Comments

403 Comments
User Icon music commented & provided feedback on uv-0.5.5-2.fc41 12 months ago

This is, formally speaking, an incompatible update, although most users should not find it disruptive. The uv package has a permanent Updates Policy exception.

User Icon music commented & provided feedback on uv-0.5.5-2.fc40 12 months ago

Upstream pull request #9424 for Windows batch script activation turned out to be incorrect. For that reason, I’m going to modify this update with a fresh build, in which the backport of that PR is reverted. As long as I’m doing a new build, this will be uv-0.5.5, which came out just after I originally created this update.

User Icon music commented & provided feedback on uv-0.5.5-2.fc41 12 months ago

Upstream pull request #9424 for Windows batch script activation turned out to be incorrect. For that reason, I’m going to modify this update with a fresh build, in which the backport of that PR is reverted. As long as I’m doing a new build, this will be uv-0.5.5, which came out just after I originally created this update.

@kkeithle

Since this depends on python-wrapt and python-deprecated, would you please consider combining it into a multi-build update from a side tag so that all three updates hit the EPEL10 repos at the same time? We just finished this process for python-wrapt and python-deprecated, as discussed in FEDORA-EPEL-2024-25d5d4e449, to create the new update FEDORA-EPEL-2024-2ce4c21cfc.

To add python-pikepdf to that update, you would need to:

  1. Unpush this update. (You can do this last instead of first, but it’s necessary in order to re-use the same build in a different update.)
  2. Run koji tag epel10.0-build-side-101152 python-pikepdf-8.8.0-4.el10_0. (You don’t have to be the one who does this, but you might as well be.)
  3. If another build in the side tag needed to rely on python-pikepdf, someone (anyone) would need to run koji wait-repo epel10.0-build-side-101152 --build=python-pikepdf-8.8.0-4.el10_0 --request, since side-tag repo regeneration is no longer automatic. In this case, this step can be skipped unless we end up trying to add python-cairocffi to the update too.
  4. Let me know that you’ve done the above, so that I can “refresh” the list of builds in FEDORA-EPEL-2024-2ce4c21cfc from the side tag, modify the description, and associate #2325908 with the update.

Here’s the replacement update: FEDORA-EPEL-2024-2ce4c21cfc

I’ll try to coordinate getting python-pikepdf added to it as well.

Thanks!

I unpushed FEDORA-EPEL-2024-4bbcbd6c6d, then I requested an EPEL10 side tag (epel10.0-build-side-101152) and ran

$ koji tag epel10.0-build-side-101152 python-wrapt-1.17.0-2.el10_0
$ koji wait-repo epel10.0-build-side-101152 --build=python-wrapt-1.17.0-2.el10_0 --request

so python-wrapt is now available for other builds in the side tag.

You can now either unpush this update and run

$ koji tag epel10.0-build-side-101152 python-deprecated-1.2.14-7.el10_0
$ koji wait-repo epel10.0-build-side-101152 --build=python-deprecated-1.2.14-7.el10_0 --request

or you can build a later NVR of python-deprecated with fedpkg build --target=epel10.0-build-side-101152. You or someone else will still need to run the koji wai-repo … command on the resulting build before it‘s available for any other builds to use, since side-tag repo regeneration isn’t automatic anymore. In this case, with a newer python-deprecated build, I think creating a new update from the side tag should obsolete this update – although unpushing it doesn’t hurt.

This update has been unpushed.

Unpushing in order to create a multi-build update with at least python-deprecated, and hopefully also python-pikepdf.

Ok, let’s do it.

I wonder if we could get @kkeithle to do the same for FEDORA-EPEL-2024-ce568eaa56, which depends on python-deprecated. Let’s get python-wrapt and python-deprecated set up first.

I’ll unpush python-wrapt, create a side tag, tag in the python-wrapt build, and follow up here.

Note that the positive karma does not re-enable auto-push to stable after negative karma disabled it. This update will need to be manually pushed to stable, or auto-push by karma and/or time will need to be manually re-enabled by @mfocko.

Either way, please ensure this reaches stable slightly after python-wrapt, so it doesn’t fail to install even briefly.

Ideally these would have been a multi-build update created from a side tag to begin with. If you want to, you could unpush this update, and I could unpush the python-wrapt update, and then I could tag both builds into a new side tag and create a combined update from that. I have done that before, and it works well. Just let me know if this is what you want.

karma

The upstream release notes for the packaged version claim it fixes the mentioned CVE’s, and the command-line tool passed a quick “smoke test” in a mock chroot.

BZ#2327536 CVE-2024-48990 needrestart: arbitrary code execution via PYTHONPATH environment variable [fedora-41]
BZ#2327541 CVE-2024-11003 needrestart: local privilege escalation via unsanitized input [fedora-41]
BZ#2327546 CVE-2024-48992 needrestart: arbitrary code execution via RUBYLIB environment variable [fedora-41]
BZ#2327553 CVE-2024-48991 needrestart: arbitrary code execution via race condition [fedora-41]

It seems like https://bugzilla.redhat.com/show_bug.cgi?id=2327553 should be associated with this update, too.

karma

The upstream release notes for the packaged version claim it fixes the mentioned CVE’s, and the command-line tool passed a quick “smoke test” in a mock chroot.

BZ#2327531 CVE-2024-48990 needrestart: arbitrary code execution via PYTHONPATH environment variable [epel-8]
BZ#2327537 CVE-2024-11003 needrestart: local privilege escalation via unsanitized input [epel-8]
BZ#2327542 CVE-2024-48992 needrestart: arbitrary code execution via RUBYLIB environment variable [epel-8]
BZ#2327549 CVE-2024-48991 needrestart: arbitrary code execution via race condition [epel-8]
karma

The upstream release notes for the packaged version claim it fixes the mentioned CVE’s, and the command-line tool passed a quick “smoke test” in a mock chroot.

BZ#2124940 Build needrestart for EPEL9
BZ#2327532 CVE-2024-48990 needrestart: arbitrary code execution via PYTHONPATH environment variable [epel-9]
BZ#2327538 CVE-2024-11003 needrestart: local privilege escalation via unsanitized input [epel-9]
BZ#2327543 CVE-2024-48992 needrestart: arbitrary code execution via RUBYLIB environment variable [epel-9]
BZ#2327550 CVE-2024-48991 needrestart: arbitrary code execution via race condition [epel-9]
karma

The upstream release notes for the packaged version claim it fixes the mentioned CVE’s, and the command-line tool passed a quick “smoke test” in a mock chroot.

BZ#2327534 CVE-2024-48990 needrestart: arbitrary code execution via PYTHONPATH environment variable [fedora-40]
BZ#2327540 CVE-2024-11003 needrestart: local privilege escalation via unsanitized input [fedora-40]
BZ#2327545 CVE-2024-48992 needrestart: arbitrary code execution via RUBYLIB environment variable [fedora-40]
BZ#2327552 CVE-2024-48991 needrestart: arbitrary code execution via race condition [fedora-40]
karma

The upstream release notes for the packaged version claim it fixes the mentioned CVE’s, and the command-line tool passed a quick “smoke test” in a mock chroot.

BZ#2327533 CVE-2024-48990 needrestart: arbitrary code execution via PYTHONPATH environment variable [fedora-39]
BZ#2327539 CVE-2024-11003 needrestart: local privilege escalation via unsanitized input [fedora-39]
BZ#2327544 CVE-2024-48992 needrestart: arbitrary code execution via RUBYLIB environment variable [fedora-39]
BZ#2327551 CVE-2024-48991 needrestart: arbitrary code execution via race condition [fedora-39]