Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.debian.maint.python > #16669 > unrolled thread
| Started by | Soren Stoutner <soren@debian.org> |
|---|---|
| First post | 2025-01-08 21:10 +0100 |
| Last post | 2025-01-08 23:50 +0100 |
| Articles | 5 — 3 participants |
Back to article view | Back to linux.debian.maint.python
Multi-Arch: same and WHEEL files Soren Stoutner <soren@debian.org> - 2025-01-08 21:10 +0100
Re: Multi-Arch: same and WHEEL files Scott Talbert <swt@techie.net> - 2025-01-08 21:50 +0100
Re: Multi-Arch: same and WHEEL files Scott Talbert <swt@techie.net> - 2025-01-08 22:10 +0100
Re: Multi-Arch: same and WHEEL files Matthias Klose <doko@debian.org> - 2025-01-08 22:10 +0100
Re: Multi-Arch: same and WHEEL files Soren Stoutner <soren@debian.org> - 2025-01-08 23:50 +0100
| From | Soren Stoutner <soren@debian.org> |
|---|---|
| Date | 2025-01-08 21:10 +0100 |
| Subject | Multi-Arch: same and WHEEL files |
| Message-ID | <K2Kxj-6Iir-3@gated-at.bofh.it> |
[Multipart message — attachments visible in raw view] — view raw
Python3-hid was marked `Multi-Arch: same` by a previous maintainer. Recently, tracker.debian.org warned that there is an issue because the WHEEL file is different in each package. https://tracker.debian.org/pkg/python-hidapi The differences are in the `tag` field in the WHEEL file. For example, amd64 has: Tag: cp312-cp312-linux_x86_64 While arm64 has: Tag: cp312-cp312-linux_aarch64 What is the recommended way to deal with this?\ 1. Should the package not be marked `Multi-Arch: same`? 2. Should I prevent the binary package from shipping the WHEEL file? 3. Is there some other recommendation? -- Soren Stoutner soren@debian.org
[toc] | [next] | [standalone]
| From | Scott Talbert <swt@techie.net> |
|---|---|
| Date | 2025-01-08 21:50 +0100 |
| Message-ID | <K2La1-6Ivi-1@gated-at.bofh.it> |
| In reply to | #16669 |
On Wed, 8 Jan 2025, Soren Stoutner wrote: > Python3-hid was marked `Multi-Arch: same` by a previous maintainer. Recently, > tracker.debian.org warned that there is an issue because the WHEEL file is > different in each package. > > https://tracker.debian.org/pkg/python-hidapi > > The differences are in the `tag` field in the WHEEL file. For example, amd64 > has: > > Tag: cp312-cp312-linux_x86_64 > > While arm64 has: > > Tag: cp312-cp312-linux_aarch64 > > What is the recommended way to deal with this?\ > > 1. Should the package not be marked `Multi-Arch: same`? > > 2. Should I prevent the binary package from shipping the WHEEL file? > > 3. Is there some other recommendation? It seems like this package probably should just not be marked Multi-Arch: same. The WHEEL file is likely important metadata. Scott
[toc] | [prev] | [next] | [standalone]
| From | Scott Talbert <swt@techie.net> |
|---|---|
| Date | 2025-01-08 22:10 +0100 |
| Message-ID | <K2Ltn-6IRr-1@gated-at.bofh.it> |
| In reply to | #16670 |
[Multipart message — attachments visible in raw view] — view raw
On Wed, 8 Jan 2025, Matthias Klose wrote: > On 08.01.25 21:36, Scott Talbert wrote: >> On Wed, 8 Jan 2025, Soren Stoutner wrote: >> >>> Python3-hid was marked `Multi-Arch: same` by a previous maintainer. >>> Recently, >>> tracker.debian.org warned that there is an issue because the WHEEL file is >>> different in each package. >>> >>> https://tracker.debian.org/pkg/python-hidapi >>> >>> The differences are in the `tag` field in the WHEEL file. For example, >>> amd64 >>> has: >>> >>> Tag: cp312-cp312-linux_x86_64 >>> >>> While arm64 has: >>> >>> Tag: cp312-cp312-linux_aarch64 >>> >>> What is the recommended way to deal with this?\ >>> >>> 1. Should the package not be marked `Multi-Arch: same`? >>> >>> 2. Should I prevent the binary package from shipping the WHEEL file? >>> >>> 3. Is there some other recommendation? >> >> It seems like this package probably should just not be marked Multi- Arch: >> same. The WHEEL file is likely important metadata. > > but why would we ship wheel files in the first place? In this case Soren's talking about the WHEEL file in the dist-info directory, not a .whl file. Scott
[toc] | [prev] | [next] | [standalone]
| From | Matthias Klose <doko@debian.org> |
|---|---|
| Date | 2025-01-08 22:10 +0100 |
| Message-ID | <K2Ltn-6IRr-3@gated-at.bofh.it> |
| In reply to | #16670 |
On 08.01.25 21:36, Scott Talbert wrote: > On Wed, 8 Jan 2025, Soren Stoutner wrote: > >> Python3-hid was marked `Multi-Arch: same` by a previous maintainer. >> Recently, >> tracker.debian.org warned that there is an issue because the WHEEL >> file is >> different in each package. >> >> https://tracker.debian.org/pkg/python-hidapi >> >> The differences are in the `tag` field in the WHEEL file. For >> example, amd64 >> has: >> >> Tag: cp312-cp312-linux_x86_64 >> >> While arm64 has: >> >> Tag: cp312-cp312-linux_aarch64 >> >> What is the recommended way to deal with this?\ >> >> 1. Should the package not be marked `Multi-Arch: same`? >> >> 2. Should I prevent the binary package from shipping the WHEEL file? >> >> 3. Is there some other recommendation? > > It seems like this package probably should just not be marked Multi- > Arch: same. The WHEEL file is likely important metadata. but why would we ship wheel files in the first place?
[toc] | [prev] | [next] | [standalone]
| From | Soren Stoutner <soren@debian.org> |
|---|---|
| Date | 2025-01-08 23:50 +0100 |
| Message-ID | <K2N29-6JMS-3@gated-at.bofh.it> |
| In reply to | #16670 |
[Multipart message — attachments visible in raw view] — view raw
On Wednesday, January 8, 2025 1:36:18 PM MST Scott Talbert wrote: > On Wed, 8 Jan 2025, Soren Stoutner wrote: > > Python3-hid was marked `Multi-Arch: same` by a previous maintainer. > > Recently, tracker.debian.org warned that there is an issue because the > > WHEEL file is different in each package. > > > > https://tracker.debian.org/pkg/python-hidapi > > > > The differences are in the `tag` field in the WHEEL file. For example, > > amd64 > > has: > > > > Tag: cp312-cp312-linux_x86_64 > > > > While arm64 has: > > > > Tag: cp312-cp312-linux_aarch64 > > > > What is the recommended way to deal with this?\ > > > > 1. Should the package not be marked `Multi-Arch: same`? > > > > 2. Should I prevent the binary package from shipping the WHEEL file? > > > > 3. Is there some other recommendation? > > It seems like this package probably should just not be marked Multi-Arch: > same. The WHEEL file is likely important metadata. Thank you for the advice. I have done so. -- Soren Stoutner soren@debian.org
[toc] | [prev] | [standalone]
Back to top | Article view | linux.debian.maint.python
csiph-web