1 2007-08-09 Ben Pfaff <blp@gnu.org>
3 Fix bug reported by Jason Stover.
4 * settings.c: Move get_termcap_viewport and in particular the
5 #include for <curses.h> to the end of the file. curses.h
6 redefines bool on some systems (e.g. OpenBSD), which causes
7 disagreement between uses of bool before its inclusion and
9 Tested by Jason Stover.
11 2007-07-29 Ben Pfaff <blp@gnu.org>
13 Provisional fix for bug #18692 and bug #20161. Reviewed by John
16 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
17 argument (never "rb" or "wb") because SUSv3 says that only those
18 modes are defined, and glibc in fact rejects other modes.
20 Open portable files with fn_open so that they can be read from
21 pipes. Fix missing fh_close call to go along with fh_open.
22 Report an error if the file close reports an error.
23 * por-file-reader.c (close_reader): New function.
24 (por_file_casereader_destroy): Use close_reader.
25 (pfm_open_reader): Open file with fn_open.
27 2007-07-28 Ben Pfaff <blp@gnu.org>
29 Make PSPP able to read all the portable files I could find on the
30 web. Thanks to John Darrington for review. Bug #17620.
31 * por-file-reader.c (struct pfm_reader): New member `line_length'.
32 (error): Print file offset in hexadecimal.
33 (warning): New function.
34 (advance): Treat lines less than 80 bytes long as padded to 80
36 (pfm_open_reader): Call read_documents if we find an "E" record.
37 (convert_format): Convert invalid formats to the default format
38 instead of aborting reading the file.
39 (read_variables): Rename duplicate variable names instead of
40 aborting reading the file.
41 (read_value_label): Allow string variables of different widths to
42 be assigned value labels in the same record. Replace duplicate
43 value labels instead of aborting.
44 (read_documents): New function.
46 * por-file-writer.c (pfm_open_writer): Call write_documents if the
47 dictionary has documents.
48 (write_documents): New function.
50 2007-07-25 Ben Pfaff <blp@gnu.org>
52 Fix bugs related to bug #17213.
54 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
55 since the former is what config.h has. Include the needed ncurses
57 (static var echo) Rename to `do_echo' because the original name is
58 the same as an ncurses identifier.
59 (get_termcap_viewport) Use error instead of msg.
61 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
62 (fn_close): Don't close stdin, stdout, stderr.
64 2007-07-26 John Darrington <john@darrington.wattle.id.au>
66 * procedure.c procedure.h: Added callbacks which get invoked whenever
67 a dataset's transformation chain changes.
69 2007-07-24 Ben Pfaff <blp@gnu.org>
72 * sys-file-writer.c (write_variable_display_parameters): Use new
73 var_default_display_width function to choose display width of
74 segments after the first one in a given variable.
75 * variable.c (var_create): Use var_default_display_width to pick
76 new variable's display width.
77 (var_default_display_width): New function.
78 Reviewed by John Darrington.
80 2007-07-24 Ben Pfaff <blp@gnu.org>
83 * por-file-writer.c (write_variables): Write weight variable.
84 Reviewed by John Darrington.
86 2007-07-23 Ben Pfaff <blp@gnu.org>
88 Improvements to system file reader and writer.
90 First, move all detailed knowledge of very long strings into
91 sys-file-private.[ch], so that this nasty stuff can be isolated.
93 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
94 (EFFECTIVE_VLS_CHUNK): New macro.
95 (min_int): New function.
96 (max_int): New function.
97 (sfm_width_to_bytes): Rewrite.
98 (sfm_width_to_octs): New function.
99 (sfm_segment_alloc_width): New function.
100 (sfm_segment_alloc_bytes): New function.
101 (sfm_segment_used_bytes): New function.
102 (sfm_segment_offset): New function.
103 (sfm_segment_effective_offset): New function.
104 (sfm_dictionary_to_sfm_vars): New function.
106 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
107 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
108 (struct sfm_var): New structure.
110 Next, improvements to the system file reader.
112 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
113 type casenumber. Added `version_major', `version_minor',
116 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
117 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
118 Change `case_cnt' to type casenumber. Removed `has_vls'.
119 (struct sfm_var): Removed.
120 (sfm_open_reader): Don't warn on wrong case size if the file was
121 written by SPSS 13, which tends to get it wrong. Use
122 sfm_dictionary_to_sfm_vars.
123 (read_header): Always output system file info.
124 (read_variable_record): Simplify code for reading missing values.
125 (read_machine_int32_info): Save version numbers from system file
126 into info struct passed as new argument.
127 (read_long_string_map): Restructured to use new sys-file-private
129 (read_value_labels): Use size_overflow_p.
130 (sys_file_casereader_read): Get rid of distinction between fast
131 and slow paths. Use information provided by sys-file-primate's
132 struct sfm_var to simplify code.
133 (skip_whole_strings): New function.
134 (read_int32): Renamed read_int. Changed return value to int.
136 (read_flt64): Renamed read_float. Changed return value to
137 double. Updated all callers.
138 (int32_to_native): Removed. Changed callers to use
140 (flt64_to_double): Removed. Changed callers to use float_convert.
142 Finally, get rid of int32, flt64 terminology and types in system
143 file writer. The former wasn't very useful since a POSIX "int"
144 can hold the whole range of int32 and we generally didn't have a
145 need for it to be exactly-32-bits, just at-least-32-bits. The
146 latter was inconvenient because we had to assume that it could be
147 different from double and thereby convert special values SYSMIS,
148 HIGHEST, LOWEST to and from it in multiple places. Instead, now
149 we just use "int" and "double" in most places, and do conversions,
150 if necessary, very close to where we do I/O. This change meant
151 that the writer code couldn't represent records in the file as C
152 structs any longer, but that's no great loss. The code actually
153 seems to be more readable without them.
155 Simplify the compression buffering code: only buffer as much as
156 necessary, which is no more than eight 8-byte units at any given
159 * sys-file-writer.c (typedef flt64): Removed.
160 (macro second_lowest_flt64): Removed.
161 (struct sysfile_header): Removed.
162 (struct sysfile_variable): Removed.
163 (struct sfm_writer): Removed `needs_translation', `has_vls',
164 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
165 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
166 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
167 compression buffering by `opcodes', `opcode_cnt', `data',
168 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
169 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
170 data writer functions instead of structures.
171 (calc_oct_idx): New function.
172 (write_header): Use simple data writer functions instead of
174 (write_format_spec): Renamed write_format. New argument.
175 (write_variable_continuation_records): New function.
176 (write_variable): Use simple data writer functions instead of
177 structures. Use write_variable_continuation_records. Write
178 entire very long string instead of requiring caller to understand
180 (write_value_labels): Use simple data writer functions instead of
182 (write_documents): Ditto.
183 (write_variable_display_parameters): Use sys-file-private
184 functions to simplify. Use simple data writer functions instead
186 (write_vls_length_table): Use simple data writer functions instead
188 (write_longvar_table): Ditto.
189 (write_rec_7_34): Break into new functions
190 write_integer_info_record, write_float_info_record. Use simple
191 data writer functions instead of structures.
192 (buf_write): Removed.
193 (append_string_max): Removed.
194 (ensure_buf_space): Removed.
195 (sys_file_casewriter_write): Get rid of the distinction between
196 fast and slow paths, which didn't seem to be too useful. Use new
197 functions write_case_uncompressed, write_case_compressed.
198 (put_instruction): Removed.
199 (put_element): Removed.
200 (write_compressed_data): Removed.
201 (close_writer): Use flush_compressed. Only write case count to
202 system file if it will fit in the field.
203 (write_case_compressed): New function.
204 (write_case_uncompressed): New function.
205 (flush_compressed): New function.
206 (put_cmp_opcode): New function.
207 (put_cmp_number): New function.
208 (write_int): New function.
209 (convert_double_to_output_format): New function.
210 (write_float): New function.
211 (write_value): New function.
212 (write_string): New function.
213 (write_bytes): New function.
214 (write_zeros): New function.
215 (write_spaces): New function.
217 Reviewed by John Darrington.
219 2007-07-22 Ben Pfaff <blp@gnu.org>
221 Don't try to write very long strings to portable files. The
222 format does not support it.
224 * por-file-writer.c (MAX_POR_WIDTH): New macro.
225 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
226 (write_format): Add arg to take width to resize format to.
227 (write_value): Limit width of value written to MAX_POR_WIDTH.
228 (write_variables): Limit width of variable and its output formats
230 Reviewed by John Darrington.
232 2007-07-22 Ben Pfaff <blp@gnu.org>
234 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
235 local variable instead of literal 5.
236 Reviewed by John Darrington.
238 2007-07-22 Ben Pfaff <blp@gnu.org>
240 Fix problems with uniqueness of short names in system files with
241 very long string variables. Now a variable may have multiple
244 * automake.mk (src_data_libdata_a_SOURCES): Add new files
245 short-names.c, short-names.h.
247 * dictionary.c (dict_clone): Clone all the short names.
248 (compare_strings): Move into short-names.c.
249 (hash_strings): Ditto.
250 (set_var_short_name_suffix): Ditto.
251 (dict_assign_short_names): Ditto, rename short_names_assign,
252 change to assign all short names.
254 * por-file-writer.c (write_variables): Use short_names_assign
255 instead of dict_assign_short_names.
257 * short-names.c: New file.
259 * short-names.h: New file.
261 * sys-file-private.c (sfm_width_to_segments): New function.
263 * sys-file-reader.c (read_long_var_name_map): Save and restore all
264 the short names, not just the first one.
266 * sys-file-writer.c (cont_var_name): Removed.
267 (sfm_open_writer): Use short_names_assign instead of
268 dict_assign_short_names. Use unique short names assigned by
269 short_names_assign instead of those generated by cont_var_name.
271 * variable.c (struct variable): Remove `short_name' member,
272 replace by `short_names' and `short_name_cnt'.
273 (var_create) Initialize new members.
274 (var_get_short_name_cnt): New function.
275 (var_get_short_name): Now takes an index argument. Changed most
277 (var_set_short_name): Ditto.
278 (var_clear_short_name): Renamed var_clear_short_names, changed to
279 clear all short names.
281 Reviewed by John Darrington.
283 2007-07-22 Ben Pfaff <blp@gnu.org>
285 * variable.c (var_set_width): Use new var_set_width function.
287 * missing-values.c (mv_n_values): Drop assertion, which was not
290 * format.c (fmt_default_for_width): New function.
291 (fmt_resize): New function.
293 Reviewed by John Darrington.
295 2007-07-18 John Darrington <john@darrington.wattle.id.au>
297 * datasheet.c (datasheet_delete_columns): Added assertion to check
298 we're not deleting outside the range of the sheet.
301 * dictionary.c dictionary.h variable.c: Added the ability for string
302 variables to be resized.
304 * vardict.h: Added some prototypes (moved from dictionary.h) as
305 these should only be called by variable.c
308 2007-07-14 John Darrington <john@darrington.wattle.id.au>
310 * sfm-reader.c: Respect case_cnt field in file header.
312 2007-07-01 John Darrington <john@darrington.wattle.id.au>
314 * transformation.c transformation.h (trns_chain_execute): Changed the
315 signature (Patch #6057)
317 2007-06-10 Ben Pfaff <blp@gnu.org>
319 * casereader-filter.c (casereader_filter_destroy): Make sure to
320 write all the remaining excluded cases to the casewriter, if any.
322 * caseinit.c (init_list_destroy): Rewrite.
323 (init_list_clear): Ditto.
325 * casegrouper.c (casegrouper_get_next_group): Always set *reader
326 to null when returning false.
328 2007-06-06 Ben Pfaff <blp@gnu.org>
330 Actually implement the new procedure code and adapt all of its
331 clients to match. Also adapt all of the other case sources and
332 sinks in the tree and their clients to use the
333 casereader/casewriter infrastructure.
335 * automake.mk: Add and remove files.
337 * any-reader.c: Change into a casereader.
338 * por-file-reader.c: Ditto.
339 * scratch-reader.c: Ditto.
340 * sys-file-reader.c: Ditto.
342 * any-writer.c: Change into a casewriter.
343 * por-file-writer.c: Ditto.
344 * scratch-writer.c: Ditto.
345 * sys-file-writer.c: Ditto.
347 * procedure.c: Change to use casereader, casewriter, caseinit, and
348 other new infrastructure.
350 * scratch-handle.c: Adapt to new infrastructure.
352 * case-sink.c: Removed, now dead code.
353 * case-sink.h: Ditto.
354 * case-source.c: Ditto.
355 * case-source.h: Ditto.
356 * casefile-factory.c: Ditto.
357 * casefile-private.h: Ditto.
360 * casefilter.c: Ditto.
361 * casefilter.h: Ditto.
364 * fastfile-factory.c: Ditto.
365 * fastfile-factory.h: Ditto.
366 * storage-stream.c: Ditto.
367 * storage-stream.h: Ditto.
369 2007-06-06 Ben Pfaff <blp@gnu.org>
373 * automake.mk: Add new files.
375 * datasheet.c: New file.
377 * datasheet.h: New file.
379 2007-06-06 Ben Pfaff <blp@gnu.org>
381 Until now, the procedure code has provided a case to the
382 case_source, which has filled in the data values that come from
383 the active file. "Left" data values that don't come from the
384 active file naturally stay the same from case to case, because the
385 procedure code keeps using that same case.
387 One of the compromises that comes with the new procedure code is
388 that the active file allocates and provides its own case, which
389 the procedure code then has to resize to provide room for any
390 other variables that should go in the case and then fill in the
391 values of "left" variables. Then, when we're done with that case,
392 we have to save the values of "left" variables to copy into the
393 next case read from the active file.
395 The caseinit code helps with this.
397 * automake.mk: Add new files.
399 * caseinit.c: New file.
401 * caseinit.h: New file.
403 2007-06-06 Ben Pfaff <blp@gnu.org>
405 * value.h (value_cnt_from_width): New function.
407 * variable.c (var_get_value_cnt): Use new function.
409 2007-06-06 Ben Pfaff <blp@gnu.org>
411 Add casegrouper, to allow cases read from a given casereader to be
412 broken into groups, each of which has its own casereader.
413 Generally cases are grouped based on having equal values for some
416 * automake.mk: Add new files.
418 * casegrouper.c: New file.
420 * casegrouper.h: New file.
422 2007-06-06 Ben Pfaff <blp@gnu.org>
424 Add interface to lexicographical ordering of cases.
426 * automake.mk: Add new files.
428 * case-ordering.c: New file.
430 * case-ordering.h: New file.
432 2007-06-06 Ben Pfaff <blp@gnu.org>
434 Add casereaders and casewriters, the basis of the new data processing
435 implementation. A casereader is a uniform interface to reading cases
436 from a data source; a casewriter is a uniform interface to writing
437 cases to a data sink.
439 * automake.mk: Add new files.
441 * casereader-filter.c: New file.
443 * casereader-provider.h: New file.
445 * casereader-translator.c: New file.
447 * casereader.c: New file.
449 * casereader.h: New file.
451 * casewriter-provider.h: New file.
453 * casewriter-translator.c: New file.
455 * casewriter.c: New file.
457 * casewriter.h: New file.
459 2007-06-06 Ben Pfaff <blp@gnu.org>
461 "casewindow" data structure that extends the deque (from libpspp)
462 of cases with the ability to dump cases to disk if we get too many
465 * automake.mk: Add new files.
467 * casewindow.c: New file.
469 * casewindow.h: New file.
471 2007-06-06 Ben Pfaff <blp@gnu.org>
473 sparse_cases data structure that augments a sparse_array of cases
474 with the ability to dump cases to disk if we get too many cases in
477 * automake.mk: Add new files.
479 * sparse-cases.c: New file.
481 * sparse-cases.h: New file.
483 2007-06-06 Ben Pfaff <blp@gnu.org>
485 Adds a low-level on-disk case array data structure.
487 * automake.mk: Add new files.
489 * case-tmpfile.c: New file.
491 * case-tmpfile.h: New file.
493 2007-06-06 Ben Pfaff <blp@gnu.org>
495 In a couple of places we calculate the maximum number of cases to
496 keep in memory based on the user-defined workspace. Enable
497 centralizing the calculation through a new function.
499 * settings.c (get_workspace_cases): New function.
501 2007-06-06 Ben Pfaff <blp@gnu.org>
503 The casenumber type is defined in transformations.h, but case.h is
504 a more sensible place. Move it.
506 * case.h (CASENUMBER_MAX): New macro.
507 (typedef casenumber): Move here, from transformations.h.
509 2007-06-03 Ben Pfaff <blp@gnu.org>
511 Slightly generalize case_to_values and case_from_values functions.
513 * case.c (case_to_values): Rename case_copy_out, change interface.
514 (case_from_values): Rename case_copy_in, change interface.
516 * fastfile.c (fastfilereader_get_next_case): Update caller.
517 (write_case_to_disk): Ditto.
519 2007-06-02 Ben Pfaff <blp@gnu.org>
521 Clean up after a forgotten part of patch #5829.
523 * casedeque.h: Remove unused file.
525 * automake.mk: Remove casedeque.h from sources.
527 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
529 * category.c: Removed redundant #include
531 2007-05-06 Ben Pfaff <blp@gnu.org>
533 Abstract the documents within a dictionary a little better.
534 Thanks to John Darrington for suggestion, initial version, and
537 * dictionary.c (struct dictionary): Change `documents' member from
538 char * to struct string.
539 (dict_clear): Destroy struct string.
540 (dict_get_documents): Convert struct string to char *.
541 (dict_set_documents): Set struct string. Pad to 80-character
543 (dict_clear_documents): New function.
544 (dict_add_document_line): New function.
545 (dict_get_document_line_cnt): New function.
546 (dict_get_document_line): New function.
548 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
550 * sys-file-reader.c (read_documents): Use new document functions.
552 2007-04-19 John Darrington <john@darrington.wattle.id.au>
554 * sys-file-reader.c: When reading a system file which has no
555 long name table, automatically create one where the long names
556 are the lower case versions of the short names.
558 2007-04-22 Ben Pfaff <blp@gnu.org>
560 * dictionary.c (dict_set_split_vars): dict_destroy expects that
561 dict_clear will free most data related to the dictionary.
562 dict_clear does a decent job, except that dict_set_split_vars on
563 some systems won't actually free the dict's "split" member.
564 Instead, it'll allocate a 1-byte region. Fix this.
566 * value.c (value_copy): New function.
567 (value_set_missing): Ditto.
569 2007-04-22 John Darrington <john@darrington.wattle.id.au>
571 * Deleted existing category.h and moved cat-routines.h into
572 category.h Encapsulated struct cat_vals better.
574 2007-04-19 John Darrington <john@darrington.wattle.id.au>
576 * sys-file-reader.c: When reading a system file which has no
577 long name table, automatically create one where the long names
578 are the lower case versions of the short names.
580 2007-04-16 John Darrington <john@darrington.wattle.id.au>
582 * sys-file-reader.c: Some versions of Other Software seem to
583 produce system files with string variables' measure set to
584 zero. We'll assume these are supposed to be nominal variables.
586 2007-03-30 Ben Pfaff <blp@gnu.org>
588 * procedure.c: Adapt to new deque data structure.
590 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
591 Ben Pfaff <blp@gnu.org>
593 * file-name.c: Mingw compatibility fixes.
594 (fn_search_path): Use ISSLASH instead of comparing against '/'
596 (fn_dir_name): Use dir_name from gnulib.
597 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
598 (fn_get_identity): Use GetFullPathName instead of canonicalize
599 from gnulib, because the latter does not fully support
600 Windows-style path names. Use this implementation based on the
601 detected presence of Windows instead of the absence of Unix, since
602 the new implementation is Windows-specific.
603 (fn_compare_file_identities): In Windows implementation, compare
604 names case-insensitively.
606 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
608 * make-file.c: Don't include mkstemp.h, because gnulib now causes
609 <stdlib.h> to have the same effect.
611 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
613 * por-file-reader.c: Add missing _() around messages.
615 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
617 * make-file.c: Include "mkstemp.h", without which linking on
620 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
622 Reduce platform dependence.
624 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
625 Everywhere we using this, we really should have just depended on
626 the shell to expand tildes.
627 (fn_search_path): Simplify, given that we don't do tilde expansion
629 (fn_normalize): Removed. Caller changed to use the canonicalize
631 (fn_get_cwd): Removed. Only user was fn_normalize.
632 (fn_is_absolute): Really only test for absolute names.
633 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
635 (fn_readlink): Removed, as it was only used fn_normalize.
636 (fn_exists): Assume the stat function is available; gnulib does.
637 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
639 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
641 * dictionary.c (dict_create_vector_assert): New function.
643 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
645 * file-name.c (fn_normalize): Correct name of function
646 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
647 for pointing this out.
649 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
651 We are using a single member in struct file_handle, the "name"
652 field, for more than one purpose. When it begins with '"', it's a
653 file name; otherwise, it's a token that can be used to identify
654 it. When that assertion fires, it's because we searched for the
655 name case-sensitively as a file name (so that there was no match),
656 and then we try to insert it case-insensitively as a token, which
657 fails because duplicates aren't allowed.
659 Solution: break the two purposes into two separate fields. This
660 fixes the problem and likely makes the code easier to read too.
662 Fixes bug #18922. Thanks to John Darrington for bug report and
665 * file-handle-def.c (struct file_handle): New `id' member.
666 (fh_from_name): Rename fh_from_id. Update all callers.
667 (create_handle): New `id' parameter. Update all callers.
668 (fh_create_file): Ditto.
669 (fh_get_id): New function.
671 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
673 * case.c (case_is_null): Change return type to bool.
675 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
679 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
680 reference counting to share data. This makes it easy for
681 valgrind, etc. to find accesses to cases that have been destroyed
682 but have been kept around by another user's ref-count. This often
683 happens when the data set is small enough to find in memory; if a
684 bigger data set that would overflow to disk were used, then data
685 corruption would occur.
687 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
691 * case.c (case_unshare): Make it check internally whether the
692 ref_cnt is greater than 1, so that the callers don't have to.
693 Update callers not to check.
695 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
697 Before, I was thinking that I might want to get rid of reference
698 counting at some point. Now, I'm pretty sure that it's here to
699 stay. Thus, because we have to store the value_cnt anyway for
700 reference-counted cases, we might as well expose it to users.
702 * case.c (case_get_value_cnt): New function.
703 (case_resize): Drop OLD_CNT argument. Update all callers. Only
704 resize case if its size actually changed.
706 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
707 instead of peeking inside struct case directly.
708 (casefile_append): Ditto.
710 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
712 Get rid of the inlines for the case functions, which made the
713 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
714 the inlines didn't speed up "make check" at all, which is not a
715 perfect benchmark but seems indicative.)
717 * case.c: Remove #ifdef DEBUGGING...#endif around many function
718 definitions. Remove some assertions on nonnull pointers that were
719 redundant with a pointer dereference soon after in the function.
721 (struct case_data): Move definition here from case.h.
725 (case_data_wr): Ditto.
727 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
729 * automake.mk: Add casedeque.h to sources.
731 * casedeque.h: New file.
733 * procedure.c: (struct dataset) Change lag_count, lag_head,
734 lag_queue member into single struct casedeque member. Update all
735 users to use the casedeque instead.
738 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
740 * procedure.c: Simplify lagged cases interface. Updated all
741 clients--well, the only client--to use the simplified interface.
742 (dataset_n_lag) Removed.
743 (dataset_set_n_lag) Removed.
744 (dataset_need_lag) New function.
746 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
748 * dictionary.c procedure.c: More changes to ensure that callbacks occur
749 whenever appropriate, but only when the dataset/dictionary is in a
752 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
754 * dictionary.c dictionary.h : Added callbacks for change of filter and
755 split variables. Refactored some code to ensure that callbacks get
756 invoked when appropriate.
758 * procedure.c (proc_cancel_temporary_transformations): Make sure that
759 replace_dict callback occurs when permanent_dict replaces the current
762 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
764 * dictionary.c dictionary.h : Added callback for when the weight
765 variable of a dictionary changes.
767 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
769 * dictionary.c dictionary.h : Added replace_source and replace_dict
770 callbacks, and functions to deal with them.
772 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
774 Simplify missing value handling.
776 * missing-values.h (enum mv_class): New type.
777 (enum mv_type): Moved definition into missing-values.c and renamed
778 each MV_* to MVT_*, to distinguish them from the exposed mv_class
779 enums. Updated all uses.
780 (struct missing_values): Changed type of `type' from `enum
781 mv_type' to `int' because the definition is no longer exposed.
783 * missing-values.c (mv_is_value_missing): Add new enum mv_class
784 parameter. Update all callers.
785 (mv_is_num_missing): Ditto.
786 (mv_is_str_missing): Ditto.
787 (mv_is_value_user_missing): Removed. Changed callers to use
789 (mv_is_num_user_missing): Removed. Changed callers to use
791 (mv_is_str_user_missing): Removed. Changed callers to use
793 (mv_is_value_system_missing): Removed. Changed callers to use
795 (mv_set_type): Removed. Changed callers to use mv_clear.
796 (mv_clear): New function.
798 * variable.c (var_is_value_missing): Add new enum mv_class
799 parameter. Update all callers.
800 (var_is_num_missing): Ditto.
801 (var_is_str_missing): Ditto.
802 (var_is_value_user_missing): Removed. Changed callers to use
803 var_is_value_missing.
804 (var_is_num_user_missing): Removed. Changed callers to use
806 (var_is_str_user_missing): Removed. Changed callers to use
808 (var_is_value_system_missing): Removed. Changed callers to use
809 var_is_value_missing.
811 * casefilter.c (struct casefilter): Use enum mv_class in place of
813 (casefilter_variable_missing): Adapt to new member.
814 (casefilter_create): Change signature to take enum mv_class,
817 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
819 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
821 * case-sink.c case-sink.h procedure.c procedure.h
822 storage-stream.c: Now uses the factory.
824 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
826 Make it possible to pull cases from the active file with a
827 function call, instead of requiring indirection through a callback
830 * case-source.h (struct case_source_class): Change ->read function
831 to return a single case, instead of calling a callback function
832 for each case. Change ->destroy function to return an error
835 * case-source.c (free_case_source): Pass along the value returned
836 by the case_source ->destroy function.
838 * procedure.c (struct write_case_data): Removed.
839 (struct dataset): Added some members to track procedure state.
840 (procedure): Optimize the trivial case at this level.
841 (internal_procedure): Re-implement in terms of proc_open,
842 proc_read, proc_close.
843 (proc_open) New function.
844 (proc_read) New function.
845 (proc_close) New function.
846 (write_case) Moved into proc_read.
847 (close_active_file) Moved closing of data source into proc_close.
849 * storage-source.c: Rewrote to conform with modified
850 case_source_class interface.
852 * transformations.c (trns_chain_execute): Added argument to allow
853 starting execution from an arbitrary transformation. Updated
856 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
858 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
860 * sys-file-reader.c (read_display_parameters): Don't assume that
861 MEASURE_* and ALIGN_* have the same values found in system files.
863 * sys-file-writer.c (write_variable_display_parameters): Ditto.
865 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
866 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
867 the value of n_MEASURES, which was off by 1 (at least from my
870 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
872 * dictionary.c dictionary.h vardict.h variable.c: Added optional
873 callbacks which are invoked when the dictionary or its
874 variables are changed.
876 * missing-values.c missing-values.h value-labels.c: Tidied up
877 consistency checks, and made some of them return false
878 instead of assert-failing.
880 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
882 * calendar.c (calendar_days_in_month): New function.
884 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
886 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
887 the value label's width.
889 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
891 * sfm-private.h: Move contents into sys-file-writer.c, which is
892 the only remaining user. Removed Borland C++-specific directives.
894 * sys-file-reader.c: Clean up and rewrite entire file. The
895 rewritten version is simpler and better abstracted, and should be
896 easier to maintain and extend. It avoids using structures to read
897 file data, which is prone to padding variations among compilers.
898 It should also handle non-IEEE 754 system files, although I
899 haven't been able to find any. It has been tested against many
900 .sav files obtained from the Web and found to produce the same
901 results as the earlier version of the code, or in some cases
902 improved results. It is more tolerant of format variations found
905 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
906 member, putting an enum integer_format in its place. New member
907 `float_format'. Changed `compressed' member to type bool.
909 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
911 * dictionary.c (dict_delete_consecutive_vars): New function.
913 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
915 * file-name.c (fn_search_path): Remove prefix arg that was unused
916 by any caller. Updated all callers.
918 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
920 * format.c (fmt_dollar_template): Use user's decimal point
921 character. Add assertion.
923 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
925 * format.c (fmt_dollar_template): New function, based on
926 dollar_format_template from var-type-dialog.c.
928 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
930 * data-out.c (output_scientific): Fix bad assumption that "buf" is
933 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
935 Finish converting struct variable to an opaque type. In this
936 phase, we add remaining setter and getter functions, convert the
937 remaining PSPP code to use them, and do a bunch of cleanup. The
938 resulting changes are pervasive but mostly trivial, and only the
939 notable changes are logged.
941 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
944 * case.c (case_data): Renamed case_data_idx.
945 (case_num): Renamed case_num_idx.
946 (case_str): Renamed case_str_idx.
947 (case_data_rw): Renamed case_data_rw_idx.
949 * case.h (case_data): New function with old name and an interface
950 that takes a variable instead of an index, which is easier to
951 use. Updated all callers to use the new interface, or to use the
952 new *_idx function (see above).
955 (case_data_rw): Ditto.
957 * category.c (cat_stored_values_destroy): Changed interface to
958 take a struct cat_vals * instead of a struct variable *.
960 * dictionary.c (dict_clone): Use new vector_clone function.
961 (dict_clear) Use new var_destroy function.
962 (add_var) New function.
963 (dict_create_var) Rewrite in terms of dict_create_var_assert.
964 (dict_create_var_assert) Rewrite in terms of add_var.
965 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
966 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
967 (dict_lookup_var) Use new var_create, var_destroy functions.
968 (dict_contains_var) Rewrite in terms of new vardict functionality.
969 (set_var_dict_index) New function.
970 (set_var_case_index) New function.
971 (reindex_vars) New function.
972 (dict_delete_var) Rewrite in terms of new vardict functionality.
973 (dict_reorder_var) Ditto.
974 (dict_reorder_vars) Ditto.
975 (rename_var) New function.
976 (dict_rename_var) Use rename_var.
977 (dict_rename_vars) Use pool to simplify code. Use rename_var.
978 (dict_get_compacted_idx_to_fv) Rename
979 dict_get_compacted_dict_index_to_case_index, update callers.
980 (dict_create_vector) Use new vector_create function.
981 (dict_clear_vectors) Use new vector_destroy function.
982 (set_var_short_name_suffix) Move here from variable.c, renamed
983 from var_set_short_name_suffix, make static, update caller.
985 * sys-file-private.c: New file.
986 (sfm_width_to_bytes) Moved here from variable.c, renamed from
987 width_to_bytes, update callers.
989 * sys-file-private.h: New file. Later it will supplant
990 sfm-private.h; for now it supplements it.
991 (macro MIN_VERY_LONG_STRING) New macro.
992 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
994 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
995 MAX_LONG_STRING was used before.
997 * sys-file-writer.c: Ditto.
999 * value-labels.c: Change the paradigm here to be that a null
1000 pointer is OK for a struct val_labs * in most cases; it just
1001 represents an empty set of value labels.
1002 (val_labs_copy) A copy of a null set is a null set.
1003 (val_labs_count) A null set has 0 labels.
1004 (val_labs_replace) Change return type to void. Rewrite for
1006 (val_labs_find) A null set does not contain the value.
1007 (value_to_string) Moved to variable.c, renamed var_get_value_name,
1008 transposed argument order, updated all callers.
1010 * value.c: New file.
1011 (value_dup) Moved here from variable.c.
1012 (compare_values) Ditto.
1015 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1016 (macro MAX_LONG_STRING) Removed, because it was only interesting
1017 for system files, not for general code.
1018 (macro MAX_VERY_LONG_STRING) Ditto.
1019 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1020 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1022 * vardict.h: New file, for an interface between variables and
1025 * variable.c: A lot of functions were moved around, for better
1027 (struct variable) Move definition here, from variable.h.
1028 (var_type_adj) Removed--makes i18n hard.
1029 (var_type_noun) Ditto.
1030 (var_create) New function.
1031 (var_clone) New function.
1032 (var_destroy) New function.
1033 (var_set_name) Assert that variable is not in a dictionary.
1034 (compare_var_names) Rename compare_vars_by_name and fix a couple
1035 of callers who thought the args were strings.
1036 (hash_var_name) Rename hash_var_by_name.
1037 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1038 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1039 (var_is_very_long_string) Removed, because it was only interesting
1040 to system file code.
1041 (var_set_missing_values) Allow the argument to be the wrong width,
1042 as long as we can resize it. Simplify callers who were doing the
1043 resizing themselves.
1044 (var_get_value_labels) New function.
1045 (var_has_value_labels) New function.
1046 (var_set_value_labels) New function.
1047 (alloc_value_labels) New function.
1048 (var_add_value_label) New function.
1049 (var_replace_value_label) New function.
1050 (var_clear_value_labels) New function.
1051 (var_lookup_value_label) New function.
1052 (var_get_value_name) Moved here from variable.c, renamed from
1053 var_get_value_name, transposed argument order, updated all
1055 (var_to_string) Moved here, from variable-label.c.
1056 (var_set_leave) New function.
1057 (var_get_leave) New function.
1058 (var_must_leave) New function.
1059 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1060 set_var_short_name_suffix.
1061 (var_get_dict_index) New function.
1062 (var_get_case_index) New function.
1063 (var_get_obs_vals) New function.
1064 (var_set_obs_vals) New function.
1065 (var_has_obs_vals) New function.
1066 (var_get_vardict) New function.
1067 (var_set_vardict) New function.
1068 (var_has_vardict) New function.
1069 (var_clear_vardict) New function.
1070 (value_dup) Moved to value.c.
1071 (compare_values) Ditto.
1074 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1075 (enum ALPHA) Rename VAR_STRING, update all users.
1077 * vector.c: New file.
1078 (struct vector) Moved here, from variable.h.
1079 (check_widths) New function.
1080 (vector_create) New function.
1081 (vector_clone) New function.
1082 (vector_destroy) New function.
1083 (vector_get_name) New function.
1084 (vector_get_var) New function.
1085 (vector_get_var_cnt) New function.
1086 (compare_vector_ptrs_by_name) New function.
1088 * vector.h: New file.
1090 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1092 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1093 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1096 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1098 * value-labels.c (destroy_atoms): New function.
1099 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1101 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1103 Thanks to Jason Stover for pointing out this problem.
1105 * data-out.c (output_number): Use gsl_finite from GSL, which is
1106 portable, instead of isfinite, which is not.
1109 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1111 * variable.c variable.h (value_dup): New function.
1113 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1115 Start converting struct variable to an opaque type. In this
1116 phase, we add a bunch of setter and getter functions and convert
1117 most of the PSPP code to use them. The resulting changes are
1118 pervasive but mostly trivial, and only the notable changes are
1121 * format.c (fmt_equal): New function.
1123 * variable.c (var_type_is_valid): New function.
1124 (measure_is_valid) Moved here, from format.c.
1125 (alignment_is_valid) Moved here, from format.c.
1126 (var_get_name) New function.
1127 (var_set_name) New function.
1128 (width_to_type) New function.
1129 (var_get_type) New function.
1130 (var_get_width) New function.
1131 (var_set_width) New function.
1132 (var_is_numeric) New function.
1133 (var_is_alpha) New function.
1134 (var_is_short_string) New function.
1135 (var_is_long_string) New function.
1136 (var_is_very_long_string) New function.
1137 (var_get_missing_values) New function.
1138 (var_set_missing_values) New function.
1139 (var_clear_missing_values) New function.
1140 (var_has_missing_values) New function.
1141 (var_is_value_missing) New function.
1142 (var_is_num_missing) New function.
1143 (var_is_str_missing) New function.
1144 (var_is_value_user_missing) New function.
1145 (var_is_num_user_missing) New function.
1146 (var_is_str_user_missing) New function.
1147 (var_is_value_system_missing) New function.
1148 (var_get_print_format) New function.
1149 (var_set_print_format) New function.
1150 (var_get_write_format) New function.
1151 (var_set_write_format) New function.
1152 (var_set_both_formats) New function.
1153 (var_get_label) New function.
1154 (var_set_label) New function.
1155 (var_clear_label) New function.
1156 (var_has_label) New function.
1157 (var_get_measure) New function.
1158 (var_set_measure) New function.
1159 (var_get_display_width) New function.
1160 (var_set_display_width) New function.
1161 (var_get_alignment) New function.
1162 (var_set_alignment) New function.
1163 (var_get_value_cnt) New function.
1164 (var_get_leave) New function.
1165 (var_get_short_name) New function.
1167 * variable.h: (struct variable) Removed "type" and "nv" members;
1168 they are now computed from "width" where needed.
1170 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1172 * missing-values.c (mv_resize): Don't write beyond end of the
1173 allocated buffer when resizing a long string.
1175 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1177 Clean up identifier code: don't require identifier enumerations to
1178 be in a particular order; make better use of string library;
1179 expose less of the internals.
1181 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1182 change interface. Updated all callers.
1183 (lex_id_match) Change interface to use struct substring, update
1185 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1186 (global array keywords[]) Make static, change form. Update all
1187 users to use lex_id_name instead.
1188 (lex_is_keyword) New function.
1189 (lex_id_to_token) Change interface to use struct substring, update
1191 (lex_id_name) New function.
1193 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1194 lex_is_keyword instead.
1195 (T_LAST_KEYWORD) Removed.
1196 (T_N_KEYWORDS) Removed.
1198 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1200 * format.c: (fmt_date_template) Distinguish characters for which a
1201 space is output and any date delimiter is allowed on input, from
1202 those for which a space is output and only a space is allowed on
1203 input. The former is represented by X, the latter by a space.
1204 Also, drop distinction between h and H, changing the former to the
1207 * data-in.c: Completely rewrite internals to conform to SPSS input
1208 formats as closely as possible.
1209 (data_in) Changed external interface by replacing the structure
1210 that was used as a single argument by a set of arguments. Updated
1212 (data_in_finite_line) Removed. Converted all callers to use plain
1214 (data_in_get_integer_format) New function.
1215 (data_in_set_integer_format) New function.
1216 (data_in_get_float_format) New function.
1217 (data_in_set_float_format) New function.
1219 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1220 (struct data_in) Removed.
1222 * data-out.c: (output_date) Drop each component from the input as
1223 it is output, to allow us to drop the distinction between h (a
1224 count of hours) and H (the hour of day) template characters.
1225 Also, handle new X template character.
1226 (output_scientific) Follow more rational rule on when to drop
1227 fraction introduced between SPSS 13 and 15. Updated test case to
1230 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1232 Fix buffer overflow reported by John Darrington.
1234 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1235 realize that DIGITS is a count of nibbles, not of bytes.
1237 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1239 * calendar.c (calendar_offset_to_gregorian) Also return the
1240 year-of-day. Change callers to new interface.
1242 * data-out.c: Completely rewrite internals to conform to SPSS
1243 output formats as completely as possible.
1244 (data_out) Change interface to put input parameters before output
1245 parameters, for consistency with the style I now prefer. Update
1247 (data_out_get_integer_format) New public function.
1248 (data_out_set_integer_format) New public function.
1249 (data_out_get_float_format) New public function.
1250 (data_out_set_float_format) New public function.
1252 * data-out.h: New file. Move prototype for data_out here, from
1255 * format.c: (fmt_step_width) Use equality comparison instead of
1256 bitwise and, for clarity.
1257 (fmt_is_string) Ditto.
1258 (fmt_input_to_output) Fix categories that are translated to F
1261 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1263 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1264 casefile-private.h: Added casefilter to assist commands with missing
1267 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1269 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1271 * settings.c: (route_errors_to_terminal) New variable.
1272 (route_errors_to_listing) New variable.
1273 (get_error_routing_to_terminal) New function.
1274 (set_error_routing_to_terminal) New function.
1275 (get_error_routing_to_listing) New function.
1276 (set_error_routing_to_listing) New function.
1278 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1280 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1282 * format.c: Completely rewrite, to achieve better abstraction.
1283 Rewrite all references to formats in other files.
1285 * format.def: Rewrite and reorganize.
1287 * settings.c: Move everything related to custom currency formats
1288 into format.[ch], changing them in form, so as to group related
1289 code and definitions better. Changed all references to use the
1291 (static var decimal) Removed.
1292 (static var grouping) Removed.
1293 (static var cc) Removed.
1294 (get_decimal) Removed.
1295 (set_decimal) Removed.
1296 (get_grouping) Removed.
1297 (set_grouping) Removed.
1301 * settings.h: (macro CC_CNT) Removed.
1302 (macro CC_WIDTH) Removed.
1303 (struct custom_currency) Removed.
1305 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1307 * data-in.c (data_in): Use switch statement instead of table, to
1308 avoid dependence on the order of the FMT_* enums.
1310 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1312 * data-out.c: (num_to_string) Removed, because it was dead code.
1314 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1316 * data-in.c (parse_trailer): Fix error message.
1318 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1320 * format.c (fmt_is_binary): New function.
1322 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1324 * procedure.c procedure.h: Encapsulated the static data into a single
1327 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1329 * casefile.c (casereader_read_xfer): Always initialize the case,
1330 even on an error condition.
1332 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1334 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1335 entire function useless.
1337 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1339 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1340 casereader method casereader_clone.
1342 * procedure.c pransformations.h: Introduced new type casenum_t
1344 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1346 * variable.c: (width_to_bytes) Rephrase code for clarify.
1348 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1350 * format.c: (fmt_type_from_string) New function.
1351 (fmt_to_string) Include decimals in output if the format has
1352 decimals, even if the format type does not. This way, we can
1353 accurately reproduce incorrect formats in user output.
1354 (check_common_specifier) Make the check for a bad format type an
1355 assertion, so we get bug reports if they show up. Fix message.
1356 Check for decimal places with a format type that doesn't allow
1358 (check_input_specifier) Remove check for FMT_X, which has been
1360 (check_output_specifier) Ditto.
1362 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1364 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1365 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1366 (enum fmt_parse_flags) Removed.
1368 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1370 * casefile.c casefile.h: Converted to an abstract base class.
1371 * casefile-private.h fastfile.c fastfile.h: New files.
1372 * automake.mk procedure.c scratch-writer.c storage-stream.c
1374 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1376 * procedure.c (internal_procedure): Create sink_case with only as
1377 many values as the compacted dictionary.
1379 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1381 Remove "debugging" code that caused plenty of false positives and
1384 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1386 * case.c: Remove all references to `this' member.
1388 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1390 Fix link error noted by Jason Stover.
1392 * storage-stream.c: Include <assert.h>.
1394 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1396 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1397 ALL) and additional underlying system file issues.
1399 Thanks to John Darrington for review.
1401 First problem: var_hash points to variables not owned by the
1402 sys-file-reader, which the caller may free or modify. Use an
1403 array of sfm_vars instead, as done earlier (e.g. CVS version
1406 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1407 members and remove all code that references it. Add vars, var_cnt
1408 members. Remove fix_specials member, which was unused.
1409 (struct sfm_var) Remove name member, which was unused.
1410 (sfm_close_reader) Free vars member instead of var_hash.
1411 (compare_var_shortnames) Removed.
1412 (hash_var_shortname) Removed.
1413 (sfm_open_reader) Fill out vars array.
1414 (compare_var_index) Removed.
1415 (sfm_read_case) Use vars instead of var_hash.
1417 Second problem: we're confused about when we actually have very
1418 long strings, causing us to choose incorrectly between slow path
1419 and fast path in sfm_read_case.
1421 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1422 have very long strings, not when we have long variable names,
1423 which is an unrelated feature.
1425 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1427 * variable.h: Move var_set and variable parsing declarations to
1428 new header, src/language/lexer/variable-parser.h. Modified lots
1429 of files to include the new header.
1431 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1433 * value-labels.c (value_to_string): When there's no value label,
1434 format the variable according to its print format, instead of
1435 always effectively using A or F format.
1437 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1439 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1442 * format.c: Removed tortological assertion.
1444 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1446 Reform string library.
1448 * file-name.c (fn_interp_vars): Change interface to take a
1449 substring as input. Updated all users.
1451 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1453 * format.c (measure_is_valid): Really return false when m >=
1456 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1458 Implement random access to casefiles, for use in GUI.
1460 * casefile.c: (struct casereader) Add `random', `file_ofs',
1461 `buffer_ofs' members.
1462 (casefile_get_random_reader) New function.
1463 (read_open_file) Break part into new function
1464 seek_and_fill_buffer().
1465 (fill_buffer) Update buffer_ofs, file_ofs.
1466 (casereader_seek) New function.
1468 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1470 * settings.c: Added call to i18n{done, init}.
1472 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1474 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1477 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1479 * casefile.c: Convert many uses of `int' to `bool'.
1481 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1483 * transformations.c (trns_chain_destroy): Destroy chain's trns
1484 member, to fix memory leak.
1486 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1488 * storage-stream.c (storage_source_decapsulate): Destroy case
1489 source to fix memory leak.
1491 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1493 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1494 case passed as argument instead of initializing the argument as a
1495 case. Fixes memory leak that showed up in
1496 tests/command/aggregate.sh with scratch files.
1498 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1500 * procedure.c (proc_done): Destroy default_dict, to fix memory
1503 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1505 Simplify procedure_with_splits().
1507 * procedure.c (struct split_aux_data): Removed case_count member.
1508 (procedure_with_splits) Don't initialize case_count.
1509 (split_procedure_case_func) Check whether prev_case is null
1510 instead of case_count.
1511 (split_procedure_end_func) Ditto.
1513 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1515 * case.c (case_move): Do nothing if dst and src are the same
1517 (case_try_create) Merge two similar cases.
1518 (case_copy) Unshare only if data must be actually copied.
1520 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1522 * data-in.c data-out.c dictionary.c sys-file-reader.c
1523 sys-file-writer.c variable.c variable.h: Reworked very long string
1524 support for better encapsulation.
1526 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1528 * value-labels.c (val_labs_can_set_width): New function.
1529 (val_labs_set_width) Clear labels if increasing width to long
1531 (val_labs_destroy) Remove unneeded test for null.
1533 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1535 * value-labels.h: Remove unneeded dependency on variable.h.
1537 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1539 Get rid of `char *c' member in union value, for cleanliness.
1541 * value.h: (union value) Remove `c' member.
1543 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1545 Make dictionary compacting functions a little more general.
1547 * sys-file-writer.c (sfm_open_writer): Use
1548 dict_compacting_would_change().
1549 (does_dict_need_translation) Removed.
1551 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1553 Make dictionary compacting functions a little more general.
1555 * dictionary.c (dict_needs_compaction): Rename
1556 dict_compacting_would_shrink(). Update all callers.
1557 (dict_compacting_would_change) New function.
1559 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1561 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1562 inverted return value (!).
1564 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1566 Continue reforming procedure execution.
1568 * procedure.c: Search and replace "vfm" by "proc". Notably:
1569 (static var vfm_source) Rename proc_source. Update all
1571 (static var vfm_sink) Rename proc_sink. Update all references.
1573 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1575 Continue reforming procedure execution. In this phase, remove
1576 PROCESS IF, which was deprecated anyway and can be easily
1577 simulated with TEMPORARY followed by SELECT IF.
1579 * procedure.c: (open_active_file) Don't call
1580 add_process_if_trns().
1581 (discard_variables) Get rid of redundant call to
1582 proc_cancel_all_transformations().
1583 (add_process_if_trns) Removed.
1584 (process_if_trns_proc) Removed.
1585 (process_if_trns_free) Removed.
1587 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1589 Continue reforming procedure execution. In this phase, add
1590 `const' to the case passed to procedure()'s callback.
1592 Updated all users of procedure() as well.
1594 * procedure.c: (struct write_case_data) Add "const" to ccase
1595 parameter for case_func member.
1596 (procedure) Add "const" to ccase parameter for proc_func
1598 (multipass_case_func) Make ccase parameter const.
1599 (internal_procedure) Add "const" to ccase parameter for case_func
1601 (split_procedure_case_func) Make ccase parameter const.
1602 (multipass_split_case_func) Make ccase parameter const.
1604 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
1606 Continue reforming procedure execution. In this phase, get rid of
1607 the output code for SPLIT FILE groups in procedure.c, which really
1608 shouldn't be doing any output. Move it into the individual
1609 procedures instead. This also adds some flexibility.
1611 Updated many users of procedure_with_splits() and
1612 multipass_procedure_with_splits() to call
1613 output_split_file_values() and to deal with increased use of
1616 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
1617 parameter to begin_func member.
1618 (procedure_with_splits) Add "const struct ccase *" parameter to
1619 begin_func parameter. Make ccase parameter const in proc_func
1621 (split_procedure_case_func) Don't dump split file group. Pass
1623 (dump_splits) Moved to language/dictionary/split-file.c as
1624 output_split_file_values().
1625 (struct multipass_split_aux_data) Add "const struct ccase *"
1626 parameter to split_func member.
1627 (multipass_procedure_with_splits) Add "const struct ccase *"
1628 parameter to split_func parameter.
1629 (multipass_split_case_func) Save new SPLIT FILE case before
1631 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
1633 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
1635 Continue reforming procedure execution. Change
1636 internal_procedure() so that it calls open_active_file() and
1637 close_active_file(), which isolates most of the actual procedure
1640 * procedure.c: (struct write_case_data) Rename `proc_func' member
1641 to `case_func' and update all references.
1642 (procedure) Rewrite as one-line wrapper around
1643 internal_procedure().
1644 (struct multipass_aux_data) New.
1645 (multipass_callback) Renamed multipass_case_func(). Use struct
1646 multipass_aux_data as auxiliary data.
1647 (multipass_end_func) New function.
1648 (multipass_procedure) Rewrite as wrapper for internal_procedure()
1649 that uses multipass_case_func, multipass_end_func.
1650 (internal_procedure) Add `end_func' argument. Move optimization
1651 of trivial case in here. Move call to open_active_file() and
1652 close_active_file() in here. Now assert that vfm_source is
1654 (procedure_with_splits_callback) Rename
1655 split_procedure_case_func().
1656 (split_procedure_end_func) New function.
1657 (multipass_split_callback) Rename multipass_split_case_func.
1658 (multipass_split_end_func) New function.
1659 (discard_variables) No need to test for nonnull vfm_source.
1661 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
1663 Continue reforming procedure execution. Get rid of unused member.
1665 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
1667 (write_case) Don't increment cases_analyzed.
1669 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
1671 Continue reforming procedure execution. In this phase, move
1672 procedure.c and procedure.h from src to src/data. Update
1673 makefiles and #includes accordingly.
1675 * procedure.c: Moved here from src/.
1677 * procedure.h: Moved here from src/.
1679 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
1681 Continue reforming procedure execution. In this phase, get rid of
1682 many global variables, consolidating procedure execution in
1683 procedure.c. Encapsulate transformations in new "struct
1684 trns_chain". Also, change implementation of N OF CASES, FILTER,
1685 and PROCESS IF from special cases to transformations.
1687 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
1690 * dictionary.c: (global variable default_dict) Move to
1693 * variable.h: (TRNS_*) Move to transformations.h.
1694 (struct transformation) Move to transformations.c.
1696 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1698 * sys-file-reader.c: Fixed invalid read problems.
1700 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
1702 * storage-stream.c: Add missing function comments.
1704 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
1706 Continue reforming procedure execution. In this phase, add some
1707 new, needed functionality to storage-stream.
1709 * storage-stream.c: (storage_source_decapsulate) New function.
1711 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
1713 * variable.c (width_to_bytes): Declarations must precede
1714 statements for C90 compliance.
1716 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
1718 * data-out.c, data-in.c, variable.c, variable.h: New functions
1719 copy_mangle and copy_demangle for reading/writing cases; emulates the
1720 way SPSS deals with strings > 255 bytes.
1722 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
1723 subtype 14 needed for strings longer than 255 bytes.
1725 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
1726 instead of literal values. Also fixed some constness issues.
1728 * format.h: Constness
1730 * sfm-private.h: Renamed the case_size identifier, since I discovered
1731 that SPSS's respect for this variable is very nominal.
1733 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
1735 Change case limit type from int to size_t.
1737 * dictionary.c: (struct dictionary) Change type of case_limit
1739 (dict_get_case_limit) Change return type.
1740 (dict_set_case_limit) Change parameter type.
1742 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
1744 * variable.h: (struct variable) Rename `reinit' member as `leave'
1745 and invert sense. Fix up all references.
1747 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
1749 Continue reforming procedure execution. In this phase, break
1750 procedure.c into multiple files.
1752 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
1754 * case-sink.c: New file.
1756 * case-sink.h: New file.
1758 * case-source.c: New file.
1760 * case-source.h: New file.
1762 * storage-stream.c: New file.
1764 * storage-stream.h: New file.
1766 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
1768 * variable.h: (struct variable) Remove `init' member and all
1769 references to it from other files. It was initialized in several
1770 places, but nothing really ever used it for anything worthwhile.
1771 Thanks to Jason Stover for pointing out how confusing this
1774 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
1776 Continue reforming error message support. In this phase, get rid
1777 of message "titles" and put the message text in `struct error'.
1778 Now `struct error' encapsulates a message more properly.
1780 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
1781 Format message ourselves.
1783 * data-in.c: (vdls_error) Ditto.
1785 * por-file-reader.c: (error) Ditto.
1787 * sys-file-reader.c: (corrupt_msg) Ditto.
1789 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
1791 GNU standards require "file name" instead of "filename" in
1792 documentation. It's nice for our code to follow the convention
1795 * casefile.c: (struct casefile) Rename `filename' member to
1796 `file_name'. Updated all references.
1798 * file-name.c: [!unix] (struct file_identity) Rename
1799 normalized_filename member to normalized_file_name. Updated all
1802 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
1804 We don't really support anything but Unix-like environments well,
1805 so we might as well de-obfuscate by writing directory and path
1806 separators explicitly.
1808 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
1810 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
1811 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
1813 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
1816 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
1818 GNU standards require "file name" instead of "filename" in
1819 documentation. It's nice for our code to follow the convention
1822 * filename.c: Rename to file-name.c.
1824 * filename.h: Rename to file-name.h. Update all inclusions.
1825 Update header guards.
1827 * automake.mk: Update file names.
1829 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
1831 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
1833 (fn_basename) Removed (dead code).
1834 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
1835 (fn_special_p) Renamed fn_is_special(), all references updated.
1836 (fn_exists_p) Renamed fn_exists(), all references updated.
1838 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
1840 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
1841 Also, now it only tilde-expands file names, not paths.
1842 (fn_search_path) Tilde-expand one directory at a time.
1844 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
1846 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
1847 get rid of non-Unixlike version of the code, which has probably
1849 (fn_prepend_dir) Removed (dead code).
1851 * filename.h: (macro DIR_SEPARATOR_STRING) New.
1852 (macro PATH_SEPARATOR_STRING) New.
1853 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
1855 Continue reforming error message support. In this phase, we get
1856 rid of VM() and the other msg() support for "verbosity", replacing
1857 it by a new function verbose_msg().
1859 * filename.c: (fn_search_path) Use verbose_msg() instead of
1862 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
1864 * sfm-private.h: Get rid of #defines after #error, which makes no
1867 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
1869 Get rid of our own int32 type in favor of the standard int32_t
1872 * sfm-private.h: (int32 macro) Don't define this anymore. Do
1875 * sys-file-reader.c: Use int32_t instead of int32 throughout.
1877 * sys-file-writer.c: Use int32_t instead of int32 throughout.
1879 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
1881 Remove ill-considered file routines that are no longer used.
1883 * filename.c: (fn_open_ext) Removed.
1884 (fn_close_ext) Removed.
1886 * filename.h: (struct file_ext) Removed.
1888 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
1890 * variable.c (var_is_valid_name): Move declarations before code
1893 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
1895 * filename.ch (fn_interp_vars): Fixed small buglet.
1897 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
1899 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
1900 so as to modify the string inline. Thus makeing it easier to
1901 destroy the results when no longer needed.
1903 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
1905 * category.c (cat_stored_values_destroy): Fixed memory leak.
1907 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
1909 Add some missing frees. Thanks to John Darrington for reporting
1912 * any-writer.c (any_writer_close): Free writer.
1914 * any-reader.c (any_reader_close): Free reader.
1916 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
1918 * por-file-reader.c: (error) Mark as NO_RETURN.
1920 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
1922 * settings.c: Changed default value of scompress to true.
1924 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
1926 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
1928 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
1931 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
1932 here from pref.h.orig.
1934 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
1936 * sys-file-reader.c: Fixed bug reading compressed files.
1938 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1940 * Numerous renames. See src/ChangeLog for details.
1942 * Moved files from src directory