1 2007-07-29 Ben Pfaff <blp@gnu.org>
3 Provisional fix for bug #18692 and bug #20161. Reviewed by John
6 * file-name.c (fn_open): Only pass "r" or "w" to popen as mode
7 argument (never "rb" or "wb") because SUSv3 says that only those
8 modes are defined, and glibc in fact rejects other modes.
10 Open portable files with fn_open so that they can be read from
11 pipes. Fix missing fh_close call to go along with fh_open.
12 Report an error if the file close reports an error.
13 * por-file-reader.c (close_reader): New function.
14 (por_file_casereader_destroy): Use close_reader.
15 (pfm_open_reader): Open file with fn_open.
17 2007-07-28 Ben Pfaff <blp@gnu.org>
19 Make PSPP able to read all the portable files I could find on the
20 web. Thanks to John Darrington for review. Bug #17620.
21 * por-file-reader.c (struct pfm_reader): New member `line_length'.
22 (error): Print file offset in hexadecimal.
23 (warning): New function.
24 (advance): Treat lines less than 80 bytes long as padded to 80
26 (pfm_open_reader): Call read_documents if we find an "E" record.
27 (convert_format): Convert invalid formats to the default format
28 instead of aborting reading the file.
29 (read_variables): Rename duplicate variable names instead of
30 aborting reading the file.
31 (read_value_label): Allow string variables of different widths to
32 be assigned value labels in the same record. Replace duplicate
33 value labels instead of aborting.
34 (read_documents): New function.
36 * por-file-writer.c (pfm_open_writer): Call write_documents if the
37 dictionary has documents.
38 (write_documents): New function.
40 2007-07-25 Ben Pfaff <blp@gnu.org>
42 Fix bugs related to bug #17213.
44 * settings.c: Use HAVE_LIBNCURSES instead of HAVE_LIBTERMCAP,
45 since the former is what config.h has. Include the needed ncurses
47 (static var echo) Rename to `do_echo' because the original name is
48 the same as an ncurses identifier.
49 (get_termcap_viewport) Use error instead of msg.
51 * file-name.c (fn_interp_vars): Fix interpolation of $VARS.
52 (fn_close): Don't close stdin, stdout, stderr.
54 2007-07-26 John Darrington <john@darrington.wattle.id.au>
56 * procedure.c procedure.h: Added callbacks which get invoked whenever
57 a dataset's transformation chain changes.
59 2007-07-24 Ben Pfaff <blp@gnu.org>
62 * sys-file-writer.c (write_variable_display_parameters): Use new
63 var_default_display_width function to choose display width of
64 segments after the first one in a given variable.
65 * variable.c (var_create): Use var_default_display_width to pick
66 new variable's display width.
67 (var_default_display_width): New function.
68 Reviewed by John Darrington.
70 2007-07-24 Ben Pfaff <blp@gnu.org>
73 * por-file-writer.c (write_variables): Write weight variable.
74 Reviewed by John Darrington.
76 2007-07-23 Ben Pfaff <blp@gnu.org>
78 Improvements to system file reader and writer.
80 First, move all detailed knowledge of very long strings into
81 sys-file-private.[ch], so that this nasty stuff can be isolated.
83 * sys-file-private.c (REAL_VLS_CHUNK): New macro.
84 (EFFECTIVE_VLS_CHUNK): New macro.
85 (min_int): New function.
86 (max_int): New function.
87 (sfm_width_to_bytes): Rewrite.
88 (sfm_width_to_octs): New function.
89 (sfm_segment_alloc_width): New function.
90 (sfm_segment_alloc_bytes): New function.
91 (sfm_segment_used_bytes): New function.
92 (sfm_segment_offset): New function.
93 (sfm_segment_effective_offset): New function.
94 (sfm_dictionary_to_sfm_vars): New function.
96 * sys-file-private.h (MIN_VERY_LONG_STRING): Removed.
97 (EFFECTIVE_LONG_STRING_LENGTH): Removed.
98 (struct sfm_var): New structure.
100 Next, improvements to the system file reader.
102 * sys-file-reader.h (struct sfm_read_info): Changed `case_cnt' to
103 type casenumber. Added `version_major', `version_minor',
106 * sys-file-reader.c (struct sfm_reader): Replaced `flt64_cnt' by
107 `oct_cnt'. Rename `vars', `var_cnt' to `sfm_vars', `sfm_var_cnt'.
108 Change `case_cnt' to type casenumber. Removed `has_vls'.
109 (struct sfm_var): Removed.
110 (sfm_open_reader): Don't warn on wrong case size if the file was
111 written by SPSS 13, which tends to get it wrong. Use
112 sfm_dictionary_to_sfm_vars.
113 (read_header): Always output system file info.
114 (read_variable_record): Simplify code for reading missing values.
115 (read_machine_int32_info): Save version numbers from system file
116 into info struct passed as new argument.
117 (read_long_string_map): Restructured to use new sys-file-private
119 (read_value_labels): Use size_overflow_p.
120 (sys_file_casereader_read): Get rid of distinction between fast
121 and slow paths. Use information provided by sys-file-primate's
122 struct sfm_var to simplify code.
123 (skip_whole_strings): New function.
124 (read_int32): Renamed read_int. Changed return value to int.
126 (read_flt64): Renamed read_float. Changed return value to
127 double. Updated all callers.
128 (int32_to_native): Removed. Changed callers to use
130 (flt64_to_double): Removed. Changed callers to use float_convert.
132 Finally, get rid of int32, flt64 terminology and types in system
133 file writer. The former wasn't very useful since a POSIX "int"
134 can hold the whole range of int32 and we generally didn't have a
135 need for it to be exactly-32-bits, just at-least-32-bits. The
136 latter was inconvenient because we had to assume that it could be
137 different from double and thereby convert special values SYSMIS,
138 HIGHEST, LOWEST to and from it in multiple places. Instead, now
139 we just use "int" and "double" in most places, and do conversions,
140 if necessary, very close to where we do I/O. This change meant
141 that the writer code couldn't represent records in the file as C
142 structs any longer, but that's no great loss. The code actually
143 seems to be more readable without them.
145 Simplify the compression buffering code: only buffer as much as
146 necessary, which is no more than eight 8-byte units at any given
149 * sys-file-writer.c (typedef flt64): Removed.
150 (macro second_lowest_flt64): Removed.
151 (struct sysfile_header): Removed.
152 (struct sysfile_variable): Removed.
153 (struct sfm_writer): Removed `needs_translation', `has_vls',
154 `flt64_cnt'. Changed `compress' to type bool and `case_cnt' to
155 type casenumber. Renamed `vars' to `sfm_vars', `var_cnt' to
156 `sfm_var_cnt'. Replaced `buf', `end', `ptr', `x', `y' for
157 compression buffering by `opcodes', `opcode_cnt', `data',
158 `data_cnt'. Renamed `var_cnt_vls' as `segment_cnt'.
159 (sfm_open_writer): Use sfm_dictionary_to_sfm_vars. Use simple
160 data writer functions instead of structures.
161 (calc_oct_idx): New function.
162 (write_header): Use simple data writer functions instead of
164 (write_format_spec): Renamed write_format. New argument.
165 (write_variable_continuation_records): New function.
166 (write_variable): Use simple data writer functions instead of
167 structures. Use write_variable_continuation_records. Write
168 entire very long string instead of requiring caller to understand
170 (write_value_labels): Use simple data writer functions instead of
172 (write_documents): Ditto.
173 (write_variable_display_parameters): Use sys-file-private
174 functions to simplify. Use simple data writer functions instead
176 (write_vls_length_table): Use simple data writer functions instead
178 (write_longvar_table): Ditto.
179 (write_rec_7_34): Break into new functions
180 write_integer_info_record, write_float_info_record. Use simple
181 data writer functions instead of structures.
182 (buf_write): Removed.
183 (append_string_max): Removed.
184 (ensure_buf_space): Removed.
185 (sys_file_casewriter_write): Get rid of the distinction between
186 fast and slow paths, which didn't seem to be too useful. Use new
187 functions write_case_uncompressed, write_case_compressed.
188 (put_instruction): Removed.
189 (put_element): Removed.
190 (write_compressed_data): Removed.
191 (close_writer): Use flush_compressed. Only write case count to
192 system file if it will fit in the field.
193 (write_case_compressed): New function.
194 (write_case_uncompressed): New function.
195 (flush_compressed): New function.
196 (put_cmp_opcode): New function.
197 (put_cmp_number): New function.
198 (write_int): New function.
199 (convert_double_to_output_format): New function.
200 (write_float): New function.
201 (write_value): New function.
202 (write_string): New function.
203 (write_bytes): New function.
204 (write_zeros): New function.
205 (write_spaces): New function.
207 Reviewed by John Darrington.
209 2007-07-22 Ben Pfaff <blp@gnu.org>
211 Don't try to write very long strings to portable files. The
212 format does not support it.
214 * por-file-writer.c (MAX_POR_WIDTH): New macro.
215 (pfm_open_writer): Limit output width to MAX_POR_WIDTH.
216 (write_format): Add arg to take width to resize format to.
217 (write_value): Limit width of value written to MAX_POR_WIDTH.
218 (write_variables): Limit width of variable and its output formats
220 Reviewed by John Darrington.
222 2007-07-22 Ben Pfaff <blp@gnu.org>
224 * sys-file-reader.c (read_variable_to_value_map): Use max_warnings
225 local variable instead of literal 5.
226 Reviewed by John Darrington.
228 2007-07-22 Ben Pfaff <blp@gnu.org>
230 Fix problems with uniqueness of short names in system files with
231 very long string variables. Now a variable may have multiple
234 * automake.mk (src_data_libdata_a_SOURCES): Add new files
235 short-names.c, short-names.h.
237 * dictionary.c (dict_clone): Clone all the short names.
238 (compare_strings): Move into short-names.c.
239 (hash_strings): Ditto.
240 (set_var_short_name_suffix): Ditto.
241 (dict_assign_short_names): Ditto, rename short_names_assign,
242 change to assign all short names.
244 * por-file-writer.c (write_variables): Use short_names_assign
245 instead of dict_assign_short_names.
247 * short-names.c: New file.
249 * short-names.h: New file.
251 * sys-file-private.c (sfm_width_to_segments): New function.
253 * sys-file-reader.c (read_long_var_name_map): Save and restore all
254 the short names, not just the first one.
256 * sys-file-writer.c (cont_var_name): Removed.
257 (sfm_open_writer): Use short_names_assign instead of
258 dict_assign_short_names. Use unique short names assigned by
259 short_names_assign instead of those generated by cont_var_name.
261 * variable.c (struct variable): Remove `short_name' member,
262 replace by `short_names' and `short_name_cnt'.
263 (var_create) Initialize new members.
264 (var_get_short_name_cnt): New function.
265 (var_get_short_name): Now takes an index argument. Changed most
267 (var_set_short_name): Ditto.
268 (var_clear_short_name): Renamed var_clear_short_names, changed to
269 clear all short names.
271 Reviewed by John Darrington.
273 2007-07-22 Ben Pfaff <blp@gnu.org>
275 * variable.c (var_set_width): Use new var_set_width function.
277 * missing-values.c (mv_n_values): Drop assertion, which was not
280 * format.c (fmt_default_for_width): New function.
281 (fmt_resize): New function.
283 Reviewed by John Darrington.
285 2007-07-18 John Darrington <john@darrington.wattle.id.au>
287 * datasheet.c (datasheet_delete_columns): Added assertion to check
288 we're not deleting outside the range of the sheet.
291 * dictionary.c dictionary.h variable.c: Added the ability for string
292 variables to be resized.
294 * vardict.h: Added some prototypes (moved from dictionary.h) as
295 these should only be called by variable.c
298 2007-07-14 John Darrington <john@darrington.wattle.id.au>
300 * sfm-reader.c: Respect case_cnt field in file header.
302 2007-07-01 John Darrington <john@darrington.wattle.id.au>
304 * transformation.c transformation.h (trns_chain_execute): Changed the
305 signature (Patch #6057)
307 2007-06-10 Ben Pfaff <blp@gnu.org>
309 * casereader-filter.c (casereader_filter_destroy): Make sure to
310 write all the remaining excluded cases to the casewriter, if any.
312 * caseinit.c (init_list_destroy): Rewrite.
313 (init_list_clear): Ditto.
315 * casegrouper.c (casegrouper_get_next_group): Always set *reader
316 to null when returning false.
318 2007-06-06 Ben Pfaff <blp@gnu.org>
320 Actually implement the new procedure code and adapt all of its
321 clients to match. Also adapt all of the other case sources and
322 sinks in the tree and their clients to use the
323 casereader/casewriter infrastructure.
325 * automake.mk: Add and remove files.
327 * any-reader.c: Change into a casereader.
328 * por-file-reader.c: Ditto.
329 * scratch-reader.c: Ditto.
330 * sys-file-reader.c: Ditto.
332 * any-writer.c: Change into a casewriter.
333 * por-file-writer.c: Ditto.
334 * scratch-writer.c: Ditto.
335 * sys-file-writer.c: Ditto.
337 * procedure.c: Change to use casereader, casewriter, caseinit, and
338 other new infrastructure.
340 * scratch-handle.c: Adapt to new infrastructure.
342 * case-sink.c: Removed, now dead code.
343 * case-sink.h: Ditto.
344 * case-source.c: Ditto.
345 * case-source.h: Ditto.
346 * casefile-factory.c: Ditto.
347 * casefile-private.h: Ditto.
350 * casefilter.c: Ditto.
351 * casefilter.h: Ditto.
354 * fastfile-factory.c: Ditto.
355 * fastfile-factory.h: Ditto.
356 * storage-stream.c: Ditto.
357 * storage-stream.h: Ditto.
359 2007-06-06 Ben Pfaff <blp@gnu.org>
363 * automake.mk: Add new files.
365 * datasheet.c: New file.
367 * datasheet.h: New file.
369 2007-06-06 Ben Pfaff <blp@gnu.org>
371 Until now, the procedure code has provided a case to the
372 case_source, which has filled in the data values that come from
373 the active file. "Left" data values that don't come from the
374 active file naturally stay the same from case to case, because the
375 procedure code keeps using that same case.
377 One of the compromises that comes with the new procedure code is
378 that the active file allocates and provides its own case, which
379 the procedure code then has to resize to provide room for any
380 other variables that should go in the case and then fill in the
381 values of "left" variables. Then, when we're done with that case,
382 we have to save the values of "left" variables to copy into the
383 next case read from the active file.
385 The caseinit code helps with this.
387 * automake.mk: Add new files.
389 * caseinit.c: New file.
391 * caseinit.h: New file.
393 2007-06-06 Ben Pfaff <blp@gnu.org>
395 * value.h (value_cnt_from_width): New function.
397 * variable.c (var_get_value_cnt): Use new function.
399 2007-06-06 Ben Pfaff <blp@gnu.org>
401 Add casegrouper, to allow cases read from a given casereader to be
402 broken into groups, each of which has its own casereader.
403 Generally cases are grouped based on having equal values for some
406 * automake.mk: Add new files.
408 * casegrouper.c: New file.
410 * casegrouper.h: New file.
412 2007-06-06 Ben Pfaff <blp@gnu.org>
414 Add interface to lexicographical ordering of cases.
416 * automake.mk: Add new files.
418 * case-ordering.c: New file.
420 * case-ordering.h: New file.
422 2007-06-06 Ben Pfaff <blp@gnu.org>
424 Add casereaders and casewriters, the basis of the new data processing
425 implementation. A casereader is a uniform interface to reading cases
426 from a data source; a casewriter is a uniform interface to writing
427 cases to a data sink.
429 * automake.mk: Add new files.
431 * casereader-filter.c: New file.
433 * casereader-provider.h: New file.
435 * casereader-translator.c: New file.
437 * casereader.c: New file.
439 * casereader.h: New file.
441 * casewriter-provider.h: New file.
443 * casewriter-translator.c: New file.
445 * casewriter.c: New file.
447 * casewriter.h: New file.
449 2007-06-06 Ben Pfaff <blp@gnu.org>
451 "casewindow" data structure that extends the deque (from libpspp)
452 of cases with the ability to dump cases to disk if we get too many
455 * automake.mk: Add new files.
457 * casewindow.c: New file.
459 * casewindow.h: New file.
461 2007-06-06 Ben Pfaff <blp@gnu.org>
463 sparse_cases data structure that augments a sparse_array of cases
464 with the ability to dump cases to disk if we get too many cases in
467 * automake.mk: Add new files.
469 * sparse-cases.c: New file.
471 * sparse-cases.h: New file.
473 2007-06-06 Ben Pfaff <blp@gnu.org>
475 Adds a low-level on-disk case array data structure.
477 * automake.mk: Add new files.
479 * case-tmpfile.c: New file.
481 * case-tmpfile.h: New file.
483 2007-06-06 Ben Pfaff <blp@gnu.org>
485 In a couple of places we calculate the maximum number of cases to
486 keep in memory based on the user-defined workspace. Enable
487 centralizing the calculation through a new function.
489 * settings.c (get_workspace_cases): New function.
491 2007-06-06 Ben Pfaff <blp@gnu.org>
493 The casenumber type is defined in transformations.h, but case.h is
494 a more sensible place. Move it.
496 * case.h (CASENUMBER_MAX): New macro.
497 (typedef casenumber): Move here, from transformations.h.
499 2007-06-03 Ben Pfaff <blp@gnu.org>
501 Slightly generalize case_to_values and case_from_values functions.
503 * case.c (case_to_values): Rename case_copy_out, change interface.
504 (case_from_values): Rename case_copy_in, change interface.
506 * fastfile.c (fastfilereader_get_next_case): Update caller.
507 (write_case_to_disk): Ditto.
509 2007-06-02 Ben Pfaff <blp@gnu.org>
511 Clean up after a forgotten part of patch #5829.
513 * casedeque.h: Remove unused file.
515 * automake.mk: Remove casedeque.h from sources.
517 2007-05-10 Jason Stover <jhs@math.gcsu.edu>
519 * category.c: Removed redundant #include
521 2007-05-06 Ben Pfaff <blp@gnu.org>
523 Abstract the documents within a dictionary a little better.
524 Thanks to John Darrington for suggestion, initial version, and
527 * dictionary.c (struct dictionary): Change `documents' member from
528 char * to struct string.
529 (dict_clear): Destroy struct string.
530 (dict_get_documents): Convert struct string to char *.
531 (dict_set_documents): Set struct string. Pad to 80-character
533 (dict_clear_documents): New function.
534 (dict_add_document_line): New function.
535 (dict_get_document_line_cnt): New function.
536 (dict_get_document_line): New function.
538 * dictionary.h (macro DOC_LINE_LENGTH): New macro.
540 * sys-file-reader.c (read_documents): Use new document functions.
542 2007-04-19 John Darrington <john@darrington.wattle.id.au>
544 * sys-file-reader.c: When reading a system file which has no
545 long name table, automatically create one where the long names
546 are the lower case versions of the short names.
548 2007-04-22 Ben Pfaff <blp@gnu.org>
550 * dictionary.c (dict_set_split_vars): dict_destroy expects that
551 dict_clear will free most data related to the dictionary.
552 dict_clear does a decent job, except that dict_set_split_vars on
553 some systems won't actually free the dict's "split" member.
554 Instead, it'll allocate a 1-byte region. Fix this.
556 * value.c (value_copy): New function.
557 (value_set_missing): Ditto.
559 2007-04-22 John Darrington <john@darrington.wattle.id.au>
561 * Deleted existing category.h and moved cat-routines.h into
562 category.h Encapsulated struct cat_vals better.
564 2007-04-19 John Darrington <john@darrington.wattle.id.au>
566 * sys-file-reader.c: When reading a system file which has no
567 long name table, automatically create one where the long names
568 are the lower case versions of the short names.
570 2007-04-16 John Darrington <john@darrington.wattle.id.au>
572 * sys-file-reader.c: Some versions of Other Software seem to
573 produce system files with string variables' measure set to
574 zero. We'll assume these are supposed to be nominal variables.
576 2007-03-30 Ben Pfaff <blp@gnu.org>
578 * procedure.c: Adapt to new deque data structure.
580 Mon Feb 19 10:53:21 2007 John McCabe-Dansted <gmatht@gmail.com>
581 Ben Pfaff <blp@gnu.org>
583 * file-name.c: Mingw compatibility fixes.
584 (fn_search_path): Use ISSLASH instead of comparing against '/'
586 (fn_dir_name): Use dir_name from gnulib.
587 (fn_is_absolute): Use IS_ABSOLUTE_FILE_NAME from gnulib.
588 (fn_get_identity): Use GetFullPathName instead of canonicalize
589 from gnulib, because the latter does not fully support
590 Windows-style path names. Use this implementation based on the
591 detected presence of Windows instead of the absence of Unix, since
592 the new implementation is Windows-specific.
593 (fn_compare_file_identities): In Windows implementation, compare
594 names case-insensitively.
596 Sun Feb 18 13:28:02 2007 Ben Pfaff <blp@gnu.org>
598 * make-file.c: Don't include mkstemp.h, because gnulib now causes
599 <stdlib.h> to have the same effect.
601 Sun Feb 18 11:20:24 2007 Ben Pfaff <blp@gnu.org>
603 * por-file-reader.c: Add missing _() around messages.
605 Sun Feb 11 20:44:13 2007 Ben Pfaff <blp@gnu.org>
607 * make-file.c: Include "mkstemp.h", without which linking on
610 Thu Feb 8 14:59:05 2007 Ben Pfaff <blp@gnu.org>
612 Reduce platform dependence.
614 * file-name.c (fn_tilde_expand): Removed, and removed calls to it.
615 Everywhere we using this, we really should have just depended on
616 the shell to expand tildes.
617 (fn_search_path): Simplify, given that we don't do tilde expansion
619 (fn_normalize): Removed. Caller changed to use the canonicalize
621 (fn_get_cwd): Removed. Only user was fn_normalize.
622 (fn_is_absolute): Really only test for absolute names.
623 (fn_is_special): Use pipe files if HAVE_POPEN, not if we're in
625 (fn_readlink): Removed, as it was only used fn_normalize.
626 (fn_exists): Assume the stat function is available; gnulib does.
627 (fn_open): Use pipe files if HAVE_POPEN, not if we're in unix.
629 Sat Feb 3 21:52:17 2007 Ben Pfaff <blp@gnu.org>
631 * dictionary.c (dict_create_vector_assert): New function.
633 Wed Feb 7 21:25:15 2007 Ben Pfaff <blp@gnu.org>
635 * file-name.c (fn_normalize): Correct name of function
636 fn_is_special. Thanks to John McCabe-Dansted <gmatht@gmail.com>
637 for pointing this out.
639 Thu Feb 1 16:53:37 2007 Ben Pfaff <blp@gnu.org>
641 We are using a single member in struct file_handle, the "name"
642 field, for more than one purpose. When it begins with '"', it's a
643 file name; otherwise, it's a token that can be used to identify
644 it. When that assertion fires, it's because we searched for the
645 name case-sensitively as a file name (so that there was no match),
646 and then we try to insert it case-insensitively as a token, which
647 fails because duplicates aren't allowed.
649 Solution: break the two purposes into two separate fields. This
650 fixes the problem and likely makes the code easier to read too.
652 Fixes bug #18922. Thanks to John Darrington for bug report and
655 * file-handle-def.c (struct file_handle): New `id' member.
656 (fh_from_name): Rename fh_from_id. Update all callers.
657 (create_handle): New `id' parameter. Update all callers.
658 (fh_create_file): Ditto.
659 (fh_get_id): New function.
661 Mon Jan 15 16:18:10 2007 Ben Pfaff <blp@gnu.org>
663 * case.c (case_is_null): Change return type to bool.
665 Mon Jan 15 10:57:28 2007 Ben Pfaff <blp@gnu.org>
669 * case.c (case_clone) [DEBUGGING]: When debugging, don't use
670 reference counting to share data. This makes it easy for
671 valgrind, etc. to find accesses to cases that have been destroyed
672 but have been kept around by another user's ref-count. This often
673 happens when the data set is small enough to find in memory; if a
674 bigger data set that would overflow to disk were used, then data
675 corruption would occur.
677 Mon Jan 15 10:55:18 2007 Ben Pfaff <blp@gnu.org>
681 * case.c (case_unshare): Make it check internally whether the
682 ref_cnt is greater than 1, so that the callers don't have to.
683 Update callers not to check.
685 Mon Jan 15 10:53:01 2007 Ben Pfaff <blp@gnu.org>
687 Before, I was thinking that I might want to get rid of reference
688 counting at some point. Now, I'm pretty sure that it's here to
689 stay. Thus, because we have to store the value_cnt anyway for
690 reference-counted cases, we might as well expose it to users.
692 * case.c (case_get_value_cnt): New function.
693 (case_resize): Drop OLD_CNT argument. Update all callers. Only
694 resize case if its size actually changed.
696 * casefile.c (casefile_append_xfer): Use case_get_value_cnt
697 instead of peeking inside struct case directly.
698 (casefile_append): Ditto.
700 Mon Jan 15 10:50:22 2007 Ben Pfaff <blp@gnu.org>
702 Get rid of the inlines for the case functions, which made the
703 header file hard to read. (Also, in testing with "-O2 -DNDEBUG",
704 the inlines didn't speed up "make check" at all, which is not a
705 perfect benchmark but seems indicative.)
707 * case.c: Remove #ifdef DEBUGGING...#endif around many function
708 definitions. Remove some assertions on nonnull pointers that were
709 redundant with a pointer dereference soon after in the function.
711 (struct case_data): Move definition here from case.h.
715 (case_data_wr): Ditto.
717 Sun Jan 14 21:41:12 2007 Ben Pfaff <blp@gnu.org>
719 * automake.mk: Add casedeque.h to sources.
721 * casedeque.h: New file.
723 * procedure.c: (struct dataset) Change lag_count, lag_head,
724 lag_queue member into single struct casedeque member. Update all
725 users to use the casedeque instead.
728 Sun Jan 14 21:43:12 2007 Ben Pfaff <blp@gnu.org>
730 * procedure.c: Simplify lagged cases interface. Updated all
731 clients--well, the only client--to use the simplified interface.
732 (dataset_n_lag) Removed.
733 (dataset_set_n_lag) Removed.
734 (dataset_need_lag) New function.
736 Tue Jan 9 07:20:05 WST 2007 John Darrington <john@darrington.wattle.id.au>
738 * dictionary.c procedure.c: More changes to ensure that callbacks occur
739 whenever appropriate, but only when the dataset/dictionary is in a
742 Sun Jan 7 08:33:04 WST 2007 John Darrington <john@darrington.wattle.id.au>
744 * dictionary.c dictionary.h : Added callbacks for change of filter and
745 split variables. Refactored some code to ensure that callbacks get
746 invoked when appropriate.
748 * procedure.c (proc_cancel_temporary_transformations): Make sure that
749 replace_dict callback occurs when permanent_dict replaces the current
752 Wed Jan 3 11:02:11 WST 2007 John Darrington <john@darrington.wattle.id.au>
754 * dictionary.c dictionary.h : Added callback for when the weight
755 variable of a dictionary changes.
757 Mon Jan 1 10:36:26 WST 2007 John Darrington <john@darrington.wattle.id.au>
759 * dictionary.c dictionary.h : Added replace_source and replace_dict
760 callbacks, and functions to deal with them.
762 Fri Dec 22 13:56:08 2006 Ben Pfaff <blp@gnu.org>
764 Simplify missing value handling.
766 * missing-values.h (enum mv_class): New type.
767 (enum mv_type): Moved definition into missing-values.c and renamed
768 each MV_* to MVT_*, to distinguish them from the exposed mv_class
769 enums. Updated all uses.
770 (struct missing_values): Changed type of `type' from `enum
771 mv_type' to `int' because the definition is no longer exposed.
773 * missing-values.c (mv_is_value_missing): Add new enum mv_class
774 parameter. Update all callers.
775 (mv_is_num_missing): Ditto.
776 (mv_is_str_missing): Ditto.
777 (mv_is_value_user_missing): Removed. Changed callers to use
779 (mv_is_num_user_missing): Removed. Changed callers to use
781 (mv_is_str_user_missing): Removed. Changed callers to use
783 (mv_is_value_system_missing): Removed. Changed callers to use
785 (mv_set_type): Removed. Changed callers to use mv_clear.
786 (mv_clear): New function.
788 * variable.c (var_is_value_missing): Add new enum mv_class
789 parameter. Update all callers.
790 (var_is_num_missing): Ditto.
791 (var_is_str_missing): Ditto.
792 (var_is_value_user_missing): Removed. Changed callers to use
793 var_is_value_missing.
794 (var_is_num_user_missing): Removed. Changed callers to use
796 (var_is_str_user_missing): Removed. Changed callers to use
798 (var_is_value_system_missing): Removed. Changed callers to use
799 var_is_value_missing.
801 * casefilter.c (struct casefilter): Use enum mv_class in place of
803 (casefilter_variable_missing): Adapt to new member.
804 (casefilter_create): Change signature to take enum mv_class,
807 Fri Dec 22 20:08:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
809 * casefile-factory.h fastfile-factory.c fastfile-factory.h: New files.
811 * case-sink.c case-sink.h procedure.c procedure.h
812 storage-stream.c: Now uses the factory.
814 Sat Dec 16 22:05:18 2006 Ben Pfaff <blp@gnu.org>
816 Make it possible to pull cases from the active file with a
817 function call, instead of requiring indirection through a callback
820 * case-source.h (struct case_source_class): Change ->read function
821 to return a single case, instead of calling a callback function
822 for each case. Change ->destroy function to return an error
825 * case-source.c (free_case_source): Pass along the value returned
826 by the case_source ->destroy function.
828 * procedure.c (struct write_case_data): Removed.
829 (struct dataset): Added some members to track procedure state.
830 (procedure): Optimize the trivial case at this level.
831 (internal_procedure): Re-implement in terms of proc_open,
832 proc_read, proc_close.
833 (proc_open) New function.
834 (proc_read) New function.
835 (proc_close) New function.
836 (write_case) Moved into proc_read.
837 (close_active_file) Moved closing of data source into proc_close.
839 * storage-source.c: Rewrote to conform with modified
840 case_source_class interface.
842 * transformations.c (trns_chain_execute): Added argument to allow
843 starting execution from an arbitrary transformation. Updated
846 * transformations.h (enum TRNS_NEXT_CASE) Renamed TRNS_END_CASE.
848 Sat Dec 16 14:09:25 2006 Ben Pfaff <blp@gnu.org>
850 * sys-file-reader.c (read_display_parameters): Don't assume that
851 MEASURE_* and ALIGN_* have the same values found in system files.
853 * sys-file-writer.c (write_variable_display_parameters): Ditto.
855 * variable.h: Change MEASURE_NOMINAL, MEASURE_ORDINAL,
856 MEASURE_SCALE to be 0-based instead of 1-based. This also fixes
857 the value of n_MEASURES, which was off by 1 (at least from my
860 Sat Dec 16 12:17:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
862 * dictionary.c dictionary.h vardict.h variable.c: Added optional
863 callbacks which are invoked when the dictionary or its
864 variables are changed.
866 * missing-values.c missing-values.h value-labels.c: Tidied up
867 consistency checks, and made some of them return false
868 instead of assert-failing.
870 Wed Dec 13 19:30:11 2006 Ben Pfaff <blp@gnu.org>
872 * calendar.c (calendar_days_in_month): New function.
874 Mon Dec 11 07:53:39 2006 Ben Pfaff <blp@gnu.org>
876 * value-labels.c (hash_int_val_lab): Only hash as many bytes as
877 the value label's width.
879 Sun Dec 10 14:21:29 2006 Ben Pfaff <blp@gnu.org>
881 * sfm-private.h: Move contents into sys-file-writer.c, which is
882 the only remaining user. Removed Borland C++-specific directives.
884 * sys-file-reader.c: Clean up and rewrite entire file. The
885 rewritten version is simpler and better abstracted, and should be
886 easier to maintain and extend. It avoids using structures to read
887 file data, which is prone to padding variations among compilers.
888 It should also handle non-IEEE 754 system files, although I
889 haven't been able to find any. It has been tested against many
890 .sav files obtained from the Web and found to produce the same
891 results as the earlier version of the code, or in some cases
892 improved results. It is more tolerant of format variations found
895 * sys-file-reader.h (struct sfm_read_info): Removed `big_endian'
896 member, putting an enum integer_format in its place. New member
897 `float_format'. Changed `compressed' member to type bool.
899 Sun Dec 10 13:48:53 2006 Ben Pfaff <blp@gnu.org>
901 * dictionary.c (dict_delete_consecutive_vars): New function.
903 Sat Dec 9 20:08:25 2006 Ben Pfaff <blp@gnu.org>
905 * file-name.c (fn_search_path): Remove prefix arg that was unused
906 by any caller. Updated all callers.
908 Sat Dec 9 20:04:22 2006 Ben Pfaff <blp@gnu.org>
910 * format.c (fmt_dollar_template): Use user's decimal point
911 character. Add assertion.
913 Sat Dec 9 20:02:25 2006 Ben Pfaff <blp@gnu.org>
915 * format.c (fmt_dollar_template): New function, based on
916 dollar_format_template from var-type-dialog.c.
918 Sat Dec 9 18:05:59 2006 Ben Pfaff <blp@gnu.org>
920 * data-out.c (output_scientific): Fix bad assumption that "buf" is
923 Sat Dec 9 17:23:23 2006 Ben Pfaff <blp@gnu.org>
925 Finish converting struct variable to an opaque type. In this
926 phase, we add remaining setter and getter functions, convert the
927 remaining PSPP code to use them, and do a bunch of cleanup. The
928 resulting changes are pervasive but mostly trivial, and only the
929 notable changes are logged.
931 * automake.mk (src_data_libdata_a_SOURCES): Add the new source
934 * case.c (case_data): Renamed case_data_idx.
935 (case_num): Renamed case_num_idx.
936 (case_str): Renamed case_str_idx.
937 (case_data_rw): Renamed case_data_rw_idx.
939 * case.h (case_data): New function with old name and an interface
940 that takes a variable instead of an index, which is easier to
941 use. Updated all callers to use the new interface, or to use the
942 new *_idx function (see above).
945 (case_data_rw): Ditto.
947 * category.c (cat_stored_values_destroy): Changed interface to
948 take a struct cat_vals * instead of a struct variable *.
950 * dictionary.c (dict_clone): Use new vector_clone function.
951 (dict_clear) Use new var_destroy function.
952 (add_var) New function.
953 (dict_create_var) Rewrite in terms of dict_create_var_assert.
954 (dict_create_var_assert) Rewrite in terms of add_var.
955 (dict_clone_var) Rewrite in terms of dict_clone_var_assert.
956 (dict_clone_var_assert) Rewrite in terms of var_clone, add_var.
957 (dict_lookup_var) Use new var_create, var_destroy functions.
958 (dict_contains_var) Rewrite in terms of new vardict functionality.
959 (set_var_dict_index) New function.
960 (set_var_case_index) New function.
961 (reindex_vars) New function.
962 (dict_delete_var) Rewrite in terms of new vardict functionality.
963 (dict_reorder_var) Ditto.
964 (dict_reorder_vars) Ditto.
965 (rename_var) New function.
966 (dict_rename_var) Use rename_var.
967 (dict_rename_vars) Use pool to simplify code. Use rename_var.
968 (dict_get_compacted_idx_to_fv) Rename
969 dict_get_compacted_dict_index_to_case_index, update callers.
970 (dict_create_vector) Use new vector_create function.
971 (dict_clear_vectors) Use new vector_destroy function.
972 (set_var_short_name_suffix) Move here from variable.c, renamed
973 from var_set_short_name_suffix, make static, update caller.
975 * sys-file-private.c: New file.
976 (sfm_width_to_bytes) Moved here from variable.c, renamed from
977 width_to_bytes, update callers.
979 * sys-file-private.h: New file. Later it will supplant
980 sfm-private.h; for now it supplements it.
981 (macro MIN_VERY_LONG_STRING) New macro.
982 (macro EFFECTIVE_LONG_STRING_LENGTH) New macro, from value.h.
984 * sys-file-reader.c: Use MIN_VERY_LONG_STRING - 1 where
985 MAX_LONG_STRING was used before.
987 * sys-file-writer.c: Ditto.
989 * value-labels.c: Change the paradigm here to be that a null
990 pointer is OK for a struct val_labs * in most cases; it just
991 represents an empty set of value labels.
992 (val_labs_copy) A copy of a null set is a null set.
993 (val_labs_count) A null set has 0 labels.
994 (val_labs_replace) Change return type to void. Rewrite for
996 (val_labs_find) A null set does not contain the value.
997 (value_to_string) Moved to variable.c, renamed var_get_value_name,
998 transposed argument order, updated all callers.
1000 * value.c: New file.
1001 (value_dup) Moved here from variable.c.
1002 (compare_values) Ditto.
1005 * value.h: (macro MAX_SHORT_STRING) Rewrote for simplicity.
1006 (macro MAX_LONG_STRING) Removed, because it was only interesting
1007 for system files, not for general code.
1008 (macro MAX_VERY_LONG_STRING) Ditto.
1009 (macro EFFECTIVE_LONG_STRING_LENGTH) Moved to sys-file-private.h.
1010 (macro MAX_ELEMS_PER_VALUE) Removed, as it was unused.
1012 * vardict.h: New file, for an interface between variables and
1015 * variable.c: A lot of functions were moved around, for better
1017 (struct variable) Move definition here, from variable.h.
1018 (var_type_adj) Removed--makes i18n hard.
1019 (var_type_noun) Ditto.
1020 (var_create) New function.
1021 (var_clone) New function.
1022 (var_destroy) New function.
1023 (var_set_name) Assert that variable is not in a dictionary.
1024 (compare_var_names) Rename compare_vars_by_name and fix a couple
1025 of callers who thought the args were strings.
1026 (hash_var_name) Rename hash_var_by_name.
1027 (compare_var_ptr_names) Rename compare_var_ptrs_by_name.
1028 (hash_var_ptr_name) Rename hash_var_ptr_by_name.
1029 (var_is_very_long_string) Removed, because it was only interesting
1030 to system file code.
1031 (var_set_missing_values) Allow the argument to be the wrong width,
1032 as long as we can resize it. Simplify callers who were doing the
1033 resizing themselves.
1034 (var_get_value_labels) New function.
1035 (var_has_value_labels) New function.
1036 (var_set_value_labels) New function.
1037 (alloc_value_labels) New function.
1038 (var_add_value_label) New function.
1039 (var_replace_value_label) New function.
1040 (var_clear_value_labels) New function.
1041 (var_lookup_value_label) New function.
1042 (var_get_value_name) Moved here from variable.c, renamed from
1043 var_get_value_name, transposed argument order, updated all
1045 (var_to_string) Moved here, from variable-label.c.
1046 (var_set_leave) New function.
1047 (var_get_leave) New function.
1048 (var_must_leave) New function.
1049 (var_set_short_name_suffix) Moved to dictionary.c, renamed
1050 set_var_short_name_suffix.
1051 (var_get_dict_index) New function.
1052 (var_get_case_index) New function.
1053 (var_get_obs_vals) New function.
1054 (var_set_obs_vals) New function.
1055 (var_has_obs_vals) New function.
1056 (var_get_vardict) New function.
1057 (var_set_vardict) New function.
1058 (var_has_vardict) New function.
1059 (var_clear_vardict) New function.
1060 (value_dup) Moved to value.c.
1061 (compare_values) Ditto.
1064 * variable.h: (enum NUMERIC) Rename VAR_NUMERIC, update all users.
1065 (enum ALPHA) Rename VAR_STRING, update all users.
1067 * vector.c: New file.
1068 (struct vector) Moved here, from variable.h.
1069 (check_widths) New function.
1070 (vector_create) New function.
1071 (vector_clone) New function.
1072 (vector_destroy) New function.
1073 (vector_get_name) New function.
1074 (vector_get_var) New function.
1075 (vector_get_var_cnt) New function.
1076 (compare_vector_ptrs_by_name) New function.
1078 * vector.h: New file.
1080 Sun Dec 10 11:32:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1082 * casefilter.c (casefilter_variable_missing): Avoided comparision of
1083 string variables to SYSMIS. Thanks to Ben Pfaff for reporting this
1086 Sat Dec 9 07:18:03 WST 2006 John Darrington <john@darrington.wattle.id.au>
1088 * value-labels.c (destroy_atoms): New function.
1089 * value-labels.c (atom_create): Call destroy_atoms in atexit handler.
1091 Thu Dec 7 17:38:26 2006 Ben Pfaff <blp@gnu.org>
1093 Thanks to Jason Stover for pointing out this problem.
1095 * data-out.c (output_number): Use gsl_finite from GSL, which is
1096 portable, instead of isfinite, which is not.
1099 Thu Dec 7 15:22:38 WST 2006 John Darrington <john@darrington.wattle.id.au>
1101 * variable.c variable.h (value_dup): New function.
1103 Mon Dec 4 22:20:17 2006 Ben Pfaff <blp@gnu.org>
1105 Start converting struct variable to an opaque type. In this
1106 phase, we add a bunch of setter and getter functions and convert
1107 most of the PSPP code to use them. The resulting changes are
1108 pervasive but mostly trivial, and only the notable changes are
1111 * format.c (fmt_equal): New function.
1113 * variable.c (var_type_is_valid): New function.
1114 (measure_is_valid) Moved here, from format.c.
1115 (alignment_is_valid) Moved here, from format.c.
1116 (var_get_name) New function.
1117 (var_set_name) New function.
1118 (width_to_type) New function.
1119 (var_get_type) New function.
1120 (var_get_width) New function.
1121 (var_set_width) New function.
1122 (var_is_numeric) New function.
1123 (var_is_alpha) New function.
1124 (var_is_short_string) New function.
1125 (var_is_long_string) New function.
1126 (var_is_very_long_string) New function.
1127 (var_get_missing_values) New function.
1128 (var_set_missing_values) New function.
1129 (var_clear_missing_values) New function.
1130 (var_has_missing_values) New function.
1131 (var_is_value_missing) New function.
1132 (var_is_num_missing) New function.
1133 (var_is_str_missing) New function.
1134 (var_is_value_user_missing) New function.
1135 (var_is_num_user_missing) New function.
1136 (var_is_str_user_missing) New function.
1137 (var_is_value_system_missing) New function.
1138 (var_get_print_format) New function.
1139 (var_set_print_format) New function.
1140 (var_get_write_format) New function.
1141 (var_set_write_format) New function.
1142 (var_set_both_formats) New function.
1143 (var_get_label) New function.
1144 (var_set_label) New function.
1145 (var_clear_label) New function.
1146 (var_has_label) New function.
1147 (var_get_measure) New function.
1148 (var_set_measure) New function.
1149 (var_get_display_width) New function.
1150 (var_set_display_width) New function.
1151 (var_get_alignment) New function.
1152 (var_set_alignment) New function.
1153 (var_get_value_cnt) New function.
1154 (var_get_leave) New function.
1155 (var_get_short_name) New function.
1157 * variable.h: (struct variable) Removed "type" and "nv" members;
1158 they are now computed from "width" where needed.
1160 Mon Dec 4 21:38:40 2006 Ben Pfaff <blp@gnu.org>
1162 * missing-values.c (mv_resize): Don't write beyond end of the
1163 allocated buffer when resizing a long string.
1165 Sat Dec 2 16:28:32 2006 Ben Pfaff <blp@gnu.org>
1167 Clean up identifier code: don't require identifier enumerations to
1168 be in a particular order; make better use of string library;
1169 expose less of the internals.
1171 * identifier.c: (lex_skip_identifier) Rename lex_id_get_length,
1172 change interface. Updated all callers.
1173 (lex_id_match) Change interface to use struct substring, update
1175 (lex_id_match_len) Removed. Update callers to use lex_id_match.
1176 (global array keywords[]) Make static, change form. Update all
1177 users to use lex_id_name instead.
1178 (lex_is_keyword) New function.
1179 (lex_id_to_token) Change interface to use struct substring, update
1181 (lex_id_name) New function.
1183 * identifier.h: (T_FIRST_KEYWORD) Removed. Changed users to call
1184 lex_is_keyword instead.
1185 (T_LAST_KEYWORD) Removed.
1186 (T_N_KEYWORDS) Removed.
1188 Sat Nov 18 20:46:35 2006 Ben Pfaff <blp@gnu.org>
1190 * format.c: (fmt_date_template) Distinguish characters for which a
1191 space is output and any date delimiter is allowed on input, from
1192 those for which a space is output and only a space is allowed on
1193 input. The former is represented by X, the latter by a space.
1194 Also, drop distinction between h and H, changing the former to the
1197 * data-in.c: Completely rewrite internals to conform to SPSS input
1198 formats as closely as possible.
1199 (data_in) Changed external interface by replacing the structure
1200 that was used as a single argument by a set of arguments. Updated
1202 (data_in_finite_line) Removed. Converted all callers to use plain
1204 (data_in_get_integer_format) New function.
1205 (data_in_set_integer_format) New function.
1206 (data_in_get_float_format) New function.
1207 (data_in_set_float_format) New function.
1209 * data-in.h: (enums DI_IGNORE_ERROR, DI_IMPLIED_DECIMALS) Removed.
1210 (struct data_in) Removed.
1212 * data-out.c: (output_date) Drop each component from the input as
1213 it is output, to allow us to drop the distinction between h (a
1214 count of hours) and H (the hour of day) template characters.
1215 Also, handle new X template character.
1216 (output_scientific) Follow more rational rule on when to drop
1217 fraction introduced between SPSS 13 and 15. Updated test case to
1220 Sat Nov 11 11:41:26 2006 Ben Pfaff <blp@gnu.org>
1222 Fix buffer overflow reported by John Darrington.
1224 * data-out.c (output_bcd_integer): In case of SYSMIS, etc.,
1225 realize that DIGITS is a count of nibbles, not of bytes.
1227 Sat Nov 4 15:59:56 2006 Ben Pfaff <blp@gnu.org>
1229 * calendar.c (calendar_offset_to_gregorian) Also return the
1230 year-of-day. Change callers to new interface.
1232 * data-out.c: Completely rewrite internals to conform to SPSS
1233 output formats as completely as possible.
1234 (data_out) Change interface to put input parameters before output
1235 parameters, for consistency with the style I now prefer. Update
1237 (data_out_get_integer_format) New public function.
1238 (data_out_set_integer_format) New public function.
1239 (data_out_get_float_format) New public function.
1240 (data_out_set_float_format) New public function.
1242 * data-out.h: New file. Move prototype for data_out here, from
1245 * format.c: (fmt_step_width) Use equality comparison instead of
1246 bitwise and, for clarity.
1247 (fmt_is_string) Ditto.
1248 (fmt_input_to_output) Fix categories that are translated to F
1251 Sun Nov 5 08:29:34 WST 2006 John Darrington <john@darrington.wattle.id.au>
1253 * casefilter.c casefilter.h (new files), casefile.c casefile.h
1254 casefile-private.h: Added casefilter to assist commands with missing
1257 Sat Nov 4 11:47:09 2006 Ben Pfaff <blp@gnu.org>
1259 Implement SET ERRORS, SHOW ERRORS. Fixes bug #17609.
1261 * settings.c: (route_errors_to_terminal) New variable.
1262 (route_errors_to_listing) New variable.
1263 (get_error_routing_to_terminal) New function.
1264 (set_error_routing_to_terminal) New function.
1265 (get_error_routing_to_listing) New function.
1266 (set_error_routing_to_listing) New function.
1268 * settings.h: (SET_ROUTE_* enums) Removed, because unused.
1270 Tue Oct 31 19:58:27 2006 Ben Pfaff <blp@gnu.org>
1272 * format.c: Completely rewrite, to achieve better abstraction.
1273 Rewrite all references to formats in other files.
1275 * format.def: Rewrite and reorganize.
1277 * settings.c: Move everything related to custom currency formats
1278 into format.[ch], changing them in form, so as to group related
1279 code and definitions better. Changed all references to use the
1281 (static var decimal) Removed.
1282 (static var grouping) Removed.
1283 (static var cc) Removed.
1284 (get_decimal) Removed.
1285 (set_decimal) Removed.
1286 (get_grouping) Removed.
1287 (set_grouping) Removed.
1291 * settings.h: (macro CC_CNT) Removed.
1292 (macro CC_WIDTH) Removed.
1293 (struct custom_currency) Removed.
1295 Tue Oct 31 19:56:19 2006 Ben Pfaff <blp@gnu.org>
1297 * data-in.c (data_in): Use switch statement instead of table, to
1298 avoid dependence on the order of the FMT_* enums.
1300 Tue Oct 31 19:35:36 2006 Ben Pfaff <blp@gnu.org>
1302 * data-out.c: (num_to_string) Removed, because it was dead code.
1304 Tue Oct 31 18:09:24 2006 Ben Pfaff <blp@gnu.org>
1306 * data-in.c (parse_trailer): Fix error message.
1308 Sat Oct 28 11:56:50 2006 Ben Pfaff <blp@gnu.org>
1310 * format.c (fmt_is_binary): New function.
1312 Thu Oct 19 22:59:56 WST 2006 John Darrington <john@darrington.wattle.id.au>
1314 * procedure.c procedure.h: Encapsulated the static data into a single
1317 Sat Oct 14 16:56:44 2006 Ben Pfaff <blp@gnu.org>
1319 * casefile.c (casereader_read_xfer): Always initialize the case,
1320 even on an error condition.
1322 Wed Sep 27 09:37:49 WST 2006 John Darrington <john@darrington.wattle.id.au>
1324 * procedure.c (case_limit_trns_proc): Fixed buglet which rendered the
1325 entire function useless.
1327 Mon Sep 25 17:11:46 WST 2006 John Darrington <john@darrington.wattle.id.au>
1329 * casefile-private.h casefile.c casefile.h fastfile.c: Created new
1330 casereader method casereader_clone.
1332 * procedure.c pransformations.h: Introduced new type casenum_t
1334 Thu Sep 21 07:00:30 2006 Ben Pfaff <blp@gnu.org>
1336 * variable.c: (width_to_bytes) Rephrase code for clarify.
1338 Sun Jul 16 19:52:03 2006 Ben Pfaff <blp@gnu.org>
1340 * format.c: (fmt_type_from_string) New function.
1341 (fmt_to_string) Include decimals in output if the format has
1342 decimals, even if the format type does not. This way, we can
1343 accurately reproduce incorrect formats in user output.
1344 (check_common_specifier) Make the check for a bad format type an
1345 assertion, so we get bug reports if they show up. Fix message.
1346 Check for decimal places with a format type that doesn't allow
1348 (check_input_specifier) Remove check for FMT_X, which has been
1350 (check_output_specifier) Ditto.
1352 * format.def: Remove FMT_T, FMT_X, FMT_DESCEND, FMT_NEWREC.
1354 * format.h: (macro FMT_TYPE_LEN_MAX) New macro.
1355 (struct fmt_desc) Use FMT_TYPE_LEN_MAX in definition.
1356 (enum fmt_parse_flags) Removed.
1358 Mon Jul 17 18:26:21 WST 2006 John Darrington <john@darrington.wattle.id.au>
1360 * casefile.c casefile.h: Converted to an abstract base class.
1361 * casefile-private.h fastfile.c fastfile.h: New files.
1362 * automake.mk procedure.c scratch-writer.c storage-stream.c
1364 Wed Jul 12 21:02:26 2006 Ben Pfaff <blp@gnu.org>
1366 * procedure.c (internal_procedure): Create sink_case with only as
1367 many values as the compacted dictionary.
1369 Wed Jul 12 21:01:00 2006 Ben Pfaff <blp@gnu.org>
1371 Remove "debugging" code that caused plenty of false positives and
1374 * case.h (struct ccase): [DEBUGGING] Remove `this' member.
1376 * case.c: Remove all references to `this' member.
1378 Thu Jul 6 19:09:53 2006 Ben Pfaff <blp@gnu.org>
1380 Fix link error noted by Jason Stover.
1382 * storage-stream.c: Include <assert.h>.
1384 Tue Jul 4 08:47:35 2006 Ben Pfaff <blp@gnu.org>
1386 Fix bug #15766 (/KEEP subcommand on SAVE doesn't fully support
1387 ALL) and additional underlying system file issues.
1389 Thanks to John Darrington for review.
1391 First problem: var_hash points to variables not owned by the
1392 sys-file-reader, which the caller may free or modify. Use an
1393 array of sfm_vars instead, as done earlier (e.g. CVS version
1396 * sys-file-reader.c (struct sfm_reader): Remove var_hash, svars
1397 members and remove all code that references it. Add vars, var_cnt
1398 members. Remove fix_specials member, which was unused.
1399 (struct sfm_var) Remove name member, which was unused.
1400 (sfm_close_reader) Free vars member instead of var_hash.
1401 (compare_var_shortnames) Removed.
1402 (hash_var_shortname) Removed.
1403 (sfm_open_reader) Fill out vars array.
1404 (compare_var_index) Removed.
1405 (sfm_read_case) Use vars instead of var_hash.
1407 Second problem: we're confused about when we actually have very
1408 long strings, causing us to choose incorrectly between slow path
1409 and fast path in sfm_read_case.
1411 * sys-file-reader.c: (sfm_open_reader) Only mark has_vls if we
1412 have very long strings, not when we have long variable names,
1413 which is an unrelated feature.
1415 Tue Jun 27 12:06:49 2006 Ben Pfaff <blp@gnu.org>
1417 * variable.h: Move var_set and variable parsing declarations to
1418 new header, src/language/lexer/variable-parser.h. Modified lots
1419 of files to include the new header.
1421 Sun Jun 25 22:39:32 2006 Ben Pfaff <blp@gnu.org>
1423 * value-labels.c (value_to_string): When there's no value label,
1424 format the variable according to its print format, instead of
1425 always effectively using A or F format.
1427 Mon Jun 19 18:05:42 WST 2006 John Darrington <john@darrington.wattle.id.au>
1429 * casefile.c (casefile_get_random_reader): Nasty hack to get around
1432 * format.c: Removed tortological assertion.
1434 Fri Jun 9 12:20:09 2006 Ben Pfaff <blp@gnu.org>
1436 Reform string library.
1438 * file-name.c (fn_interp_vars): Change interface to take a
1439 substring as input. Updated all users.
1441 Fri Jun 9 12:11:24 2006 Ben Pfaff <blp@gnu.org>
1443 * format.c (measure_is_valid): Really return false when m >=
1446 Tue Jun 6 18:46:26 2006 Ben Pfaff <blp@gnu.org>
1448 Implement random access to casefiles, for use in GUI.
1450 * casefile.c: (struct casereader) Add `random', `file_ofs',
1451 `buffer_ofs' members.
1452 (casefile_get_random_reader) New function.
1453 (read_open_file) Break part into new function
1454 seek_and_fill_buffer().
1455 (fill_buffer) Update buffer_ofs, file_ofs.
1456 (casereader_seek) New function.
1458 Tue May 30 19:52:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1460 * settings.c: Added call to i18n{done, init}.
1462 Tue May 9 21:09:17 2006 Ben Pfaff <blp@gnu.org>
1464 * procedure.h: Add WARN_UNUSED_RESULT to procedure function
1467 Tue May 9 21:08:05 2006 Ben Pfaff <blp@gnu.org>
1469 * casefile.c: Convert many uses of `int' to `bool'.
1471 Sat May 6 22:49:43 2006 Ben Pfaff <blp@gnu.org>
1473 * transformations.c (trns_chain_destroy): Destroy chain's trns
1474 member, to fix memory leak.
1476 Sat May 6 22:48:30 2006 Ben Pfaff <blp@gnu.org>
1478 * storage-stream.c (storage_source_decapsulate): Destroy case
1479 source to fix memory leak.
1481 Sat May 6 22:46:47 2006 Ben Pfaff <blp@gnu.org>
1483 * scratch-reader.c (scratch_reader_read_case): Copy into existing
1484 case passed as argument instead of initializing the argument as a
1485 case. Fixes memory leak that showed up in
1486 tests/command/aggregate.sh with scratch files.
1488 Sat May 6 22:45:55 2006 Ben Pfaff <blp@gnu.org>
1490 * procedure.c (proc_done): Destroy default_dict, to fix memory
1493 Sat May 6 22:44:44 2006 Ben Pfaff <blp@gnu.org>
1495 Simplify procedure_with_splits().
1497 * procedure.c (struct split_aux_data): Removed case_count member.
1498 (procedure_with_splits) Don't initialize case_count.
1499 (split_procedure_case_func) Check whether prev_case is null
1500 instead of case_count.
1501 (split_procedure_end_func) Ditto.
1503 Sat May 6 22:42:23 2006 Ben Pfaff <blp@gnu.org>
1505 * case.c (case_move): Do nothing if dst and src are the same
1507 (case_try_create) Merge two similar cases.
1508 (case_copy) Unshare only if data must be actually copied.
1510 Sun May 7 10:04:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1512 * data-in.c data-out.c dictionary.c sys-file-reader.c
1513 sys-file-writer.c variable.c variable.h: Reworked very long string
1514 support for better encapsulation.
1516 Sat May 6 19:02:00 2006 Ben Pfaff <blp@gnu.org>
1518 * value-labels.c (val_labs_can_set_width): New function.
1519 (val_labs_set_width) Clear labels if increasing width to long
1521 (val_labs_destroy) Remove unneeded test for null.
1523 Sat May 6 16:14:08 2006 Ben Pfaff <blp@gnu.org>
1525 * value-labels.h: Remove unneeded dependency on variable.h.
1527 Sat May 6 15:58:36 2006 Ben Pfaff <blp@gnu.org>
1529 Get rid of `char *c' member in union value, for cleanliness.
1531 * value.h: (union value) Remove `c' member.
1533 Sat May 6 15:36:59 2006 Ben Pfaff <blp@gnu.org>
1535 Make dictionary compacting functions a little more general.
1537 * sys-file-writer.c (sfm_open_writer): Use
1538 dict_compacting_would_change().
1539 (does_dict_need_translation) Removed.
1541 Sat May 6 15:35:42 2006 Ben Pfaff <blp@gnu.org>
1543 Make dictionary compacting functions a little more general.
1545 * dictionary.c (dict_needs_compaction): Rename
1546 dict_compacting_would_shrink(). Update all callers.
1547 (dict_compacting_would_change) New function.
1549 Sat May 6 14:25:49 2006 Ben Pfaff <blp@gnu.org>
1551 * sys-file-writer.c: (does_dict_need_translation) Fix bug:
1552 inverted return value (!).
1554 Sat May 6 13:37:52 2006 Ben Pfaff <blp@gnu.org>
1556 Continue reforming procedure execution.
1558 * procedure.c: Search and replace "vfm" by "proc". Notably:
1559 (static var vfm_source) Rename proc_source. Update all
1561 (static var vfm_sink) Rename proc_sink. Update all references.
1563 Sat May 6 12:38:55 2006 Ben Pfaff <blp@gnu.org>
1565 Continue reforming procedure execution. In this phase, remove
1566 PROCESS IF, which was deprecated anyway and can be easily
1567 simulated with TEMPORARY followed by SELECT IF.
1569 * procedure.c: (open_active_file) Don't call
1570 add_process_if_trns().
1571 (discard_variables) Get rid of redundant call to
1572 proc_cancel_all_transformations().
1573 (add_process_if_trns) Removed.
1574 (process_if_trns_proc) Removed.
1575 (process_if_trns_free) Removed.
1577 Sat May 6 10:58:05 2006 Ben Pfaff <blp@gnu.org>
1579 Continue reforming procedure execution. In this phase, add
1580 `const' to the case passed to procedure()'s callback.
1582 Updated all users of procedure() as well.
1584 * procedure.c: (struct write_case_data) Add "const" to ccase
1585 parameter for case_func member.
1586 (procedure) Add "const" to ccase parameter for proc_func
1588 (multipass_case_func) Make ccase parameter const.
1589 (internal_procedure) Add "const" to ccase parameter for case_func
1591 (split_procedure_case_func) Make ccase parameter const.
1592 (multipass_split_case_func) Make ccase parameter const.
1594 Sat May 6 10:30:33 2006 Ben Pfaff <blp@gnu.org>
1596 Continue reforming procedure execution. In this phase, get rid of
1597 the output code for SPLIT FILE groups in procedure.c, which really
1598 shouldn't be doing any output. Move it into the individual
1599 procedures instead. This also adds some flexibility.
1601 Updated many users of procedure_with_splits() and
1602 multipass_procedure_with_splits() to call
1603 output_split_file_values() and to deal with increased use of
1606 * procedure.c: (struct split_aux_data) Add "const struct ccase *"
1607 parameter to begin_func member.
1608 (procedure_with_splits) Add "const struct ccase *" parameter to
1609 begin_func parameter. Make ccase parameter const in proc_func
1611 (split_procedure_case_func) Don't dump split file group. Pass
1613 (dump_splits) Moved to language/dictionary/split-file.c as
1614 output_split_file_values().
1615 (struct multipass_split_aux_data) Add "const struct ccase *"
1616 parameter to split_func member.
1617 (multipass_procedure_with_splits) Add "const struct ccase *"
1618 parameter to split_func parameter.
1619 (multipass_split_case_func) Save new SPLIT FILE case before
1621 (multipass_split_output) Pass saved SPLIT FILE case to split_func.
1623 Fri May 5 22:48:50 2006 Ben Pfaff <blp@gnu.org>
1625 Continue reforming procedure execution. Change
1626 internal_procedure() so that it calls open_active_file() and
1627 close_active_file(), which isolates most of the actual procedure
1630 * procedure.c: (struct write_case_data) Rename `proc_func' member
1631 to `case_func' and update all references.
1632 (procedure) Rewrite as one-line wrapper around
1633 internal_procedure().
1634 (struct multipass_aux_data) New.
1635 (multipass_callback) Renamed multipass_case_func(). Use struct
1636 multipass_aux_data as auxiliary data.
1637 (multipass_end_func) New function.
1638 (multipass_procedure) Rewrite as wrapper for internal_procedure()
1639 that uses multipass_case_func, multipass_end_func.
1640 (internal_procedure) Add `end_func' argument. Move optimization
1641 of trivial case in here. Move call to open_active_file() and
1642 close_active_file() in here. Now assert that vfm_source is
1644 (procedure_with_splits_callback) Rename
1645 split_procedure_case_func().
1646 (split_procedure_end_func) New function.
1647 (multipass_split_callback) Rename multipass_split_case_func.
1648 (multipass_split_end_func) New function.
1649 (discard_variables) No need to test for nonnull vfm_source.
1651 Fri May 5 21:34:02 2006 Ben Pfaff <blp@gnu.org>
1653 Continue reforming procedure execution. Get rid of unused member.
1655 * procedure.c: (struct write_case_data) Remove `cases_analyzed'
1657 (write_case) Don't increment cases_analyzed.
1659 Thu May 4 21:50:11 2006 Ben Pfaff <blp@gnu.org>
1661 Continue reforming procedure execution. In this phase, move
1662 procedure.c and procedure.h from src to src/data. Update
1663 makefiles and #includes accordingly.
1665 * procedure.c: Moved here from src/.
1667 * procedure.h: Moved here from src/.
1669 Wed May 3 22:42:12 2006 Ben Pfaff <blp@gnu.org>
1671 Continue reforming procedure execution. In this phase, get rid of
1672 many global variables, consolidating procedure execution in
1673 procedure.c. Encapsulate transformations in new "struct
1674 trns_chain". Also, change implementation of N OF CASES, FILTER,
1675 and PROCESS IF from special cases to transformations.
1677 * automake.mk: (src_data_libdata_a_SOURCES) Add transformations.c,
1680 * dictionary.c: (global variable default_dict) Move to
1683 * variable.h: (TRNS_*) Move to transformations.h.
1684 (struct transformation) Move to transformations.c.
1686 Thu May 4 13:47:06 WST 2006 John Darrington <john@darrington.wattle.id.au>
1688 * sys-file-reader.c: Fixed invalid read problems.
1690 Tue May 2 15:57:10 2006 Ben Pfaff <blp@gnu.org>
1692 * storage-stream.c: Add missing function comments.
1694 Tue May 2 15:50:21 2006 Ben Pfaff <blp@gnu.org>
1696 Continue reforming procedure execution. In this phase, add some
1697 new, needed functionality to storage-stream.
1699 * storage-stream.c: (storage_source_decapsulate) New function.
1701 Tue May 2 15:43:36 2006 Ben Pfaff <blp@gnu.org>
1703 * variable.c (width_to_bytes): Declarations must precede
1704 statements for C90 compliance.
1706 Tue May 2 10:42:05 WST 2006 John Darrington <john@darrington.wattle.id.au>
1708 * data-out.c, data-in.c, variable.c, variable.h: New functions
1709 copy_mangle and copy_demangle for reading/writing cases; emulates the
1710 way SPSS deals with strings > 255 bytes.
1712 * sys-file-reader.c sys-file-writer.c: Added support for Record 7,
1713 subtype 14 needed for strings longer than 255 bytes.
1715 * dictionary.c, format.def, value.c : Updated to use MAX_STRING
1716 instead of literal values. Also fixed some constness issues.
1718 * format.h: Constness
1720 * sfm-private.h: Renamed the case_size identifier, since I discovered
1721 that SPSS's respect for this variable is very nominal.
1723 Mon May 1 15:45:42 2006 Ben Pfaff <blp@gnu.org>
1725 Change case limit type from int to size_t.
1727 * dictionary.c: (struct dictionary) Change type of case_limit
1729 (dict_get_case_limit) Change return type.
1730 (dict_set_case_limit) Change parameter type.
1732 Wed Apr 26 20:01:19 2006 Ben Pfaff <blp@gnu.org>
1734 * variable.h: (struct variable) Rename `reinit' member as `leave'
1735 and invert sense. Fix up all references.
1737 Wed Apr 26 19:39:28 2006 Ben Pfaff <blp@gnu.org>
1739 Continue reforming procedure execution. In this phase, break
1740 procedure.c into multiple files.
1742 * automake.mk: (src_data_libdata_a_SOURCES) Add all the new files.
1744 * case-sink.c: New file.
1746 * case-sink.h: New file.
1748 * case-source.c: New file.
1750 * case-source.h: New file.
1752 * storage-stream.c: New file.
1754 * storage-stream.h: New file.
1756 Wed Apr 26 14:55:19 2006 Ben Pfaff <blp@gnu.org>
1758 * variable.h: (struct variable) Remove `init' member and all
1759 references to it from other files. It was initialized in several
1760 places, but nothing really ever used it for anything worthwhile.
1761 Thanks to Jason Stover for pointing out how confusing this
1764 Sun Apr 23 22:04:45 2006 Ben Pfaff <blp@gnu.org>
1766 Continue reforming error message support. In this phase, get rid
1767 of message "titles" and put the message text in `struct error'.
1768 Now `struct error' encapsulates a message more properly.
1770 * casefile.c: (io_error) Use err_msg() instead of err_vmsg().
1771 Format message ourselves.
1773 * data-in.c: (vdls_error) Ditto.
1775 * por-file-reader.c: (error) Ditto.
1777 * sys-file-reader.c: (corrupt_msg) Ditto.
1779 Sun Apr 16 18:49:51 2006 Ben Pfaff <blp@gnu.org>
1781 GNU standards require "file name" instead of "filename" in
1782 documentation. It's nice for our code to follow the convention
1785 * casefile.c: (struct casefile) Rename `filename' member to
1786 `file_name'. Updated all references.
1788 * file-name.c: [!unix] (struct file_identity) Rename
1789 normalized_filename member to normalized_file_name. Updated all
1792 Sun Apr 16 18:35:33 2006 Ben Pfaff <blp@gnu.org>
1794 We don't really support anything but Unix-like environments well,
1795 so we might as well de-obfuscate by writing directory and path
1796 separators explicitly.
1798 * file-name.h: (macro DIR_SEPARATOR) Removed. Changed all usages
1800 (macro PATH_SEPARATOR) Removed. Changed all usages to just ':'.
1801 (macro DIR_SEPARATOR_STRING) Removed. Changed all usages to just
1803 (macro PATH_SEPARATOR_STRING) Removed. Changed all usages to just
1806 Sun Apr 16 18:28:35 2006 Ben Pfaff <blp@gnu.org>
1808 GNU standards require "file name" instead of "filename" in
1809 documentation. It's nice for our code to follow the convention
1812 * filename.c: Rename to file-name.c.
1814 * filename.h: Rename to file-name.h. Update all inclusions.
1815 Update header guards.
1817 * automake.mk: Update file names.
1819 Sun Apr 16 16:42:47 2006 Ben Pfaff <blp@gnu.org>
1821 * filename.c: (fn_dirname) Renamed fn_dir_name(), all references
1823 (fn_basename) Removed (dead code).
1824 (fn_absolute_p) Renamed fn_is_absolute(), all references updated.
1825 (fn_special_p) Renamed fn_is_special(), all references updated.
1826 (fn_exists_p) Renamed fn_exists(), all references updated.
1828 Sun Apr 16 16:33:58 2006 Ben Pfaff <blp@gnu.org>
1830 * filename.c: (fn_tilde_expand) Rewrite for cleaner code.
1831 Also, now it only tilde-expands file names, not paths.
1832 (fn_search_path) Tilde-expand one directory at a time.
1834 Sun Apr 16 16:28:06 2006 Ben Pfaff <blp@gnu.org>
1836 * filename.c: (fn_search_path) rewrite for cleaner code. Also,
1837 get rid of non-Unixlike version of the code, which has probably
1839 (fn_prepend_dir) Removed (dead code).
1841 * filename.h: (macro DIR_SEPARATOR_STRING) New.
1842 (macro PATH_SEPARATOR_STRING) New.
1843 Sun Apr 16 16:05:28 2006 Ben Pfaff <blp@gnu.org>
1845 Continue reforming error message support. In this phase, we get
1846 rid of VM() and the other msg() support for "verbosity", replacing
1847 it by a new function verbose_msg().
1849 * filename.c: (fn_search_path) Use verbose_msg() instead of
1852 Sat Apr 15 19:53:19 2006 Ben Pfaff <blp@gnu.org>
1854 * sfm-private.h: Get rid of #defines after #error, which makes no
1857 Sat Apr 15 19:48:57 2006 Ben Pfaff <blp@gnu.org>
1859 Get rid of our own int32 type in favor of the standard int32_t
1862 * sfm-private.h: (int32 macro) Don't define this anymore. Do
1865 * sys-file-reader.c: Use int32_t instead of int32 throughout.
1867 * sys-file-writer.c: Use int32_t instead of int32 throughout.
1869 Sat Apr 15 19:36:47 2006 Ben Pfaff <blp@gnu.org>
1871 Remove ill-considered file routines that are no longer used.
1873 * filename.c: (fn_open_ext) Removed.
1874 (fn_close_ext) Removed.
1876 * filename.h: (struct file_ext) Removed.
1878 Mon Apr 3 13:22:39 2006 Ben Pfaff <blp@gnu.org>
1880 * variable.c (var_is_valid_name): Move declarations before code
1883 Tue Apr 4 15:28:40 WST 2006 John Darrington <john@darrington.wattle.id.au>
1885 * filename.ch (fn_interp_vars): Fixed small buglet.
1887 Tue Mar 28 13:47:16 WST 2006 John Darrington <john@darrington.wattle.id.au>
1889 * filename.[ch] (fn_interp_vars): Changed the signature and semantics
1890 so as to modify the string inline. Thus makeing it easier to
1891 destroy the results when no longer needed.
1893 2006-03-25 Jason Stover <jhs@math.gcsu.edu>
1895 * category.c (cat_stored_values_destroy): Fixed memory leak.
1897 Fri Mar 24 18:15:41 2006 Ben Pfaff <blp@gnu.org>
1899 Add some missing frees. Thanks to John Darrington for reporting
1902 * any-writer.c (any_writer_close): Free writer.
1904 * any-reader.c (any_reader_close): Free reader.
1906 Mon Mar 20 16:33:53 2006 Ben Pfaff <blp@gnu.org>
1908 * por-file-reader.c: (error) Mark as NO_RETURN.
1910 Sat Mar 11 15:06:07 WST 2006 John Darrington <john@darrington.wattle.id.au>
1912 * settings.c: Changed default value of scompress to true.
1914 Sat Mar 4 13:22:51 2006 Ben Pfaff <blp@gnu.org>
1916 * sfm-private.h: Include variable.h, to get SHORT_NAME_LEN.
1918 * value.h: Remove check on MAX_SHORT_STRING, which I don't think
1921 * variable.h: Move definition of SHORT_NAME_LEN, LONG_NAME_LEN
1922 here from pref.h.orig.
1924 Sat Mar 4 12:50:48 WST 2006 John Darrington <john@darrington.wattle.id.au>
1926 * sys-file-reader.c: Fixed bug reading compressed files.
1928 Thu Mar 2 08:40:33 WST 2006 John Darrington <john@darrington.wattle.id.au>
1930 * Numerous renames. See src/ChangeLog for details.
1932 * Moved files from src directory