libgfortran/ChangeLog

   1 2011-02-14  Jakub Jelinek  <jakub@redhat.com>
   2
   3         PR fortran/47642
   4         * io/write_float.def (DTOAQ): Use quadmath_snprintf instead of
   5         quadmath_flt128tostr.
   6         * io/transfer128.c (tmp2): Initialize to quadmath_snprintf instead
   7         of quadmath_flt128tostr.
   8
   9 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
  10
  11         * Makefile.in: Regenerate.
  12         * aclocal.m4: Likewise.
  13         * configure: Likewise.
  14
  15 2011-02-05  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
  16
  17         PR libgfortran/47567
  18         * io/write_float.def (output_float): Eliminate some redundant code.
  19         Adjust width for case of F0.X for values of zero and all other values.
  20         Expand cases where '*' is set to give cleaner results.
  21
  22 2011-02-05  Janne Blomqvist  <jb@gcc.gnu.org>
  23
  24         PR libfortran/47571
  25         * intrinsics/time_1.h (GF_CLOCK_MONOTONIC): Move to system_clock.c.
  26         (weak_gettime): Likewise.
  27         (gf_gettime): Change API, move weak_gettime() usage to
  28         system_clock.c
  29         * intrinsics/system_clock.c (GTHREAD_USE_WEAK): Define.
  30         (gf_gettime_mono): New function.
  31         (system_clock_4): Use gf_gettime_mono().
  32         (system_clock_8): Likewise.
  33         * intrinsics/date_and_time.c (date_and_time): Update gf_gettime()
  34         usage.
  35
  36 2011-02-02  Janne Blomqvist  <jb@gcc.gnu.org>
  37
  38         PR libfortran/47571
  39         * configure: Regenerated.
  40         * configure.ac: Don't add librt to LIBS.
  41         * intrinsics/time_1.h (weak_gettime): Weakref trickery for
  42         clock_gettime().
  43         (gf_gettime): Use weak_gettime() instead of clock_gettime().
  44
  45 2011-02-01  Janne Blomqvist  <jb@gcc.gnu.org>
  46
  47         * intrinsics/time_1.h: Include errno.h needed by fallbacks.
  48
  49 2011-01-31  Janne Blomqvist  <jb@gcc.gnu.org>
  50
  51         * intrinsics/time_1.h: Fix definition of GF_CLOCK_MONOTONIC macro.
  52
  53 2011-01-31  Janne Blomqvist  <jb@gcc.gnu.org>
  54
  55         * configure.ac: Check for clock_gettime().
  56         * configure: Regenerated.
  57         * config.h.in: Regenerated.
  58         * intrinsics/time_1.h (__time_1): Rename to gf_cputime, add
  59         times() fallback.
  60         (gf_gettime): New function.
  61         * intrinsics/cpu_time.c (__cpu_time_1): Update to call gf_cputime.
  62         * intrinsics/date_and_time.c (date_and_time): Use gf_gettime.
  63         * intrinsics/dtime.c (dtime_sub): Use gf_cputime.
  64         * intrinsics/etime.c (etime_sub): Use gf_cputime.
  65         * intrinsics/system_clock.c (system_clock_4): Use gf_gettime.
  66         (system_clock_8): Use gf_gettime, increase count rate to allow
  67         nanosecond precision, remove overflow prone branch.
  68
  69 2011-01-29  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
  70
  71         PR libgfortran/47434
  72         * io/write_float.def (write_infnan): Use calculate_sign to determine
  73         if the sign should be given and check field widths accordingly.
  74
  75 2011-01-29  Kai Tietz  <kai.tietz@onevision.com>
  76
  77         * intrinsics/ctime.c (ctime_r): Improve implementation.
  78
  79 2011-01-27  Janne Blomqvist  <jb@gcc.gnu.org>
  80
  81         PR libfortran/47431
  82         * config.h.in: Regenerated.
  83         * configure: Regenerated.
  84         * configure.ac: Add check for ctime_r().
  85         * intrinsics/ctime.c (ctime_r): Fallback implementation.
  86         (fdate): Use ctime_r() instead of ctime().
  87         (fdate_sub): Likewise.
  88         (ctime): Likewise.
  89         (ctime_sub): Likewise.
  90
  91 2011-01-27  Janne Blomqvist  <jb@gcc.gnu.org>
  92
  93         PR libfortran/47432
  94         * config.h.in: Regenerated.
  95         * configure: Regenerated.
  96         * configure.ac: Add check for ttyname_r.
  97         * io/unix.h: Add TTY_NAME_MAX, change stream_ttyname prototype.
  98         * io/unix.c (stream_ttyname): Use ttyname_r if available, conform
  99         to new prototype.
 100         * io/inquire.c (inquire_via_unit): Use changed stream_ttyname.
 101         * io/intrinsics.c (ttynam_sub): Likewise.
 102         (ttynam): Likewise.
 103
 104 2011-01-27  Janne Blomqvist  <jb@gcc.gnu.org>
 105
 106         PR libfortran/47491
 107         * configure.ac: Call AC_USE_SYSTEM_EXTENSIONS to enable common
 108         extensions.
 109         * config.h.in: Regenerate.
 110         * configure: Regenerate.
 111
 112 2011-01-26  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 113
 114         PR libgfortran/47285
 115         * io/write_float.def (write_infnan): Adjust processor selected width
 116         to 3 if NaN.
 117
 118 2011-01-26  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 119
 120         PR libgfortran/47285
 121         * io/write_float.def (output_float): Return SUCCESS or FAILURE and use
 122         the result to set the padding.
 123
 124 2011-01-26  Kai Tietz  <kai.tietz@onevision.com>
 125
 126         * intrinsics/getlog.c (getlog): Fix label/statement issue.
 127
 128 2011-01-25  Janne Blomqvist  <jb@gcc.gnu.org>
 129
 130         PR libfortran/47375
 131         * config.h.in: Regenerated.
 132         * configure: Regenerated.
 133         * configure.ac: Add check for getpwuid_r.
 134         * intrinsics/getlog.c (getlog): Use getpwuid_r() if available.
 135
 136 2011-01-22  Janne Blomqvist  <jb@gcc.gnu.org>
 137
 138         PR libfortran/46267
 139         * config.h.in: Regenerated.
 140         * configure: Regenerated.
 141         * configure.ac: Check presence of strerror_r.
 142         * intrinsics/gerror.c (gerror): Use gf_strerror, modify logic.
 143         * io/unix.c (get_oserror): Remove.
 144         * libgfortran.h (gf_strerror): Add prototype.
 145         (get_oserror): Remove prototype.
 146         * runtime/error.c (gf_strerror): New function.
 147         (os_error): Use gf_strerror instead of get_oserror.
 148         (generate_errror): Likewise.
 149
 150 2011-01-17  Janne Blomqvist  <jb@gcc.gnu.org>
 151
 152         PR libfortran/47296
 153         * io/unix.c (tempfile): Set opp->file and opp->file_len also if an
 154         error occurs.
 155
 156 2011-01-16  Jakub Jelinek  <jakub@redhat.com>
 157
 158         PR fortran/46625
 159         * io/write_float.def (DTOAQ): Use quadmath_flt128tostr
 160         instead of quadmath_dtoa.
 161         * io/transfer128.c (tmp1, tmp2): New variables, bring in
 162         strtoflt128 and quadmath_flt128tostr.
 163         (transfer_real128, transfer_real128_write, transfer_complex128,
 164         transfer_complex128_write): Remove tmp1/tmp2 variables.
 165         * io/read.c (convert_real): Use strtoflt128 instead of
 166         quadmath_strtopQ, adjust for the changed arguments and return
 167         value.
 168
 169 2011-01-14  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 170
 171         PR libgfortran/47296
 172         * io/unix.c (unpack_filename): Return non-zero if the filename passed
 173         in is NULL.
 174
 175 2011-01-04  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
 176
 177         PR libgfortran/47154
 178         * io/list_read.c (namelist_read): Remove calls to hit_eof to avoid the
 179         duplicate calls via next_record.
 180
 181 \f
 182 Copyright (C) 2011 Free Software Foundation, Inc.
 183
 184 Copying and distribution of this file, with or without modification,
 185 are permitted in any medium without royalty provided the copyright
 186 notice and this notice are preserved.