• RE: Debian - Release Cadence Options

    From Greg McPherran@3:633/10 to All on Thursday, October 16, 2025 18:10:01
    Hi,

    I certainly understand the stable model. However, if I may suggest considering a cadence option similar to Ubuntu and Fedora that provide an official release every 6 months or so.

    I prefer more up-to-date kernels and packages (without requiring backports).

    Simply a friendly suggestion. I like Debian and your minimal installation options such as mini.iso and netinst. (I originally developed web engine on Debian in 2014).

    Have Great Day,
    Greg McPherran
    (McPherran Web Engine)





    --- PyGate Linux v1.0
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Mechtilde Stehmann@3:633/10 to All on Thursday, October 16, 2025 18:40:02
    Hello Greg,

    Am 16.10.25 um 18:00 schrieb Greg McPherran:
    Hi,

    I certainly understand the stable model. However, if I may suggest considering a cadence option similar to Ubuntu and Fedora that provide an official release every 6 months or so.

    Yes you can have a stable release nearly every two years with support
    for three years.

    You can get more up-to-date version of packages using backports. So the packages you want to have more up-to-date work at your stable system

    Or you can get recent packages in testing with all the risks on your side.


    I prefer more up-to-date kernels and packages (without requiring backports).

    Simply a friendly suggestion. I like Debian and your minimal installation options such as mini.iso and netinst. (I originally developed web engine on Debian in 2014).

    Have Great Day,
    Greg McPherran
    (McPherran Web Engine)
    Have a nice day

    --
    Mechtilde Stehmann
    ## PGP encryption welcome
    ## F0E3 7F3D C87A 4998 2899 39E7 F287 7BBA 141A AD7F



    --- PyGate Linux v1.0
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Colin Watson@3:633/10 to All on Friday, October 17, 2025 14:40:01
    On Thu, Oct 16, 2025 at 12:00:11PM -0400, Greg McPherran wrote:
    I certainly understand the stable model. However, if I may suggest >considering a cadence option similar to Ubuntu and Fedora that provide
    an official release every 6 months or so.

    The release cadence is basically driven by what's practical given the resources Debian has. We've managed a two-year cadence pretty reliably
    for 20 years now. It's easy to think of benefits of a faster cadence,
    but a faster cadence would also incur at least the following extra
    problems:

    * it would be more difficult to organize larger projects that need to
    be completed in a single release cycle (e.g. the 64-bit time_t
    transition in trixie)

    * freezes would need to be much quicker in order that developers have
    time to do ordinary work in non-frozen periods, which would require
    more people to fix release-critical bugs even more aggressively than
    we already do

    * teams that are heavily involved in the release process (release,
    archive management, installer, etc.) would have more frequent work to
    do and would get fewer breaks

    * offering an equivalent security support period with more frequent
    stable releases would multiply the work that the security team has to
    do, since more stable releases would have to be supported at once

    Canonical solves some of these problems for Ubuntu by paying more people
    to do things (although I can certainly tell you from personal experience
    that this doesn't make the resulting burnout go away). I assume Red Hat
    does similarly for Fedora. There are people paid to work at least some
    of the time on Debian, but it isn't really the same model and Debian's
    own resources would not stretch to very much of that on an ongoing
    basis. So you're pretty much left with the established compromise that Debian's core teams have found to be sustainable over the long haul.
    I'm not sure what it would take to change that, but I doubt it would be
    any single simple change.

    Regards,

    --
    Colin Watson (he/him) [cjwatson@debian.org]

    --- PyGate Linux v1.0
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Jonathan Dowland@3:633/10 to All on Friday, October 17, 2025 15:50:01
    We currently enjoy 3 years of project support for a release on the
    current 2-year cadence.

    If we were cutting new releases every 6 months, what kind of support
    window would you want?

    I wouldn't want to be forced to upgrade more often than I do now.


    --
    Jonathan Dowland
    jmtd@debian.org
    https://jmtd.net

    --- PyGate Linux v1.0
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From thomas@3:633/10 to All on Friday, October 17, 2025 15:50:01


    On Oct 16, 2025 18:09, Greg McPherran <gm@mcpherranweb.com> wrote:

    Hi,

    I certainly understand the stable model. However, if I may suggest considering a cadence option similar to Ubuntu and Fedora that provide an official release every 6 months or so.
    Hell, no ! It's already too demanding to upgrade servers every 2 years, plus backporting security patches is already too much work that we don't want to impose on ouraeves 4 times more.

    Thomas Goirand (zigo)


    --- PyGate Linux v1.0
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Greg McPherran@3:633/10 to All on Friday, October 17, 2025 17:30:01
    Hell, no ! It's already...
    Language please. I specifically indicated that it was merely a _friendly_ suggestion. :-) Impressed with your work and others' work to provide Debian Stable and I'm also aware of shiny new toy syndrome. :-) For server, I would use stable! But for desktop and development, I like the latest available as it's constantly improved, and with Wayland support, this is especially important. It's not the days of decades of X11 and minor updates to it. We're in an important transition phase to a new desktop environment platform.

    I am considering using unstable, and would be happy to provide feedback on any findings. If someone could point me to the best contact for such feedback, so that I may be helpful, I would be grateful. (Introduce myself: I'm particularly focused on Rust now, but I have a background in kernels and C/C++/C# [30 yrs exp.]. I also have BSEE + post-grad MS courses).

    Sincerely,
    Greg McPherran

    On 10/17/2025 8:50 AM EDT thomas@goirand.fr wrote:




    On Oct 16, 2025 18:09, Greg McPherran <gm@mcpherranweb.com> wrote:

    Hi,

    I certainly understand the stable model. However, if I may suggest considering a cadence option similar to Ubuntu and Fedora that provide an official release every 6 months or so.

    Hell, no ! It's already too demanding to upgrade servers every 2 years, plus backporting security patches is already too much work that we don't want to impose on ouraeves 4 times more.

    Thomas Goirand (zigo)




    --- PyGate Linux v1.0
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Russ Allbery@3:633/10 to All on Friday, October 17, 2025 19:00:01
    Greg McPherran <gm@mcpherranweb.com> writes:

    I am considering using unstable, and would be happy to provide feedback
    on any findings. If someone could point me to the best contact for such feedback, so that I may be helpful, I would be grateful.

    If you run into problems with unstable or testing, you should report them
    as bugs against the relevant package with reportbug. That way we can track
    them and try to get them fixed for the next release.

    I would recommend testing over unstable unless you're very familiar with
    Linux and comfortable with your ability to roll back to previous kernels, downgrade and pin packages, and fix weird problems. Unstable doesn't break
    that often, but it is prone to more low-level breakage than testing is.

    --
    Russ Allbery (rra@debian.org) <https://www.eyrie.org/~eagle/>

    --- PyGate Linux v1.0
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Antoine Le Gonidec@3:633/10 to All on Saturday, October 18, 2025 00:40:01
    Le Fri, Oct 17, 2025 at 09:49:20AM -0700, Russ Allbery a ‚crit :
    Greg McPherran <gm@mcpherranweb.com> writes:
    I am considering using unstable, and would be happy to provide feedback
    on any findings. If someone could point me to the best contact for such feedback, so that I may be helpful, I would be grateful.
    (?)
    I would recommend testing over unstable unless you're very familiar with Linux and comfortable with your ability to roll back to previous kernels, downgrade and pin packages, and fix weird problems. Unstable doesn't break that often, but it is prone to more low-level breakage than testing is.
    And I would recommend unstable over testing, for security reasons, less
    risks of packages disappearing, and bugs actually being fixed in a
    timely manner.
    In my 15 years of user support, a huge majority of reported problems
    were with testing. But the sample might be biased by testing being the
    one chosen by less experienced users, wrongly thinking it would be some
    kind of middle ground between stable and unstable.


    --- PyGate Linux v1.0
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From tomas@3:633/10 to All on Saturday, October 18, 2025 07:00:01
    On Sat, Oct 18, 2025 at 12:17:53AM +0200, Antoine Le Gonidec wrote:
    Le Fri, Oct 17, 2025 at 09:49:20AM -0700, Russ Allbery a ‚crit :
    Greg McPherran <gm@mcpherranweb.com> writes:
    I am considering using unstable, and would be happy to provide feedback on any findings. If someone could point me to the best contact for such feedback, so that I may be helpful, I would be grateful.
    (?)
    I would recommend testing over unstable unless you're very familiar with Linux and comfortable with your ability to roll back to previous kernels, downgrade and pin packages, and fix weird problems. Unstable doesn't break that often, but it is prone to more low-level breakage than testing is.

    And I would recommend unstable over testing, for security reasons, less
    risks of packages disappearing, and bugs actually being fixed in a
    timely manner.

    In my 15 years of user support, a huge majority of reported problems
    were with testing. But the sample might be biased by testing being the
    one chosen by less experienced users, wrongly thinking it would be some
    kind of middle ground between stable and unstable.
    In a way, it is: when packages land in testing they have passed through
    a rough first test. So in a way, you get less surprises in testing.
    In some other way... it's the other way around: packages "disappear" from testing when some bigger issue is found, but they stay on stable. Plus,
    since both don't have a security policy, security issues get (arguably)
    fixed faster on testing.
    If you (try to) install packages from a future suite (FrankenDebian),
    then testing is more often the better bet. If you run a future suite
    wholesale, the situation is not so clear. The difference to a rolling
    release is that in the latter case, more effort is put to avoid
    breakage.
    Cheers
    --
    tom s


    --- PyGate Linux v1.0
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Russ Allbery@3:633/10 to All on Saturday, October 18, 2025 12:40:02
    Antoine Le Gonidec <vv221@debian.org> writes:

    In my 15 years of user support, a huge majority of reported problems
    were with testing. But the sample might be biased by testing being the
    one chosen by less experienced users, wrongly thinking it would be some
    kind of middle ground between stable and unstable.

    I think so, but of course I don't really know.

    All I want to be sure the original poster realizes is that if you use
    unstable, your system *will* break sometimes in ways that you will have to manually fix. I have a pretty light footprint on a system and things tend
    to weirdly break less for me than for other people, and I still have to semi-manually unbreak something at least a couple of times a year. (I also
    run a testing system and have basically never had a problem with it, but
    this is just one possibly unrepresentative data point.)

    If you know how to do the things that I mentioned, unstable can be a great experience. I've used unstable on some of my systems for something like 20 years now. But I'm pretty unphased by having to manually clean up
    something or file a bug report and downgrade a package.

    --
    Russ Allbery (rra@debian.org) <https://www.eyrie.org/~eagle/>

    --- PyGate Linux v1.5
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Marc Haber@3:633/10 to All on Saturday, October 18, 2025 12:40:02
    On Fri, Oct 17, 2025 at 10:40:53AM -0400, Greg McPherran wrote:
    I am considering using unstable, and would be happy to provide feedback on any findings. If someone could point me to the best contact for such feedback, so that I may be helpful, I would be grateful. (Introduce myself: I'm particularly focused on Rust now, but I have a background in kernels and C/C++/C# [30 yrs exp.]. I also have BSEE + post-grad MS courses).

    If you can fix a broken linux, then using unstable would also be "giving
    back" to the project. We desperately need experienced user using our unreleased development version and report bugs with high quality so we
    can fix them before the release.

    I have been using unstable on my personal workstations for decades with
    good experiences. Have a rescue system like grml available, have a
    backup, maybe even a spare machine if things burn down badly (never
    happned for me). One of my machines has btrfs and I do snapshots before upgrading, and it has always been easier to select out broken packages
    and pull those from testing to fix breakage.

    I do not recommend testing unless in the very last weeks before a
    release for security reasons, and when you're using testing you cannot
    easily go back to older package versions because there is only
    snapshot.d.o to pull from.

    Greetings
    Marc

    -- ----------------------------------------------------------------------------- Marc Haber | "I don't trust Computers. They | Mailadresse im Header Leimen, Germany | lose things." Winona Ryder | Fon: *49 6224 1600402 Nordisch by Nature | How to make an American Quilt | Fax: *49 6224 1600421

    --- PyGate Linux v1.5
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Greg McPherran@3:633/10 to All on Saturday, October 18, 2025 16:10:01

    There's no language problem here.
    It's generally a religious term, and the religious meaning is probably what gives it such emphasis, so, out of respect for religion(s), I generally avoid such terms. Seems to be a pattern in GNU'ish Linux:

    GNOME: Adwaita - Sanskrit - oneness - with the Supreme
    GNOME: epiphany - manifestation of a divine being
    "Bourne Again" shell
    St. Ignucious
    ...
    ...

    I've coresponded with Mr. Stallman over the years. He appears to be an excellent software engineer.

    If Linux wants to grow and compete in the actual professional marketplace, I suggest professional terminology that stays away from using religous terminology.

    It was a friendly opinion. Please assume good faith.
    Yes, I basically understood, although when one has never met someone, it can add a bit of harshness to the opinion. I am simply sharing yet another suggestion here. It's a free country, speak as you wish, and we also have freedom to ignore/disengage.

    Rightly. Or testing maybe?
    Sincerely and respectfully to you and all Debian, I'm typing from Arch using GNOME Web (same base as Safari). The latest GNOME is excellent!

    It dawned on me that Debian's main market is probably server, and absolutely, as I had said, the stability for server is to praised and appreciated. I always wanted to have the same distribution for desktop and server but I see that may not be neccesary. On my dedicated server, I still have Trixie. One thing I don't like is different package managers, though.

    Cheers,
    Thanks for responding, and Cheers to you :-)

    --- PyGate Linux v1.5
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Martin Dosch@3:633/10 to All on Saturday, October 18, 2025 16:10:01
    On Sat Oct 18, 2025 at 00:17 CEST, Antoine Le Gonidec wrote:
    In my 15 years of user support, a huge majority of reported problems
    were with testing. But the sample might be biased by testing being the
    one chosen by less experienced users, wrongly thinking it would be some
    kind of middle ground between stable and unstable.
    It sort of is a middle ground between stable and unstable. You have
    newer versions than stable and a sort of rolling release but the chance
    of getting hit by breaking bugs is a lot lower compared to using
    unstable.
    I am using testing for ~15 years on my main laptop and it never happened
    that my system totally broke or was unbootable, which sometimes can
    happen in unstable. If there is a security issue or a bug, there often
    is already a fixed version in unstable which I could easily cherrypick
    (I always have testing and unstable in my sources.list with a pinning to
    prefer testing).
    Best regards,
    Martin


    --- PyGate Linux v1.5
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Antoine Le Gonidec@3:633/10 to All on Saturday, October 18, 2025 16:10:01
    Le Sat, Oct 18, 2025 at 04:23:07AM +0200, tomas@tuxteam.de a ‚crit :
    (?) Plus,
    since both don't have a security policy, security issues get (arguably)
    fixed faster on testing.
    Beware, if you?re talking unstable vs. testing here, testing is the one
    with the worst security support (= none at all) of all Debian branches. Security fixes are provided in unstable as part of packaging new
    upstream releases, and cherry-picked to stable through the dedicated
    security channel. On testing, security fixes do not go through a special channel: they trickle down from unstable following the usual rules, so
    with at least a 3 to 5 days delay that can grow to weeks or even months if you?re unlucky.


    --- PyGate Linux v1.5
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Andrey Rakhmatullin@3:633/10 to All on Saturday, October 18, 2025 17:00:01
    On Sat, Oct 18, 2025 at 04:23:07AM +0200, tomas@tuxteam.de wrote:
    since both don't have a security policy, security issues get (arguably)
    fixed faster on testing.
    I think this is either some unusual phrasing or a mistake.

    --
    WBR, wRAR


    --- PyGate Linux v1.5
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From tomas@3:633/10 to All on Saturday, October 18, 2025 17:00:01
    On Sat, Oct 18, 2025 at 02:24:40PM +0200, Antoine Le Gonidec wrote:
    Le Sat, Oct 18, 2025 at 04:23:07AM +0200, tomas@tuxteam.de a ‚crit :
    (?) Plus,
    since both don't have a security policy, security issues get (arguably) fixed faster on testing.
    Uh... sorry. I wanted to write unstable :-/
    Beware, if you?re talking unstable vs. testing here, testing is the one
    with the worst security support (= none at all) of all Debian branches.
    Yes. That's what I meant to write, sorry.
    Security fixes are provided in unstable as part of packaging new
    upstream releases, and cherry-picked to stable through the dedicated
    security channel. On testing, security fixes do not go through a special channel: they trickle down from unstable following the usual rules, so
    with at least a 3 to 5 days delay that can grow to weeks or even months if you?re unlucky.
    100% agree :-)
    Cheers
    --
    t


    --- PyGate Linux v1.5
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From tomas@3:633/10 to All on Saturday, October 18, 2025 18:40:02
    On Sat, Oct 18, 2025 at 07:33:33PM +0500, Andrey Rakhmatullin wrote:
    On Sat, Oct 18, 2025 at 04:23:07AM +0200, tomas@tuxteam.de wrote:
    since both don't have a security policy, security issues get (arguably) fixed faster on testing.

    I think this is either some unusual phrasing or a mistake.
    A mistake indeed -- see my other post :)
    Cheers
    --
    t


    --- PyGate Linux v1.5
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)
  • From Mathias Behrle@3:633/10 to All on Sunday, October 19, 2025 02:30:01
    * Antoine Le Gonidec: " Re: Debian - Release Cadence Options" (Sat, 18 Oct 2025
    14:24:40 +0200):
    Le Sat, Oct 18, 2025 at 04:23:07AM +0200, tomas@tuxteam.de a ‚crit :
    (?) Plus,
    since both don't have a security policy, security issues get (arguably) fixed faster on testing.

    Beware, if you?re talking unstable vs. testing here, testing is the one
    with the worst security support (= none at all) of all Debian branches.

    Security fixes are provided in unstable as part of packaging new
    upstream releases, and cherry-picked to stable through the dedicated
    security channel. On testing, security fixes do not go through a special channel: they trickle down from unstable following the usual rules, so
    with at least a 3 to 5 days delay that can grow to weeks or even months if you?re unlucky.
    I am using testing with debsecan which for me combines the best of both worlds.

    --
    Mathias Behrle
    PGP/GnuPG key availabable from any keyserver, ID: 0xD6D09BE48405BBF6
    AC29 7E5C 46B9 D0B6 1C71 7681 D6D0 9BE4 8405 BBF6


    --- PyGate Linux v1.5
    * Origin: Dragon's Lair, PyGate NNTP<>Fido Gate (3:633/10)