1 2020-05-15 H.J. Lu <hongjiu.lu@intel.com>
4 * configure: Regenerated.
6 2020-05-14 Thomas Koenig <tkoenig@gcc.gnu.org>
9 * io/close.c (close_status): Add CLOSE_INVALID.
10 (st_close): Return early on invalid STATUS parameter.
12 2020-05-14 H.J. Lu <hongjiu.lu@intel.com>
14 * configure: Regenerated.
16 2020-05-06 Uroš Bizjak <ubizjak@gmail.com>
18 * config/fpu-387.h (__math_force_eval): Remove.
19 (__math_force_eval_div): New define.
20 (local_feraiseexcept): Use __math_force_eval_div to use
21 generic division to generate INVALID, DIVZERO and INEXACT
23 (struct fenv): Define named struct instead of typedef.
25 2020-05-01 Uroš Bizjak <ubizjak@gmail.com>
27 * config/fpu-387.h (__math_force_eval): New define.
28 (local_feraiseexcept): Use __math_force_eval to evaluate
29 generic division to generate INVALID and DIVZERO exceptions.
31 2020-04-22 Fritz Reese <foreese@gcc.gnu.org>
33 * intrinsics/trigd.c, intrinsics/trigd_lib.inc, intrinsics/trigd.inc:
34 Guard against unavailable math functions.
35 Use suffixes from kinds.h based on the REAL kind.
37 2020-04-22 Jakub Jelinek <jakub@redhat.com>
41 * configure.ac: Add math func checks for fmaf, fma and fmal. Add
42 HAVE_INLINE_BUILTIN_COPYSIGN check.
43 * c99_protos.h (copysign, fmaf, fma, fmal): Provide fallback
45 (HAVE_COPYSIGN, HAVE_FMAF, HAVE_FMA, HAVE_FMAL): Define if not
46 defined and fallback version is provided.
47 * intrinsics/c99_functions.c (copysign, fmaf, fma, fmal): Provide
48 fallback implementations if possible
49 * configure: Regenerated.
50 * config.h.in: Regenerated.
52 2020-04-19 Uroš Bizjak <ubizjak@gmail.com>
54 * config/fpu-387.h (local_feraiseexcept) [__SSE_MATH__]:
55 Remove unneeded assignments to volatile memory.
57 2020-04-01 Fritz Reese <foreese@gcc.gnu.org>
58 Steven G. Kargl <kargl@gcc.gnu.org>
61 * Makefile.am, Makefile.in: New make rule for intrinsics/trigd.c.
62 * gfortran.map: New routines for {sind, cosd, tand}X{r4, r8, r10, r16}.
63 * intrinsics/trigd.c, intrinsics/trigd_lib.inc, intrinsics/trigd.inc:
64 New files. Defines native degree-valued trig functions.
66 2020-02-18 Thomas Koenig <tkoenig@gcc.gnu.org>
69 * io/async.c (destroy_adv_cond): Do not destroy lock.
70 (async_io): Make sure au->lock is locked for finishing of thread.
71 Do not lock/unlock around signalling emptysignal. Unlock au->lock
73 (init_adv_cond): Do not initialize lock.
74 (enqueue_transfer): Unlock after signal.
75 (enqueue_done_id): Likewise.
76 (enqueue_done): Likewise.
77 (enqueue_close): Likewise.
78 (enqueue_data_transfer): Likewise.
79 (async_wait_id): Do not lock/unlock around signalling au->work.
80 (async_wait): Unlock after signal.
81 * io/async.h (SIGNAL): Add comment about needed au->lock.
82 Remove locking/unlocking of advcond->lock.
83 (WAIT_SIGNAL_MUTEX): Add comment. Remove locking/unlocking of
84 advcond->lock. Unlock mutex only at the end. Loop on
85 __ghread_cond_wait returning zero.
86 (REVOKE_SIGNAL): Add comment. Remove locking/unlocking of
88 (struct adv_cond): Remove mutex from struct.
90 2020-02-12 Sandra Loosemore <sandra@codesourcery.com>
95 * configure: Regenerated.
97 2020-01-24 Maciej W. Rozycki <macro@wdc.com>
99 * configure.ac: Handle `--with-toolexeclibdir='.
100 * Makefile.in: Regenerate.
101 * aclocal.m4: Regenerate.
102 * configure: Regenerate.
104 2020-01-17 Jerry DeLisle <jvdelisle@gcc.gnu.org>
107 * io/unit.c (set_internal_unit): Set round and sign flags
110 2020-01-17 Jerry DeLisle <jvdelisle@gcc.gnu.org>
113 * io/format.c (parse_format_list): Zero width not allowed with
115 * io/write_float.def (build_float_string): Include range of
116 higher exponent values that require wider width.
118 2020-01-01 Jerry DeLisle <jvdelisle@gcc.gnu.org>
121 * io/format.c (parse_format_list): Implement the E0 exponent
122 width to provide smallest possible width for exponent fields.
123 Refactor code for correct parsing and better readability of the
125 * io/io.h (write_real_w0): Change interface to pass in pointer
127 * io/transfer.c: Update all calls to write_real_w0 to use the
129 * io/write.c ((write_real_w0): Use the new interface with fnode
130 to access both the decimal precision and exponent widths used in
132 * io/write_float.def (build_float_string): Use the passed in
133 exponent width to calculate the used width in the case of E0.
135 2020-01-01 Jakub Jelinek <jakub@redhat.com>
137 Update copyright years.
139 Copyright (C) 2020 Free Software Foundation, Inc.
141 Copying and distribution of this file, with or without modification,
142 are permitted in any medium without royalty provided the copyright
143 notice and this notice are preserved.