Re: scikit-learn testing migration
- To: debian-science@lists.debian.org
- Subject: Re: scikit-learn testing migration
- From: Christian Kastner <ckk@debian.org>
- Date: Tue, 16 Aug 2022 19:59:49 +0200
- Message-id: <[🔎] 0c2c9442-69e8-d3d3-5a63-4a2e4cbec1cb@debian.org>
- In-reply-to: <CAM8zJQtKjmsPs4=TXHDaXrJA_0M_7i48AefNjTHRWhZoixTnJw@mail.gmail.com>
- References: <20220716154631.44yj3ct5em6epbg3@debian> <32951d554eb1018b453dacb0630ee19f28613f39.camel@gmail.com> <Ytfk7vzFaqWaNEBR@an3as.eu> <7a65d5c8-fda2-7fe1-e019-a774b9c6e722@debian.org> <YthWMB8pVSio6ozi@an3as.eu> <YuDvv0GCo0CAc2U7@an3as.eu> <f8c903a0e7bfb36ba830d7936ff02cc53e25aa2e.camel@riseup.net> <CAM8zJQtKjmsPs4=TXHDaXrJA_0M_7i48AefNjTHRWhZoixTnJw@mail.gmail.com>
On 2022-07-28 09:15, Graham Inggs wrote:
> On Wed, 27 Jul 2022 at 17:57, M. Zhou <lumin@debian.org> wrote:
>> The previous segfault on armel becomes Bus Error on armel and armhf.
>> I can build it on Power9, but it seems that the test fails on power8 (our buildd).
>
> In #1003165, one of the arm porters wrote they are happy to look at
> the bus errors, but the baseline issue should be fixed first.
FWIW, a few months ago, I looked at a core dump of this and documented
the results here:
https://github.com/scikit-learn/scikit-learn/issues/16443#issuecomment-671063747
Unless I'm misremembering something, the scikit-learn Cython code looks
OK, so the problem probably stems from the code generated from that. It
hat been suggested that some kind of unaligned access is produced.
I stopped debugging there, as it's been a while since I've done
low-level stuff like that.
Best,
Christian
Reply to: