OSDN Git Service

d53329daae6f6a7defd6a1d38c0b98ad58d88d8b
[pf3gnuchains/gcc-fork.git] / libgfortran / ChangeLog
1 2011-02-19  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
2
3         PR libgfortran/47567
4         * io/write_float.def (output_float): Adjust width for F0.d to
5         allow space for negative signs on zero.
6
7 2011-02-16  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
8
9         PR libgfortran/47667
10         * io/list_read.c (read_logical): Check for end of line before calling
11         eat_line. (read_integer): Likewise. (parse_real): Don't unget the
12         separator. Check for end of line before calling eat_line.
13         (read_complex): Allow line-end before and after parenthesis and comma.
14         Check for end of line before calling eat_line. (read_real): Check for
15         end of line before calling eat_line.
16
17 2011-02-16  Jakub Jelinek  <jakub@redhat.com>
18
19         PR libfortran/47757
20         * gfortran.map (GFORTRAN_1.4): Export
21         _gfortran_{m,s}i{all,any,parity}_i{1,2,4,8,16} and
22         _gfortran_{cshift0,eoshift{0,2}}_16_char4.
23
24 2011-02-15  Tobias Burnus  <burnus@net-b.de>
25
26         PR fortran/47716
27         PR fortran/47648
28         * acinclude.m4 (LIBGFOR_CHECK_FLOAT128): Use check from
29         libquadmath, which uses more features.
30         * configure: Regenerate.
31
32 2011-02-14  Jakub Jelinek  <jakub@redhat.com>
33
34         PR fortran/47642
35         * io/write_float.def (DTOAQ): Use quadmath_snprintf instead of
36         quadmath_flt128tostr.
37         * io/transfer128.c (tmp2): Initialize to quadmath_snprintf instead
38         of quadmath_flt128tostr.
39
40 2011-02-13  Ralf Wildenhues  <Ralf.Wildenhues@gmx.de>
41
42         * Makefile.in: Regenerate.
43         * aclocal.m4: Likewise.
44         * configure: Likewise.
45
46 2011-02-05  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
47
48         PR libgfortran/47567
49         * io/write_float.def (output_float): Eliminate some redundant code.
50         Adjust width for case of F0.X for values of zero and all other values.
51         Expand cases where '*' is set to give cleaner results.
52
53 2011-02-05  Janne Blomqvist  <jb@gcc.gnu.org>
54
55         PR libfortran/47571
56         * intrinsics/time_1.h (GF_CLOCK_MONOTONIC): Move to system_clock.c.
57         (weak_gettime): Likewise.
58         (gf_gettime): Change API, move weak_gettime() usage to
59         system_clock.c
60         * intrinsics/system_clock.c (GTHREAD_USE_WEAK): Define.
61         (gf_gettime_mono): New function.
62         (system_clock_4): Use gf_gettime_mono().
63         (system_clock_8): Likewise.
64         * intrinsics/date_and_time.c (date_and_time): Update gf_gettime()
65         usage.
66
67 2011-02-02  Janne Blomqvist  <jb@gcc.gnu.org>
68
69         PR libfortran/47571
70         * configure: Regenerated.
71         * configure.ac: Don't add librt to LIBS.
72         * intrinsics/time_1.h (weak_gettime): Weakref trickery for
73         clock_gettime().
74         (gf_gettime): Use weak_gettime() instead of clock_gettime().
75
76 2011-02-01  Janne Blomqvist  <jb@gcc.gnu.org>
77
78         * intrinsics/time_1.h: Include errno.h needed by fallbacks.
79
80 2011-01-31  Janne Blomqvist  <jb@gcc.gnu.org>
81
82         * intrinsics/time_1.h: Fix definition of GF_CLOCK_MONOTONIC macro.
83
84 2011-01-31  Janne Blomqvist  <jb@gcc.gnu.org>
85
86         * configure.ac: Check for clock_gettime().
87         * configure: Regenerated.
88         * config.h.in: Regenerated.
89         * intrinsics/time_1.h (__time_1): Rename to gf_cputime, add
90         times() fallback.
91         (gf_gettime): New function.
92         * intrinsics/cpu_time.c (__cpu_time_1): Update to call gf_cputime.
93         * intrinsics/date_and_time.c (date_and_time): Use gf_gettime.
94         * intrinsics/dtime.c (dtime_sub): Use gf_cputime.
95         * intrinsics/etime.c (etime_sub): Use gf_cputime.
96         * intrinsics/system_clock.c (system_clock_4): Use gf_gettime.
97         (system_clock_8): Use gf_gettime, increase count rate to allow
98         nanosecond precision, remove overflow prone branch.
99
100 2011-01-29  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
101
102         PR libgfortran/47434
103         * io/write_float.def (write_infnan): Use calculate_sign to determine
104         if the sign should be given and check field widths accordingly.
105
106 2011-01-29  Kai Tietz  <kai.tietz@onevision.com>
107
108         * intrinsics/ctime.c (ctime_r): Improve implementation.
109
110 2011-01-27  Janne Blomqvist  <jb@gcc.gnu.org>
111
112         PR libfortran/47431
113         * config.h.in: Regenerated.
114         * configure: Regenerated.
115         * configure.ac: Add check for ctime_r().
116         * intrinsics/ctime.c (ctime_r): Fallback implementation.
117         (fdate): Use ctime_r() instead of ctime().
118         (fdate_sub): Likewise.
119         (ctime): Likewise.
120         (ctime_sub): Likewise.
121
122 2011-01-27  Janne Blomqvist  <jb@gcc.gnu.org>
123
124         PR libfortran/47432
125         * config.h.in: Regenerated.
126         * configure: Regenerated.
127         * configure.ac: Add check for ttyname_r.
128         * io/unix.h: Add TTY_NAME_MAX, change stream_ttyname prototype.
129         * io/unix.c (stream_ttyname): Use ttyname_r if available, conform
130         to new prototype.
131         * io/inquire.c (inquire_via_unit): Use changed stream_ttyname.
132         * io/intrinsics.c (ttynam_sub): Likewise.
133         (ttynam): Likewise.
134
135 2011-01-27  Janne Blomqvist  <jb@gcc.gnu.org>
136
137         PR libfortran/47491
138         * configure.ac: Call AC_USE_SYSTEM_EXTENSIONS to enable common
139         extensions.
140         * config.h.in: Regenerate.
141         * configure: Regenerate.
142
143 2011-01-26  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
144
145         PR libgfortran/47285
146         * io/write_float.def (write_infnan): Adjust processor selected width
147         to 3 if NaN.
148
149 2011-01-26  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
150
151         PR libgfortran/47285
152         * io/write_float.def (output_float): Return SUCCESS or FAILURE and use
153         the result to set the padding.
154
155 2011-01-26  Kai Tietz  <kai.tietz@onevision.com>
156
157         * intrinsics/getlog.c (getlog): Fix label/statement issue.
158
159 2011-01-25  Janne Blomqvist  <jb@gcc.gnu.org>
160
161         PR libfortran/47375
162         * config.h.in: Regenerated.
163         * configure: Regenerated.
164         * configure.ac: Add check for getpwuid_r.
165         * intrinsics/getlog.c (getlog): Use getpwuid_r() if available.
166
167 2011-01-22  Janne Blomqvist  <jb@gcc.gnu.org>
168
169         PR libfortran/46267
170         * config.h.in: Regenerated.
171         * configure: Regenerated.
172         * configure.ac: Check presence of strerror_r.
173         * intrinsics/gerror.c (gerror): Use gf_strerror, modify logic.
174         * io/unix.c (get_oserror): Remove.
175         * libgfortran.h (gf_strerror): Add prototype.
176         (get_oserror): Remove prototype.
177         * runtime/error.c (gf_strerror): New function.
178         (os_error): Use gf_strerror instead of get_oserror.
179         (generate_errror): Likewise.
180
181 2011-01-17  Janne Blomqvist  <jb@gcc.gnu.org>
182
183         PR libfortran/47296
184         * io/unix.c (tempfile): Set opp->file and opp->file_len also if an
185         error occurs.
186
187 2011-01-16  Jakub Jelinek  <jakub@redhat.com>
188
189         PR fortran/46625
190         * io/write_float.def (DTOAQ): Use quadmath_flt128tostr
191         instead of quadmath_dtoa.
192         * io/transfer128.c (tmp1, tmp2): New variables, bring in
193         strtoflt128 and quadmath_flt128tostr.
194         (transfer_real128, transfer_real128_write, transfer_complex128,
195         transfer_complex128_write): Remove tmp1/tmp2 variables.
196         * io/read.c (convert_real): Use strtoflt128 instead of
197         quadmath_strtopQ, adjust for the changed arguments and return
198         value.
199
200 2011-01-14  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
201
202         PR libgfortran/47296
203         * io/unix.c (unpack_filename): Return non-zero if the filename passed
204         in is NULL.
205
206 2011-01-04  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
207
208         PR libgfortran/47154
209         * io/list_read.c (namelist_read): Remove calls to hit_eof to avoid the
210         duplicate calls via next_record.
211
212 \f
213 Copyright (C) 2011 Free Software Foundation, Inc.
214
215 Copying and distribution of this file, with or without modification,
216 are permitted in any medium without royalty provided the copyright
217 notice and this notice are preserved.