1 2007-07-28 Ben Pfaff <blp@gnu.org>
3 Make PSPP able to read all the portable files I could find on the
4 web. Thanks to John Darrington for review. Bug #17620.
5 * por-file-reader.c (struct pfm_reader): New member `line_length'.
6 (error): Print file offset in hexadecimal.
7 (warning): New function.
8 (advance): Treat lines less than 80 bytes long as padded to 80
10 (pfm_open_reader): Call read_documents if we find an "E" record.
11 (convert_format): Convert invalid formats to the default format
12 instead of aborting reading the file.
13 (read_variables): Rename duplicate variable names instead of
14 aborting reading the file.
15 (read_value_label): Allow string variables of different widths to
16 be assigned value labels in the same record. Replace duplicate
17 value labels instead of aborting.
18 (read_documents): New function.
20 * por-file-writer.c (pfm_open_writer): Call write_documents if the
21 dictionary has documents.
22 (write_documents): New function.
24 2007-07-25 Ben Pfaff <blp@gnu.org>
26 Fix bugs related to bug #17213.
28 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
29 since the former is what config.h has. Include the needed ncurses
31 (static var echo) Rename to `do_echo' because the original name is
32 the same as an ncurses identifier.
33 (get_termcap_viewport) Use error instead of msg.
35 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
36 (fn_close): Don't close stdin, stdout, stderr.
38 2007-07-26 John Darrington <john@darrington.wattle.id.au>
40 * procedure.c procedure.h: Added callbacks which get invoked whenever
41 a dataset's transformation chain changes.
43 2007-07-24 Ben Pfaff <blp@gnu.org>
46 * sys-file-writer.c (write_variable_display_parameters): Use new
47 var_default_display_width function to choose display width of
48 segments after the first one in a given variable.
49 * variable.c (var_create): Use var_default_display_width to pick
50 new variable's display width.
51 (var_default_display_width): New function.
52 Reviewed by John Darrington.
54 2007-07-24 Ben Pfaff <blp@gnu.org>
57 * por-file-writer.c (write_variables): Write weight variable.
58 Reviewed by John Darrington.
60 2007-07-23 Ben Pfaff <blp@gnu.org>
62 Improvements to system file reader and writer.
64 First, move all detailed knowledge of very long strings into
65 sys-file-private.[ch], so that this nasty stuff can be isolated.
67 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
68 (EFFECTIVE_VLS_CHUNK): New macro.
69 (min_int): New function.
70 (max_int): New function.
71 (sfm_width_to_bytes): Rewrite.
72 (sfm_width_to_octs): New function.
73 (sfm_segment_alloc_width): New function.
74 (sfm_segment_alloc_bytes): New function.
75 (sfm_segment_used_bytes): New function.
76 (sfm_segment_offset): New function.
77 (sfm_segment_effective_offset): New function.
78 (sfm_dictionary_to_sfm_vars): New function.
80 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
81 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
82 (struct sfm_var): New structure.
84 Next, improvements to the system file reader.
86 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
87 type casenumber. Added `version_major', `version_minor',
90 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
91 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
92 Change `case_cnt' to type casenumber. Removed `has_vls'.
93 (struct sfm_var): Removed.
94 (sfm_open_reader): Don't warn on wrong case size if the file was
95 written by SPSS 13, which tends to get it wrong. Use
96 sfm_dictionary_to_sfm_vars.
97 (read_header): Always output system file info.
98 (read_variable_record): Simplify code for reading missing values.
99 (read_machine_int32_info): Save version numbers from system file
100 into info struct passed as new argument.
101 (read_long_string_map): Restructured to use new sys-file-private
103 (read_value_labels): Use size_overflow_p.
104 (sys_file_casereader_read): Get rid of distinction between fast
105 and slow paths. Use information provided by sys-file-primate's
106 struct sfm_var to simplify code.
107 (skip_whole_strings): New function.
108 (read_int32): Renamed read_int. Changed return value to int.
110 (read_flt64): Renamed read_float. Changed return value to
111 double. Updated all callers.
112 (int32_to_native): Removed. Changed callers to use
114 (flt64_to_double): Removed. Changed callers to use float_convert.
116 Finally, get rid of int32, flt64 terminology and types in system
117 file writer. The former wasn't very useful since a POSIX "int"
118 can hold the whole range of int32 and we generally didn't have a
119 need for it to be exactly-32-bits, just at-least-32-bits. The
120 latter was inconvenient because we had to assume that it could be
121 different from double and thereby convert special values SYSMIS,
122 HIGHEST, LOWEST to and from it in multiple places. Instead, now
123 we just use "int" and "double" in most places, and do conversions,
124 if necessary, very close to where we do I/O. This change meant
125 that the writer code couldn't represent records in the file as C
126 structs any longer, but that's no great loss. The code actually
127 seems to be more readable without them.
129 Simplify the compression buffering code: only buffer as much as
130 necessary, which is no more than eight 8-byte units at any given
133 * sys-file-writer.c (typedef flt64): Removed.
134 (macro second_lowest_flt64): Removed.
135 (struct sysfile_header): Removed.
136 (struct sysfile_variable): Removed.
137 (struct sfm_writer): Removed `needs_translation', `has_vls',
138 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
139 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
140 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
141 compression buffering by `opcodes', `opcode_cnt', `data',
142 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
143 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
144 data writer functions instead of structures.
145 (calc_oct_idx): New function.
146 (write_header): Use simple data writer functions instead of
148 (write_format_spec): Renamed write_format. New argument.
149 (write_variable_continuation_records): New function.
150 (write_variable): Use simple data writer functions instead of
151 structures. Use write_variable_continuation_records. Write
152 entire very long string instead of requiring caller to understand
154 (write_value_labels): Use simple data writer functions instead of
156 (write_documents): Ditto.
157 (write_variable_display_parameters): Use sys-file-private
158 functions to simplify. Use simple data writer functions instead
160 (write_vls_length_table): Use simple data writer functions instead
162 (write_longvar_table): Ditto.
163 (write_rec_7_34): Break into new functions
164 write_integer_info_record, write_float_info_record. Use simple
165 data writer functions instead of structures.
166 (buf_write): Removed.
167 (append_string_max): Removed.
168 (ensure_buf_space): Removed.
169 (sys_file_casewriter_write): Get rid of the distinction between
170 fast and slow paths, which didn't seem to be too useful. Use new
171 functions write_case_uncompressed, write_case_compressed.
172 (put_instruction): Removed.
173 (put_element): Removed.
174 (write_compressed_data): Removed.
175 (close_writer): Use flush_compressed. Only write case count to
176 system file if it will fit in the field.
177 (write_case_compressed): New function.
178 (write_case_uncompressed): New function.
179 (flush_compressed): New function.
180 (put_cmp_opcode): New function.
181 (put_cmp_number): New function.
182 (write_int): New function.
183 (convert_double_to_output_format): New function.
184 (write_float): New function.
185 (write_value): New function.
186 (write_string): New function.
187 (write_bytes): New function.
188 (write_zeros): New function.
189 (write_spaces): New function.
191 Reviewed by John Darrington.
193 2007-07-22 Ben Pfaff <blp@gnu.org>
195 Don't try to write very long strings to portable files. The
196 format does not support it.
198 * por-file-writer.c (MAX_POR_WIDTH): New macro.
199 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
200 (write_format): Add arg to take width to resize format to.
201 (write_value): Limit width of value written to MAX_POR_WIDTH.
202 (write_variables): Limit width of variable and its output formats
204 Reviewed by John Darrington.
206 2007-07-22 Ben Pfaff <blp@gnu.org>
208 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
209 local variable instead of literal 5.
210 Reviewed by John Darrington.
212 2007-07-22 Ben Pfaff <blp@gnu.org>
214 Fix problems with uniqueness of short names in system files with
215 very long string variables. Now a variable may have multiple
218 * automake.mk (src_data_libdata_a_SOURCES): Add new files
219 short-names.c, short-names.h.
221 * dictionary.c (dict_clone): Clone all the short names.
222 (compare_strings): Move into short-names.c.
223 (hash_strings): Ditto.
224 (set_var_short_name_suffix): Ditto.
225 (dict_assign_short_names): Ditto, rename short_names_assign,
226 change to assign all short names.
228 * por-file-writer.c (write_variables): Use short_names_assign
229 instead of dict_assign_short_names.
231 * short-names.c: New file.
233 * short-names.h: New file.
235 * sys-file-private.c (sfm_width_to_segments): New function.
237 * sys-file-reader.c (read_long_var_name_map): Save and restore all
238 the short names, not just the first one.
240 * sys-file-writer.c (cont_var_name): Removed.
241 (sfm_open_writer): Use short_names_assign instead of
242 dict_assign_short_names. Use unique short names assigned by
243 short_names_assign instead of those generated by cont_var_name.
245 * variable.c (struct variable): Remove `short_name' member,
246 replace by `short_names' and `short_name_cnt'.
247 (var_create) Initialize new members.
248 (var_get_short_name_cnt): New function.
249 (var_get_short_name): Now takes an index argument. Changed most
251 (var_set_short_name): Ditto.
252 (var_clear_short_name): Renamed var_clear_short_names, changed to
253 clear all short names.
255 Reviewed by John Darrington.
257 2007-07-22 Ben Pfaff <blp@gnu.org>
259 * variable.c (var_set_width): Use new var_set_width function.
261 * missing-values.c (mv_n_values): Drop assertion, which was not
264 * format.c (fmt_default_for_width): New function.
265 (fmt_resize): New function.
267 Reviewed by John Darrington.
269 2007-07-18 John Darrington <john@darrington.wattle.id.au>
271 * datasheet.c (datasheet_delete_columns): Added assertion to check
272 we're not deleting outside the range of the sheet.
275 * dictionary.c dictionary.h variable.c: Added the ability for string
276 variables to be resized.
278 * vardict.h: Added some prototypes (moved from dictionary.h) as
279 these should only be called by variable.c
282 2007-07-14 John Darrington <john@darrington.wattle.id.au>
284 * sfm-reader.c: Respect case_cnt field in file header.
286 2007-07-01 John Darrington <john@darrington.wattle.id.au>
288 * transformation.c transformation.h (trns_chain_execute): Changed the
289 signature (Patch #6057)
291 2007-06-10 Ben Pfaff <blp@gnu.org>
293 * casereader-filter.c (casereader_filter_destroy): Make sure to
294 write all the remaining excluded cases to the casewriter, if any.
296 * caseinit.c (init_list_destroy): Rewrite.
297 (init_list_clear): Ditto.
299 * casegrouper.c (casegrouper_get_next_group): Always set *reader
300 to null when returning false.
302 2007-06-06 Ben Pfaff <blp@gnu.org>
304 Actually implement the new procedure code and adapt all of its
305 clients to match. Also adapt all of the other case sources and
306 sinks in the tree and their clients to use the
307 casereader/casewriter infrastructure.
309 * automake.mk: Add and remove files.
311 * any-reader.c: Change into a casereader.
312 * por-file-reader.c: Ditto.
313 * scratch-reader.c: Ditto.
314 * sys-file-reader.c: Ditto.
316 * any-writer.c: Change into a casewriter.
317 * por-file-writer.c: Ditto.
318 * scratch-writer.c: Ditto.
319 * sys-file-writer.c: Ditto.
321 * procedure.c: Change to use casereader, casewriter, caseinit, and
322 other new infrastructure.
324 * scratch-handle.c: Adapt to new infrastructure.
326 * case-sink.c: Removed, now dead code.
327 * case-sink.h: Ditto.
328 * case-source.c: Ditto.
329 * case-source.h: Ditto.
330 * casefile-factory.c: Ditto.
331 * casefile-private.h: Ditto.
334 * casefilter.c: Ditto.
335 * casefilter.h: Ditto.
338 * fastfile-factory.c: Ditto.
339 * fastfile-factory.h: Ditto.
340 * storage-stream.c: Ditto.
341 * storage-stream.h: Ditto.
343 2007-06-06 Ben Pfaff <blp@gnu.org>
347 * automake.mk: Add new files.
349 * datasheet.c: New file.
351 * datasheet.h: New file.
353 2007-06-06 Ben Pfaff <blp@gnu.org>
355 Until now, the procedure code has provided a case to the
356 case_source, which has filled in the data values that come from
357 the active file. "Left" data values that don't come from the
358 active file naturally stay the same from case to case, because the
359 procedure code keeps using that same case.
361 One of the compromises that comes with the new procedure code is
362 that the active file allocates and provides its own case, which
363 the procedure code then has to resize to provide room for any
364 other variables that should go in the case and then fill in the
365 values of "left" variables. Then, when we're done with that case,
366 we have to save the values of "left" variables to copy into the
367 next case read from the active file.
369 The caseinit code helps with this.
371 * automake.mk: Add new files.
373 * caseinit.c: New file.
375 * caseinit.h: New file.
377 2007-06-06 Ben Pfaff <blp@gnu.org>
379 * value.h (value_cnt_from_width): New function.
381 * variable.c (var_get_value_cnt): Use new function.
383 2007-06-06 Ben Pfaff <blp@gnu.org>
385 Add casegrouper, to allow cases read from a given casereader to be
386 broken into groups, each of which has its own casereader.
387 Generally cases are grouped based on having equal values for some
390 * automake.mk: Add new files.
392 * casegrouper.c: New file.
394 * casegrouper.h: New file.
396 2007-06-06 Ben Pfaff <blp@gnu.org>
398 Add interface to lexicographical ordering of cases.
400 * automake.mk: Add new files.
402 * case-ordering.c: New file.
404 * case-ordering.h: New file.
406 2007-06-06 Ben Pfaff <blp@gnu.org>
408 Add casereaders and casewriters, the basis of the new data processing
409 implementation. A casereader is a uniform interface to reading cases
410 from a data source; a casewriter is a uniform interface to writing
411 cases to a data sink.
413 * automake.mk: Add new files.
415 * casereader-filter.c: New file.
417 * casereader-provider.h: New file.
419 * casereader-translator.c: New file.
421 * casereader.c: New file.
423 * casereader.h: New file.
425 * casewriter-provider.h: New file.
427 * casewriter-translator.c: New file.
429 * casewriter.c: New file.
431 * casewriter.h: New file.
433 2007-06-06 Ben Pfaff <blp@gnu.org>
435 "casewindow" data structure that extends the deque (from libpspp)
436 of cases with the ability to dump cases to disk if we get too many
439 * automake.mk: Add new files.
441 * casewindow.c: New file.
443 * casewindow.h: New file.
445 2007-06-06 Ben Pfaff <blp@gnu.org>
447 sparse_cases data structure that augments a sparse_array of cases
448 with the ability to dump cases to disk if we get too many cases in
451 * automake.mk: Add new files.
453 * sparse-cases.c: New file.
455 * sparse-cases.h: New file.
457 2007-06-06 Ben Pfaff <blp@gnu.org>
459 Adds a low-level on-disk case array data structure.
461 * automake.mk: Add new files.
463 * case-tmpfile.c: New file.
465 * case-tmpfile.h: New file.
467 2007-06-06 Ben Pfaff <blp@gnu.org>
469 In a couple of places we calculate the maximum number of cases to
470 keep in memory based on the user-defined workspace. Enable
471 centralizing the calculation through a new function.
473 * settings.c (get_workspace_cases): New function.
475 2007-06-06 Ben Pfaff <blp@gnu.org>
477 The casenumber type is defined in transformations.h, but case.h is
478 a more sensible place. Move it.
480 * case.h (CASENUMBER_MAX): New macro.
481 (typedef casenumber): Move here, from transformations.h.
483 2007-06-03 Ben Pfaff <blp@gnu.org>
485 Slightly generalize case_to_values and case_from_values functions.
487 * case.c (case_to_values): Rename case_copy_out, change interface.
488 (case_from_values): Rename case_copy_in, change interface.
490 * fastfile.c (fastfilereader_get_next_case): Update caller.
491 (write_case_to_disk): Ditto.
493 2007-06-02 Ben Pfaff <blp@gnu.org>
495 Clean up after a forgotten part of patch #5829.
497 * casedeque.h: Remove unused file.
499 * automake.mk: Remove casedeque.h from sources.
501 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
503 * category.c: Removed redundant #include
505 2007-05-06 Ben Pfaff <blp@gnu.org>
507 Abstract the documents within a dictionary a little better.
508 Thanks to John Darrington for suggestion, initial version, and
511 * dictionary.c (struct dictionary): Change `documents' member from
512 char * to struct string.
513 (dict_clear): Destroy struct string.
514 (dict_get_documents): Convert struct string to char *.
515 (dict_set_documents): Set struct string. Pad to 80-character
517 (dict_clear_documents): New function.
518 (dict_add_document_line): New function.
519 (dict_get_document_line_cnt): New function.
520 (dict_get_document_line): New function.
522 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
524 * sys-file-reader.c (read_documents): Use new document functions.
526 2007-04-19 John Darrington <john@darrington.wattle.id.au>
528 * sys-file-reader.c: When reading a system file which has no
529 long name table, automatically create one where the long names
530 are the lower case versions of the short names.
532 2007-04-22 Ben Pfaff <blp@gnu.org>
534 * dictionary.c (dict_set_split_vars): dict_destroy expects that
535 dict_clear will free most data related to the dictionary.
536 dict_clear does a decent job, except that dict_set_split_vars on
537 some systems won't actually free the dict's "split" member.
538 Instead, it'll allocate a 1-byte region. Fix this.
540 * value.c (value_copy): New function.
541 (value_set_missing): Ditto.
543 2007-04-22 John Darrington <john@darrington.wattle.id.au>
545 * Deleted existing category.h and moved cat-routines.h into
546 category.h Encapsulated struct cat_vals better.
548 2007-04-19 John Darrington <john@darrington.wattle.id.au>
550 * sys-file-reader.c: When reading a system file which has no
551 long name table, automatically create one where the long names
552 are the lower case versions of the short names.
554 2007-04-16 John Darrington <john@darrington.wattle.id.au>
556 * sys-file-reader.c: Some versions of Other Software seem to
557 produce system files with string variables' measure set to
558 zero. We'll assume these are supposed to be nominal variables.
560 2007-03-30 Ben Pfaff <blp@gnu.org>
562 * procedure.c: Adapt to new deque data structure.
564 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
565 Ben Pfaff <blp@gnu.org>
567 * file-name.c: Mingw compatibility fixes.
568 (fn_search_path): Use ISSLASH instead of comparing against '/'
570 (fn_dir_name): Use dir_name from gnulib.
571 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
572 (fn_get_identity): Use GetFullPathName instead of canonicalize
573 from gnulib, because the latter does not fully support
574 Windows-style path names. Use this implementation based on the
575 detected presence of Windows instead of the absence of Unix, since
576 the new implementation is Windows-specific.
577 (fn_compare_file_identities): In Windows implementation, compare
578 names case-insensitively.
580 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
582 * make-file.c: Don't include mkstemp.h, because gnulib now causes
583 <stdlib.h> to have the same effect.
585 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
587 * por-file-reader.c: Add missing _() around messages.
589 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
591 * make-file.c: Include "mkstemp.h", without which linking on
594 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
596 Reduce platform dependence.
598 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
599 Everywhere we using this, we really should have just depended on
600 the shell to expand tildes.
601 (fn_search_path): Simplify, given that we don't do tilde expansion
603 (fn_normalize): Removed. Caller changed to use the canonicalize
605 (fn_get_cwd): Removed. Only user was fn_normalize.
606 (fn_is_absolute): Really only test for absolute names.
607 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
609 (fn_readlink): Removed, as it was only used fn_normalize.
610 (fn_exists): Assume the stat function is available; gnulib does.
611 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
613 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
615 * dictionary.c (dict_create_vector_assert): New function.
617 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
619 * file-name.c (fn_normalize): Correct name of function
620 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
621 for pointing this out.
623 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
625 We are using a single member in struct file_handle, the "name"
626 field, for more than one purpose. When it begins with '"', it's a
627 file name; otherwise, it's a token that can be used to identify
628 it. When that assertion fires, it's because we searched for the
629 name case-sensitively as a file name (so that there was no match),
630 and then we try to insert it case-insensitively as a token, which
631 fails because duplicates aren't allowed.
633 Solution: break the two purposes into two separate fields. This
634 fixes the problem and likely makes the code easier to read too.
636 Fixes bug #18922. Thanks to John Darrington for bug report and
639 * file-handle-def.c (struct file_handle): New `id' member.
640 (fh_from_name): Rename fh_from_id. Update all callers.
641 (create_handle): New `id' parameter. Update all callers.
642 (fh_create_file): Ditto.
643 (fh_get_id): New function.
645 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
647 * case.c (case_is_null): Change return type to bool.
649 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
653 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
654 reference counting to share data. This makes it easy for
655 valgrind, etc. to find accesses to cases that have been destroyed
656 but have been kept around by another user's ref-count. This often
657 happens when the data set is small enough to find in memory; if a
658 bigger data set that would overflow to disk were used, then data
659 corruption would occur.
661 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
665 * case.c (case_unshare): Make it check internally whether the
666 ref_cnt is greater than 1, so that the callers don't have to.
667 Update callers not to check.
669 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
671 Before, I was thinking that I might want to get rid of reference
672 counting at some point. Now, I'm pretty sure that it's here to
673 stay. Thus, because we have to store the value_cnt anyway for
674 reference-counted cases, we might as well expose it to users.
676 * case.c (case_get_value_cnt): New function.
677 (case_resize): Drop OLD_CNT argument. Update all callers. Only
678 resize case if its size actually changed.
680 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
681 instead of peeking inside struct case directly.
682 (casefile_append): Ditto.
684 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
686 Get rid of the inlines for the case functions, which made the
687 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
688 the inlines didn't speed up "make check" at all, which is not a
689 perfect benchmark but seems indicative.)
691 * case.c: Remove #ifdef DEBUGGING...#endif around many function
692 definitions. Remove some assertions on nonnull pointers that were
693 redundant with a pointer dereference soon after in the function.
695 (struct case_data): Move definition here from case.h.
699 (case_data_wr): Ditto.
701 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
703 * automake.mk: Add casedeque.h to sources.
705 * casedeque.h: New file.
707 * procedure.c: (struct dataset) Change lag_count, lag_head,
708 lag_queue member into single struct casedeque member. Update all
709 users to use the casedeque instead.
712 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
714 * procedure.c: Simplify lagged cases interface. Updated all
715 clients--well, the only client--to use the simplified interface.
716 (dataset_n_lag) Removed.
717 (dataset_set_n_lag) Removed.
718 (dataset_need_lag) New function.
720 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
722 * dictionary.c procedure.c: More changes to ensure that callbacks occur
723 whenever appropriate, but only when the dataset/dictionary is in a
726 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
728 * dictionary.c dictionary.h : Added callbacks for change of filter and
729 split variables. Refactored some code to ensure that callbacks get
730 invoked when appropriate.
732 * procedure.c (proc_cancel_temporary_transformations): Make sure that
733 replace_dict callback occurs when permanent_dict replaces the current
736 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
738 * dictionary.c dictionary.h : Added callback for when the weight
739 variable of a dictionary changes.
741 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
743 * dictionary.c dictionary.h : Added replace_source and replace_dict
744 callbacks, and functions to deal with them.
746 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
748 Simplify missing value handling.
750 * missing-values.h (enum mv_class): New type.
751 (enum mv_type): Moved definition into missing-values.c and renamed
752 each MV_* to MVT_*, to distinguish them from the exposed mv_class
753 enums. Updated all uses.
754 (struct missing_values): Changed type of `type' from `enum
755 mv_type' to `int' because the definition is no longer exposed.
757 * missing-values.c (mv_is_value_missing): Add new enum mv_class
758 parameter. Update all callers.
759 (mv_is_num_missing): Ditto.
760 (mv_is_str_missing): Ditto.
761 (mv_is_value_user_missing): Removed. Changed callers to use
763 (mv_is_num_user_missing): Removed. Changed callers to use
765 (mv_is_str_user_missing): Removed. Changed callers to use
767 (mv_is_value_system_missing): Removed. Changed callers to use
769 (mv_set_type): Removed. Changed callers to use mv_clear.
770 (mv_clear): New function.
772 * variable.c (var_is_value_missing): Add new enum mv_class
773 parameter. Update all callers.
774 (var_is_num_missing): Ditto.
775 (var_is_str_missing): Ditto.
776 (var_is_value_user_missing): Removed. Changed callers to use
777 var_is_value_missing.
778 (var_is_num_user_missing): Removed. Changed callers to use
780 (var_is_str_user_missing): Removed. Changed callers to use
782 (var_is_value_system_missing): Removed. Changed callers to use
783 var_is_value_missing.
785 * casefilter.c (struct casefilter): Use enum mv_class in place of
787 (casefilter_variable_missing): Adapt to new member.
788 (casefilter_create): Change signature to take enum mv_class,
791 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
793 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
795 * case-sink.c case-sink.h procedure.c procedure.h
796 storage-stream.c: Now uses the factory.
798 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
800 Make it possible to pull cases from the active file with a
801 function call, instead of requiring indirection through a callback
804 * case-source.h (struct case_source_class): Change ->read function
805 to return a single case, instead of calling a callback function
806 for each case. Change ->destroy function to return an error
809 * case-source.c (free_case_source): Pass along the value returned
810 by the case_source ->destroy function.
812 * procedure.c (struct write_case_data): Removed.
813 (struct dataset): Added some members to track procedure state.
814 (procedure): Optimize the trivial case at this level.
815 (internal_procedure): Re-implement in terms of proc_open,
816 proc_read, proc_close.
817 (proc_open) New function.
818 (proc_read) New function.
819 (proc_close) New function.
820 (write_case) Moved into proc_read.
821 (close_active_file) Moved closing of data source into proc_close.
823 * storage-source.c: Rewrote to conform with modified
824 case_source_class interface.
826 * transformations.c (trns_chain_execute): Added argument to allow
827 starting execution from an arbitrary transformation. Updated
830 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
832 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
834 * sys-file-reader.c (read_display_parameters): Don't assume that
835 MEASURE_* and ALIGN_* have the same values found in system files.
837 * sys-file-writer.c (write_variable_display_parameters): Ditto.
839 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
840 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
841 the value of n_MEASURES, which was off by 1 (at least from my
844 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
846 * dictionary.c dictionary.h vardict.h variable.c: Added optional
847 callbacks which are invoked when the dictionary or its
848 variables are changed.
850 * missing-values.c missing-values.h value-labels.c: Tidied up
851 consistency checks, and made some of them return false
852 instead of assert-failing.
854 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
856 * calendar.c (calendar_days_in_month): New function.
858 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
860 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
861 the value label's width.
863 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
865 * sfm-private.h: Move contents into sys-file-writer.c, which is
866 the only remaining user. Removed Borland C++-specific directives.
868 * sys-file-reader.c: Clean up and rewrite entire file. The
869 rewritten version is simpler and better abstracted, and should be
870 easier to maintain and extend. It avoids using structures to read
871 file data, which is prone to padding variations among compilers.
872 It should also handle non-IEEE 754 system files, although I
873 haven't been able to find any. It has been tested against many
874 .sav files obtained from the Web and found to produce the same
875 results as the earlier version of the code, or in some cases
876 improved results. It is more tolerant of format variations found
879 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
880 member, putting an enum integer_format in its place. New member
881 `float_format'. Changed `compressed' member to type bool.
883 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
885 * dictionary.c (dict_delete_consecutive_vars): New function.
887 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
889 * file-name.c (fn_search_path): Remove prefix arg that was unused
890 by any caller. Updated all callers.
892 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
894 * format.c (fmt_dollar_template): Use user's decimal point
895 character. Add assertion.
897 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
899 * format.c (fmt_dollar_template): New function, based on
900 dollar_format_template from var-type-dialog.c.
902 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
904 * data-out.c (output_scientific): Fix bad assumption that "buf" is
907 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
909 Finish converting struct variable to an opaque type. In this
910 phase, we add remaining setter and getter functions, convert the
911 remaining PSPP code to use them, and do a bunch of cleanup. The
912 resulting changes are pervasive but mostly trivial, and only the
913 notable changes are logged.
915 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
918 * case.c (case_data): Renamed case_data_idx.
919 (case_num): Renamed case_num_idx.
920 (case_str): Renamed case_str_idx.
921 (case_data_rw): Renamed case_data_rw_idx.
923 * case.h (case_data): New function with old name and an interface
924 that takes a variable instead of an index, which is easier to
925 use. Updated all callers to use the new interface, or to use the
926 new *_idx function (see above).
929 (case_data_rw): Ditto.
931 * category.c (cat_stored_values_destroy): Changed interface to
932 take a struct cat_vals * instead of a struct variable *.
934 * dictionary.c (dict_clone): Use new vector_clone function.
935 (dict_clear) Use new var_destroy function.
936 (add_var) New function.
937 (dict_create_var) Rewrite in terms of dict_create_var_assert.
938 (dict_create_var_assert) Rewrite in terms of add_var.
939 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
940 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
941 (dict_lookup_var) Use new var_create, var_destroy functions.
942 (dict_contains_var) Rewrite in terms of new vardict functionality.
943 (set_var_dict_index) New function.
944 (set_var_case_index) New function.
945 (reindex_vars) New function.
946 (dict_delete_var) Rewrite in terms of new vardict functionality.
947 (dict_reorder_var) Ditto.
948 (dict_reorder_vars) Ditto.
949 (rename_var) New function.
950 (dict_rename_var) Use rename_var.
951 (dict_rename_vars) Use pool to simplify code. Use rename_var.
952 (dict_get_compacted_idx_to_fv) Rename
953 dict_get_compacted_dict_index_to_case_index, update callers.
954 (dict_create_vector) Use new vector_create function.
955 (dict_clear_vectors) Use new vector_destroy function.
956 (set_var_short_name_suffix) Move here from variable.c, renamed
957 from var_set_short_name_suffix, make static, update caller.
959 * sys-file-private.c: New file.
960 (sfm_width_to_bytes) Moved here from variable.c, renamed from
961 width_to_bytes, update callers.
963 * sys-file-private.h: New file. Later it will supplant
964 sfm-private.h; for now it supplements it.
965 (macro MIN_VERY_LONG_STRING) New macro.
966 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
968 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
969 MAX_LONG_STRING was used before.
971 * sys-file-writer.c: Ditto.
973 * value-labels.c: Change the paradigm here to be that a null
974 pointer is OK for a struct val_labs * in most cases; it just
975 represents an empty set of value labels.
976 (val_labs_copy) A copy of a null set is a null set.
977 (val_labs_count) A null set has 0 labels.
978 (val_labs_replace) Change return type to void. Rewrite for
980 (val_labs_find) A null set does not contain the value.
981 (value_to_string) Moved to variable.c, renamed var_get_value_name,
982 transposed argument order, updated all callers.
985 (value_dup) Moved here from variable.c.
986 (compare_values) Ditto.
989 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
990 (macro MAX_LONG_STRING) Removed, because it was only interesting
991 for system files, not for general code.
992 (macro MAX_VERY_LONG_STRING) Ditto.
993 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
994 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
996 * vardict.h: New file, for an interface between variables and
999 * variable.c: A lot of functions were moved around, for better
1001 (struct variable) Move definition here, from variable.h.
1002 (var_type_adj) Removed--makes i18n hard.
1003 (var_type_noun) Ditto.
1004 (var_create) New function.
1005 (var_clone) New function.
1006 (var_destroy) New function.
1007 (var_set_name) Assert that variable is not in a dictionary.
1008 (compare_var_names) Rename compare_vars_by_name and fix a couple
1009 of callers who thought the args were strings.
1010 (hash_var_name) Rename hash_var_by_name.
1011 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1012 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1013 (var_is_very_long_string) Removed, because it was only interesting
1014 to system file code.
1015 (var_set_missing_values) Allow the argument to be the wrong width,
1016 as long as we can resize it. Simplify callers who were doing the
1017 resizing themselves.
1018 (var_get_value_labels) New function.
1019 (var_has_value_labels) New function.
1020 (var_set_value_labels) New function.
1021 (alloc_value_labels) New function.
1022 (var_add_value_label) New function.
1023 (var_replace_value_label) New function.
1024 (var_clear_value_labels) New function.
1025 (var_lookup_value_label) New function.
1026 (var_get_value_name) Moved here from variable.c, renamed from
1027 var_get_value_name, transposed argument order, updated all
1029 (var_to_string) Moved here, from variable-label.c.
1030 (var_set_leave) New function.
1031 (var_get_leave) New function.
1032 (var_must_leave) New function.
1033 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1034 set_var_short_name_suffix.
1035 (var_get_dict_index) New function.
1036 (var_get_case_index) New function.
1037 (var_get_obs_vals) New function.
1038 (var_set_obs_vals) New function.
1039 (var_has_obs_vals) New function.
1040 (var_get_vardict) New function.
1041 (var_set_vardict) New function.
1042 (var_has_vardict) New function.
1043 (var_clear_vardict) New function.
1044 (value_dup) Moved to value.c.
1045 (compare_values) Ditto.
1048 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1049 (enum ALPHA) Rename VAR_STRING, update all users.
1051 * vector.c: New file.
1052 (struct vector) Moved here, from variable.h.
1053 (check_widths) New function.
1054 (vector_create) New function.
1055 (vector_clone) New function.
1056 (vector_destroy) New function.
1057 (vector_get_name) New function.
1058 (vector_get_var) New function.
1059 (vector_get_var_cnt) New function.
1060 (compare_vector_ptrs_by_name) New function.
1062 * vector.h: New file.
1064 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1066 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1067 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1070 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1072 * value-labels.c (destroy_atoms): New function.
1073 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1075 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1077 Thanks to Jason Stover for pointing out this problem.
1079 * data-out.c (output_number): Use gsl_finite from GSL, which is
1080 portable, instead of isfinite, which is not.
1083 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1085 * variable.c variable.h (value_dup): New function.
1087 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1089 Start converting struct variable to an opaque type. In this
1090 phase, we add a bunch of setter and getter functions and convert
1091 most of the PSPP code to use them. The resulting changes are
1092 pervasive but mostly trivial, and only the notable changes are
1095 * format.c (fmt_equal): New function.
1097 * variable.c (var_type_is_valid): New function.
1098 (measure_is_valid) Moved here, from format.c.
1099 (alignment_is_valid) Moved here, from format.c.
1100 (var_get_name) New function.
1101 (var_set_name) New function.
1102 (width_to_type) New function.
1103 (var_get_type) New function.
1104 (var_get_width) New function.
1105 (var_set_width) New function.
1106 (var_is_numeric) New function.
1107 (var_is_alpha) New function.
1108 (var_is_short_string) New function.
1109 (var_is_long_string) New function.
1110 (var_is_very_long_string) New function.
1111 (var_get_missing_values) New function.
1112 (var_set_missing_values) New function.
1113 (var_clear_missing_values) New function.
1114 (var_has_missing_values) New function.
1115 (var_is_value_missing) New function.
1116 (var_is_num_missing) New function.
1117 (var_is_str_missing) New function.
1118 (var_is_value_user_missing) New function.
1119 (var_is_num_user_missing) New function.
1120 (var_is_str_user_missing) New function.
1121 (var_is_value_system_missing) New function.
1122 (var_get_print_format) New function.
1123 (var_set_print_format) New function.
1124 (var_get_write_format) New function.
1125 (var_set_write_format) New function.
1126 (var_set_both_formats) New function.
1127 (var_get_label) New function.
1128 (var_set_label) New function.
1129 (var_clear_label) New function.
1130 (var_has_label) New function.
1131 (var_get_measure) New function.
1132 (var_set_measure) New function.
1133 (var_get_display_width) New function.
1134 (var_set_display_width) New function.
1135 (var_get_alignment) New function.
1136 (var_set_alignment) New function.
1137 (var_get_value_cnt) New function.
1138 (var_get_leave) New function.
1139 (var_get_short_name) New function.
1141 * variable.h: (struct variable) Removed "type" and "nv" members;
1142 they are now computed from "width" where needed.
1144 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1146 * missing-values.c (mv_resize): Don't write beyond end of the
1147 allocated buffer when resizing a long string.
1149 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1151 Clean up identifier code: don't require identifier enumerations to
1152 be in a particular order; make better use of string library;
1153 expose less of the internals.
1155 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1156 change interface. Updated all callers.
1157 (lex_id_match) Change interface to use struct substring, update
1159 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1160 (global array keywords[]) Make static, change form. Update all
1161 users to use lex_id_name instead.
1162 (lex_is_keyword) New function.
1163 (lex_id_to_token) Change interface to use struct substring, update
1165 (lex_id_name) New function.
1167 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1168 lex_is_keyword instead.
1169 (T_LAST_KEYWORD) Removed.
1170 (T_N_KEYWORDS) Removed.
1172 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1174 * format.c: (fmt_date_template) Distinguish characters for which a
1175 space is output and any date delimiter is allowed on input, from
1176 those for which a space is output and only a space is allowed on
1177 input. The former is represented by X, the latter by a space.
1178 Also, drop distinction between h and H, changing the former to the
1181 * data-in.c: Completely rewrite internals to conform to SPSS input
1182 formats as closely as possible.
1183 (data_in) Changed external interface by replacing the structure
1184 that was used as a single argument by a set of arguments. Updated
1186 (data_in_finite_line) Removed. Converted all callers to use plain
1188 (data_in_get_integer_format) New function.
1189 (data_in_set_integer_format) New function.
1190 (data_in_get_float_format) New function.
1191 (data_in_set_float_format) New function.
1193 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1194 (struct data_in) Removed.
1196 * data-out.c: (output_date) Drop each component from the input as
1197 it is output, to allow us to drop the distinction between h (a
1198 count of hours) and H (the hour of day) template characters.
1199 Also, handle new X template character.
1200 (output_scientific) Follow more rational rule on when to drop
1201 fraction introduced between SPSS 13 and 15. Updated test case to
1204 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1206 Fix buffer overflow reported by John Darrington.
1208 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1209 realize that DIGITS is a count of nibbles, not of bytes.
1211 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1213 * calendar.c (calendar_offset_to_gregorian) Also return the
1214 year-of-day. Change callers to new interface.
1216 * data-out.c: Completely rewrite internals to conform to SPSS
1217 output formats as completely as possible.
1218 (data_out) Change interface to put input parameters before output
1219 parameters, for consistency with the style I now prefer. Update
1221 (data_out_get_integer_format) New public function.
1222 (data_out_set_integer_format) New public function.
1223 (data_out_get_float_format) New public function.
1224 (data_out_set_float_format) New public function.
1226 * data-out.h: New file. Move prototype for data_out here, from
1229 * format.c: (fmt_step_width) Use equality comparison instead of
1230 bitwise and, for clarity.
1231 (fmt_is_string) Ditto.
1232 (fmt_input_to_output) Fix categories that are translated to F
1235 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1237 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1238 casefile-private.h: Added casefilter to assist commands with missing
1241 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1243 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1245 * settings.c: (route_errors_to_terminal) New variable.
1246 (route_errors_to_listing) New variable.
1247 (get_error_routing_to_terminal) New function.
1248 (set_error_routing_to_terminal) New function.
1249 (get_error_routing_to_listing) New function.
1250 (set_error_routing_to_listing) New function.
1252 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1254 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1256 * format.c: Completely rewrite, to achieve better abstraction.
1257 Rewrite all references to formats in other files.
1259 * format.def: Rewrite and reorganize.
1261 * settings.c: Move everything related to custom currency formats
1262 into format.[ch], changing them in form, so as to group related
1263 code and definitions better. Changed all references to use the
1265 (static var decimal) Removed.
1266 (static var grouping) Removed.
1267 (static var cc) Removed.
1268 (get_decimal) Removed.
1269 (set_decimal) Removed.
1270 (get_grouping) Removed.
1271 (set_grouping) Removed.
1275 * settings.h: (macro CC_CNT) Removed.
1276 (macro CC_WIDTH) Removed.
1277 (struct custom_currency) Removed.
1279 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1281 * data-in.c (data_in): Use switch statement instead of table, to
1282 avoid dependence on the order of the FMT_* enums.
1284 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1286 * data-out.c: (num_to_string) Removed, because it was dead code.
1288 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1290 * data-in.c (parse_trailer): Fix error message.
1292 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1294 * format.c (fmt_is_binary): New function.
1296 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1298 * procedure.c procedure.h: Encapsulated the static data into a single
1301 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1303 * casefile.c (casereader_read_xfer): Always initialize the case,
1304 even on an error condition.
1306 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1308 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1309 entire function useless.
1311 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1313 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1314 casereader method casereader_clone.
1316 * procedure.c pransformations.h: Introduced new type casenum_t
1318 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1320 * variable.c: (width_to_bytes) Rephrase code for clarify.
1322 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1324 * format.c: (fmt_type_from_string) New function.
1325 (fmt_to_string) Include decimals in output if the format has
1326 decimals, even if the format type does not. This way, we can
1327 accurately reproduce incorrect formats in user output.
1328 (check_common_specifier) Make the check for a bad format type an
1329 assertion, so we get bug reports if they show up. Fix message.
1330 Check for decimal places with a format type that doesn't allow
1332 (check_input_specifier) Remove check for FMT_X, which has been
1334 (check_output_specifier) Ditto.
1336 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1338 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1339 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1340 (enum fmt_parse_flags) Removed.
1342 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1344 * casefile.c casefile.h: Converted to an abstract base class.
1345 * casefile-private.h fastfile.c fastfile.h: New files.
1346 * automake.mk procedure.c scratch-writer.c storage-stream.c
1348 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1350 * procedure.c (internal_procedure): Create sink_case with only as
1351 many values as the compacted dictionary.
1353 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1355 Remove "debugging" code that caused plenty of false positives and
1358 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1360 * case.c: Remove all references to `this' member.
1362 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1364 Fix link error noted by Jason Stover.
1366 * storage-stream.c: Include <assert.h>.
1368 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1370 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1371 ALL) and additional underlying system file issues.
1373 Thanks to John Darrington for review.
1375 First problem: var_hash points to variables not owned by the
1376 sys-file-reader, which the caller may free or modify. Use an
1377 array of sfm_vars instead, as done earlier (e.g. CVS version
1380 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1381 members and remove all code that references it. Add vars, var_cnt
1382 members. Remove fix_specials member, which was unused.
1383 (struct sfm_var) Remove name member, which was unused.
1384 (sfm_close_reader) Free vars member instead of var_hash.
1385 (compare_var_shortnames) Removed.
1386 (hash_var_shortname) Removed.
1387 (sfm_open_reader) Fill out vars array.
1388 (compare_var_index) Removed.
1389 (sfm_read_case) Use vars instead of var_hash.
1391 Second problem: we're confused about when we actually have very
1392 long strings, causing us to choose incorrectly between slow path
1393 and fast path in sfm_read_case.
1395 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1396 have very long strings, not when we have long variable names,
1397 which is an unrelated feature.
1399 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1401 * variable.h: Move var_set and variable parsing declarations to
1402 new header, src/language/lexer/variable-parser.h. Modified lots
1403 of files to include the new header.
1405 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1407 * value-labels.c (value_to_string): When there's no value label,
1408 format the variable according to its print format, instead of
1409 always effectively using A or F format.
1411 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1413 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1416 * format.c: Removed tortological assertion.
1418 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1420 Reform string library.
1422 * file-name.c (fn_interp_vars): Change interface to take a
1423 substring as input. Updated all users.
1425 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1427 * format.c (measure_is_valid): Really return false when m >=
1430 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1432 Implement random access to casefiles, for use in GUI.
1434 * casefile.c: (struct casereader) Add `random', `file_ofs',
1435 `buffer_ofs' members.
1436 (casefile_get_random_reader) New function.
1437 (read_open_file) Break part into new function
1438 seek_and_fill_buffer().
1439 (fill_buffer) Update buffer_ofs, file_ofs.
1440 (casereader_seek) New function.
1442 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1444 * settings.c: Added call to i18n{done, init}.
1446 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1448 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1451 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1453 * casefile.c: Convert many uses of `int' to `bool'.
1455 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1457 * transformations.c (trns_chain_destroy): Destroy chain's trns
1458 member, to fix memory leak.
1460 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1462 * storage-stream.c (storage_source_decapsulate): Destroy case
1463 source to fix memory leak.
1465 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1467 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1468 case passed as argument instead of initializing the argument as a
1469 case. Fixes memory leak that showed up in
1470 tests/command/aggregate.sh with scratch files.
1472 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1474 * procedure.c (proc_done): Destroy default_dict, to fix memory
1477 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1479 Simplify procedure_with_splits().
1481 * procedure.c (struct split_aux_data): Removed case_count member.
1482 (procedure_with_splits) Don't initialize case_count.
1483 (split_procedure_case_func) Check whether prev_case is null
1484 instead of case_count.
1485 (split_procedure_end_func) Ditto.
1487 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1489 * case.c (case_move): Do nothing if dst and src are the same
1491 (case_try_create) Merge two similar cases.
1492 (case_copy) Unshare only if data must be actually copied.
1494 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1496 * data-in.c data-out.c dictionary.c sys-file-reader.c
1497 sys-file-writer.c variable.c variable.h: Reworked very long string
1498 support for better encapsulation.
1500 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1502 * value-labels.c (val_labs_can_set_width): New function.
1503 (val_labs_set_width) Clear labels if increasing width to long
1505 (val_labs_destroy) Remove unneeded test for null.
1507 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1509 * value-labels.h: Remove unneeded dependency on variable.h.
1511 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1513 Get rid of `char *c' member in union value, for cleanliness.
1515 * value.h: (union value) Remove `c' member.
1517 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1519 Make dictionary compacting functions a little more general.
1521 * sys-file-writer.c (sfm_open_writer): Use
1522 dict_compacting_would_change().
1523 (does_dict_need_translation) Removed.
1525 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1527 Make dictionary compacting functions a little more general.
1529 * dictionary.c (dict_needs_compaction): Rename
1530 dict_compacting_would_shrink(). Update all callers.
1531 (dict_compacting_would_change) New function.
1533 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1535 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1536 inverted return value (!).
1538 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1540 Continue reforming procedure execution.
1542 * procedure.c: Search and replace "vfm" by "proc". Notably:
1543 (static var vfm_source) Rename proc_source. Update all
1545 (static var vfm_sink) Rename proc_sink. Update all references.
1547 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1549 Continue reforming procedure execution. In this phase, remove
1550 PROCESS IF, which was deprecated anyway and can be easily
1551 simulated with TEMPORARY followed by SELECT IF.
1553 * procedure.c: (open_active_file) Don't call
1554 add_process_if_trns().
1555 (discard_variables) Get rid of redundant call to
1556 proc_cancel_all_transformations().
1557 (add_process_if_trns) Removed.
1558 (process_if_trns_proc) Removed.
1559 (process_if_trns_free) Removed.
1561 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1563 Continue reforming procedure execution. In this phase, add
1564 `const' to the case passed to procedure()'s callback.
1566 Updated all users of procedure() as well.
1568 * procedure.c: (struct write_case_data) Add "const" to ccase
1569 parameter for case_func member.
1570 (procedure) Add "const" to ccase parameter for proc_func
1572 (multipass_case_func) Make ccase parameter const.
1573 (internal_procedure) Add "const" to ccase parameter for case_func
1575 (split_procedure_case_func) Make ccase parameter const.
1576 (multipass_split_case_func) Make ccase parameter const.
1578 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
1580 Continue reforming procedure execution. In this phase, get rid of
1581 the output code for SPLIT FILE groups in procedure.c, which really
1582 shouldn't be doing any output. Move it into the individual
1583 procedures instead. This also adds some flexibility.
1585 Updated many users of procedure_with_splits() and
1586 multipass_procedure_with_splits() to call
1587 output_split_file_values() and to deal with increased use of
1590 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
1591 parameter to begin_func member.
1592 (procedure_with_splits) Add "const struct ccase *" parameter to
1593 begin_func parameter. Make ccase parameter const in proc_func
1595 (split_procedure_case_func) Don't dump split file group. Pass
1597 (dump_splits) Moved to language/dictionary/split-file.c as
1598 output_split_file_values().
1599 (struct multipass_split_aux_data) Add "const struct ccase *"
1600 parameter to split_func member.
1601 (multipass_procedure_with_splits) Add "const struct ccase *"
1602 parameter to split_func parameter.
1603 (multipass_split_case_func) Save new SPLIT FILE case before
1605 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
1607 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
1609 Continue reforming procedure execution. Change
1610 internal_procedure() so that it calls open_active_file() and
1611 close_active_file(), which isolates most of the actual procedure
1614 * procedure.c: (struct write_case_data) Rename `proc_func' member
1615 to `case_func' and update all references.
1616 (procedure) Rewrite as one-line wrapper around
1617 internal_procedure().
1618 (struct multipass_aux_data) New.
1619 (multipass_callback) Renamed multipass_case_func(). Use struct
1620 multipass_aux_data as auxiliary data.
1621 (multipass_end_func) New function.
1622 (multipass_procedure) Rewrite as wrapper for internal_procedure()
1623 that uses multipass_case_func, multipass_end_func.
1624 (internal_procedure) Add `end_func' argument. Move optimization
1625 of trivial case in here. Move call to open_active_file() and
1626 close_active_file() in here. Now assert that vfm_source is
1628 (procedure_with_splits_callback) Rename
1629 split_procedure_case_func().
1630 (split_procedure_end_func) New function.
1631 (multipass_split_callback) Rename multipass_split_case_func.
1632 (multipass_split_end_func) New function.
1633 (discard_variables) No need to test for nonnull vfm_source.
1635 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
1637 Continue reforming procedure execution. Get rid of unused member.
1639 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
1641 (write_case) Don't increment cases_analyzed.
1643 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
1645 Continue reforming procedure execution. In this phase, move
1646 procedure.c and procedure.h from src to src/data. Update
1647 makefiles and #includes accordingly.
1649 * procedure.c: Moved here from src/.
1651 * procedure.h: Moved here from src/.
1653 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
1655 Continue reforming procedure execution. In this phase, get rid of
1656 many global variables, consolidating procedure execution in
1657 procedure.c. Encapsulate transformations in new "struct
1658 trns_chain". Also, change implementation of N OF CASES, FILTER,
1659 and PROCESS IF from special cases to transformations.
1661 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
1664 * dictionary.c: (global variable default_dict) Move to
1667 * variable.h: (TRNS_*) Move to transformations.h.
1668 (struct transformation) Move to transformations.c.
1670 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1672 * sys-file-reader.c: Fixed invalid read problems.
1674 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
1676 * storage-stream.c: Add missing function comments.
1678 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
1680 Continue reforming procedure execution. In this phase, add some
1681 new, needed functionality to storage-stream.
1683 * storage-stream.c: (storage_source_decapsulate) New function.
1685 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
1687 * variable.c (width_to_bytes): Declarations must precede
1688 statements for C90 compliance.
1690 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
1692 * data-out.c, data-in.c, variable.c, variable.h: New functions
1693 copy_mangle and copy_demangle for reading/writing cases; emulates the
1694 way SPSS deals with strings > 255 bytes.
1696 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
1697 subtype 14 needed for strings longer than 255 bytes.
1699 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
1700 instead of literal values. Also fixed some constness issues.
1702 * format.h: Constness
1704 * sfm-private.h: Renamed the case_size identifier, since I discovered
1705 that SPSS's respect for this variable is very nominal.
1707 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
1709 Change case limit type from int to size_t.
1711 * dictionary.c: (struct dictionary) Change type of case_limit
1713 (dict_get_case_limit) Change return type.
1714 (dict_set_case_limit) Change parameter type.
1716 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
1718 * variable.h: (struct variable) Rename `reinit' member as `leave'
1719 and invert sense. Fix up all references.
1721 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
1723 Continue reforming procedure execution. In this phase, break
1724 procedure.c into multiple files.
1726 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
1728 * case-sink.c: New file.
1730 * case-sink.h: New file.
1732 * case-source.c: New file.
1734 * case-source.h: New file.
1736 * storage-stream.c: New file.
1738 * storage-stream.h: New file.
1740 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
1742 * variable.h: (struct variable) Remove `init' member and all
1743 references to it from other files. It was initialized in several
1744 places, but nothing really ever used it for anything worthwhile.
1745 Thanks to Jason Stover for pointing out how confusing this
1748 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
1750 Continue reforming error message support. In this phase, get rid
1751 of message "titles" and put the message text in `struct error'.
1752 Now `struct error' encapsulates a message more properly.
1754 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
1755 Format message ourselves.
1757 * data-in.c: (vdls_error) Ditto.
1759 * por-file-reader.c: (error) Ditto.
1761 * sys-file-reader.c: (corrupt_msg) Ditto.
1763 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
1765 GNU standards require "file name" instead of "filename" in
1766 documentation. It's nice for our code to follow the convention
1769 * casefile.c: (struct casefile) Rename `filename' member to
1770 `file_name'. Updated all references.
1772 * file-name.c: [!unix] (struct file_identity) Rename
1773 normalized_filename member to normalized_file_name. Updated all
1776 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
1778 We don't really support anything but Unix-like environments well,
1779 so we might as well de-obfuscate by writing directory and path
1780 separators explicitly.
1782 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
1784 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
1785 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
1787 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
1790 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
1792 GNU standards require "file name" instead of "filename" in
1793 documentation. It's nice for our code to follow the convention
1796 * filename.c: Rename to file-name.c.
1798 * filename.h: Rename to file-name.h. Update all inclusions.
1799 Update header guards.
1801 * automake.mk: Update file names.
1803 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
1805 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
1807 (fn_basename) Removed (dead code).
1808 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
1809 (fn_special_p) Renamed fn_is_special(), all references updated.
1810 (fn_exists_p) Renamed fn_exists(), all references updated.
1812 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
1814 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
1815 Also, now it only tilde-expands file names, not paths.
1816 (fn_search_path) Tilde-expand one directory at a time.
1818 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
1820 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
1821 get rid of non-Unixlike version of the code, which has probably
1823 (fn_prepend_dir) Removed (dead code).
1825 * filename.h: (macro DIR_SEPARATOR_STRING) New.
1826 (macro PATH_SEPARATOR_STRING) New.
1827 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
1829 Continue reforming error message support. In this phase, we get
1830 rid of VM() and the other msg() support for "verbosity", replacing
1831 it by a new function verbose_msg().
1833 * filename.c: (fn_search_path) Use verbose_msg() instead of
1836 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
1838 * sfm-private.h: Get rid of #defines after #error, which makes no
1841 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
1843 Get rid of our own int32 type in favor of the standard int32_t
1846 * sfm-private.h: (int32 macro) Don't define this anymore. Do
1849 * sys-file-reader.c: Use int32_t instead of int32 throughout.
1851 * sys-file-writer.c: Use int32_t instead of int32 throughout.
1853 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
1855 Remove ill-considered file routines that are no longer used.
1857 * filename.c: (fn_open_ext) Removed.
1858 (fn_close_ext) Removed.
1860 * filename.h: (struct file_ext) Removed.
1862 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
1864 * variable.c (var_is_valid_name): Move declarations before code
1867 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
1869 * filename.ch (fn_interp_vars): Fixed small buglet.
1871 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
1873 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
1874 so as to modify the string inline. Thus makeing it easier to
1875 destroy the results when no longer needed.
1877 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
1879 * category.c (cat_stored_values_destroy): Fixed memory leak.
1881 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
1883 Add some missing frees. Thanks to John Darrington for reporting
1886 * any-writer.c (any_writer_close): Free writer.
1888 * any-reader.c (any_reader_close): Free reader.
1890 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
1892 * por-file-reader.c: (error) Mark as NO_RETURN.
1894 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
1896 * settings.c: Changed default value of scompress to true.
1898 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
1900 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
1902 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
1905 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
1906 here from pref.h.orig.
1908 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
1910 * sys-file-reader.c: Fixed bug reading compressed files.
1912 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1914 * Numerous renames. See src/ChangeLog for details.
1916 * Moved files from src directory