From: Igor Munkin via Tarantool-patches <tarantool-patches@dev.tarantool.org> To: Sergey Kaplun <skaplun@tarantool.org> Cc: tarantool-patches@dev.tarantool.org Subject: Re: [Tarantool-patches] [PATCH luajit] x86/x64: Fix math.ceil(-0.9) result sign. Date: Thu, 23 Nov 2023 06:31:24 +0000 [thread overview] Message-ID: <ZV7xvB9MTFJCv8qj@tarantool.org> (raw) In-Reply-To: <20231102123616.10527-1-skaplun@tarantool.org> Sergey, I've checked the patchset into all long-term branches in tarantool/luajit and bumped a new version in master, release/2.11 and release/2.10. On 02.11.23, Sergey Kaplun via Tarantool-patches wrote: > From: Mike Pall <mike> > > Reported by minoki. > > (cherry-picked from commit 674afcd4e21d0cf64de3219d347557a0aed8ecc7) > > The `ceil` (`floor`) math function implementation calculates (|x| + > 2^52) - 2^52 for its argument to determine the fractional part of x , so > it will be rounded to the nearest integer and its sign is restored. > After that, if the original value is < (>) than the result, the -1 (1) > is subtracted from it. Take a look at the `ceil()` case. The result of > the operation `-1 - (-1)` is +0 for FP arithmetic, against -0 expected > as a result. > > This patch changes the `- (-1)` operation to `+ 1` and restores sign > after it again. > > NB: Since in DUALNUM mode on x86/x64 all results are tried to be > converted to integers the sign of 0 is neglected. > > Sergey Kaplun: > * added the description and the test for the problem > > Part of tarantool/tarantool#9145 > --- > > Branch: https://github.com/tarantool/luajit/tree/skaplun/lj-859-math-ceil-sign > Tarantool PR: https://github.com/tarantool/tarantool/pull/9326 > Related issues: > * https://github.com/LuaJIT/LuaJIT/issues/859 > * https://github.com/tarantool/tarantool/issues/9145 > > src/vm_x64.dasc | 13 ++++++------ > src/vm_x86.dasc | 13 ++++++------ > .../lj-859-math-ceil-sign.test.lua | 20 +++++++++++++++++++ > 3 files changed, 32 insertions(+), 14 deletions(-) > create mode 100644 test/tarantool-tests/lj-859-math-ceil-sign.test.lua > <snipped> > -- > 2.42.0 > -- Best regards, IM
prev parent reply other threads:[~2023-11-23 6:37 UTC|newest] Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top 2023-11-02 12:36 Sergey Kaplun via Tarantool-patches 2023-11-07 9:44 ` Maxim Kokryashkin via Tarantool-patches 2023-11-07 10:21 ` Sergey Kaplun via Tarantool-patches 2023-11-07 10:50 ` Maxim Kokryashkin via Tarantool-patches 2023-11-09 13:26 ` Sergey Bronnikov via Tarantool-patches 2023-11-09 14:32 ` Sergey Kaplun via Tarantool-patches 2023-11-09 16:04 ` Sergey Bronnikov via Tarantool-patches 2023-11-23 6:31 ` Igor Munkin via Tarantool-patches [this message]
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=ZV7xvB9MTFJCv8qj@tarantool.org \ --to=tarantool-patches@dev.tarantool.org \ --cc=imun@tarantool.org \ --cc=skaplun@tarantool.org \ --subject='Re: [Tarantool-patches] [PATCH luajit] x86/x64: Fix math.ceil(-0.9) result sign.' \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: link
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox