testsuite: Restore lp64 for some gcc.target/powerpc testcases.
[official-gcc.git] / libgfortran / ChangeLog
blob0c38d4db5498d1f666e8505826bf86589003c845
1 2020-06-22  David Edelsohn  <dje.gcc@gmail.com>
3         * Makefile.am: Use -include.
4         * Makefile.in: Regenerate.
6 2020-06-21  David Edelsohn  <dje.gcc@gmail.com>
8         * Makefile.am (tmake_file): Build and install AIX-style FAT libraries.
9         * Makefile.in: Regenerate.
10         * configure.ac (tmake_file): Substitute.
11         * configure: Regenerate.
12         * configure.host: Add system configury stanza. Define tmake_file.
13         * config/t-aix: New file.
15 2020-06-13  Thomas Koenig  <tkoenig@gcc.gnu.org>
17         PR libfortran/95313
18         * io/write.c (ztoa_big): Disable -Wstringop-overflow for one
19         line.
21 2020-06-08  Harald Anlauf  <anlauf@gmx.de>
23         PR fortran/95091
24         * io/transfer.c (finalize_transfer): Fix type in error message.
26 2020-06-01  Uroš Bizjak  <ubizjak@gmail.com>
28         PR libfortran/95418
29         * config/fpu-387.h (struct fenv): Add __attribute__ ((gcc_struct)).
31 2020-05-29  H.J. Lu  <hjl.tools@gmail.com>
33         PR bootstrap/95413
34         * configure: Regenerated.
36 2020-05-29  Jakub Jelinek  <jakub@redhat.com>
38         PR libfortran/95390
39         * Makefile.am (i_findloc0_c): Add findloc0_i10.c.
40         (i_findloc1_c): Add findloc1_i10.c.
41         * gfortran.map (GFORTRAN_10.2): New symbol version, export
42         _gfortran_{,m,s}findloc{0,1}_c10 symbols.
43         * Makefile.in: Regenerated.
44         * generated/findloc0_c10.c: Generated.
45         * generated/findloc1_c10.c: Generated.
47 2020-05-28  Harald Anlauf  <anlauf@gmx.de>
49         PR libfortran/95104
50         * io/unit.c (unlock_unit): Guard by check for NULL pointer.
52 2020-05-26  Harald Anlauf  <anlauf@gmx.de>
53             Steven G. Kargl   <kargl@gcc.gnu.org>
55         PR libfortran/95104
56         * io/transfer.c (st_wait_async): Do not dereference NULL pointer.
58 2020-05-26  Harald Anlauf  <anlauf@gmx.de>
60         PR fortran/95195
61         * io/transfer.c (finalize_transfer): Generate runtime error for
62         namelist input/output to unformatted file.
64 2020-05-23  Thomas Koenig  <tkoenig@gcc.gnu.org>
66         PR libfortran/95191
67         * io/async.c (async_wait_id): Generate error if ID is higher
68         than the highest current ID.
69         * runtime/error.c (translate_error): Handle LIBERROR_BAD_WAIT_ID.
71 2020-05-21  H.J. Lu  <hongjiu.lu@intel.com>
73         * m4/matmul.m4: Don't include <config/i386/cpuinfo.h>.  Use
74         __builtin_cpu_is/__builtin_cpu_supports
75         * generated/matmul_c10.c: Regenerated.
76         * generated/matmul_c16.c: Likewise.
77         * generated/matmul_c4.c: Likewise.
78         * generated/matmul_c8.c: Likewise.
79         * generated/matmul_i1.c: Likewise.
80         * generated/matmul_i16.c: Likewise.
81         * generated/matmul_i2.c: Likewise.
82         * generated/matmul_i4.c: Likewise.
83         * generated/matmul_i8.c: Likewise.
84         * generated/matmul_r10.c: Likewise.
85         * generated/matmul_r16.c: Likewise.
86         * generated/matmul_r4.c: Likewise.
87         * generated/matmul_r8.c: Likewise.
89 2020-05-15  H.J. Lu  <hongjiu.lu@intel.com>
91         PR bootstrap/95147
92         * configure: Regenerated.
94 2020-05-14  Thomas Koenig  <tkoenig@gcc.gnu.org>
96         PR libfortran/95119
97         * io/close.c (close_status): Add CLOSE_INVALID.
98         (st_close): Return early on invalid STATUS parameter.
100 2020-05-14  H.J. Lu  <hongjiu.lu@intel.com>
102         * configure: Regenerated.
104 2020-05-06  Uroš Bizjak  <ubizjak@gmail.com>
106         * config/fpu-387.h (__math_force_eval): Remove.
107         (__math_force_eval_div): New define.
108         (local_feraiseexcept): Use __math_force_eval_div to use
109         generic division to generate INVALID, DIVZERO and INEXACT
110         exceptions.
111         (struct fenv): Define named struct instead of typedef.
113 2020-05-01  Uroš Bizjak  <ubizjak@gmail.com>
115         * config/fpu-387.h (__math_force_eval): New define.
116         (local_feraiseexcept): Use __math_force_eval to evaluate
117         generic division to generate INVALID and DIVZERO exceptions.
119 2020-04-22  Fritz Reese  <foreese@gcc.gnu.org>
121         * intrinsics/trigd.c, intrinsics/trigd_lib.inc, intrinsics/trigd.inc:
122         Guard against unavailable math functions.
123         Use suffixes from kinds.h based on the REAL kind.
125 2020-04-22  Jakub Jelinek  <jakub@redhat.com>
127         PR libfortran/94694
128         PR libfortran/94586
129         * configure.ac: Add math func checks for fmaf, fma and fmal.  Add
130         HAVE_INLINE_BUILTIN_COPYSIGN check.
131         * c99_protos.h (copysign, fmaf, fma, fmal): Provide fallback
132         prototypes.
133         (HAVE_COPYSIGN, HAVE_FMAF, HAVE_FMA, HAVE_FMAL): Define if not
134         defined and fallback version is provided.
135         * intrinsics/c99_functions.c (copysign, fmaf, fma, fmal): Provide
136         fallback implementations if possible
137         * configure: Regenerated.
138         * config.h.in: Regenerated.
140 2020-04-19  Uroš Bizjak  <ubizjak@gmail.com>
142         * config/fpu-387.h (local_feraiseexcept) [__SSE_MATH__]:
143         Remove unneeded assignments to volatile memory.
145 2020-04-01  Fritz Reese  <foreese@gcc.gnu.org>
146             Steven G. Kargl  <kargl@gcc.gnu.org>
148         PR fortran/93871
149         * Makefile.am, Makefile.in: New make rule for intrinsics/trigd.c.
150         * gfortran.map: New routines for {sind, cosd, tand}X{r4, r8, r10, r16}.
151         * intrinsics/trigd.c, intrinsics/trigd_lib.inc, intrinsics/trigd.inc:
152         New files. Defines native degree-valued trig functions.
154 2020-02-18  Thomas Koenig  <tkoenig@gcc.gnu.org>
156         PR fortran/93599
157         * io/async.c (destroy_adv_cond): Do not destroy lock.
158         (async_io): Make sure au->lock is locked for finishing of thread.
159         Do not lock/unlock around signalling emptysignal. Unlock au->lock
160         before return.
161         (init_adv_cond): Do not initialize lock.
162         (enqueue_transfer): Unlock after signal.
163         (enqueue_done_id): Likewise.
164         (enqueue_done): Likewise.
165         (enqueue_close): Likewise.
166         (enqueue_data_transfer): Likewise.
167         (async_wait_id): Do not lock/unlock around signalling au->work.
168         (async_wait): Unlock after signal.
169         * io/async.h (SIGNAL): Add comment about needed au->lock.
170         Remove locking/unlocking of advcond->lock.
171         (WAIT_SIGNAL_MUTEX): Add comment. Remove locking/unlocking of
172         advcond->lock.  Unlock mutex only at the end.  Loop on
173         __ghread_cond_wait returning zero.
174         (REVOKE_SIGNAL): Add comment. Remove locking/unlocking of
175         advcond->lock.
176         (struct adv_cond): Remove mutex from struct.
178 2020-02-12  Sandra Loosemore  <sandra@codesourcery.com>
180         PR libstdc++/79193
181         PR libstdc++/88999
183         * configure: Regenerated.
185 2020-01-24  Maciej W. Rozycki  <macro@wdc.com>
187         * configure.ac: Handle `--with-toolexeclibdir='.
188         * Makefile.in: Regenerate.
189         * aclocal.m4: Regenerate.
190         * configure: Regenerate.
192 2020-01-17  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
194         PR libfortran/93234
195         * io/unit.c (set_internal_unit): Set round and sign flags
196         correctly.
198 2020-01-17  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
200         PR libfortran/90374
201         * io/format.c (parse_format_list): Zero width not allowed with
202         FMT_D.
203         * io/write_float.def (build_float_string): Include range of
204         higher exponent values that require wider width.
206 2020-01-01  Jerry DeLisle  <jvdelisle@gcc.gnu.org>
208         PR libfortran/90374
209         * io/format.c (parse_format_list): Implement the E0 exponent
210         width to provide smallest possible width for exponent fields.
211         Refactor code for correct parsing and better readability of the
212         code.
213         * io/io.h (write_real_w0): Change interface to pass in pointer
214         to fnode.
215         * io/transfer.c: Update all calls to write_real_w0 to use the
216         new interface.
217         * io/write.c ((write_real_w0): Use the new interface with fnode
218         to access both the decimal precision and exponent widths used in
219         build_float_string.
220         * io/write_float.def (build_float_string): Use the passed in
221         exponent width to calculate the used width in the case of E0.
223 2020-01-01  Jakub Jelinek  <jakub@redhat.com>
225         Update copyright years.
227 Copyright (C) 2020 Free Software Foundation, Inc.
229 Copying and distribution of this file, with or without modification,
230 are permitted in any medium without royalty provided the copyright
231 notice and this notice are preserved.