All of lore.kernel.org
 help / color / mirror / Atom feed
From: Paolo Wattebled <paolo.wattebled@savoirfairelinux.com>
To: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: bitbake-devel <bitbake-devel@lists.openembedded.org>
Subject: Re: [bitbake-devel] [PATCH] toaster: Add recipe variables view
Date: Wed, 19 Aug 2026 15:16:07 -0400 (EDT)	[thread overview]
Message-ID: <1319358081.1239871.1787166966997.JavaMail.zimbra@savoirfairelinux.com> (raw)
In-Reply-To: <58e087a5908a808a3557f92d3844f79b1cce61cd.camel@linuxfoundation.org>

Hi Richard,

Thank you for the feedback, and sorry—the previous approach was too intrusive
in BitBake’s parsing and cache machinery.

I have reworked it so that no recipe-variable data is added to BitBake’s parse
cache or dependency events. Toaster now records the recipe files selected for
the runqueue and, after the build completes, acts as a client of the BitBake
server. It calls a new read-only command for each recipe, which uses BitBake’s
normal parser to create a temporary recipe datastore and returns the resolved
variables to Toaster. Toaster then compresses and stores the result in its own
database.

The history-based selection has also been removed, so indirect changes such as
X = "${Y}" are represented correctly. Functions and values requiring inline
Python expansion are currently omitted to avoid execution side effects, so the
result is a safely expanded subset rather than a byte-for-byte equivalent of
bitbake -e.

This approach does require reparsing the runqueue recipes after the build. In a
preliminary cached imx-image-core test, it added approximately 42 seconds
compared with unmodified Toaster. I am continuing to evaluate that cost.

I hope this provides a better compromise by keeping the data out of BitBake’s
normal cache and build flow while allowing Toaster to query BitBake through a
defined API. I do not currently see a less intrusive way for Toaster to obtain
build-specific recipe values.

If the general approach seems viable to you, I can send the incremental changes
for review.

Thank you again for your feedback.

Have a great day,  
Paolo


      reply	other threads:[~2026-08-19 19:16 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-08-12 15:04 [PATCH] toaster: Add recipe variables view Paolo Wattebled
2026-08-12 20:34 ` [bitbake-devel] " Richard Purdie
2026-08-13 15:31   ` Paolo Wattebled
2026-08-13 15:58     ` Richard Purdie
2026-08-14 14:31       ` Paolo Wattebled
2026-08-14 20:56         ` Richard Purdie
2026-08-17 19:48           ` Paolo Wattebled
2026-08-17 21:08             ` Richard Purdie
2026-08-18 18:33               ` Paolo Wattebled
2026-08-18 19:04                 ` Richard Purdie
2026-08-19 19:16                   ` Paolo Wattebled [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=1319358081.1239871.1787166966997.JavaMail.zimbra@savoirfairelinux.com \
    --to=paolo.wattebled@savoirfairelinux.com \
    --cc=bitbake-devel@lists.openembedded.org \
    --cc=richard.purdie@linuxfoundation.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.