1 2016-10-09 Janne Blomqvist <jb@gcc.gnu.org>
4 * io/io.h: TEMP_FAILURE_RETRY: Define macro if not found.
5 * io/unix.c (raw_read): Handle EINTR.
6 (raw_write): Check for return value -1.
7 (raw_seek): Handle EINTR.
10 (raw_truncate): Likewise.
11 (raw_close): Likewise.
12 (buf_flush): Call raw_seek instead of lseek.
14 (buf_write): Likewise.
15 (fd_to_stream): Handle EINTR.
16 (tempfile_open): Likewise.
17 (regular_file2): Likewise.
18 (compare_file_filename): Likewise.
19 (find_file): Likewise.
20 (inquire_sequential): Likewise.
21 (inquire_direct): Likewise.
22 (inquire_formatted): Likewise.
24 2016-10-05 Jerry DeLisle <jvdelisle@gcc.gnu.org>
27 * io/inquire.c (inquire_via_unit): NULL guard the check for
28 internal unit passed into child IO procedure.
30 2016-10-04 Jerry DeLisle <jvdelisle@gcc.gnu.org>
32 * io/inquire.c (inquire_via_unit): Add check for internal unit
33 passed into child IO procedure.
35 2016-10-01 Andre Vehreschild <vehre@gcc.gnu.org>
38 * caf/single.c (caf_internal_error): Fix not terminating va-list.
39 (_gfortran_caf_register): Free memory also when other allocs failed.
40 (_gfortran_caf_get_by_ref): Fixed style.
41 (send_by_ref): Token is now stored at the correct position preventing
42 inaccessible tokens, memory loss and possibly crashes.
44 2016-09-28 Jerry DeLisle <jvdelisle@gcc.gnu.org>
47 io/transfer.c (next_record): Flush before calculating next_record.
50 2016-09-23 Jerry DeLisle <jvdelisle@gcc.gnu.org>
53 * io/inquire.c (inquire_via_unit): Adjust error check for the
54 two possible internal unit KINDs.
55 * io/io.h: Adjust defines for is_internal_unit and
56 is_char4_unit. (gfc_unit): Add internal unit data to structure.
57 (get_internal_unit): Change declaration to set_internal_unit.
58 (free_internal_unit): Change name to stash_internal_unit_number.
59 (get_unique_unit_number): Adjust parameter argument.
60 Define IOPARM_DT_HAS_UDTIO. (gfc_saved_unit): New structure.
61 * io/list_read.c (next_char_internal): Use is_char4_unit.
62 * io/open.c (st_open): Adjust call to get_unique_unit_number.
63 * io/transfer.c (write_block): Use is_char4_unit.
64 (data_transfer_init): Update check for unit numbers.
65 (st_read_done): Free the various allocated memories used for the
66 internal units and stash the negative unit number and pointer to unit
67 structure to allow reuse. (st_write_done): Likewise stash the freed
69 * io/unit.c: Create a fixed size buffer of 16 gfc_saved_unit's to use
70 as a stack to save newunit unit numbers and unit structure for reuse.
71 (get_external_unit): Change name to get_gfc_unit to better
72 reflect what it does. (find_unit): Change call to get_gfc_unit.
73 (find_or_create_unit): Likewise. (get_internal_unit): Change
74 name to set_internal_unit. Move internal unit from the dtp
75 structure to the gfc_unit structure so that it can be passed to
76 child I/O statements through the UNIT.
77 (free_internal_unit): Change name to stash_internal_unit_number.
78 Push the common.unit number onto the newunit stack, saving it
79 for possible reuse later. (get_unit): Set the internal unit
80 KIND. Use get_unique_unit_number to get a negative unit number
81 for the internal unit. Use get_gfc_unit to get the unit structure
82 and use set_internal_unit to initialize it.
83 (init_units): Initialize the newunit stack.
84 (get_unique_unit_number): Check the stack for an available unit
85 number and use it. If none there get the next most negative
86 number. (close_units): Free any unit structures pointed to from the save
89 2016-09-21 Janne Blomqvist <jb@gcc.gnu.org>
91 * intrinsics/random.c (getosrandom): Use rand_s() on
92 MinGW-w64. Fix bounds overflow in fallback code.
94 2016-09-19 Andre Vehreschild <vehre@gcc.gnu.org>
96 * caf/libcaf.h: Add caf_reference_type.
97 * caf/mpi.c: Adapted signature of caf_register().
98 * caf/single.c (struct caf_single_token): Added to keep the pointer
99 to the memory registered and array descriptor.
100 (caf_internal_error): Added convenience interface.
101 (_gfortran_caf_register): Adapted to work with caf_single_token and
102 return memory in the array descriptor.
103 (_gfortran_caf_deregister): Same.
104 (assign_char1_from_char4): Fixed style.
105 (convert_type): Fixed incorrect conversion.
106 (_gfortran_caf_get): Adapted to work with caf_single_token.
107 (_gfortran_caf_send): Same.
108 (_gfortran_caf_sendget): Same.
109 (copy_data): Added to stop repeating it in all _by_ref functions.
110 (get_for_ref): Recursive getting of coarray data using a chain of
112 (_gfortran_caf_get_by_ref): Driver for computing the memory needed for
113 the get and checking properties of the operation.
114 (send_by_ref): Same as get_for_ref but for sending data.
115 (_gfortran_caf_send_by_ref): Same like caf_get_by_ref but for sending.
116 (_gfortran_caf_sendget_by_ref): Uses get_by_ref and send_by_ref to
117 implement sendget for reference chains.
118 (_gfortran_caf_atomic_define): Adapted to work with caf_single_token.
119 (_gfortran_caf_atomic_ref): Likewise.
120 (_gfortran_caf_atomic_cas): Likewise.
121 (_gfortran_caf_atomic_op): Likewise.
122 (_gfortran_caf_event_post): Likewise.
123 (_gfortran_caf_event_wait): Likewise.
124 (_gfortran_caf_event_query): Likewise.
125 (_gfortran_caf_lock): Likewise.
126 (_gfortran_caf_unlock): Likewise.
128 2016-09-09 Steven G. Kargl <kargl@gcc.gnu.org>
131 * ieee/ieee_arithmetic.F90 (IEEE_VALUE_4,IEEE_VALUE_8,IEEE_VALULE_10,
132 IEEE_VALUE_16): Use correct keyword.
134 2016-09-06 Jerry DeLisle <jvdelisle@gcc.gnu.org>
137 * io/write_float.def (build_float_string): Recognize when the
138 result will not fit in the user provided, star fill, and exit
141 2016-08-31 Jerry DeLisle <jvdelisle@gcc.gnu.org>
144 * io/write.c (kind_from_size): New function to calculate required buffer
145 size based on kind type. (select_buffer, select_string): Use new
146 function. (write_float_0, write_real, write_real_g0, write_complex):
147 Adjust calls to pass parameters needed by new function.
149 2016-08-31 Jerry DeLisle <jvdelisle@gcc.gnu.org>
150 Paul Thomas <pault@gcc.gnu.org>
153 * gfortran.map : Flag _st_set_nml_dtio_var and
154 _gfortran_transfer_derived.
155 * io/format.c (format_lex): Detect DTIO formatting.
156 (parse_format_list): Parse the DTIO format.
157 (next_format): Include FMT_DT.
158 * io/format.h : Likewise. Add structure 'udf' to structure
159 'fnode' to carry the IOTYPE string and the 'vlist'.
160 * io/io.h : Add prototypes for the two types of DTIO subroutine
161 and a typedef for gfc_class. Also, add to 'namelist_type'
162 fields for the pointer to the DTIO procedure and the vtable.
163 Add fields to struct st_parameter_dt for pointers to the two
164 types of DTIO subroutine. Add to gfc_unit DTIO specific fields.
165 (internal_proto): Add prototype for 'read_user_defined' and
166 'write_user_defined'.
167 * io/list_read.c (check_buffers): Use the 'current_unit' field.
168 (unget_char): Likewise.
169 (eat_spaces): Likewise.
170 (list_formatted_read_scalar): For case BT_CLASS, call the DTIO
172 (nml_get_obj_data): Likewise when DTIO procedure is present,.
173 * io/transfer.c : Export prototypes for 'transfer_derived' and
174 'transfer_derived_write'.
175 (unformatted_read): For case BT_CLASS, call the DTIO procedure.
176 (unformatted_write): Likewise.
177 (formatted_transfer_scalar_read): Likewise.
178 (formatted_transfer_scalar_write: Likewise.
179 (transfer_derived): New function.
180 (data_transfer_init): Set last_char if no child_dtio.
181 (finalize_transfer): Return if child_dtio set.
182 (st_write_done): Add condition for child_dtio not set.
183 Add extra arguments for st_set_nml_var prototype.
184 (set_nml_var): New function that contains the contents of the
185 old version of st_set_nml_var. Also sets the 'dtio_sub' and
186 'vtable' fields of the 'nml' structure.
187 (st_set_nml_var): Now just calls set_nml_var with 'dtio_sub'
189 (st_set_nml_dtio_var): New function that calls set_nml_var.
190 * io/unit.c (get_external_unit): If the found unit child_dtio
191 is non zero, don't do any mutex locking/unlocking. Just
193 * io/unix.c (tempfile_open): Revert to C style comment.
194 * io/write.c (list_formatted_write_scalar): Do the DTIO call.
195 (nml_write_obj): Add BT_CLASS and do the DTIO call.
197 2016-08-29 Nathan Sidwell <nathan@acm.org>
199 * configure.ac (nvptx-*): Hardwire newlib.
200 * configure: Rebuilt.
202 2016-08-19 Janne Blomqvist <jb@gcc.gnu.org>
204 * intrinsics/random.c (xor_keys): New array with "secret" keys.
205 (scramble_seed): XOR given seed with xor_keys array rather than
207 (unscramble_seed): Remove function.
208 (random_seed_i4): Use new scramble_seed.
209 (random_seed_i8): Likewise.
211 2016-08-19 Janne Blomqvist <jb@gcc.gnu.org>
213 * intrinsics/random.c (master_init): New variable.
214 (init_rand_state): Move below getosrandom (), maybe initialize
216 (random_seed_i4): If called with no arguments, set master_init to
217 false, and reinitialize. If called with PUT=, set master_init to
219 (random_seed_i8): Likewise.
221 2016-08-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
223 * intrinsics/random.c: Include <stdlib.h>.
225 2016-08-11 Janne Blomqvist <jb@gcc.gnu.org>
227 * intrinsics/random.c: Replace KISS with xorshift1024* using
229 * runtime/main.c (init): Don't call random_seed_i4.
231 2016-07-22 Andre Vehreschild <vehre@gcc.gnu.org>
233 * caf/libcaf.h: Add parameter stat to caf_get() and
234 caf_send()'s function prototypes.
235 * caf/single.c (_gfortran_caf_get): Implement reporting
236 error using stat instead of abort().
237 (_gfortran_caf_send): Same.
238 (_gfortran_caf_sendget): Use NULL for stat when calling
241 2016-06-23 Jerry DeLisle <jvdelisle@gcc.gnu.org>
244 * io/write.c: Cleaned up whitespace.
245 (write_d, write_e, write_f, write_es, write_en): Use new helper
246 function write_float_0.
247 (write_float_0): New helper function.
248 (get_precision, select_buffer, select_string, write_float_string): New
249 helper functions used in remaining float writing functions.
250 Helper function write_float_string now contains code for writing
251 to kind=4 character internal units.
252 (write_real): Modified to establish working buffers at this level
253 and to use new helper functions.
254 (write_real_g0): Likewise modified.
255 (write_complex): Likewise modified. Gets both float strings before
256 output so that final lengths can be determined which allows right
257 justifying the complex number with no intervening spaces.
258 * io/write_float.def (build_float_string): Renamed from previosly
259 output_float, modified to use buffers passed in from higher functions,
260 builds a null terminated string of the floating point value. Character
261 kind=4 code eliminated.
262 (write_infnan): Likewise modified to use incoming buffers and eliminate
264 (OUTPUT_FLOAT_FMT_G): Deleted, functionality moved into FORMAT_FLOAT.
265 (FORMAT_FLOAT): Renamed macro from WRITE_FLOAT. Use build_float_string.
266 (get_float_string): Renamed from write_float, uses FORMAT_FLOAT macro.
267 Buffer allocation removed, now at higher level.
269 2016-05-23 Jerry DeLisle <jvdelisle@gcc.gnu.org>
272 * io/list_read (eat_spaces): Eat '\r' as part of spaces.
274 2016-04-19 Jerry DeLisle <jvdelisle@gcc.gnu.org>
277 * io/list_read (check_buffers): Add '\r' to check for end of line.
279 2016-03-30 Jerry DeLisle <jvdelisle@gcc.gnu.org>
280 Dominique d'Humieres <dominiq@lps.ens.fr>
283 * io/write_float.def: Fix PF format for negative values of the scale
286 2016-03-28 Alessandro Fanfarillo <fanfarillo.gcc@gmail.com>
288 * caf/libcaf.h: caf_stop_numeric and caf_stop_str prototype.
289 * caf/single.c: _gfortran_caf_stop_numeric and
290 _gfortran_caf_stop_str implementation.
292 2016-02-23 Jerry DeLisle <jvdelisle@gcc.gnu.org>
295 * io/list_read.c (read_real): If digit is missing from exponent issue
296 an error. (parse_real): Likewise and adjusted error message to clarify
297 it is part of a complex number.
298 (nml_read_obj): Bump item count and add comment that this is used to
299 identify which item in a namelist read has a problem.
301 2016-02-17 Jerry DeLisle <jvdelisle@gcc.gnu.org>
304 * io/list_read.c (push_char4): Fix the pointer usage for xrealloc.
306 2016-02-15 Jerry DeLisle <jvdelisle@gcc.gnu.org>
309 * io/list_read.c: Entire file trailing spaces removed.
310 (CASE_SEPARATORS): Remove '!'.
311 (is_separator): Add namelist mode as condition with '!'.
312 (push_char): Remove un-needed memset. (push_char4): Likewise and remove
313 'new' pointer. (eat_separator): Remove un-needed use of notify_std.
314 (read_logical): If '!' bang encountered when not in namelist mode got
315 bad_logical to give an error. (read_integer): Likewise reject '!'.
316 (read_character): Remove condition testing c = '!' which is now inside
317 the is_separator macro. (parse_real): Reject '!' unless in namelist mode.
318 (read_complex): Reject '!' unless in namelist mode. (read_real): Likewise
321 2016-02-12 Jerry DeLisle <jvdelisle@gcc.gnu.org>
324 * io/list_read.c (read_character): Remove code related to DELIM_NONE.
326 2016-01-23 John David Anglin <danglin@gcc.gnu.org>
329 * runtime/backtrace.c: Include gthr.h.
330 (show_backtrace): Use __gthread_active_p() to determine whether threads
331 are active. Return if lbstate is NULL.
333 2016-01-15 Jakub Jelinek <jakub@redhat.com>
335 * intrinsics/execute_command_line.c (set_cmdstat): Use "%s", msg
336 instead of msg to avoid -Wformat-security warning.
338 2016-01-04 Jakub Jelinek <jakub@redhat.com>
340 Update copyright years.
342 Copyright (C) 2016 Free Software Foundation, Inc.
344 Copying and distribution of this file, with or without modification,
345 are permitted in any medium without royalty provided the copyright
346 notice and this notice are preserved.