mirrors/phobos
1
0
Fork 0
mirror of https://github.com/dlang/phobos.git synced 2025-04-29 22:50:38 +03:00
Code Issues Projects Releases Packages Wiki Activity Actions
19785 commits 7 branches 511 tags 95 MiB
Commit graph

59 commits

Author SHA1 Message Date
Iain Buclaw
2a62b4c798 std.math.exponential: Add deprecations for calling log family functions with an integral argument 2022-12-13 00:56:49 +01:00
Iain Buclaw
8e799c14dc std.math.exponential: Implement log/log2/log10 for double- and float-precision 2022-12-12 19:59:05 +01:00
Iain Buclaw
5a9919c802 std.math.hardware: Use forceOp in another ieeeFlags test 2022-12-04 20:26:28 +01:00
Iain Buclaw
2e1f8de817 std.math.hardware: Fix broken IEEE FP flags tests 
1. The `pragma(inline, false)` doesn't do what the author thinks it does
   (Issue 23520).
2. Change that introduced `blockopt` seems to have a slight
   misunderstanding about what constant propagation is (`a /= 0.0L` can
   always be constant propagated regardless of what surrounds it).
3. Undocument tests that depend on the need for forcing operations to
   disarm the optimizer, and replace with simple tests behind StdDdoc.
 2022-12-03 00:17:36 +01:00
Iain Buclaw
d957f13db3 std.math: Faster floor(double) implementation 2022-07-13 00:16:51 +00:00
Iain Buclaw
ff16d73e4c std.math.exponential: Fix coefficients for log2 and log10 2022-07-05 23:24:50 +00:00
Iain Buclaw
0497537e22 std.math.exponential: Split generic logN implementations to own impl templates 2022-07-05 10:37:31 +02:00
Iain Buclaw
682e16d3ec std.math.exponential: Implement pure logb() for double and float precision 2022-07-04 23:48:56 +00:00
Ast-x64
0c56468d19 RISCV: Fix wrong ExceptionMask values 
According to RISCV specifications
https://riscv.org/wp-content/uploads/2019/12/riscv-spec-20191213.pdf,
figure 11.2, the mask flags should be NX, UF, OF, DZ, NV in order.
 2022-06-16 20:22:10 +00:00
Ast-x64
b2d5da9c43 Extend floating point precision in some tests 
These tests may work better with quadruple-presicion real types now.
 2022-06-16 19:53:02 +00:00
Luís Ferreira
a56287dd8b chore(math/exponential): avoid import constants if not InlineAsm_X87 
Signed-off-by: Luís Ferreira <contact@lsferreira.net>
 2022-05-09 18:44:14 +00:00
Luís Ferreira
19c6e8d251 chore(math/hardware): align spaces by 4 instead of by 3 
Signed-off-by: Luís Ferreira <contact@lsferreira.net>
 2022-05-09 18:42:23 +00:00
Petar Kirov
dfaaf8fa53 Merge remote-tracking branch 'upstream/stable' into merge_stable 2021-11-09 10:30:28 +02:00
Brian Callahan
8a42656f67 Fix Issue 22416 - Unify polyImpl implementations 2021-10-18 05:40:34 +00:00
Brian Callahan
36eeb42292 OpenBSD: Add polyImpl implementation for x86 
Fix Issue 22393 - OpenBSD: Add polyImpl implementation for x86
 2021-10-16 08:05:02 -04:00
Per Nordlöw
46d416fa9d Qualify some function parameters under std.format as const to reduce template bloat 2021-09-22 23:12:15 +00:00
Iain Buclaw
1cbbb845af std.math: Reference Issue 5305 in unittest taking pointer to function 2021-08-25 19:27:16 +00:00
Luís Ferreira
62e01fc442
std.math.trigonometry: fix tan asm x86 implementation for ldc compiler 
Somehow, separating asm blocks in small pieces produces additional
`mov`s and other instructions on LDC that breaks the tanAsm logic and
therefore returns wrong values.

Signed-off-by: Luís Ferreira <contact@lsferreira.net>
 2021-08-21 17:29:15 +01:00
Luís Ferreira
8514636059 std.math.algebraic: use fabs from druntime 
Signed-off-by: Luís Ferreira <contact@lsferreira.net>
 2021-08-19 09:56:55 +00:00
Martin Nowak
3fff60bc50 Merge remote-tracking branch 'upstream/stable' into merge_stable 2021-07-13 12:24:06 +02:00
Walter Bright
13d84801e7 declaring templates inside static foreach is not reliable 2021-06-08 10:57:08 +00:00
The Dlang Bot
19003dd639
Merge pull request #8127 from kinke/bitpattern 
[stable] std.math.operations: Slightly revise extractBitpattern()

Signed-off-by: Razvan Nitu <RazvanN7@users.noreply.github.com>
Merged-on-behalf-of: Razvan Nitu <RazvanN7@users.noreply.github.com>
 2021-06-07 18:22:22 +00:00
Martin Kinkelin
06037a25b4 std.math.exponential: Minimally relax a pow() unittest for targets with 64-bit real 
Required since #7783.
 2021-06-03 07:56:03 +00:00
Martin Kinkelin
8946209a9b Revert #7823 - Faster pow implementation for integral exponents 
There are multiple issues with that commit, incl. out-of-bounds read
on x86 systems and a significant performance decrease for qualified
floating-point types; see #8124 for a full list of flaws.

I'm questioning the tackled goal in the first place, because I expect
users to use that `pow` mainly with small integral exponents, *not*
expecting underflows to 0 or overflows to ±infinity, so optimizing
for these cases and sacrificing performance for the regular case doesn't
make a lot of sense to me.
 2021-06-03 09:57:57 +09:00
Martin Kinkelin
3d90677e51 std.math.operations: Fix extractBitpattern() unittest 2021-06-02 21:22:27 +02:00
Martin Kinkelin
4d8be9050b std.math.operations: Slightly simplify extractBitpattern() 2021-06-02 20:45:13 +02:00
Martin Kinkelin
5a46c9ecf6 std.math.operations: Make extractBitpattern() @trusted 
To enable unsafe reinterpret-casts directly, without extra lambdas.
 2021-06-02 20:42:09 +02:00
Martin Kinkelin
ddd044d6c7 std.math.operations: Fix out-of-bounds read in extractBitpattern() 
For 32-bit x86 targets with `real.sizeof == 12`.
 2021-06-02 20:38:25 +02:00
Martin Kinkelin
a5b66d6614 std.math.hardware: Remove unused InlineAsm_X87* version declarations 2021-05-29 14:50:32 +02:00
Martin Kinkelin
0a685c19ac std.math: Fix missing imports for non-x87 real after split-up 2021-05-29 14:48:17 +02:00
Iain Buclaw
79bf8c3c32 std.math.algebraic: Convert hypot to a template 2021-05-11 12:22:13 +02:00
Iain Buclaw
52699fc020 std.math: Adjust whitespace of a few unittests 2021-05-11 09:14:14 +02:00
Iain Buclaw
543a3d0cd6 std.math.algebraic: Support more float types in hypot 2021-05-11 02:59:57 +02:00
berni44
e53d010def extractBitpattern: Add implied bit to mantissa and subtract bias from exponent. 2021-05-07 09:37:18 +02:00
Razvan Nitu
9909b6ebe3
Merge pull request #8030 from berni44/fixup_7775 
Add transition note from approxEqual to isClose.
 2021-05-07 15:00:23 +08:00
Nathan Sashihara
efced87ae8 Again use core.math intrinsics instead of std.math wrappers 
As in PR #7821 & PR #7825. Mostly changes imports
from std.math being divided into submodules.
 2021-05-04 15:59:34 +02:00
berni44
8408ce3108 Add transition note from approxEqual to isClose. 2021-05-03 11:41:44 +02:00
berni44
8d97da17d0 Move bitpattern extraction from std.format to std.math. 2021-04-30 09:37:22 +02:00
Iain Buclaw
1d0da90d82 std.math.exponential: Fix imports for non-INLINE_YL2X path 2021-04-30 02:15:43 +02:00
berni44
661c7a83b1 Clean up std.math.package 2021-04-20 10:04:22 +02:00
berni44
7f42447274 std.math: Move powmod to exponential submodule. 2021-04-20 04:54:59 +02:00
berni44
0c4dcb32a0 Narrow imports in std.math. 2021-04-19 22:32:55 +02:00
berni44
943cf022ea std.math: Move some stranded unittests to correct place. 2021-04-19 18:22:33 +02:00
berni44
4a4af2d78e std.math: Move hardware support into hardware submodule. 2021-04-19 10:46:14 +02:00
berni44
e9014c4342 std.math: Move floating point operations into operations submodule. 2021-04-17 21:52:27 +02:00
Iain Buclaw
7a258f293f std.math.algebraic: Add overload for three-dimensional hypotenuse 2021-04-17 18:13:07 +02:00
berni44
bb2432fdb4 Replace equalsDigit with isClose 2021-04-17 16:59:34 +02:00
berni44
64398c058f std.math: Move exponential functions into exponential submodule. 2021-04-14 16:18:20 +02:00
berni44
9812f41efd std.math: Move introspection functions into submodule. 2021-04-14 14:16:52 +02:00
berni44
66324313a8 Remove std.math.integral. 2021-04-14 14:15:38 +02:00