Re: git: 0e0919db966a - stable/14 - periodic.conf: remove long deprecated security_daily_compat_var()

From: Enji Cooper <yaneurabeya_at_gmail.com>
Date: Thu, 06 Jun 2024 14:24:52 UTC
> On Jun 6, 2024, at 00:20, Michael Osipov <michaelo@freebsd.org> wrote:
> 
> On 2024-06-05 17:55, Enji Cooper wrote:
>>>> On Jun 5, 2024, at 1:07 AM, Michael Osipov <michaelo@FreeBSD.org> wrote:
>>> 
>>> The branch stable/14 has been updated by michaelo:
>>> 
>>> URL: https://cgit.FreeBSD.org/src/commit/?id=0e0919db966a83efc696d19b39a5962500e6ad46
>>> 
>>> commit 0e0919db966a83efc696d19b39a5962500e6ad46
>>> Author:     Michael Osipov <michaelo@FreeBSD.org>
>>> AuthorDate: 2024-04-15 11:05:17 +0000
>>> Commit:     Michael Osipov <michaelo@FreeBSD.org>
>>> CommitDate: 2024-06-05 08:05:54 +0000
>>> 
>>>    periodic.conf: remove long deprecated security_daily_compat_var()
>>> 
>>>    This function is documented to be gone in after 11. Time to remove this
>>>    compat shim.
>>> 
>>>    PR:             275296
>>>    Reviewed by:    jrm (mentor)
>>>    MFC after:      1 month
>>>    Differential Revision:  https://reviews.freebsd.org/D44796
>>> 
>>>    (cherry picked from commit aa48259f337100e79933d660fec8856371f761ed)
>> It’s kind of unfortunate that this didn’t just break periodic for folks that rely on these compat variables prior to their removal.
> 
> Enji,
> 
> we talked about this with Baptiste a couple of weeks ago. He fixed pkg and deep-analyzed all ports for possible issues. All clean and we so no issue backporting it. Also the removed function has been issuing a warning on every call, did you miss it?

Hi Michael,
    I haven’t missed these messages in the past (I pay WAYYY too much attention to console noise 😂). My concern is that I’ve seen warnings like these be missed by others (in particular when FreeBSD is used in a dev/embedded environment), since issues like this are lost in the overall “sea of noise” that gets dumped out to the console (periodic output doesn’t even go to the console by default since it’s invoked by cron, so it’s sort of worse…).
    One way to mitigate this issue (in my mind) is to document this removal in the Release Notes, so maybe (hopefully!) someone might read the RNs, notice the change, and adjust their configs to use the updated variables 🙂.
Cheers!
-Enji