Details
Diff Detail
- Repository
- R9 FreeBSD doc repository
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
| website/content/en/releases/14.4R/errata.adoc | ||
|---|---|---|
| 66–68 | Can we put the simplest and clearest overview of the behavior people will see up front? something like "Systems using kernels_autodetect will boot and get a fallback kernel if the default kernel is a symlink", but I'm not sure if I understand the issue. The bug thread can describe the mechanics, but relnote should to be most accessible because everybody is expected to read these. | |
| 67 | Is automatic kernel detection broken for all use cases or disable automatic kernel detection only if they encounter the behavior? | |
| website/content/en/releases/14.4R/errata.adoc | ||
|---|---|---|
| 66–68 | here is just suggestion how to describe from the user point of view: | |
| 67 | also suggestion how to clarify: | |
Vladlen's suggested text would address my concerns. @leres can you do this before the release announcement goes out in ten hours? If I hear nothing and no one objects, I will commit the one I wrote already in a few hours.
| website/content/en/releases/14.4R/errata.adoc | ||
|---|---|---|
| 66–68 |
Your version leaves out "regression" which I included because this use worked before. | |
| 67 |
It's broken when symlinks are used. | |
| 67 |
I don't understand "If this behavior is encountered", the workaround applies to the issue being described. | |
Your version leaves out "regression" which I included because this use worked before.
I'm not attached to my version, we just need to have something that's clear. It's fine to include the word regression, however the sole point of this section is listing regressions.
It's broken when symlinks are used.
Then it needs to say that clearly, like I wrote several different ways.
I don't understand "If this behavior is encountered", the workaround applies to the issue being described.
The target audience for this doc is "everyone who may install this freebsd". So it needs to be very specific. Without that, you are telling everyone using kernels_autodetect to disable it so they can install the system. This is too heavy handed and increases the cost of deploying this release for no reason.
My version already has feedback address via email:
On 3/8/26 11:21, Colin Percival wrote:
I would say "A regression in the 14.4-RELEASE boot loader
scripts" (assuming I
have that detail correct?) to clarify for people who might not have the
boot
loader, kernel, and userland all in sync.
It's broken when symlinks are used.
Then it needs to say that clearly, like I wrote several different ways.
Is "use of a symlink to select the kernel tree to boot" not clear?
I don't understand "If this behavior is encountered", the workaround applies to the issue being described.
The target audience for this doc is "everyone who may install this freebsd". So it needs to be very specific. Without that, you are telling everyone using kernels_autodetect to disable it so they can install the system. This is too heavy handed and increases the cost of deploying this release for no reason.
The workaround is a part of an issue in the Open Issues section, I guess I don't see how anyone would apply it to installation when it doesn't apply (no symlink for /boot/kernel or the wrong kernel being booted -- which is how I discovered the issue).
Is "use of a symlink to select the kernel tree to boot" not clear?
Craig, you have the word boot four times in one sentence.
I am open to other ways to simplify this, the text you have currently is not acceptable to me for user facing doc.
The workaround is a part of an issue in the Open Issues section,
One example of a reasonable workflow for someone who isn't a high level engineer is to read the Open Issues section and apply the workarounds if possible to cover their asses so they don't break the installation.
If someone doesn't understand the problem you're talking about because you're speaking more technically than they understand, they better apply the workaround or it's their fault if it explodes, does that make sense?
I think that's a bit harsh. I agree that it's a bit wordy but there's a gulf between "not worded as clearly as possible" and "not acceptable".
@ziaee I'm thinking of rewording that as
2. **man:loader_lua[8]** A regression in the boot loader scripts causes the wrong kernel to be selected of /boot/kernel is a symlink +
Does that sound ok to you?
Does that sound ok to you?
Yes, thank you so much. That is even simpler and clearer than what I wrote, which is all I wanted.
Note, there's a typo s/of/if/
| website/content/en/releases/14.4R/errata.adoc | ||
|---|---|---|
| 71 | I'm not sure what the point of this macro is, but we do have this as well. | |
Thanks, [.filename] highlights the text in green and avoids the need to have a space between the path and the terminating period.
Note that https://bugs.freebsd.org/293654 has been solved by D55713. (I'm not clear on how to insure it is included in 15.1)
We'll make sure that gets MFCed to stable/15 before the end of April. And we should be able to get an Errata Notice for 14.4 as well.
(I started with periods but then copied the style of https://www.freebsd.org/releases/15.0R/errata/#open-issues and removed them)