Bug#918341: transition: jemalloc
Control: tags -1 confirmed
On 05/01/2019 11:29, Faidon Liambotis wrote:
> Package: release.debian.org
> Severity: normal
> User: release.debian.org@packages.debian.org
> Usertags: transition
>
> Release Team,
>
> This has been pending for a long time, and while the pieces have been
> mostly there, I've dropped the ball and did not proceed with the
> transition earlier. Apologies for this and for the last minute
> submission!
>
> So, I'd like to ask for permission to upload jemalloc 5.1.0-2 to sid:
>
> - stretch/buster have 3.6.0-11, a pretty old upstream version, and also
> presenting some long-standing issues like #843926 (hard-coded page
> sizes at build time).
>
> - 5.1.0-1 has been in experimental since May 2018, and has successfully
> been built on all official architectures (and all debian-ports except
> hurd-i386 and riscv64). I have not made an upload to unstable, but can
> do so as soon as you give me the green light.
>
> - This newer version uses a slightly different ABI and thus has a new
> SONAME and binary package, libjemalloc2 (vs. libjemalloc1), so this
> requires a transition. API is (mostly) the same -- the exception being
> what was an experimental API that has been now dropped and replaced by
> a stable one.
>
> - Adam Borowski (Cc'ed), who was interested in this and pinged me about
> it recently, ran a rebuild of all rdepds and encountered and reported
> only one FTBFS among packages in testing, #914814 affecting spades,
> which was using the aforementioned experimental API and -as of a few
> minutes ago- has a patch attached.
>
> - Ubuntu apparently imported 5.1.0-1 to their distribution back in May,
> and shipped their Cosmic stable release with it ([1] says "Copied from
> debian experimental in Primary Archive for Debian GNU/Linux by Steve
> Langasek"), apparently with no ill effects(?)
>
> Let me know if there are any questions and thanks for your work and
> consideration!
Please go ahead.
Emilio
Reply to: