From: Richard Purdie <richard.purdie@linuxfoundation.org>
To: Rusty Howell <rustyhowell@gmail.com>
Cc: poky@lists.yoctoproject.org
Subject: Re: [poky] Ability to customize the prserver connection?
Date: Thu, 04 May 2023 13:58:28 +0100 [thread overview]
Message-ID: <380fbc73bede37c82b64ef24b45358f8beed2fa7.camel@linuxfoundation.org> (raw)
In-Reply-To: <CAE+BM3kiMRJa+_TwfP2TP_R6MehgDOaVk1+giJr18sJgCdESBA@mail.gmail.com>
On Wed, 2023-05-03 at 13:06 -0600, Rusty Howell wrote:
> My company has need of proper package feeds that support multiple
> active OS releases. That is, we need to have the ability to work on
> OSv1.1 and OSv2 simultaneously. The problem we are having is that
> PRAUTO responses from the prserver do not have a concept of OS
> release. The later query gets a higher number. So if we build OSv1.1
> after OSv2, that means that OSv1.1 has a package with a greater
> PRAUTO value than it's counterpart in OSv2. In that scenario, our OS
> dist-upgrade will be incomplete. I solved this initially by hacking
> the PRserver to include an int os_version, based on a variable in our
> site.conf. And then I changed the prserver to support this extra
> param with another column in the database. Now queries to the
> prserver consist of a tuple like ("base-files-1.0-r0", "aarch64",
> "12345678978", "11") where "11" is the new os_version param, and the
> OSv2 counterpart ("base-files-1.0-r0", "aarch64", "12345678978",
> "20") where "20" is the new param. The SELECT queries to the
> database find the max value for that package/arch and os_version <=X.
>
> In this manner, any increase in the OSv1.1 package, will cause the
> OSv2 package to get bumped on the next build automatically,
> regardless of a checksum match or not.
> Recently with our Honister upgrade, we had the added complexity of
> changes to the transport protocol (xmlrpc to socket+json) as well as
> simple changes to package.bbclass has caused us redo some things. And
> hence the idea of being able to provide our own custom prserver
> client for bitbake to consume.
> I hope this explains our needs well enough.
That makes sense. It could be done with alternate ports but I'd take a
new field which is used as an ID to make this possible with one port.
Patches welcome.
Cheers,
Richard
prev parent reply other threads:[~2023-05-04 12:58 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-03 17:03 Ability to customize the prserver connection? Rusty Howell
2023-05-03 18:03 ` [poky] " Richard Purdie
2023-05-03 19:06 ` Rusty Howell
2023-05-04 12:58 ` Richard Purdie [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=380fbc73bede37c82b64ef24b45358f8beed2fa7.camel@linuxfoundation.org \
--to=richard.purdie@linuxfoundation.org \
--cc=poky@lists.yoctoproject.org \
--cc=rustyhowell@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).