Loading today’s stories
Build1 distinct publisher3 min readPublished
The release-notes table read on 26 August 2026 puts Unagi on Scarthgap and Vimba on Wrynose. One has about 20 months of upstream fixes left; the other has 44.
The Engineer · Build desk
Compiled by The EngineerSomething wrong?How this is made
Every one of those series is named after a Yocto release, and the name carries a schedule with it. A distribution assembled with OpenEmbedded inherits its recipe set, its toolchain and its CVE-patching cadence from the Yocto release it sits on, and when that release goes end of life the recipes stop receiving fixes while the backporting work moves to whoever is shipping the product [10].
So the number worth writing down is not the version string but the remainder. Scarthgap 5.0 arrived in April 2024 with long-term support to April 2028 [11]. Read on 26 August 2026 [1], that leaves roughly 20 months of upstream fixes [16]. Wrynose 6.0, released April 2026 and supported to April 2030 [12], leaves about 44 [17], more than twice as much [18]. Both windows are 48 months wide [19], so picking a branch does not buy a longer window, only an earlier place in one. Kirkstone marks the far end: released May 2022, last documented point release 4.0.33 in February 2026, now flagged end of life [13], about 45 months of service [21].
Then there is what "Latest Stable Release" means on that page. Vibrant Vimba holds the heading [5], but it is fetched with `repo init -b master` plus a milestone manifest, and unlike Unagi and Quillback it has no release branch of its own [7]. Its two documented builds are numbered v21.91.0 and v21.92.0 while the Jira filter linked from both reads "Vibrant Vimba 22.0.0" [6]. The manifest still pins revisions, so the build reproduces; what it does not give you is a line that has stopped moving underneath you, which Unagi does [c7b]. Nor is Unagi stale: four point releases are documented, and v21.0.3 tracks scarthgap 5.0.19 and adds the SiFive hifive-premier-p550 board [8]. The shorter window is upstream arithmetic, not neglect.
The old-stable series has a different defect. AGL's build-artifacts note for Quirky Quillback reads "Uses Yocto 4.0.35 (YP kirkstone)" [9], while the Yocto releases table read the same day stops Kirkstone at 4.0.33 [13]: two point releases of daylight between the two pages [20]. Either reading puts a 17.x programme on a dead baseline, so the practical answer does not change. What changes is your confidence that one page is a record of the other.
The non-LTS entries are why the LTS distinction is worth checking rather than assuming. Walnascar 5.2 has been end of life since November 2025, and Whinlatter 5.3 is finished as well [14]. A series named after either would have looked current in an announcement and been unsupported inside a year of shipping.
None of this makes the structure wrong. Large open source projects carry several series at once deliberately [22]. It does mean the release notes hand you one third of a production baseline. The other two thirds, by the definition on that same page, are validation evidence and a maintenance plan naming who fixes what and for how long [15]. The manifest you can fetch this afternoon. The maintenance plan is the document a supplier signs, and the Yocto dates set its floor.
Ranked by verification strength, evidence, and original report placement.
The two baselines an engineer could land on have Yocto support windows that end two years apart, and a vehicle programme outlives both.
Vibrant Vimba is fetched with repo init -b master -m vimba_21.92.0.xml and has no release branch of its own, whereas Unagi is fetched with -b unagi and Quillback with -b quillback.
The branch Vimba sits on is still the one under active development, whereas Unagi gives both a manifest and a branch that is not moving underneath you.
A distribution built with OpenEmbedded inherits its recipe set, its toolchain and its CVE-patching cadence from the Yocto release it is based on; when that release goes end of life the recipes stop receiving fixes and the burden of backporting moves to whoever is shipping the product.
As of 26 August 2026 the Automotive Grade Linux release-notes page documents three listed release series on three different Yocto Project releases: Ultimate Unagi on Scarthgap 5.0.x, Vibrant Vimba on Wrynose 6.0.x, and old-stable Quirky Quillback on Kirkstone 4.0.x, which the Yocto Project now marks end of life.
The release-notes wiki page at wiki.automotivelinux.org/agl-distro/release-notes records its own last modification as 24 August 2026.
Follow any of these and your For You feed starts watching them — no settings page required.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Primary artefacts quoted with read dates, but one publisher and one unverified hook
Most factual load is carried by verbatim strings from two primary artefacts (the AGL release-notes wiki, self-reported last modified 24 August 2026, and the Yocto Project releases table) each with an explicit 26 August 2026 read date, plus reproduced repo init command lines. That is unusually checkable for a single-author post. It is capped by there being exactly one publisher in the cluster, by the central channel-divergence assertion never being shown, and by an internal inconsistency the piece does not reconcile (Kirkstone 4.0.35 versus 4.0.33).
Documented release cadence across three series, no deployment or usage data
There is real, dated release-artefact evidence: four Unagi point releases tracking Scarthgap 5.0.18/5.0.19, two Vimba milestone builds on Wrynose, a still-listed Quillback old-stable series, and the corresponding upstream Yocto release/EOL states. What is entirely absent is downstream adoption: no OEM, tier-one or programme is named, no shipment, deployment or usage disclosure appears, and the assertion that Quillback-validated programmes remain in the field is unquantified. Scored low rather than insufficient because release and lifecycle observations exist and are dated.
Mildly overstated hook on an otherwise self-limiting body
The body is notably restrained: it declines to treat multi-series maintenance as a fault, and it explicitly qualifies that the dates are upstream Yocto horizons only, saying nothing about AGL series maintenance, BSP vendor support or programme requirements. The overstatement is localised to the framing device - the title and second section assert that the press release and product page point engineers at a different baseline than the release-notes table, and that specific conflict is never evidenced in the supplied text. Small positive gap rather than large, because the quantitative core (20 versus 44 months, equal 48-month windows) is arithmetic from quoted dates.
Practitioner-authored explainer with self-referential linking and vendor-adjacent authorship
Moderate and visible rather than hidden. The piece is a developer-platform post whose shell prompts carry a training/consultancy domain ([email protected]) and which cross-links the author's own companion article on kernel LTS planning, both of which reward the 'you need help planning baselines' conclusion. Against that, no product, service or paid support offering is pitched in the supplied body, no AGL or Yocto affiliation is claimed, and the article repeatedly narrows its own claims. Nothing in the sources indicates sponsorship or vendor payment, so this is scored as ordinary authority-building incentive.
Checkable artefact quotes, single publisher, truncated body
Confidence is middling. The artefact-level facts and the arithmetic derived from them are quotable and internally consistent, and the read dates make them re-verifiable, which supports the technical core. But the cluster has one publisher and no corroboration, the supplied body is cut off mid-sentence at 'On 14 May 202' so the 'what has moved since May' material cannot be assessed, one internal version discrepancy is unresolved, and the framing claim about conflicting AGL channels is unsupported.
build
A cleanup commit deleted the sanitizer. Five days later a scanner cashed it in.1 distinct publisher
build
The stability step is a branch, not a pipeline: inside one team's release-candidate discipline1 distinct publisher
build
Atlassian's full AI opt-out costs an Enterprise upgrade: three of four tiers can't switch metadata off1 distinct publisher
build
The release board is the schedule: what 74% hybrid adoption actually measures1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
dev.to
1 article · August 27, 2026