[PATCH] pypi: set SRC_URI with =, not with +=

[PATCH] pypi: set SRC_URI with =, not with +=

[Alexander Kanavin]

This did not cause problems in builds, but broke some devtool
workflows such as version upgrades or checking the latest version
from upstream. Tarballs should come first, not the patches.

Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
---
 meta/classes/pypi.bbclass | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/meta/classes/pypi.bbclass b/meta/classes/pypi.bbclass
index 87b4c85fc0..899126e9e8 100644
--- a/meta/classes/pypi.bbclass
+++ b/meta/classes/pypi.bbclass
@@ -19,7 +19,7 @@ PYPI_SRC_URI ?= "${@pypi_src_uri(d)}"
 
 HOMEPAGE ?= "https://pypi.python.org/pypi/${PYPI_PACKAGE}/"
 SECTION = "devel/python"
-SRC_URI += "${PYPI_SRC_URI}"
+SRC_URI = "${PYPI_SRC_URI}"
 S = "${WORKDIR}/${PYPI_PACKAGE}-${PV}"
 
 UPSTREAM_CHECK_URI ?= "https://pypi.org/project/${PYPI_PACKAGE}/"
-- 
2.31.1

Re: [PATCH] pypi: set SRC_URI with =, not with +=

[Konrad Weihmann]

[-- Attachment #1: Type: text/plain, Size: 387 bytes --]

That doesn't look completely right to me.

You're absolutely right about devtool expecting the tarball in as the 1st item, but this patch here breaks existing recipes like

SRC_URI += "file://some.patch"

inherit pypi

as the patch will now be dropped.

I would suggest to go for SRC_URI_prepend = "${PYPI_SRC_URI} " which fixes the issue *and* doesn't break exisiting setups

[-- Attachment #2: Type: text/html, Size: 446 bytes --]

Re: [OE-core] [PATCH] pypi: set SRC_URI with =, not with +=

[Alexander Kanavin]

[-- Attachment #1: Type: text/plain, Size: 552 bytes --]

Right, I sent an updated patch.

Alex

On Wed, 23 Jun 2021 at 09:01, Konrad Weihmann <kweihmann@outlook.com> wrote:

> That doesn't look completely right to me.
>
> You're absolutely right about devtool expecting the tarball in as the 1st
> item, but this patch here breaks existing recipes like
>
> SRC_URI += "file://some.patch"
>
> inherit pypi
>
> as the patch will now be dropped.
>
>
>
> I would suggest to go for SRC_URI_prepend = "${PYPI_SRC_URI} " which fixes
> the issue *and* doesn't break exisiting setups
>
> 
>
>

[-- Attachment #2: Type: text/html, Size: 913 bytes --]

Re: [OE-core] [PATCH] pypi: set SRC_URI with =, not with +=

[Peter Kjellerstedt]

[-- Attachment #1: Type: text/plain, Size: 1224 bytes --]

I suggest you use =+ instead of _prepend to maintain expansion order. I.e., if someone does:

inherit pypi

SRC_URI = "http://some_url ${PYPI_SRC_URI}"

However, any way you change it, you risk breaking someone’s use case, but I believe using =+ instead of += should be the least invasive change compared to the current version.

//Peter

From: openembedded-core@lists.openembedded.org <openembedded-core@lists.openembedded.org> On Behalf Of Alexander Kanavin
Sent: den 23 juni 2021 10:34
To: Konrad Weihmann <kweihmann@outlook.com>
Cc: OE-core <openembedded-core@lists.openembedded.org>
Subject: Re: [OE-core] [PATCH] pypi: set SRC_URI with =, not with +=

Right, I sent an updated patch.

Alex

On Wed, 23 Jun 2021 at 09:01, Konrad Weihmann <kweihmann@outlook.com<mailto:kweihmann@outlook.com>> wrote:

That doesn't look completely right to me.

You're absolutely right about devtool expecting the tarball in as the 1st item, but this patch here breaks existing recipes like

SRC_URI += "file://some.patch"

inherit pypi

as the patch will now be dropped.



I would suggest to go for SRC_URI_prepend = "${PYPI_SRC_URI} " which fixes the issue *and* doesn't break exisiting setups



[-- Attachment #2: Type: text/html, Size: 5083 bytes --]

Re: [OE-core] [PATCH] pypi: set SRC_URI with =, not with +=

[Alexander Kanavin]

[-- Attachment #1: Type: text/plain, Size: 1650 bytes --]

Frankly guys, I think maybe we should just break stuff in this case, and
have it fixed up properly, then try to guess what workaround might work.

From that perspective the original patch is the most clear.

Alex

On Wed, 23 Jun 2021 at 11:26, Peter Kjellerstedt <
peter.kjellerstedt@axis.com> wrote:

> I suggest you use =+ instead of _prepend to maintain expansion order.
> I.e., if someone does:
>
>
>
> inherit pypi
>
>
>
> SRC_URI = "http://some_url ${PYPI_SRC_URI}"
>
>
>
> However, any way you change it, you risk breaking someone’s use case, but
> I believe using =+ instead of += should be the least invasive change
> compared to the current version.
>
>
>
> //Peter
>
>
>
> *From:* openembedded-core@lists.openembedded.org <
> openembedded-core@lists.openembedded.org> *On Behalf Of *Alexander Kanavin
> *Sent:* den 23 juni 2021 10:34
> *To:* Konrad Weihmann <kweihmann@outlook.com>
> *Cc:* OE-core <openembedded-core@lists.openembedded.org>
> *Subject:* Re: [OE-core] [PATCH] pypi: set SRC_URI with =, not with +=
>
>
>
> Right, I sent an updated patch.
>
>
>
> Alex
>
>
>
> On Wed, 23 Jun 2021 at 09:01, Konrad Weihmann <kweihmann@outlook.com>
> wrote:
>
> That doesn't look completely right to me.
>
> You're absolutely right about devtool expecting the tarball in as the 1st
> item, but this patch here breaks existing recipes like
>
> SRC_URI += "file://some.patch"
>
> inherit pypi
>
> as the patch will now be dropped.
>
>
>
> I would suggest to go for SRC_URI_prepend = "${PYPI_SRC_URI} " which fixes
> the issue *and* doesn't break exisiting setups
>
>
>
>

[-- Attachment #2: Type: text/html, Size: 4628 bytes --]

Re: [OE-core] [PATCH] pypi: set SRC_URI with =, not with +=

[Andre McCurdy]

On Wed, Jun 23, 2021 at 3:18 AM Alexander Kanavin
<alex.kanavin@gmail.com> wrote:
>
> Frankly guys, I think maybe we should just break stuff in this case, and have it fixed up properly, then try to guess what workaround might work.
>
> From that perspective the original patch is the most clear.

The typical ordering of variables etc in recipes is to put SRC_URI
before inherit.

From the perspective of trying to get recipes to stick to a standard
ordering of variables etc, _prepend seems to be the best solution.

Re: [OE-core] [PATCH] pypi: set SRC_URI with =, not with +=

[Alexander Kanavin]

[-- Attachment #1: Type: text/plain, Size: 793 bytes --]

On Thu, 24 Jun 2021 at 09:21, Andre McCurdy <armccurdy@gmail.com> wrote:

> On Wed, Jun 23, 2021 at 3:18 AM Alexander Kanavin
> <alex.kanavin@gmail.com> wrote:
> >
> > Frankly guys, I think maybe we should just break stuff in this case, and
> have it fixed up properly, then try to guess what workaround might work.
> >
> > From that perspective the original patch is the most clear.
>
> The typical ordering of variables etc in recipes is to put SRC_URI
> before inherit.
>
> From the perspective of trying to get recipes to stick to a standard
> ordering of variables etc, _prepend seems to be the best solution.
>

That's fine with me, we should probably fix up gnomebase.bbclass as well,
as using = there adds friction to how SRC_URI must be set in the recipes.

Alex

[-- Attachment #2: Type: text/html, Size: 1265 bytes --]