libgfortran/ChangeLog

   1 2011-07-13  Janne Blomqvist  <jb@gcc.gnu.org>
   2
   3         PR libfortran/49296
   4         * io/list_read.c (read_logical): Don't error out if a valid value
   5         is followed by EOF instead of a normal separator.
   6         (read_integer): Likewise.
   7
   8 2011-07-09  Tobias Burnus  <burnus@net-b.de>
   9
  10         * runtime/error.c (sys_abort): Change argument list
  11         from "()" to "(void)".
  12
  13 2011-07-09  Tobias Burnus  <burnus@net-b.de>
  14             Daniel Carrera  <dcarrera@gmail.com>
  15
  16         * caf/mpi.c (runtime_error): New function.
  17         (_gfortran_caf_register): Use it.
  18         (_gfortran_caf_sync_all): Use it, add STAT_STOPPED_IMAGE
  19         as possible status value.
  20         (_gfortran_caf_sync_images): Ditto.
  21
  22 2011-07-07  Tobias Burnus  <burnus@net-b.de>
  23
  24         * libcaf.h (__attribute__, unlikely, likely): New macros.
  25         (caf_register_t): Update comment.
  26         (_gfortran_caf_register): Add stat, errmsg, errmsg_len arguments.
  27         * single.c (_gfortran_caf_register): Ditto; add error diagnostics.
  28         * mpi.c (_gfortran_caf_register): Ditto.
  29         (caf_is_finalized): New global variable.
  30         (_gfortran_caf_finalize): Use it.
  31
  32 2011-07-05  Thomas Koenig  <tkoenig@gcc.gnu.org>
  33
  34         * runtime/memory.c (internal_malloc_size):  If size is zero,
  35         allocate a single byte.
  36         * m4/pack.m4 (pack_'rtype_code`):  Don't check for zero size
  37         for the argument of internal_malloc_size.
  38         * m4/spread.m4 (spread_'rtype_code`):  Likewise.
  39         * m4/eoshift1.m4 (eoshift1):  Don't allocate twice.  Don't check
  40         for zero size for the argument of internal_malloc_size.
  41         * m4/eoshift3.m4: Don't check for zero size for the argument of
  42         internal_malloc_size.
  43         * intrinsics/pack_generic.c (pack_internal):  Likewise.
  44         (pack_s_internal):  Likewise.
  45         * intrinsics/cshift0.c (cshift0):  Likewise.
  46         * intrinsics/spread_generic.c (spread_internal):  Likewise.
  47         * intrinsics/eoshift0.c (eoshift0):  Likewise.
  48         * intrinsics/eoshift2.c (eoshift2):  Likewise.
  49         * generated/eoshift1_16.c: Regenerated.
  50         * generated/eoshift1_4.c: Regenerated.
  51         * generated/eoshift1_8.c: Regenerated.
  52         * generated/eoshift3_16.c: Regenerated.
  53         * generated/eoshift3_4.c: Regenerated.
  54         * generated/eoshift3_8.c: Regenerated.
  55         * generated/pack_c10.c: Regenerated.
  56         * generated/pack_c16.c: Regenerated.
  57         * generated/pack_c4.c: Regenerated.
  58         * generated/pack_c8.c: Regenerated.
  59         * generated/pack_i16.c: Regenerated.
  60         * generated/pack_i1.c: Regenerated.
  61         * generated/pack_i2.c: Regenerated.
  62         * generated/pack_i4.c: Regenerated.
  63         * generated/pack_i8.c: Regenerated.
  64         * generated/pack_r10.c: Regenerated.
  65         * generated/pack_r16.c: Regenerated.
  66         * generated/pack_r4.c: Regenerated.
  67         * generated/pack_r8.c: Regenerated.
  68         * generated/spread_c10.c: Regenerated.
  69         * generated/spread_c16.c: Regenerated.
  70         * generated/spread_c4.c: Regenerated.
  71         * generated/spread_c8.c: Regenerated.
  72         * generated/spread_i16.c: Regenerated.
  73         * generated/spread_i1.c: Regenerated.
  74         * generated/spread_i2.c: Regenerated.
  75         * generated/spread_i4.c: Regenerated.
  76         * generated/spread_i8.c: Regenerated.
  77         * generated/spread_r10.c: Regenerated.
  78         * generated/spread_r16.c: Regenerated.
  79         * generated/spread_r4.c: Regenerated.
  80         * generated/spread_r8.c: Regenerated.
  81
  82 2011-06-28  Thomas Koenig  <tkoenig@gcc.gnu.org>
  83
  84         PR fortran/49479
  85         * m4/reshape.m4: If source allocation is smaller than one, set it
  86         to one.
  87         * intrinsics/reshape_generic.c:  Likewise.
  88         * generated/reshape_r16.c: Regenerated.
  89         * generated/reshape_c4.c: Regenerated.
  90         * generated/reshape_c16.c: Regenerated.
  91         * generated/reshape_c8.c: Regenerated.
  92         * generated/reshape_r4.c: Regenerated.
  93         * generated/reshape_i4.c: Regenerated.
  94         * generated/reshape_r10.c: Regenerated.
  95         * generated/reshape_r8.c: Regenerated.
  96         * generated/reshape_c10.c: Regenerated.
  97         * generated/reshape_i8.c: Regenerated.
  98         * generated/reshape_i16.c: Regenerated.
  99
 100 2011-06-18  Janne Blomqvist  <jb@gcc.gnu.org>
 101
 102         PR libfortran/49296
 103         * io/list_read.c (read_character): Accept EOF as a separator when
 104         reading string.
 105
 106 2011-06-17  Daniel Carrera  <dcarrera@gmail.com>
 107
 108         * caf/single.c (_gfortran_caf_register): Store the address
 109         of all static coarrays in a linked list.
 110         (_gfortran_caf_finalize): Free memory of staic coarrays.
 111         * caf/mpi.c (_gfortran_caf_register): Store the address
 112         of all static coarrays in a linked list. Initialize MPI
 113         if necessary.
 114         (_gfortran_caf_finalize): Free memory of staic coarrays.
 115         (_gfortran_caf_init): Check if MPI is already initialized
 116         before initializing again.
 117         * caf/libcaf.h: Add a type to caf_register_t to distinguish
 118         static coarrays and add the type caf_static_t to make the
 119         linked list of static coarrays.
 120
 121 2011-06-11  Janne Blomqvist  <jb@gcc.gnu.org>
 122
 123         * io/unix.c (buf_seek): Return error if file is not seekable.
 124         (buf_tell): Call buf_seek.
 125
 126 2011-06-11  Janne Blomqvist  <jb@gcc.gnu.org>
 127
 128         * io/unix.c (fd_to_stream): Figure out if a fd is seekable by
 129         trying lseek().
 130
 131 2011-06-10  Daniel Carrera  <dcarrera@gmail.com>
 132
 133         * caf/mpi.c (_gfortran_caf_sync_all,
 134         _gfortran_caf_sync_images): Functions have void return type
 135         and move status into parameter list.
 136         * caf/single.c (_gfortran_caf_sync_all,
 137         _gfortran_caf_sync_images): Functions have void return type
 138         and move status into parameter list.
 139         * caf/libcaf.h (_gfortran_caf_sync_all,
 140         _gfortran_caf_sync_images): Functions have void return type
 141         and move status into parameter list.
 142
 143 2011-06-03  Richard Henderson  <rth@redhat.com>
 144             Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 145
 146         * config/fpu-387.h [__sun__ && __svr4__] (sigill_hdlr): Correct
 147         insn, insn size.
 148         (has_sse) [!__x86_64__ && __sun__ && __svr4__]: Use movaps.
 149
 150 2011-05-29  Janne Blomqvist  <jb@gcc.gnu.org>
 151
 152         PR libfortran/48931
 153         * libgfortran.h (find_addr2line): New prototype.
 154         * runtime/backtrace.c (show_backtrace): Use async-signal-safe
 155         execve and stored path of addr2line.
 156         * runtime/compile_options.c (maybe_find_addr2line): New function.
 157         (set_options): Call maybe_find_addr2line if backtracing is enabled.
 158         * runtime/main.c (find_addr2line): New function.
 159         (init): Call find_addr2line if backtracing is enabled.
 160         (cleanup): Free addr2line_path.
 161
 162 2011-05-29  Janne Blomqvist  <jb@gcc.gnu.org>
 163
 164         PR libfortran/49214
 165         * runtime/backtrace.c (fd_gets): Return NULL if nothing was read.
 166
 167 2011-05-29  Janne Blomqvist  <jb@gcc.gnu.org>
 168
 169         PR libfortran/19155
 170         * io/read.c (convert_real): Check for invalid input by comparing
 171         endptr instead of EINVAL.
 172         (read_f): Fixup floating point input without significand.
 173
 174 2011-05-22  Tobias Burnus  <burnus@net-b.de>
 175
 176         * runtime/stop.c (stop_string,error_stop_string): Minor cleanup.
 177
 178 2011-05-22  Janne Blomqvist  <jb@gcc.gnu.org>
 179
 180         PR libfortran/48931
 181         * configure.ac: Check for backtrace_symbols_fd instead of
 182         backtrace_symbols, check for readlink.
 183         * config.h.in: Regenerated.
 184         * configure: Regenerated.
 185         * runtime/backtrace.c (local_strcasestr): Remove.
 186         (bt_header): New function.
 187         (dump_glibc_backtrace): Remove.
 188         (fd_gets): New function.
 189         (show_backtrace): Rework to use backtrace_symbols_fd and pipes
 190         avoiding functions that are not async-signal-safe, reformat
 191         output.
 192         * runtime/main.c (store_exe_path): Try to check /proc/self/exe
 193         first.
 194
 195 2011-05-20  Janne Blomqvist  <jb@gcc.gnu.org>
 196             Uros Bizjak  <ubizjak@gmail.com>
 197
 198         PR libfortran/48977
 199         * configure.host: Swap order of glibc and x86 tests.
 200
 201 2011-05-20  Janne Blomqvist  <jb@gcc.gnu.org>
 202
 203         * config/fpu-387.h (set_fpu): Use renamed inexact macro.
 204         * config/fpu-aix.h (set_fpu): Clarify error messages, use renamed
 205         inexact macro, set TRP_INEXACT for inexact exception instead of
 206         TRP_UNDERFLOW.
 207         * config/fpu-generic.h (set_fpu): Clarify error messages, use
 208         renamed inexact macro.
 209         * config/fpu-glibc.h (set_fpu): Likewise.
 210         * config/fpu-sysv.h (set_fpu): Likewise.
 211
 212 2011-05-14  Tobias Burnus  <burnus@net-b.de>
 213
 214         * runtime/stop.c (error_stop_string, error_stop_numeric):
 215         Do not backtrace for ERROR STOP.
 216
 217 2011-05-14  Janne Blomqvist  <jb@gcc.gnu.org>
 218
 219         PR libfortran/48915
 220         * intrinsics/abort.c (abort): Call sys_abort().
 221         * io/unix.c (st_vprintf): Call sys_abort().
 222         * libgfortran.h (options_t): Remove dump_core member.
 223         (handler): Rename to backtrace_handler.
 224         (compile_options_t): Remove dump_core member.
 225         (sys_exit): Remove.
 226         (sys_abort): New function.
 227         * runtime/backtrace.c (show_backtrace): Print a message if
 228         backtrace is not available on target, update filter symbol name
 229         due to backtrace_handler renaming.
 230         * runtime/compile_options.c (backtrace_handler): New
 231         implementation based on example in Glibc manual.
 232         (set_options): Remove dump_core handling, make signal handler
 233         handle more signals.
 234         (init_compile_options): Remove dump_core.
 235         * runtime/environ.c (variable_table[]): Remove
 236         GFORTRAN_ERROR_DUMPCORE element.
 237         (show_variables): Update name, call exit().
 238         * runtime/error.c (sys_exit): Remove.
 239         (sys_abort): New function.
 240         (recursion_check): Call sys_abort().
 241         (os_error): Call exit().
 242         (runtime_error): Likewise.
 243         (runtime_error_at): Likewise.
 244         (internal_error): Likewise.
 245         (generate_error): Likewise.
 246         (notify_std): Likewise.
 247         * runtime/stop.c (stop_numeric): Call exit().
 248         (stop_numeric_f08): Likewise.
 249         (stop_string): Likewise.
 250         (error_stop_string): Call sys_abort().
 251         (error_stop_numeric): Print backtrace, exit with provided code.
 252
 253 2011-05-14  Janne Blomqvist  <jb@gcc.gnu.org>
 254
 255         * io/unix.c (st_vprintf,st_printf): Move to runtime/error.c.
 256         * libgfortran.h (struct options_t): Remove use_stderr field.
 257         (st_vprintf,st_printf): Move prototypes.
 258         (estr_write): New prototype.
 259         * runtime/error.c (sys_exit): Use estr_write instead of st_printf.
 260         (estr_write): New function.
 261         (st_vprintf): Move from io/unix.c, use stack allocated buffer,
 262         always output to stderr.
 263         (st_printf): Move from io/unix.c.
 264         (show_locus): Use a local variable instead of static.
 265         (os_error): Use estr_write instead of st_printf.
 266         (runtime_error): Likewise.
 267         (runtime_error_at): Likewise.
 268         (runtime_warning_at): Likewise.
 269         (internal_error): Likewise.
 270         (generate_error): Likewise.
 271         (generate_warning): Likewise.
 272         (notify_std): Likewise.
 273         * runtime/pause.c (do_pause): Likewise.
 274         (pause_string): Likewise.
 275         * runtime/stop.c (stop_string): Likewise.
 276         (error_stop_string): Likewise.
 277         * config/fpu_aix.h (set_fpu): Likewise.
 278         * config/fpu_generic.h (set_fpu): Likewise.
 279         * config/fpu_glibc.h (set_fpu): Likewise.
 280         * config/fpu-sysv.h (set_fpu): Likewise.
 281         * runtime/backtrace.c (dump_glibc_backtrace): Likewise.
 282         (show_backtrace): Likewise.
 283         * runtime/environ.c (print_spaces): Likewise.
 284         (show_string): Likewise.
 285         (show_variables): Likewise.
 286         (variable_table[]): Remove GFORTRAN_USE_STDERR entry.
 287
 288 2011-05-14  Tobias Burnus  <burnus@net-b.de>
 289
 290         PR fortran/48961
 291         * intrinsics/execute_command_line.c (set_cmdstat): Don't abort if
 292         synchronously executing with WAIT=.false.
 293         (execute_command_line): Fix setting of cmdstat and exitstat.
 294
 295 2011-05-06  Tobias Burnus  <burnus@net-b.de>
 296
 297         PR fortran/18918
 298         * caf/libcaf.h: Cleanup headers.
 299         (_gfortran_caf_critical, _gfortran_caf_end_critical): Make stub.
 300         (caf_register_t): New enum.
 301         (_gfortran_caf_register, _gfortran_caf_deregister): New prototype.
 302         * caf/single.c (_gfortran_caf_critical,
 303         _gfortran_caf_end_critical): Remove.
 304         (_gfortran_caf_register, _gfortran_caf_deregister): New functions.
 305         * caf/mpi.c (_gfortran_caf_critical,
 306         _gfortran_caf_end_critical): Remove.
 307         (_gfortran_caf_register, _gfortran_caf_deregister): New functions.
 308         (caf_world_window): Remove global variable.
 309         (_gfortran_caf_init): Fix off-by-one error of this_image.
 310
 311 2011-05-04  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 312
 313         PR libgfortran/48787
 314         * io/write_float.def (output_float): Adjust up and down rounding for
 315         cases where 'd' = 0. Gather common code to one location.
 316
 317 2011-05-01  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 318
 319         PR libgfortran/48787
 320         * io/write_float.def (output_float): Gather up integer declarations and
 321         add new 'p' for scale factor. Use 'p' in place of the 'dtp' reference
 322         everywhere. For ROUND_UP scan the digit string and only perform
 323         rounding if something other than '0' is found.
 324
 325 2011-04-29  Janne Blomqvist  <jb@gcc.gnu.org>
 326
 327         * io/unix.c (min): New macro.
 328         (unpack_filename): Return errno number for errors.
 329         (regular_file): Use appropriately sized buffer for path.
 330         (compare_file_filename): Likewise.
 331         (find_file): Likewise.
 332         (delete_file): Likewise.
 333         (file_exists): Likewise.
 334         (file_size): Likewise.
 335         (inquire_sequential): Likewise.
 336         (inquire_direct): Likewise.
 337         (inquire_formatted): Likewise.
 338         (inquire_access): Likewise.
 339
 340 2011-04-29  Janne Blomqvist  <jb@gcc.gnu.org>
 341
 342         PR libfortran/48488
 343         * io/write.c (write_real, write_real_g0): Update comments.
 344
 345 2011-04-29  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 346             Janne Blomqvist  <jb@gcc.gnu.org>
 347
 348         PR libgfortran/48488
 349         PR libgfortran/48602
 350         PR libgfortran/48615
 351         PR libgfortran/48684
 352         PR libgfortran/48787
 353         * io/write.c (write_d, write_e, write_f, write_en,
 354         write_es): Add precision compemsation parameter to call.
 355         (set_fnode_default): Adjust default widths to assure
 356         round trip on write and read. (write_real): Adjust call to write_float.
 357         (write_real_g0): Calculate compensation for extra precision and adjust
 358         call to write_float.
 359         * io/write_float.def (output_float_FMT_G_): Use volatile rather than
 360         asm volatile to avoid optimization issue. Correctly calculate the
 361         number of blanks (nb) to be appended and simplify calculation logic.
 362         (write_float): Increase MIN_FIELD_WIDTH by one to accomodate the new
 363         default widths. Eliminate the code that attempted to reduce the
 364         the precision used in later sprintf functions.  Add call parameter to
 365         compensate for extra precision.
 366
 367 2011-04-20  Jim Meyering  <meyering@redhat.com>
 368
 369         * intrinsics/move_alloc.c (move_alloc): Remove useless
 370         if-before-free test.
 371         * io/fbuf.c (fbuf_destroy): Likewise.
 372         * io/format.c (save_parsed_format): Likewise.
 373         * io/open.c (already_open): Likewise.
 374         * io/unit.c (free_internal_unit, close_unit_1): Likewise.
 375         * io/unix.c (mem_close): Likewise.
 376
 377 2011-04-19  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 378
 379         PR libgfortran/48602
 380         * io/write_float.def (output_float_FMT_G): Fix reversal in conditional.
 381         Use asm volatile to mark temp variable, avoiding optimization errors.
 382
 383 2011-04-17  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 384
 385         PR libgfortran/48602
 386         * io/write_float.def (output_float_FMT_G): Use current rounding mode
 387         to set the rounding parameters. (output_float): Skip rounding
 388         if value is zero.
 389
 390 2011-04-16  Janne Blomqvist  <jb@gcc.gnu.org>
 391
 392         * intrinsics/date_and_time.c (date_and_time): Remove sprintf CPP
 393         branch.
 394         * io/format.c (format_error): Use snprintf instead of sprintf.
 395         * io/list_read.c: Move snprintf fallback macro to libgfortran.h.
 396         (convert_integer): Use snprintf instead of sprintf.
 397         (parse_repeat): Likewise.
 398         (read_logical): Likewise.
 399         (read_integer): Likewise.
 400         (read_character): Likewise.
 401         (parse_real): Likewise.
 402         (read_complex): Likewise.
 403         (read_real): Likewise.
 404         (check_type): Likewise.
 405         (nml_parse_qualifier): Add string length argument, use snprintf
 406         instead of sprintf.
 407         (nml_get_obj_data): Use snprintf instead of sprintf.
 408         * io/open.c (new_unit): Remove sprintf CPP branch, use snprintf
 409         instead of sprintf.
 410         * io/transfer.c (require_type): Use snprintf instead of sprintf.
 411         * io/unix.c (tempfile): Likewise.
 412         * io/write.c (nml_write_obj): Likewise.
 413         * io/write_float.def (output_float): Remove sprintf CPP branch,
 414         use snprintf instead of sprintf.
 415         * libgfortran.h: Add fallback snprintf macro from io/list_read.c.
 416         * runtime/backtrace.c (show_backtrace): Remove sprintf CPP branch.
 417         * runtime/main.c (store_exe_path): Use snprintf instead of
 418         sprintf.
 419
 420 2011-04-15  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 421
 422         PR libgfortran/48589
 423         * io/write_float.def (write_infnan): Set width properly for G0.
 424
 425 2011-04-15  Tobias Burnus  <burnus@net-b.de>
 426
 427         PR fortran/18918
 428         * Makefile.am: Build and install caf/single.c as
 429         libcaf_single.a.
 430         * Makefile.in: Regenerate.
 431
 432 2011-04-15  Janne Blomqvist  <jb@gcc.gnu.org>
 433
 434         PR libfortran/47571
 435         * configure: Regenerated.
 436         * config.h.in: Regenerated.
 437         * acinclude.m4: Add alpha*-dec-osf* to gthread blacklist.
 438         * configure.ac: Use separate symbol for clock_gettime in librt.
 439         * intrinsics/system_clock.c: Use weakrefs only when needed and
 440         supported.
 441
 442 2011-04-12  Janne Blomqvist  <jb@gcc.gnu.org>
 443
 444         * configure.ac: Use AC_TYPE_* to make sure we have (u)intptr_t,
 445         check for ptrdiff_t.
 446         * libgfortran.h: Remove (u)intptr_t definitions, use ptrdiff_t for
 447         index_type, change cshift0* prototypes.
 448         * config.h.in: Regenerated.
 449         * configure: Regenerated.
 450         * intrinsics/cshift0.c (cshift0): Use ptrdiff_t instead of ssize_t.
 451         * io/io.h (array_loop_spec): Use index_type instead of ssize_t.
 452         * io/list_read.c (nml_parse_qualifier): Likewise.
 453         * io/write.c (nml_write_obj): Likewise.
 454         * m4/cshift0.c (cshift0_'rtype_code`): Likewise.
 455         * generated/cshift0_*.c: Regenerated.
 456
 457 2011-04-12  Janne Blomqvist  <jb@gcc.gnu.org>
 458
 459         * Makefile.am: Remove _GNU_SOURCE from AM_CPPFLAGS.
 460         * Makefile.in: Regenerated.
 461
 462 2011-04-06  Tobias Burnus  <burnus@net-b.de>
 463
 464         PR fortran/18918
 465         * caf/mpi.c (_gfortran_caf_init, _gfortran_caf_finalize):
 466         Add global variable caf_mpi_initialized and use it for when
 467         finalizing.
 468
 469 2011-04-04  Tobias Burnus  <burnus@net-b.de>
 470
 471         * unix.c: Adapt stat DEFINEs since MinGW64 supports LFS.
 472         (fallback_access, open_internal4, compare_file_filename,
 473         find_file, file_size, inquire_sequential, inquire_direct,
 474         inquire_formatted): Use "struct stat" instead of gfstat_t.
 475
 476 2011-03-27  Tobias Burnus  <burnus@net-b.de>
 477
 478         PR fortran/18918
 479         * caf/libcaf.h: New - coarray communication library.
 480         * caf/mpi.c: New.
 481         * caf/single.c: New.
 482
 483 2011-03-23  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 484
 485         PR libgfortran/48030
 486         * io/read.c (read_x): Re-implement using fbuf_getc.
 487
 488 2011-03-21  Rainer Orth  <ro@CeBiTec.Uni-Bielefeld.DE>
 489
 490         PR bootstrap/48135
 491         * configure.ac (gfortran_use_symver): Handle --disable-symvers.
 492         * configure: Regenerate.
 493
 494 2011-03-19  John David Anglin  <dave.anglin@nrc-cnrc.gc.ca>
 495
 496         PR fortran/35667
 497         * io/io.h (convert_infnan): Declare.
 498         * io/read.c (convert_infnan): New.
 499         (read_f): Use convert_infnan to convert INFs and NANs.
 500         * list_read.c (parse_real, read_real): Likewise.
 501
 502 2011-03-19  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>
 503
 504         PR libfortran/47439
 505         * io/unix.c (tempfile): Work around poor mktemp() implementations.
 506
 507 2011-03-16  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>
 508
 509         PR libfortran/47883
 510         * acinclude.m4: Don't use AC_TRY_ macros.
 511         * configure.ac: Likewise.
 512         * configure: Regenerate.
 513
 514 2011-03-12  Thomas Koenig  <tkoenig@gcc.gnu.org>
 515
 516         PR libfortran/48066
 517         * m4/ifunction.m4:  If return array is empty, return.
 518         * m4/ifunction_logical.m4:  Likewise.
 519         * generated/all_l16.c: Regenerated.
 520         * generated/all_l1.c: Regenerated.
 521         * generated/all_l2.c: Regenerated.
 522         * generated/all_l4.c: Regenerated.
 523         * generated/all_l8.c: Regenerated.
 524         * generated/any_l16.c: Regenerated.
 525         * generated/any_l1.c: Regenerated.
 526         * generated/any_l2.c: Regenerated.
 527         * generated/any_l4.c: Regenerated.
 528         * generated/any_l8.c: Regenerated.
 529         * generated/count_16_l.c: Regenerated.
 530         * generated/count_1_l.c: Regenerated.
 531         * generated/count_2_l.c: Regenerated.
 532         * generated/count_4_l.c: Regenerated.
 533         * generated/count_8_l.c: Regenerated.
 534         * generated/maxloc1_16_i16.c: Regenerated.
 535         * generated/maxloc1_16_i1.c: Regenerated.
 536         * generated/maxloc1_16_i2.c: Regenerated.
 537         * generated/maxloc1_16_i4.c: Regenerated.
 538         * generated/maxloc1_16_i8.c: Regenerated.
 539         * generated/maxloc1_16_r10.c: Regenerated.
 540         * generated/maxloc1_16_r16.c: Regenerated.
 541         * generated/maxloc1_16_r4.c: Regenerated.
 542         * generated/maxloc1_16_r8.c: Regenerated.
 543         * generated/maxloc1_4_i16.c: Regenerated.
 544         * generated/maxloc1_4_i1.c: Regenerated.
 545         * generated/maxloc1_4_i2.c: Regenerated.
 546         * generated/maxloc1_4_i4.c: Regenerated.
 547         * generated/maxloc1_4_i8.c: Regenerated.
 548         * generated/maxloc1_4_r10.c: Regenerated.
 549         * generated/maxloc1_4_r16.c: Regenerated.
 550         * generated/maxloc1_4_r4.c: Regenerated.
 551         * generated/maxloc1_4_r8.c: Regenerated.
 552         * generated/maxloc1_8_i16.c: Regenerated.
 553         * generated/maxloc1_8_i1.c: Regenerated.
 554         * generated/maxloc1_8_i2.c: Regenerated.
 555         * generated/maxloc1_8_i4.c: Regenerated.
 556         * generated/maxloc1_8_i8.c: Regenerated.
 557         * generated/maxloc1_8_r10.c: Regenerated.
 558         * generated/maxloc1_8_r16.c: Regenerated.
 559         * generated/maxloc1_8_r4.c: Regenerated.
 560         * generated/maxloc1_8_r8.c: Regenerated.
 561         * generated/maxval_i16.c: Regenerated.
 562         * generated/maxval_i1.c: Regenerated.
 563         * generated/maxval_i2.c: Regenerated.
 564         * generated/maxval_i4.c: Regenerated.
 565         * generated/maxval_i8.c: Regenerated.
 566         * generated/maxval_r10.c: Regenerated.
 567         * generated/maxval_r16.c: Regenerated.
 568         * generated/maxval_r4.c: Regenerated.
 569         * generated/maxval_r8.c: Regenerated.
 570         * generated/minloc1_16_i16.c: Regenerated.
 571         * generated/minloc1_16_i1.c: Regenerated.
 572         * generated/minloc1_16_i2.c: Regenerated.
 573         * generated/minloc1_16_i4.c: Regenerated.
 574         * generated/minloc1_16_i8.c: Regenerated.
 575         * generated/minloc1_16_r10.c: Regenerated.
 576         * generated/minloc1_16_r16.c: Regenerated.
 577         * generated/minloc1_16_r4.c: Regenerated.
 578         * generated/minloc1_16_r8.c: Regenerated.
 579         * generated/minloc1_4_i16.c: Regenerated.
 580         * generated/minloc1_4_i1.c: Regenerated.
 581         * generated/minloc1_4_i2.c: Regenerated.
 582         * generated/minloc1_4_i4.c: Regenerated.
 583         * generated/minloc1_4_i8.c: Regenerated.
 584         * generated/minloc1_4_r10.c: Regenerated.
 585         * generated/minloc1_4_r16.c: Regenerated.
 586         * generated/minloc1_4_r4.c: Regenerated.
 587         * generated/minloc1_4_r8.c: Regenerated.
 588         * generated/minloc1_8_i16.c: Regenerated.
 589         * generated/minloc1_8_i1.c: Regenerated.
 590         * generated/minloc1_8_i2.c: Regenerated.
 591         * generated/minloc1_8_i4.c: Regenerated.
 592         * generated/minloc1_8_i8.c: Regenerated.
 593         * generated/minloc1_8_r10.c: Regenerated.
 594         * generated/minloc1_8_r16.c: Regenerated.
 595         * generated/minloc1_8_r4.c: Regenerated.
 596         * generated/minloc1_8_r8.c: Regenerated.
 597         * generated/minval_i16.c: Regenerated.
 598         * generated/minval_i1.c: Regenerated.
 599         * generated/minval_i2.c: Regenerated.
 600         * generated/minval_i4.c: Regenerated.
 601         * generated/minval_i8.c: Regenerated.
 602         * generated/minval_r10.c: Regenerated.
 603         * generated/minval_r16.c: Regenerated.
 604         * generated/minval_r4.c: Regenerated.
 605         * generated/minval_r8.c: Regenerated.
 606         * generated/product_c10.c: Regenerated.
 607         * generated/product_c16.c: Regenerated.
 608         * generated/product_c4.c: Regenerated.
 609         * generated/product_c8.c: Regenerated.
 610         * generated/product_i16.c: Regenerated.
 611         * generated/product_i1.c: Regenerated.
 612         * generated/product_i2.c: Regenerated.
 613         * generated/product_i4.c: Regenerated.
 614         * generated/product_i8.c: Regenerated.
 615         * generated/product_r10.c: Regenerated.
 616         * generated/product_r16.c: Regenerated.
 617         * generated/product_r4.c: Regenerated.
 618         * generated/product_r8.c: Regenerated.
 619         * generated/sum_c10.c: Regenerated.
 620         * generated/sum_c16.c: Regenerated.
 621         * generated/sum_c4.c: Regenerated.
 622         * generated/sum_c8.c: Regenerated.
 623         * generated/sum_i16.c: Regenerated.
 624         * generated/sum_i1.c: Regenerated.
 625         * generated/sum_i2.c: Regenerated.
 626         * generated/sum_i4.c: Regenerated.
 627         * generated/sum_i8.c: Regenerated.
 628         * generated/sum_r10.c: Regenerated.
 629         * generated/sum_r16.c: Regenerated.
 630         * generated/sum_r4.c: Regenerated.
 631         * generated/sum_r8.c: Regenerated.
 632
 633 2011-03-11  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 634
 635         PR libgfortran/48047
 636         * io/write_float.def (write_float): Change MIN_FIELD_WIDTH to 48.
 637
 638 2011-03-04  Janne Blomqvist  <jb@gcc.gnu.org>
 639
 640         PR libfortran/47802
 641         * intrinsics/ctime.c (strctime): Use builtins to check localtime_r
 642         return type.
 643
 644 2011-03-04  Janne Blomqvist  <jb@gcc.gnu.org>
 645
 646         PR libfortran/47802
 647         * intrinsics/ctime.c (strctime): Don't use return value of
 648         localtime_r.
 649
 650 2011-02-28  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 651
 652         PR libgfortran/47567
 653         * io/write_float.def (output_float): Move handling of w = 0 to after
 654         output rounding. Check for zero and set zero_flag accordingly. Set
 655         width according to zero_flag. Add better comments.
 656
 657 2011-02-27  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 658
 659         PR libgfortran/47778
 660         * io/list_read.c (namelist_read): Intialize the error string buffere.
 661         If pprev_nl was used during the previous namelist read and the rank
 662         was zero, reset the pointer to NULL for the next namelist read.
 663
 664 2011-02-26  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>
 665
 666         PR libfortran/45165
 667         * unix.c (fallback_access): Fix file descriptor leaks.
 668
 669 2011-02-25  Francois-Xavier Coudert  <fxcoudert@gcc.gnu.org>
 670
 671         * acinclude.m4 (LIBGFOR_CHECK_FPSETMASK): Set shell variable
 672         tested in configure.host.
 673         * configure: Regenerate.
 674
 675 2011-02-24  Janne Blomqvist  <jb@gcc.gnu.org>
 676
 677         PR libfortran/47802
 678         * config.h.in: Regenerated.
 679         * configure: Regenerated.
 680         * configure.ac: Remove checks for ctime and ctime_r, add check for
 681         strftime.
 682         * intrinsics/date_and_time.c (localtime_r): Move fallback
 683         implementation to time_1.h.
 684         * intrinsics/time_1.h (localtime_r): Fallback implementation.
 685         * intrinsics/ctime.c: Include time_1.h.
 686         (ctime_r): Remove fallback implementation.
 687         (strctime): New function.
 688         (fdate): Use strctime instead of ctime_r.
 689         (fdate_sub): Likewise.
 690         (ctime): Likewise.
 691         (ctime_sub): Likewise.
 692
 693 2011-02-24  Jakub Jelinek  <jakub@redhat.com>
 694
 695         PR fortran/47878
 696         * io/transfer.c (read_sf): Call fbuf_getptr only at the end,
 697         and subtract n, dtp->u.p.sf_seen_eor and seen_comma from it.
 698
 699 2011-02-24  Janne Blomqvist  <jb@gcc.gnu.org>
 700
 701         PR libfortran/47802
 702         * configure.ac: Add test for POSIX getpwuid_r.
 703         * intrinsics/getlog.c (getlog): CPP test for
 704         HAVE_POSIX_GETPWUID_R.
 705         * config.h.in: Regenerated.
 706         * configure: Regenerated.
 707
 708 2011-02-23  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 709
 710         PR libgfortran/47567
 711         * io/write_float.def (output_float): Remove special case handling of
 712         zero with width 1.
 713
 714 2011-02-23  Janne Blomqvist  <jb@gcc.gnu.org>
 715             Jerry DeLisle    <jvdelisle@gcc.gnu.org>
 716
 717         PR libfortran/47694
 718         * io/fbuf.h (fbuf_getptr): New inline function.
 719         * io/transfer.c (read_sf): Use fbuf_getptr and fbuf_getc to scan
 720         through the string instead of fbuf_read.
 721
 722 2011-02-22  Tobias Burnus  <burnus@net-b.de>
 723             Kai-Uwe Eckhardt  <kuehro@gmx.de>
 724
 725         PR libfortran/47830
 726         * intrinsics/c99_functions.c (roundl): Make C valid for
 727         HAVE_NEXTAFTERL.
 728
 729 2011-02-19  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 730
 731         PR libgfortran/47567
 732         * io/write_float.def (output_float): Adjust width for F0.d to
 733         allow space for negative signs on zero.
 734
 735 2011-02-16  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 736
 737         PR libgfortran/47667
 738         * io/list_read.c (read_logical): Check for end of line before calling
 739         eat_line. (read_integer): Likewise. (parse_real): Don't unget the
 740         separator. Check for end of line before calling eat_line.
 741         (read_complex): Allow line-end before and after parenthesis and comma.
 742         Check for end of line before calling eat_line. (read_real): Check for
 743         end of line before calling eat_line.
 744
 745 2011-02-16  Jakub Jelinek  <jakub@redhat.com>
 746
 747         PR libfortran/47757
 748         * gfortran.map (GFORTRAN_1.4): Export
 749         _gfortran_{m,s}i{all,any,parity}_i{1,2,4,8,16} and
 750         _gfortran_{cshift0,eoshift{0,2}}_16_char4.
 751
 752 2011-02-15  Tobias Burnus  <burnus@net-b.de>
 753
 754         PR fortran/47716
 755         PR fortran/47648
 756         * acinclude.m4 (LIBGFOR_CHECK_FLOAT128): Use check from
 757         libquadmath, which uses more features.
 758         * configure: Regenerate.
 759
 760 2011-02-14  Jakub Jelinek  <jakub@redhat.com>
 761
 762         PR fortran/47642
 763         * io/write_float.def (DTOAQ): Use quadmath_snprintf instead of
 764         quadmath_flt128tostr.
 765         * io/transfer128.c (tmp2): Initialize to quadmath_snprintf instead
 766         of quadmath_flt128tostr.
 767
 768 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
 769
 770         * Makefile.in: Regenerate.
 771         * aclocal.m4: Likewise.
 772         * configure: Likewise.
 773
 774 2011-02-05  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 775
 776         PR libgfortran/47567
 777         * io/write_float.def (output_float): Eliminate some redundant code.
 778         Adjust width for case of F0.X for values of zero and all other values.
 779         Expand cases where '*' is set to give cleaner results.
 780
 781 2011-02-05  Janne Blomqvist  <jb@gcc.gnu.org>
 782
 783         PR libfortran/47571
 784         * intrinsics/time_1.h (GF_CLOCK_MONOTONIC): Move to system_clock.c.
 785         (weak_gettime): Likewise.
 786         (gf_gettime): Change API, move weak_gettime() usage to
 787         system_clock.c
 788         * intrinsics/system_clock.c (GTHREAD_USE_WEAK): Define.
 789         (gf_gettime_mono): New function.
 790         (system_clock_4): Use gf_gettime_mono().
 791         (system_clock_8): Likewise.
 792         * intrinsics/date_and_time.c (date_and_time): Update gf_gettime()
 793         usage.
 794
 795 2011-02-02  Janne Blomqvist  <jb@gcc.gnu.org>
 796
 797         PR libfortran/47571
 798         * configure: Regenerated.
 799         * configure.ac: Don't add librt to LIBS.
 800         * intrinsics/time_1.h (weak_gettime): Weakref trickery for
 801         clock_gettime().
 802         (gf_gettime): Use weak_gettime() instead of clock_gettime().
 803
 804 2011-02-01  Janne Blomqvist  <jb@gcc.gnu.org>
 805
 806         * intrinsics/time_1.h: Include errno.h needed by fallbacks.
 807
 808 2011-01-31  Janne Blomqvist  <jb@gcc.gnu.org>
 809
 810         * intrinsics/time_1.h: Fix definition of GF_CLOCK_MONOTONIC macro.
 811
 812 2011-01-31  Janne Blomqvist  <jb@gcc.gnu.org>
 813
 814         * configure.ac: Check for clock_gettime().
 815         * configure: Regenerated.
 816         * config.h.in: Regenerated.
 817         * intrinsics/time_1.h (__time_1): Rename to gf_cputime, add
 818         times() fallback.
 819         (gf_gettime): New function.
 820         * intrinsics/cpu_time.c (__cpu_time_1): Update to call gf_cputime.
 821         * intrinsics/date_and_time.c (date_and_time): Use gf_gettime.
 822         * intrinsics/dtime.c (dtime_sub): Use gf_cputime.
 823         * intrinsics/etime.c (etime_sub): Use gf_cputime.
 824         * intrinsics/system_clock.c (system_clock_4): Use gf_gettime.
 825         (system_clock_8): Use gf_gettime, increase count rate to allow
 826         nanosecond precision, remove overflow prone branch.
 827
 828 2011-01-29  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 829
 830         PR libgfortran/47434
 831         * io/write_float.def (write_infnan): Use calculate_sign to determine
 832         if the sign should be given and check field widths accordingly.
 833
 834 2011-01-29  Kai Tietz  <kai.tietz@onevision.com>
 835
 836         * intrinsics/ctime.c (ctime_r): Improve implementation.
 837
 838 2011-01-27  Janne Blomqvist  <jb@gcc.gnu.org>
 839
 840         PR libfortran/47431
 841         * config.h.in: Regenerated.
 842         * configure: Regenerated.
 843         * configure.ac: Add check for ctime_r().
 844         * intrinsics/ctime.c (ctime_r): Fallback implementation.
 845         (fdate): Use ctime_r() instead of ctime().
 846         (fdate_sub): Likewise.
 847         (ctime): Likewise.
 848         (ctime_sub): Likewise.
 849
 850 2011-01-27  Janne Blomqvist  <jb@gcc.gnu.org>
 851
 852         PR libfortran/47432
 853         * config.h.in: Regenerated.
 854         * configure: Regenerated.
 855         * configure.ac: Add check for ttyname_r.
 856         * io/unix.h: Add TTY_NAME_MAX, change stream_ttyname prototype.
 857         * io/unix.c (stream_ttyname): Use ttyname_r if available, conform
 858         to new prototype.
 859         * io/inquire.c (inquire_via_unit): Use changed stream_ttyname.
 860         * io/intrinsics.c (ttynam_sub): Likewise.
 861         (ttynam): Likewise.
 862
 863 2011-01-27  Janne Blomqvist  <jb@gcc.gnu.org>
 864
 865         PR libfortran/47491
 866         * configure.ac: Call AC_USE_SYSTEM_EXTENSIONS to enable common
 867         extensions.
 868         * config.h.in: Regenerate.
 869         * configure: Regenerate.
 870
 871 2011-01-26  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 872
 873         PR libgfortran/47285
 874         * io/write_float.def (write_infnan): Adjust processor selected width
 875         to 3 if NaN.
 876
 877 2011-01-26  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 878
 879         PR libgfortran/47285
 880         * io/write_float.def (output_float): Return SUCCESS or FAILURE and use
 881         the result to set the padding.
 882
 883 2011-01-26  Kai Tietz  <kai.tietz@onevision.com>
 884
 885         * intrinsics/getlog.c (getlog): Fix label/statement issue.
 886
 887 2011-01-25  Janne Blomqvist  <jb@gcc.gnu.org>
 888
 889         PR libfortran/47375
 890         * config.h.in: Regenerated.
 891         * configure: Regenerated.
 892         * configure.ac: Add check for getpwuid_r.
 893         * intrinsics/getlog.c (getlog): Use getpwuid_r() if available.
 894
 895 2011-01-22  Janne Blomqvist  <jb@gcc.gnu.org>
 896
 897         PR libfortran/46267
 898         * config.h.in: Regenerated.
 899         * configure: Regenerated.
 900         * configure.ac: Check presence of strerror_r.
 901         * intrinsics/gerror.c (gerror): Use gf_strerror, modify logic.
 902         * io/unix.c (get_oserror): Remove.
 903         * libgfortran.h (gf_strerror): Add prototype.
 904         (get_oserror): Remove prototype.
 905         * runtime/error.c (gf_strerror): New function.
 906         (os_error): Use gf_strerror instead of get_oserror.
 907         (generate_errror): Likewise.
 908
 909 2011-01-17  Janne Blomqvist  <jb@gcc.gnu.org>
 910
 911         PR libfortran/47296
 912         * io/unix.c (tempfile): Set opp->file and opp->file_len also if an
 913         error occurs.
 914
 915 2011-01-16  Jakub Jelinek  <jakub@redhat.com>
 916
 917         PR fortran/46625
 918         * io/write_float.def (DTOAQ): Use quadmath_flt128tostr
 919         instead of quadmath_dtoa.
 920         * io/transfer128.c (tmp1, tmp2): New variables, bring in
 921         strtoflt128 and quadmath_flt128tostr.
 922         (transfer_real128, transfer_real128_write, transfer_complex128,
 923         transfer_complex128_write): Remove tmp1/tmp2 variables.
 924         * io/read.c (convert_real): Use strtoflt128 instead of
 925         quadmath_strtopQ, adjust for the changed arguments and return
 926         value.
 927
 928 2011-01-14  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 929
 930         PR libgfortran/47296
 931         * io/unix.c (unpack_filename): Return non-zero if the filename passed
 932         in is NULL.
 933
 934 2011-01-04  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 935
 936         PR libgfortran/47154
 937         * io/list_read.c (namelist_read): Remove calls to hit_eof to avoid the
 938         duplicate calls via next_record.
 939
 940 \f
 941 Copyright (C) 2011 Free Software Foundation, Inc.
 942
 943 Copying and distribution of this file, with or without modification,
 944 are permitted in any medium without royalty provided the copyright
 945 notice and this notice are preserved.