 (var_is_value_missing) New function.
1093 (var_is_num_missing) New function.
1094 (var_is_str_missing) New function.
1095 (var_is_value_user_missing) New function.
1096 (var_is_num_user_missing) New function.
1097 (var_is_str_user_missing) New function.
1098 (var_is_value_system_missing) New function.
1099 (var_get_print_format) New function.
1100 (var_set_print_format) New function.
1101 (var_get_write_format) New function.
1102 (var_set_write_format) New function.
1103 (var_set_both_formats) New function.
1104 (var_get_label) New function.
1105 (var_set_label) New function.
1106 (var_clear_label) New function.
1107 (var_has_label) New function.
1108 (var_get_measure) New function.
1109 (var_set_measure) New function.
1110 (var_get_display_width) New function.
1111 (var_set_display_width) New function.
1112 (var_get_alignment) New function.
1113 (var_set_alignment) New function.
1114 (var_get_value_cnt) New function.
1115 (var_get_leave) New function.
1116 (var_get_short_name) New function.
1118 * variable.h: (struct variable) Removed "type" and "nv" members;
1119 they are now computed from "width" where needed.
1121 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1123 * missing-values.c (mv_resize): Don't write beyond end of the
1124 allocated buffer when resizing a long string.
1126 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1128 Clean up identifier code: don't require identifier enumerations to
1129 be in a particular order; make better use of string library;
1130 expose less of the internals.
1132 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1133 change interface. Updated all callers.
1134 (lex_id_match) Change interface to use struct substring, update
1136 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1137 (global array keywords[]) Make static, change form. Update all
1138 users to use lex_id_name instead.
1139 (lex_is_keyword) New function.
1140 (lex_id_to_token) Change interface to use struct substring, update
1142 (lex_id_name) New function.
1144 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1145 lex_is_keyword instead.
1146 (T_LAST_KEYWORD) Removed.
1147 (T_N_KEYWORDS) Removed.
1149 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1151 * format.c: (fmt_date_template) Distinguish characters for which a
1152 space is output and any date delimiter is allowed on input, from
1153 those for which a space is output and only a space is allowed on
1154 input. The former is represented by X, the latter by a space.
1155 Also, drop distinction between h and H, changing the former to the
1158 * data-in.c: Completely rewrite internals to conform to SPSS input
1159 formats as closely as possible.
1160 (data_in) Changed external interface by replacing the structure
1161 that was used as a single argument by a set of arguments. Updated
1163 (data_in_finite_line) Removed. Converted all callers to use plain
1165 (data_in_get_integer_format) New function.
1166 (data_in_set_integer_format) New function.
1167 (data_in_get_float_format) New function.
1168 (data_in_set_float_format) New function.
1170 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1171 (struct data_in) Removed.
1173 * data-out.c: (output_date) Drop each component from the input as
1174 it is output, to allow us to drop the distinction between h (a
1175 count of hours) and H (the hour of day) template characters.
1176 Also, handle new X template character.
1177 (output_scientific) Follow more rational rule on when to drop
1178 fraction introduced between SPSS 13 and 15. Updated test case to
1181 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1183 Fix buffer overflow reported by John Darrington.
1185 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1186 realize that DIGITS is a count of nibbles, not of bytes.
1188 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1190 * calendar.c (calendar_offset_to_gregorian) Also return the
1191 year-of-day. Change callers to new interface.
1193 * data-out.c: Completely rewrite internals to conform to SPSS
1194 output formats as completely as possible.
1195 (data_out) Change interface to put input parameters before output
1196 parameters, for consistency with the style I now prefer. Update
1198 (data_out_get_integer_format) New public function.
1199 (data_out_set_integer_format) New public function.
1200 (data_out_get_float_format) New public function.
1201 (data_out_set_float_format) New public function.
1203 * data-out.h: New file. Move prototype for data_out here, from
1206 * format.c: (fmt_step_width) Use equality comparison instead of
1207 bitwise and, for clarity.
1208 (fmt_is_string) Ditto.
1209 (fmt_input_to_output) Fix categories that are translated to F
1212 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1214 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1215 casefile-private.h: Added casefilter to assist commands with missing
1218 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1220 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1222 * settings.c: (route_errors_to_terminal) New variable.
1223 (route_errors_to_listing) New variable.
1224 (get_error_routing_to_terminal) New function.
1225 (set_error_routing_to_terminal) New function.
1226 (get_error_routing_to_listing) New function.
1227 (set_error_routing_to_listing) New function.
1229 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1231 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1233 * format.c: Completely rewrite, to achieve better abstraction.
1234 Rewrite all references to formats in other files.
1236 * format.def: Rewrite and reorganize.
1238 * settings.c: Move everything related to custom currency formats
1239 into format.[ch], changing them in form, so as to group related
1240 code and definitions better. Changed all references to use the
1242 (static var decimal) Removed.
1243 (static var grouping) Removed.
1244 (static var cc) Removed.
1245 (get_decimal) Removed.
1246 (set_decimal) Removed.
1247 (get_grouping) Removed.
1248 (set_grouping) Removed.
1252 * settings.h: (macro CC_CNT) Removed.
1253 (macro CC_WIDTH) Removed.
1254 (struct custom_currency) Removed.
1256 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1258 * data-in.c (data_in): Use switch statement instead of table, to
1259 avoid dependence on the order of the FMT_* enums.
1261 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1263 * data-out.c: (num_to_string) Removed, because it was dead code.
1265 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1267 * data-in.c (parse_trailer): Fix error message.
1269 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1271 * format.c (fmt_is_binary): New function.
1273 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1275 * procedure.c procedure.h: Encapsulated the static data into a single
1278 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1280 * casefile.c (casereader_read_xfer): Always initialize the case,
1281 even on an error condition.
1283 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1285 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1286 entire function useless.
1288 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1290 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1291 casereader method casereader_clone.
1293 * procedure.c pransformations.h: Introduced new type casenum_t
1295 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1297 * variable.c: (width_to_bytes) Rephrase code for clarify.
1299 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1301 * format.c: (fmt_type_from_string) New function.
1302 (fmt_to_string) Include decimals in output if the format has
1303 decimals, even if the format type does not. This way, we can
1304 accurately reproduce incorrect formats in user output.
1305 (check_common_specifier) Make the check for a bad format type an
1306 assertion, so we get bug reports if they show up. Fix message.
1307 Check for decimal places with a format type that doesn't allow
1309 (check_input_specifier) Remove check for FMT_X, which has been
1311 (check_output_specifier) Ditto.
1313 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1315 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1316 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1317 (enum fmt_parse_flags) Removed.
1319 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1321 * casefile.c casefile.h: Converted to an abstract base class.
1322 * casefile-private.h fastfile.c fastfile.h: New files.
1323 * automake.mk procedure.c scratch-writer.c storage-stream.c
1325 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1327 * procedure.c (internal_procedure): Create sink_case with only as
1328 many values as the compacted dictionary.
1330 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1332 Remove "debugging" code that caused plenty of false positives and
1335 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1337 * case.c: Remove all references to `this' member.
1339 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1341 Fix link error noted by Jason Stover.
1343 * storage-stream.c: Include <assert.h>.
1345 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1347 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1348 ALL) and additional underlying system file issues.
1350 Thanks to John Darrington for review.
1352 First problem: var_hash points to variables not owned by the
1353 sys-file-reader, which the caller may free or modify. Use an
1354 array of sfm_vars instead, as done earlier (e.g. CVS version
1357 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1358 members and remove all code that references it. Add vars, var_cnt
1359 members. Remove fix_specials member, which was unused.
1360 (struct sfm_var) Remove name member, which was unused.
1361 (sfm_close_reader) Free vars member instead of var_hash.
1362 (compare_var_shortnames) Removed.
1363 (hash_var_shortname) Removed.
1364 (sfm_open_reader) Fill out vars array.
1365 (compare_var_index) Removed.
1366 (sfm_read_case) Use vars instead of var_hash.
1368 Second problem: we're confused about when we actually have very
1369 long strings, causing us to choose incorrectly between slow path
1370 and fast path in sfm_read_case.
1372 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1373 have very long strings, not when we have long variable names,
1374 which is an unrelated feature.
1376 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1378 * variable.h: Move var_set and variable parsing declarations to
1379 new header, src/language/lexer/variable-parser.h. Modified lots
1380 of files to include the new header.
1382 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1384 * value-labels.c (value_to_string): When there's no value label,
1385 format the variable according to its print format, instead of
1386 always effectively using A or F format.
1388 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1390 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1393 * format.c: Removed tortological assertion.
1395 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1397 Reform string library.
1399 * file-name.c (fn_interp_vars): Change interface to take a
1400 substring as input. Updated all users.
1402 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1404 * format.c (measure_is_valid): Really return false when m >=
1407 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1409 Implement random access to casefiles, for use in GUI.
1411 * casefile.c: (struct casereader) Add `random', `file_ofs',
1412 `buffer_ofs' members.
1413 (casefile_get_random_reader) New function.
1414 (read_open_file) Break part into new function
1415 seek_and_fill_buffer().
1416 (fill_buffer) Update buffer_ofs, file_ofs.
1417 (casereader_seek) New function.
1419 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1421 * settings.c: Added call to i18n{done, init}.
1423 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1425 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1428 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1430 * casefile.c: Convert many uses of `int' to `bool'.
1432 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1434 * transformations.c (trns_chain_destroy): Destroy chain's trns
1435 member, to fix memory leak.
1437 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1439 * storage-stream.c (storage_source_decapsulate): Destroy case
1440 source to fix memory leak.
1442 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1444 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1445 case passed as argument instead of initializing the argument as a
1446 case. Fixes memory leak that showed up in
1447 tests/command/aggregate.sh with scratch files.
1449 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1451 * procedure.c (proc_done): Destroy default_dict, to fix memory
1454 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1456 Simplify procedure_with_splits().
1458 * procedure.c (struct split_aux_data): Removed case_count member.
1459 (procedure_with_splits) Don't initialize case_count.
1460 (split_procedure_case_func) Check whether prev_case is null
1461 instead of case_count.
1462 (split_procedure_end_func) Ditto.
1464 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1466 * case.c (case_move): Do nothing if dst and src are the same
1468 (case_try_create) Merge two similar cases.
1469 (case_copy) Unshare only if data must be actually copied.
1471 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1473 * data-in.c data-out.c dictionary.c sys-file-reader.c
1474 sys-file-writer.c variable.c variable.h: Reworked very long string
1475 support for better encapsulation.
1477 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1479 * value-labels.c (val_labs_can_set_width): New function.
1480 (val_labs_set_width) Clear labels if increasing width to long
1482 (val_labs_destroy) Remove unneeded test for null.
1484 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1486 * value-labels.h: Remove unneeded dependency on variable.h.
1488 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1490 Get rid of `char *c' member in union value, for cleanliness.
1492 * value.h: (union value) Remove `c' member.
1494 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1496 Make dictionary compacting functions a little more general.
1498 * sys-file-writer.c (sfm_open_writer): Use
1499 dict_compacting_would_change().
1500 (does_dict_need_translation) Removed.
1502 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1504 Make dictionary compacting functions a little more general.
1506 * dictionary.c (dict_needs_compaction): Rename
1507 dict_compacting_would_shrink(). Update all callers.
1508 (dict_compacting_would_change) New function.
1510 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1512 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1513 inverted return value (!).
1515 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1517 Continue reforming procedure execution.
1519 * procedure.c: Search and replace "vfm" by "proc". Notably:
1520 (static var vfm_source) Rename proc_source. Update all
1522 (static var vfm_sink) Rename proc_sink. Update all references.
1524 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1526 Continue reforming procedure execution. In this phase, remove
1527 PROCESS IF, which was deprecated anyway and can be easily
1528 simulated with TEMPORARY followed by SELECT IF.
1530 * procedure.c: (open_active_file) Don't call
1531 add_process_if_trns().
1532 (discard_variables) Get rid of redundant call to
1533 proc_cancel_all_transformations().
1534 (add_process_if_trns) Removed.
1535 (process_if_trns_proc) Removed.
1536 (process_if_trns_free) Removed.
1538 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1540 Continue reforming procedure execution. In this phase, add
1541 `const' to the case passed to procedure()'s callback.
1543 Updated all users of procedure() as well.
1545 * procedure.c: (struct write_case_data) Add "const" to ccase
1546 parameter for case_func member.
1547 (procedure) Add "const" to ccase parameter for proc_func
1549 (multipass_case_func) Make ccase parameter const.
1550 (internal_procedure) Add "const" to ccase parameter for case_func
1552 (split_procedure_case_func) Make ccase parameter const.
1553 (multipass_split_case_func) Make ccase parameter const.
1555 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
1557 Continue reforming procedure execution. In this phase, get rid of
1558 the output code for SPLIT FILE groups in procedure.c, which really
1559 shouldn't be doing any output. Move it into the individual
1560 procedures instead. This also adds some flexibility.
1562 Updated many users of procedure_with_splits() and
1563 multipass_procedure_with_splits() to call
1564 output_split_file_values() and to deal with increased use of
1567 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
1568 parameter to begin_func member.
1569 (procedure_with_splits) Add "const struct ccase *" parameter to
1570 begin_func parameter. Make ccase parameter const in proc_func
1572 (split_procedure_case_func) Don't dump split file group. Pass
1574 (dump_splits) Moved to language/dictionary/split-file.c as
1575 output_split_file_values().
1576 (struct multipass_split_aux_data) Add "const struct ccase *"
1577 parameter to split_func member.
1578 (multipass_procedure_with_splits) Add "const struct ccase *"
1579 parameter to split_func parameter.
1580 (multipass_split_case_func) Save new SPLIT FILE case before
1582 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
1584 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
1586 Continue reforming procedure execution. Change
1587 internal_procedure() so that it calls open_active_file() and
1588 close_active_file(), which isolates most of the actual procedure
1591 * procedure.c: (struct write_case_data) Rename `proc_func' member
1592 to `case_func' and update all references.
1593 (procedure) Rewrite as one-line wrapper around
1594 internal_procedure().
1595 (struct multipass_aux_data) New.
1596 (multipass_callback) Renamed multipass_case_func(). Use struct
1597 multipass_aux_data as auxiliary data.
1598 (multipass_end_func) New function.
1599 (multipass_procedure) Rewrite as wrapper for internal_procedure()
1600 that uses multipass_case_func, multipass_end_func.
1601 (internal_procedure) Add `end_func' argument. Move optimization
1602 of trivial case in here. Move call to open_active_file() and
1603 close_active_file() in here. Now assert that vfm_source is
1605 (procedure_with_splits_callback) Rename
1606 split_procedure_case_func().
1607 (split_procedure_end_func) New function.
1608 (multipass_split_callback) Rename multipass_split_case_func.
1609 (multipass_split_end_func) New function.
1610 (discard_variables) No need to test for nonnull vfm_source.
1612 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
1614 Continue reforming procedure execution. Get rid of unused member.
1616 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
1618 (write_case) Don't increment cases_analyzed.
1620 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
1622 Continue reforming procedure execution. In this phase, move
1623 procedure.c and procedure.h from src to src/data. Update
1624 makefiles and #includes accordingly.
1626 * procedure.c: Moved here from src/.
1628 * procedure.h: Moved here from src/.
1630 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
1632 Continue reforming procedure execution. In this phase, get rid of
1633 many global variables, consolidating procedure execution in
1634 procedure.c. Encapsulate transformations in new "struct
1635 trns_chain". Also, change implementation of N OF CASES, FILTER,
1636 and PROCESS IF from special cases to transformations.
1638 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
1641 * dictionary.c: (global variable default_dict) Move to
1644 * variable.h: (TRNS_*) Move to transformations.h.
1645 (struct transformation) Move to transformations.c.
1647 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1649 * sys-file-reader.c: Fixed invalid read problems.
1651 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
1653 * storage-stream.c: Add missing function comments.
1655 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
1657 Continue reforming procedure execution. In this phase, add some
1658 new, needed functionality to storage-stream.
1660 * storage-stream.c: (storage_source_decapsulate) New function.
1662 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
1664 * variable.c (width_to_bytes): Declarations must precede
1665 statements for C90 compliance.
1667 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
1669 * data-out.c, data-in.c, variable.c, variable.h: New functions
1670 copy_mangle and copy_demangle for reading/writing cases; emulates the
1671 way SPSS deals with strings > 255 bytes.
1673 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
1674 subtype 14 needed for strings longer than 255 bytes.
1676 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
1677 instead of literal values. Also fixed some constness issues.
1679 * format.h: Constness
1681 * sfm-private.h: Renamed the case_size identifier, since I discovered
1682 that SPSS's respect for this variable is very nominal.
1684 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
1686 Change case limit type from int to size_t.
1688 * dictionary.c: (struct dictionary) Change type of case_limit
1690 (dict_get_case_limit) Change return type.
1691 (dict_set_case_limit) Change parameter type.
1693 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
1695 * variable.h: (struct variable) Rename `reinit' member as `leave'
1696 and invert sense. Fix up all references.
1698 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
1700 Continue reforming procedure execution. In this phase, break
1701 procedure.c into multiple files.
1703 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
1705 * case-sink.c: New file.
1707 * case-sink.h: New file.
1709 * case-source.c: New file.
1711 * case-source.h: New file.
1713 * storage-stream.c: New file.
1715 * storage-stream.h: New file.
1717 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
1719 * variable.h: (struct variable) Remove `init' member and all
1720 references to it from other files. It was initialized in several
1721 places, but nothing really ever used it for anything worthwhile.
1722 Thanks to Jason Stover for pointing out how confusing this
1725 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
1727 Continue reforming error message support. In this phase, get rid
1728 of message "titles" and put the message text in `struct error'.
1729 Now `struct error' encapsulates a message more properly.
1731 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
1732 Format message ourselves.
1734 * data-in.c: (vdls_error) Ditto.
1736 * por-file-reader.c: (error) Ditto.
1738 * sys-file-reader.c: (corrupt_msg) Ditto.
1740 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
1742 GNU standards require "file name" instead of "filename" in
1743 documentation. It's nice for our code to follow the convention
1746 * casefile.c: (struct casefile) Rename `filename' member to
1747 `file_name'. Updated all references.
1749 * file-name.c: [!unix] (struct file_identity) Rename
1750 normalized_filename member to normalized_file_name. Updated all
1753 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
1755 We don't really support anything but Unix-like environments well,
1756 so we might as well de-obfuscate by writing directory and path
1757 separators explicitly.
1759 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
1761 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
1762 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
1764 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
1767 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
1769 GNU standards require "file name" instead of "filename" in
1770 documentation. It's nice for our code to follow the convention
1773 * filename.c: Rename to file-name.c.
1775 * filename.h: Rename to file-name.h. Update all inclusions.
1776 Update header guards.
1778 * automake.mk: Update file names.
1780 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
1782 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
1784 (fn_basename) Removed (dead code).
1785 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
1786 (fn_special_p) Renamed fn_is_special(), all references updated.
1787 (fn_exists_p) Renamed fn_exists(), all references updated.
1789 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
1791 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
1792 Also, now it only tilde-expands file names, not paths.
1793 (fn_search_path) Tilde-expand one directory at a time.
1795 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
1797 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
1798 get rid of non-Unixlike version of the code, which has probably
1800 (fn_prepend_dir) Removed (dead code).
1802 * filename.h: (macro DIR_SEPARATOR_STRING) New.
1803 (macro PATH_SEPARATOR_STRING) New.
1804 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
1806 Continue reforming error message support. In this phase, we get
1807 rid of VM() and the other msg() support for "verbosity", replacing
1808 it by a new function verbose_msg().
1810 * filename.c: (fn_search_path) Use verbose_msg() instead of
1813 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
1815 * sfm-private.h: Get rid of #defines after #error, which makes no
1818 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
1820 Get rid of our own int32 type in favor of the standard int32_t
1823 * sfm-private.h: (int32 macro) Don't define this anymore. Do
1826 * sys-file-reader.c: Use int32_t instead of int32 throughout.
1828 * sys-file-writer.c: Use int32_t instead of int32 throughout.
1830 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
1832 Remove ill-considered file routines that are no longer used.
1834 * filename.c: (fn_open_ext) Removed.
1835 (fn_close_ext) Removed.
1837 * filename.h: (struct file_ext) Removed.
1839 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
1841 * variable.c (var_is_valid_name): Move declarations before code
1844 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
1846 * filename.ch (fn_interp_vars): Fixed small buglet.
1848 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
1850 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
1851 so as to modify the string inline. Thus makeing it easier to
1852 destroy the results when no longer needed.
1854 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
1856 * category.c (cat_stored_values_destroy): Fixed memory leak.
1858 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
1860 Add some missing frees. Thanks to John Darrington for reporting
1863 * any-writer.c (any_writer_close): Free writer.
1865 * any-reader.c (any_reader_close): Free reader.
1867 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
1869 * por-file-reader.c: (error) Mark as NO_RETURN.
1871 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
1873 * settings.c: Changed default value of scompress to true.
1875 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
1877 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
1879 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
1882 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
1883 here from pref.h.orig.
1885 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
1887 * sys-file-reader.c: Fixed bug reading compressed files.
1889 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1891 * Numerous renames. See src/ChangeLog for details.
1893 * Moved files from src directory