All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Yann E. MORIN" <yann.morin.1998@free.fr>
To: "Guillaume W. Bres" <guillaume.bressaix@gmail.com>
Cc: buildroot@busybox.net, Ralf Gommers <ralf.gommers@gmail.com>
Subject: Re: [Buildroot] [RESEND v1 1/1] package/python-scipy: bump version to 1.9.1
Date: Sat, 1 Oct 2022 20:19:38 +0200	[thread overview]
Message-ID: <20221001181938.GO30794@scaer> (raw)
In-Reply-To: <20220928193437.13407-1-guillaume.bressaix@gmail.com>

Guillaume, All,

On 2022-09-28 21:34 +0200, Guillaume W. Bres spake thusly:
> Signed-off-by: Guillaume W. Bres <guillaume.bressaix@gmail.com>
> ---

All that is below the triple-dash line will be dropped by git when we
apply the patch. However, you provided interesting information in that
post-commit-log message...

> Scipy is expected to move to a meson build infrastructure
> by the next release
> 
> scipy-.tar.xz is no longer released.

This should be part of the commit log proper, to explain why you
switched to the .tar.gz archive.

> For some reason, a new "Python.h" missing header
> build time failure arised. I solve it with a -Ipython include.

This should also be part of the commit log 

> Then test script failed, with linker/runtime missing symbols error,
> all related to the linear algebra portion (which we do test).
> Errors like "undefined reference to __atomic_compare_exchange 8"
> regarding
> "site-/scipy/optimize/_highs/_highs_wrapper.cpython-310-arm-linux-gnueabi.so"
> to be specific. I solved it with a -latomic flag.

I see you pass it via CFLAGS; that's not correct, it should be passed
via LDFLAGS. If that doesnt work, it should be explained why.

Also, we can't arbitrary link agains -latomic, because not all
toolchains have it. See cairo.mk for how it should be conditionally
used.

Regards,
Yann E. MORIN.

> Once 1.9.2 is out, we know we already have my V3-1.9.1 patch
> serie that builds "fine" with meson, it involved a minor tweak
> (hopefuly solved by this time). The test script no longer did pass though,
> this serie is not 100% ready to go
> ---
>  package/python-scipy/python-scipy.hash | 2 +-
>  package/python-scipy/python-scipy.mk   | 7 +++++--
>  2 files changed, 6 insertions(+), 3 deletions(-)
> 
> diff --git a/package/python-scipy/python-scipy.hash b/package/python-scipy/python-scipy.hash
> index 0274679c43..1ba239aabf 100644
> --- a/package/python-scipy/python-scipy.hash
> +++ b/package/python-scipy/python-scipy.hash
> @@ -1,5 +1,5 @@
>  # Locally generated
> -sha256  0140dd24d14ce194fdb3df496c2190f0896ca84ac98bf5d69c3b3fb7bde1a1e9  scipy-1.8.1.tar.xz
> +sha256  26d28c468900e6d5fdb37d2812ab46db0ccd22c63baa095057871faa3a498bc9  scipy-1.9.1.tar.gz
>  sha256  96599f8ec69a2cc609265eed13fc75090aacb28742260e55ce623788a33da19c  LICENSE.txt
>  sha256  37e64a498894ac7c3b070023e3689e954a8ecf8a23b90968d09a455f1b4f7b35  scipy/linalg/src/lapack_deprecations/LICENSE
>  sha256  606209a000716c5f66e33e180ce08434b96ed17db4975ab9723c6b5fbcc89609  scipy/ndimage/LICENSE.txt
> diff --git a/package/python-scipy/python-scipy.mk b/package/python-scipy/python-scipy.mk
> index a8c23c6c88..1619b9f40a 100644
> --- a/package/python-scipy/python-scipy.mk
> +++ b/package/python-scipy/python-scipy.mk
> @@ -4,8 +4,8 @@
>  #
>  ################################################################################
>  
> -PYTHON_SCIPY_VERSION = 1.8.1
> -PYTHON_SCIPY_SOURCE = scipy-$(PYTHON_SCIPY_VERSION).tar.xz
> +PYTHON_SCIPY_VERSION = 1.9.1
> +PYTHON_SCIPY_SOURCE = scipy-$(PYTHON_SCIPY_VERSION).tar.gz
>  PYTHON_SCIPY_SITE = https://github.com/scipy/scipy/releases/download/v$(PYTHON_SCIPY_VERSION)
>  PYTHON_SCIPY_LICENSE = \
>  	BSD-3-Clause, \
> @@ -36,6 +36,8 @@ PYTHON_SCIPY_INSTALL_STAGING = YES
>  PYTHON_SCIPY_SETUP_TYPE = setuptools
>  PYTHON_SCIPY_BUILD_OPTS = config_fc --fcompiler=gnu95
>  
> +PYTHON_SCIPY_CFLAGS = \
> +	-I$(STAGING_DIR)/usr/include/python$(PYTHON3_VERSION_MAJOR)
>  PYTHON_SCIPY_LDFLAGS = $(TARGET_LDFLAGS) -shared \
>  	-L$(PYTHON3_PATH)/site-packages/numpy/core/lib
>  # -lnpyrandom localization
> @@ -44,6 +46,7 @@ PYTHON_SCIPY_LDFLAGS += \
>  
>  PYTHON_SCIPY_ENV = \
>  	F90=$(TARGET_FC) \
> +	CFLAGS="$(PYTHON_SCIPY_CFLAGS) -latomic" \
>  	LDFLAGS="$(PYTHON_SCIPY_LDFLAGS)"
>  
>  # Provide system configuration options to numpy distutils extensions, telling
> -- 
> 2.20.1
> 
> _______________________________________________
> buildroot mailing list
> buildroot@buildroot.org
> https://lists.buildroot.org/mailman/listinfo/buildroot

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

  reply	other threads:[~2022-10-01 18:19 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-28 19:34 [Buildroot] [RESEND v1 1/1] package/python-scipy: bump version to 1.9.1 Guillaume W. Bres
2022-10-01 18:19 ` Yann E. MORIN [this message]
2022-10-02  8:37   ` Guillaume Bres

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=20221001181938.GO30794@scaer \
    --to=yann.morin.1998@free.fr \
    --cc=buildroot@busybox.net \
    --cc=guillaume.bressaix@gmail.com \
    --cc=ralf.gommers@gmail.com \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.