devxy/build-cran-binaries
Watch
1
0
Fork
You've already forked build-cran-binaries
 0
Code Issues 12 Pull requests Releases Activity

chore: remove plans and specs design docs (#252)

Browse source 
Removes the internal `plans/` and `specs/` design documents from the repo.

- Delete the 2 plans and 3 specs.
- Drop the README link to the package-patching spec.
- Drop their markdownlint ignores.

No behaviour change.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Reviewed-on: #252
This commit is contained in:
Patrick Schratz 2026-11-14 12:09:25 +00:00 • committed by Patrick Schratz
commit 56dc06734c
7 changed files with 0 additions and 2146 deletions
Download patch file Download diff file Expand all files Collapse all files

3
.markdownlint-cli2.yaml
View file

@ -2,6 +2,3 @@
ignores:
- LICENSE.md
- docs/superpowers/**
# Internal design docs (specs/plans) are not user-facing reference material.
- specs/**
- plans/**

2
README.md
View file

@ -73,8 +73,6 @@ Patching uses a two-tier approach:
The system is implemented in `bincraft`: when a package needs patching, `bincraft` pre-builds it with the patch and serves the patched binary to `pak`, ensuring transitive dependents receive the fixed package.
This way, the fix cascades to all packages that depend on it.
For the design rationale and architecture, see `specs/2026-06-30-package-patching-design.md`.
## Build Environment
Binaries are built on a mixed-architecture Kubernetes cluster using CI.

1477
plans/2026-06-30-package-patching-implementation.md
View file

File diff suppressed because it is too large Load diff

166
plans/2026-08-07-per-minor-edge-routing.md
View file

@ -1,166 +0,0 @@
# Per-R-minor edge routing implementation plan
Spec: `specs/2026-08-07-per-minor-edge-routing-design.md`
**Goal:** let a stock `install.packages()` see the per-minor packages by routing `PACKAGES*` requests to `…/src/contrib/<x.y>/`, where `bincraft` publishes a union index.
**Architecture:** the union is built in `bincraft`; the edge script only redirects index requests, gated on a `UNION_SLOTS` script variable; the script lives in this repo and is applied by OpenTofu.
**Tech stack:** Deno / TypeScript (Bunny Edge Scripting, SDK 0.12), OpenTofu with `BunnyWay/bunnynet` 0.17, R (bincraft).
## Global constraints
- Redirect only `PACKAGES`, `PACKAGES.gz` and `PACKAGES.rds`; never a tarball, because the union index already carries the correct tarball URL for both classes of package.
- Every redirect carries `Cache-Control: no-store`; redirect targets stay UA-independent.
- `UNION_SLOTS` is empty by default, so deploying the script is a no-op until a slot is backfilled.
- A slot is `<arch>/<os>`, e.g. `amd64/alpine324`.
- Verified prerequisites: `PACKAGES*` is served `cdn-cache: BYPASS`, so the script sees every index request; `Deno.env.get()` reads script variables; the SDK local server listens on `127.0.0.1:8080`.
---
## Task 1: Edge script and its test matrix
**Files:**
- Create: `edge/rpkgs-router.ts`
- Create: `edge/rpkgs-router.test.ts`
- Modify: `justfile` (add `edge-test`)
**Produces:** a single-file script deployable as `bunnynet_compute_script.content`, reading `UNION_SLOTS` from the environment.
- [ ] **Step 1: write the test matrix first**
`edge/rpkgs-router.test.ts` spawns `deno run -A edge/rpkgs-router.ts` with `UNION_SLOTS=amd64/alpine324`, waits for `127.0.0.1:8080`, and issues requests with `redirect: "manual"`.
Cases, asserted on the `location` header (or its absence):
| # | path | User-Agent | expectation |
| --- | ------------------------------------------------------ | --------------------------------------------- | ------------------------------------------------------------ |
| 1 | `/amd64/alpine324/latest/src/contrib/PACKAGES.gz` | `R (4.5.3 x86_64-pc-linux-musl …)` | 302 → `…/src/contrib/4.5/PACKAGES.gz` |
| 2 | same | `R (4.6.0 …)` | 302 → `…/src/contrib/4.6/PACKAGES.gz` |
| 3 | same, but slot `amd64/noble` | `R (4.5.3 …)` | no redirect (slot not in `UNION_SLOTS`) |
| 4 | `…/src/contrib/curl_7.1.0.tar.gz` | `R (4.5.3 …)` | no redirect |
| 5 | `…/src/contrib/4.5/PACKAGES.gz` | `R (4.5.3 …)` | no redirect (loop guard) |
| 6 | `…/src/contrib/PACKAGES.gz` | `curl/8.0` | no redirect (no R minor) |
| 7 | `/src/contrib/PACKAGES.gz` | alpine UA with `Alpine Linux … 3.24` | 302 → `/amd64/alpine324/latest/src/contrib/4.5/PACKAGES.gz` |
| 8 | `/src/contrib/PACKAGES.gz` | `R (4.5.3 x86_64-pc-linux-musl …)`, no distro | 302 → `cran.r-project.org`, **not** a `linux-musl` slot |
| 9 | `/src/contrib/foo_1.0.tar.gz` | `R (4.5.1 aarch64-apple-darwin20 …)` | 302 → `/bin/macosx/big-sur-arm64/contrib/4.5/foo_1.0.tar.gz` |
| 10 | `/bin/macosx/big-sur-arm64/contrib/4.5/foo_1.0.tar.gz` | any | 302 → `cran.r-project.org` |
| 11 | any redirect above | — | `cache-control: no-store` |
- [ ] **Step 2: run the tests and watch them fail**
`just edge-test` → every case fails, because `edge/rpkgs-router.ts` does not exist.
- [ ] **Step 3: write `edge/rpkgs-router.ts`**
Order of evaluation in `onOriginRequest`:
1. normalise `//` runs in the path
2. darwin `/src/contrib/*` → `/bin/macosx/<flavour>/contrib/<x.y>/`
3. `/bin/macosx/**` → CRAN
4. `/{arch}/{os}/latest/src/contrib/<rest>`: pass through if `rest` already starts with `<x.y>/`, or is not an index file, or the slot is not in `UNION_SLOTS`, or the UA has no R minor; otherwise redirect into `<x.y>/`
5. `/`, `/src/contrib`, `/src/contrib/**`: resolve arch+os from the UA, redirect to CRAN when the distro is unidentifiable, otherwise redirect to the qualified path, adding `<x.y>/` under the same index-file rule
6. anything else: pass through
The R minor comes from either `R/4.5.3` or `R (4.5.3 …)`, so a stock UA is enough. The `linux-gnu` / `linux-musl` fallback in `parseUserAgent` is deleted: those are not slot names.
- [ ] **Step 4: run the tests until they pass**
`just edge-test`
- [ ] **Step 5: commit**
```bash
git add edge/rpkgs-router.ts edge/rpkgs-router.test.ts justfile
git commit -m "feat(edge): route PACKAGES requests to the per-R-minor slot"
```
---
## Task 2: Manage the script from OpenTofu
**Files:**
- Modify: `cdn.tf`
**Consumes:** `edge/rpkgs-router.ts` from Task 1.
- [ ] **Step 1: add the resources**
```terraform
resource "bunnynet_compute_script" "rpkgs_router" {
type = "middleware"
name = "rpkgs-router"
content = file("${path.module}/edge/rpkgs-router.ts")
}
resource "bunnynet_compute_script_variable" "rpkgs_router_union_slots" {
script = bunnynet_compute_script.rpkgs_router.id
name = "UNION_SLOTS"
default_value = ""
required = false
}
```
and replace `middleware_script = 29277` with `middleware_script = bunnynet_compute_script.rpkgs_router.id`.
- [ ] **Step 2: validate**
`tofu init -backend=false && tofu validate`
- [ ] **Step 3: import the existing script (needs `BUNNYNET_API_KEY`)**
```bash
tofu import bunnynet_compute_script.rpkgs_router 29277
tofu plan
```
The plan must show an in-place `content` update and **no** replacement of the pull zone. A replacement means the import did not take.
- [ ] **Step 4: commit**
```bash
git add cdn.tf
git commit -m "feat(cdn): manage the edge middleware script from this repo"
```
---
## Task 3: Union index writer in bincraft
**Files (repo `codefloe.com/rpkgs/bincraft`):**
- Modify: `R/package_index.R`
- Test: `tests/testthat/test-package_index.R`
**Produces:** `write_union_index(flat_records, minor_records)` returning the merged records, called from `upload_package_index()` when `r_minor` is set.
- [ ] **Step 1: write the failing tests**
- a package present in both slots keeps the per-minor record, with `Path = "4.5"`
- a package only in the flat slot survives with no `Path`
- a package only in the per-minor slot survives with `Path = "4.5"`
- a union smaller than the flat input raises an error rather than returning
- [ ] **Step 2: run them and watch them fail**
`Rscript -e 'testthat::test_file("tests/testthat/test-package_index.R")'`
- [ ] **Step 3: implement `write_union_index()` and call it from `upload_package_index()`**
After `update_PACKAGES()` has written the per-minor index, read the flat slot's `PACKAGES.rds`, set `Path = <r_minor>` on the per-minor records, drop the flat records for packages the per-minor slot already has, and rewrite `PACKAGES`, `PACKAGES.gz` and `PACKAGES.rds` in the per-minor slot.
- [ ] **Step 4: run the tests until they pass**
- [ ] **Step 5: commit and open the PR against bincraft**
---
## Task 4: Roll out slot by slot
- [ ] Re-index one slot (`amd64/alpine324`, R 4.5) and confirm the union index lists both `curl` (per-minor, `Path: 4.5`) and `jsonlite` (flat, no `Path`).
- [ ] Set `UNION_SLOTS = "amd64/alpine324"` and confirm in `reg.devxy.io/r/r-alpine:4.5-3.24` that `available.packages()` returns the union count and `"curl" %in% rownames(...)`.
- [ ] Add `arm64/alpine324`, then the remaining slots.
`install.packages("curl")` will still fail to build on `alpine324` until that slot's source tarballs are replaced with real binaries. That is tracked separately.

157
specs/2026-06-30-package-patching-design.md
View file

@ -1,157 +0,0 @@
# Design: Dynamic per-package patching during binary builds
Date: 2026-06-30
Status: Approved (pending spec review)
## Problem
Some CRAN packages fail to compile on specific build platforms due to compiler- or OS-specific issues that have nothing to do with the package being built.
The canonical example is `RcppParallel`: its bundled Intel TBB sources fail to build on musl (Alpine) and on newer OS/compiler combinations.
Observed failure on `ubuntu-2604` ("resolute") with `g++ 15.2.0`:
```
../build/common.inc:74: *** "" is not supported. Add build/.inc file with os-specific settings . Stop.
make: *** [Makevars:163: tbb] Error 2
ERROR: compilation failed for package 'RcppParallel'
```
Because `RcppParallel` is a dependency of many packages, a single such failure cascades: every dependent package (e.g. `rts2`) also fails, even though nothing is wrong with the dependent itself.
Today there is no way to intervene.
A package can only be **excluded** (`local/excluded-packages.json`), which is all-or-nothing and does not help dependents.
## Goal
Allow a curated set of packages to be "patched" — via lightweight build-time overrides or, when necessary, real source diffs — **before** they are installed, whether the package is a direct build target or a transitive dependency pulled in by `pak`.
## Key constraint that drives the design
When `RcppParallel` fails here, it is being installed as a **transitive dependency** by `pak`, inside `bincraft::build_binary_package()`.
`pak` downloads, configures, and compiles it in one subprocess; this repo never touches that source.
For a fix to reach a dependency-of-a-dependency, the fixed package must be visible to `pak` itself, where `pak`'s repositories/sources are configured — which is inside `bincraft`.
Decisions taken during brainstorming:
- **Mechanism lives in `bincraft`** (the engine), because only there can transitive deps be influenced.
- **Patch tiers: both, env-overrides first.** Support cheap per-package env vars / configure args / Makevars (version-independent) *and* true source diffs (version-pinned), preferring the lightweight override.
- **Registry data lives in this repo** (`build-cran-binaries`) and is passed into `bincraft`, keeping `bincraft` as pure mechanism and the frequently-changing policy data with operational config.
## Approaches considered
| Approach | How pak sees the fix | Verdict |
|---|---|---|
| A. Patched **source** repo — drop patched `.tar.gz` source into a local repo, prepend it | pak recompiles from your source | Simple, but env-tier overrides leak globally (one subprocess builds everything) and the dep recompiles on every dependent build |
| **B. Pre-built patched binary repo (chosen)** | pak installs a ready binary by repo priority | Per-package scoping is free; no recompile; the binary is a cacheable/uploadable artifact that fits the existing system |
| C. pkgdepends per-build hook | intercept each build | No clean per-package pre-compile hook exists; fragile |
Chosen: **B**.
## Architecture
### Registry (this repo)
```
local/patches/
registry.json # the manifest
RcppParallel/
fix.patch # optional source diff, referenced by an entry
```
`registry.json` is an array of entries:
```json
[
{
"package": "RcppParallel",
"versions": "*",
"platforms": ["alpine", "ubuntu-2604"],
"env": { "RCPP_PARALLEL_USE_TBB": "0" },
"configure_args": [],
"makevars": {},
"patch": null,
"reason": "bundled TBB fails to build on musl / newer compilers"
}
]
```
Field semantics:
- `package` (string, required): CRAN package name.
- `versions` (string, required): `"*"` for any, a constraint such as `">=5.1.0"`, or an exact version `"5.1.11-2"`.
Env-tier fixes are typically `"*"`; source diffs are normally exact or lower-bounded because a diff is pinned to the source it was generated against.
- `platforms` (array of strings, required): matched against the running build's platform tokens — distro family (`alpine`, `ubuntu`, `redhat`), codename (`ubuntu-2604`, `alpine-324`), and arch (`amd64`, `arm64`).
An entry matches if any listed token matches any build token.
`["*"]` matches all platforms.
- `env` (object, optional): environment variables exported only for this package's isolated build.
- `configure_args` (array, optional): passed as `--configure-args` to the isolated build.
- `makevars` (object, optional): key/value pairs written into a package-local Makevars for the isolated build.
- `patch` (string or null, optional): path (relative to `local/patches/`) to a unified diff applied to the unpacked CRAN source before building.
- `reason` (string, required): human explanation, surfaced in logs and metadata.
A fix is any combination of `env`, `configure_args`, `makevars`, and `patch`.
"Env-first" is an authoring guideline (prefer the lightweight override) and an ordering of effort, not a runtime branch — all present fields are applied together for the isolated build.
### Flow (inside bincraft, around existing pak resolution)
1. **Resolve** the dependency set (dry-run) to learn the concrete versions `pak` will install.
Reuse bincraft's existing resolution where possible (e.g. a `pkgdepends` proposal: `$resolve()` → inspect resolution → ... → `$solve()` / `$install()` after the local repo is prepended).
2. For each resolved package that matches a registry entry (name + `versions` + `platforms`): obtain a **patched binary** for the exact `version × platform × arch × R-minor`:
- **Cache hit** (local `/mnt/cache/patched-binaries/` or S3): fetch it into the local repo.
- **Cache miss**: download the CRAN **source** for that version, apply the source `patch` (if any) to the unpacked tree, build the binary in isolation with `env` / `configure_args` / `makevars` applied, then place the binary in the local repo and write it to the cache (and S3 if uploading is enabled).
3. **Prepend** the local binary repo (`file://…`) to `pak`'s repo list, and regenerate its `PACKAGES` index.
4. Run the **normal install**.
`pak` resolves the patched binary for the matched package — direct or transitive — because it wins on repo priority for an equal version, and installs it without recompiling.
### Caching (essential)
`RcppParallel` is a dependency of dozens of packages; without caching the fix would be rebuilt on every dependent build.
Patched binaries are keyed by:
```
<package>_<version>_<platform>_<arch>_<rminor>_<patchhash>
```
`patchhash` is a hash of the normalized registry entry plus the referenced diff file contents.
Editing a patch therefore changes the hash and auto-invalidates stale cached binaries.
- Local cache: `/mnt/cache/patched-binaries/`.
- Optional S3 cache for cross-build reuse: a dedicated `…/patched/` slot under the existing arch/codename structure, mirroring how normal binaries are stored.
### S3 upload
Patched binaries **are** uploaded to S3 (in addition to the local cache) so they are reused across CI jobs and machines, not just within one container.
They live in a separate `patched/` slot and are not published into the user-facing `src/contrib` index — they are an internal build accelerator, not a distributed artifact.
## Error handling
- **Source diff fails to apply** (CRAN moved past the pinned version): log a clear warning, skip that entry, and proceed.
The package builds unpatched (status quo) and may fail.
The skipped/failed-to-apply patch is surfaced in build metadata.
- **Pre-build of the patched binary fails**: log a warning, skip, proceed.
- **No version or platform match**: skip silently (the entry simply does not apply to this build).
- **Overlapping entries for one package**: the most specific entry wins (a concrete `platforms`/`versions` beats `"*"`).
Genuine ambiguity (two equally specific, conflicting entries) is a validation error reported before the build.
## Observability
- One log line per applied patch, e.g.: `Applying patch to RcppParallel 5.1.11-2 [env: RCPP_PARALLEL_USE_TBB=0]: bundled TBB fails on musl / newer compilers`
- The set of applied patches (package, version, `patchhash`) is recorded in the Postgres build-metadata row for the build, so it is queryable later.
## Testing
- **Unit (registry):** parsing and matching — version constraints, platform token matching, precedence/specificity, and detection of ambiguous overlaps.
- **Unit (cache key):** `patchhash` changes when the entry or diff changes; is stable otherwise.
- **Integration:** `RcppParallel` on `resolute` (and/or Alpine) fails to build without a registry entry and succeeds with one; a dependent package such as `rts2` succeeds once the dependency is patched.
- **Failure path:** an entry pinned to an old version against a newer CRAN release → graceful skip with a warning, build continues.
## Out of scope
- Shipping a default registry inside `bincraft` (registry is repo-local for now; a baseline-in-engine + repo-override model can come later if needed).
- Publishing patched binaries into the public `src/contrib` index.
- Automatic detection of which packages need patches — entries are curated by hand.
## Split of work
- **bincraft:** the mechanism — registry ingestion, resolution hook, isolated patched-binary build, caching/upload, local-repo prepend, logging, metadata recording.
A new `patches` argument on `build_binary_package()`.
- **build-cran-binaries (this repo):** the `local/patches/` registry and diffs, passing `patches = "local/patches"` through `build-one.R` / `build-all.R`, and documentation.

174
specs/2026-08-07-per-minor-edge-routing-design.md
View file

@ -1,174 +0,0 @@
# Design: Routing clients to per-R-minor binary slots
Date: 2026-08-07
Status: Approved (pending spec review)
## Problem
`bincraft` routes ABI-"risky" packages to a per-minor slot `…/latest/src/contrib/<x.y>/` and indexes every directory independently (`upload_package_index()` calls `cranlike::update_PACKAGES()` on one prefix at a time).
Nothing unions those indices, and `contrib.url()` only ever yields `<repos>/src/contrib`, so no value of `options(repos)` can address a per-minor slot.
Only `uvr` resolves per-minor URLs, which means the per-minor slots are invisible to `install.packages()` by construction.
Measured on 2026-08-07:
| slot | flat `src/contrib` | `src/contrib/4.5` | unique packages only in the per-minor slot |
| ----------------- | ------------------ | ----------------- | ------------------------------------------ |
| `amd64/alpine324` | 21 640 | 3 310 | 2 886 |
| `amd64/noble` | 24 495 | 398 | 23 |
This is what issue #63 records as "missing binaries" on `alpine324`.
The packages are not missing; they are in a directory base R cannot reach.
The user-visible symptom in `reg.devxy.io/r/r-alpine:4.5-3.24` is:
```
> install.packages("curl")
Warning message:
package 'curl' is not available for this version of R
```
A second, unrelated defect exists on the same slot and is **out of scope here**: many `alpine324` tarballs are byte-identical CRAN _source_ tarballs that the index nevertheless stamps `Built: R 4.5.3; …-linux-musl`.
Routing exposes `curl`; only a rebuild of that slot makes it install.
## Goal
Let a stock `install.packages()` see one complete package list for its own R minor, without duplicating tarballs and without an R-version-varying cache key anywhere in the CDN.
## Key constraint that drives the design
R resolves a package's download URL from the index, not from the request path, and it keeps the `contriburl` it _asked for_ rather than the one it was redirected to.
Measured with `options(repos = …/latest)` against a middleware that redirects the index into `4.5/`:
```
curl available: TRUE
curl repo: …/latest/src/contrib # the flat URL, not the 4.5 one it was served from
```
So the union index is always addressed relative to the **flat** directory, whatever path it was fetched from.
`available.packages()` honours a `Path:` field and folds it into the `Repository` column, which gives the whole routing for free:
- a per-minor record carries `Path: <x.y>`, so its tarball is fetched from `…/src/contrib/<x.y>/`
- a flat record carries no `Path`, so its tarball is fetched from `…/src/contrib/`
Verified end to end against the live CDN with a locally built union index for `amd64/alpine324` (31 507 records):
```
curl: 7.1.0 -> …/latest/src/contrib/4.5 -> curl_7.1.0.tar.gz 717 725 B
jsonlite: 2.0.0 -> …/latest/src/contrib -> jsonlite_2.0.0.tar.gz 1 055 849 B
```
The corollary is that the edge script must **not** rewrite tarball requests: every tarball URL is already correct when it leaves the client, and redirecting one into `<x.y>/` would break exactly the flat packages the union is meant to preserve.
The complementary trick does not work: R's `gzcon()` reads only the first member of a concatenated gzip stream (10 291 of an expected 31 931 records), so an edge-side merge would have to fully decompress and recompress both indices and additionally 404 `PACKAGES.rds` to stop R preferring it.
That is why the union is produced in `bincraft`, not at the edge.
## Approaches considered
| Approach | Where the union lives | Verdict |
| ----------------------------------------------------------- | ------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------- |
| **A. Union index written by `bincraft` (chosen)** | per-minor `PACKAGES*`, per-minor entries carry `Path: <x.y>` | Edge does one redirect; `PACKAGES.rds` stays correct; no duplication |
| B. Merge at the edge | middleware fetches both indices, recompresses | ~2 MB decompress/recompress per cache fill, cache key must include the R minor, breaks R's `.rds` fast path |
| C. Move the minor up the path (`latest/<x.y>/src/contrib/`) | addressable by `options(repos)` directly | No edge logic at all, but a full layout migration and breaks the published URL contract |
Chosen: **A**.
## Architecture
### bincraft: union index (separate PR)
After writing a per-minor index, republish it as a union of that slot and the flat slot:
1. Read the flat slot's `PACKAGES.rds` and the per-minor slot's own records.
2. Set `Path: <x.y>` on every per-minor record, so its tarball resolves into the per-minor directory.
3. Drop every flat record whose package is already present in the per-minor slot, so the per-minor build always wins, and leave the survivors without a `Path`.
4. Write the merged `PACKAGES`, `PACKAGES.gz` and `PACKAGES.rds` into `…/src/contrib/<x.y>/`.
Guard: refuse to publish a union with fewer records than the flat index it was built from.
A truncated union is worse than no union, because it silently removes packages from every client on that minor.
### Edge script (this repo)
The script routes `PACKAGES`, `PACKAGES.gz` and `PACKAGES.rds` requests, and nothing else.
```
normalize path
parseClient(UA) -> { rMinor, arch, os } # rMinor from "R (4.5.3 …)" or "R/4.5.3"
darwin branches # unchanged
if path is /{arch}/{os}/latest/src/contrib/PACKAGES*
already under /<x.y>/ ? pass through # loop guard
rMinor known && slot in UNION_SLOTS ? 302 -> …/src/contrib/<rMinor>/PACKAGES*
else pass through # flat slot, today's behaviour
if path is /src/contrib/… # bare root
resolve arch + os; unknown -> 302 to CRAN
then apply the same PACKAGES* rule
else pass through
```
Redirects carry `Cache-Control: no-store`.
Every cacheable URL is therefore UA-independent, and no cache key has to vary by R version.
### Repaired bare-root detection
The bare `https://cran.rpkgs.com` form is currently broken for every Linux client that uses a stock R user agent.
`ALPINE_REGEX`, `UBUNTU_REGEX` and `RHEL_REGEX` only match a Posit-style user agent the user has to set by hand; stock R never carries the distro, so the script falls through to `extractOs()` and redirects to a slot that does not exist:
```
UA: R (4.5.3 x86_64-pc-linux-musl …) -> 302 /amd64/linux-musl/latest/… (404)
UA: R (4.5.3 x86_64-pc-linux-gnu …) -> 302 /amd64/linux-gnu/latest/… (404)
```
The fallback to a phantom `linux-musl` / `linux-gnu` slot is removed.
An unidentifiable distro redirects to CRAN, which is the existing behaviour for an unparseable user agent.
The R _minor_ is always present in a stock user agent, so per-minor routing itself does not depend on distro detection.
### Rollout gate
`UNION_SLOTS` is a `bunnynet_compute_script_variable` listing the slots whose per-minor index is already a union.
It is empty by default, so deploying the script changes nothing until `bincraft` has backfilled a slot, and a rollback is a variable edit rather than a code deploy.
All slots currently carry `4.4`, `4.5` and `4.6`; a client on any other minor falls through to the flat slot.
### Deployment from this repo
The script is a file in the repo, applied by the existing OpenTofu configuration:
```
edge/rpkgs-router.ts # the script
edge/rpkgs-router.test.ts # UA x path -> expected Location matrix
cdn.tf # bunnynet_compute_script + _variable
```
Provider `BunnyWay/bunnynet` v0.17.0 (already pinned) ships `bunnynet_compute_script` with `content` loadable via `file()`, plus `bunnynet_compute_script_variable`.
`middleware_script = bunnynet_compute_script.rpkgs_router.id` replaces the hard-coded `29277`, after a one-time `tofu import` of the existing script.
## Error handling
- Unknown R minor, or a slot not listed in `UNION_SLOTS`: pass through to the flat slot.
The client sees exactly today's behaviour.
- Unparseable distro on the bare-root form: redirect to CRAN.
- A request already under `…/src/contrib/<x.y>/`: pass through, so a redirect can never loop.
- A per-minor slot that does not exist for a listed minor: the client gets the origin's 404.
`UNION_SLOTS` is the operator's assertion that the slot is ready, so this is a configuration error, not a runtime condition to paper over.
## Testing
Local, before any apply: `deno run -A edge/rpkgs-router.ts` serves the middleware against the real origin, so `edge/rpkgs-router.test.ts` drives the whole matrix against that local server.
- User agent matrix: R 4.4 / 4.5 / 4.6 on musl and gnu, both arches, Posit-style and stock forms, plus a darwin UA and a non-R UA.
- Path matrix: `PACKAGES`, `PACKAGES.gz`, `PACKAGES.rds`, a tarball, a path already under `4.5/`, and `/src/contrib/…` on the bare root.
- Assertion is the `Location` header (or its absence), not the body.
After apply, a smoke test against `cran.rpkgs.com`:
- `available.packages()` inside `reg.devxy.io/r/r-alpine:4.5-3.24` returns the union count, and `"curl" %in% rownames(...)` is `TRUE`.
- A flat-slot package still downloads from `…/src/contrib/`, and a per-minor package downloads from `…/src/contrib/<x.y>/`.
## Out of scope
- `Meta/archive.rds` stays flat-only, so `remotes::install_version()` does not see per-minor archives.
- The `alpine324` source-tarball defect: that slot serves CRAN sources stamped as binaries, and needs a rebuild independent of this work.
- Any change to how `uvr` resolves per-minor URLs; it already addresses the slots directly.
## Split of work
1. `bincraft`: union index writer plus its guard, and a re-index of one slot to validate.
2. This repo: `edge/rpkgs-router.ts`, its test matrix, and the `cdn.tf` resources with `UNION_SLOTS` empty.
3. Enable `UNION_SLOTS` slot by slot as `bincraft` backfills them.

167
specs/2026-08-12-shard-weekly-rebuild-design.md
View file

@ -1,167 +0,0 @@
# Design: Sharding and resuming the weekly rebuild
Date: 2026-08-12
Status: Approved (pending spec review)
## Problem
`weekly-rebuild-missing` runs one job per `<os>-<arch>` and walks that slot's rebuild list serially in a single `R -q -e` invocation (`.crow/weekly-rebuild-missing.yaml:165`).
Until 2026-08-09 that was cheap, because every source fallback was skipped as "already built" and the list was effectively empty.
Since bincraft #105/#106/#107 and build-cran-binaries #159 the gate works, and the lists are now large.
Share of records whose object is byte-identical to CRAN's source, measured against `cran.r-project.org` MD5s on 2026-08-12:
| slot | records | source-served | share |
| ------------------ | ------: | ------------: | -----------: |
| `amd64/resolute` | 24 212 | 15 023 | 62.1% |
| `arm64/resolute` | 24 291 | 13 670 | 56.3% |
| `arm64/alpine324` | 24 328 | 9 514 | 39.2% |
| `amd64/alpine324` | 24 343 | 8 917 | 36.7% |
| `arm64/rhel10` | 24 695 | 5 384 | 21.9% |
| `amd64/rhel10` | 24 881 | 4 712 | 19.2% |
| 12 remaining slots | ~24 700 | 850 to 2 130 | 3.5% to 8.7% |
A single serial job cannot absorb that.
Pipeline 10910 (`weekly_rebuild_missing:alpine-324-amd64`) started on 2026-08-09, ran for roughly two days, reached `[8692/23885] cholera`, and was killed there.
Two distinct failures follow from that shape.
**No parallelism.** The work is embarrassingly parallel across packages, but one job does all of it.
**No resumability, and no clean stopping point.** The loop has no terminating condition other than exhausting the list, so the only way to stop it is a kill.
A restarted run re-reads the same list and walks it from the first entry.
It skips completed packages via `check_s3_root_package()`, but that costs a CRAN version resolution and an S3 `HEAD` per package, thousands of times, before it reaches new work.
Worse, a kill is not a pipeline failure: the `Purge CDN cache` step is guarded by `when: status: [success, failure]` (`.crow/weekly-rebuild-missing.yaml:206-207`), and on 10910 it produced no output at all.
So the ~4 600 binaries that run did publish stayed hidden behind stale edge copies.
## Goal
Turn each slot's rebuild into bounded, parallel, restartable units, without introducing state that can disagree with the bucket.
## Design
### 1. Shard the matrix three ways
Each of the 18 `OS`/`ARCH` rows in `.crow/weekly-rebuild-missing.yaml` gains `SPLIT_INTO: 3` and `SPLIT_INDEX: 1|2|3`, giving 54 rows.
This mirrors `.crow/build-all-versions.yaml:57-98`, which already shards its matrix four ways per arch.
Routing needs no change.
The cron filter `cron: weekly-rebuild-missing-${OS}-${ARCH}` and the manual `evaluate: weekly_rebuild_missing == "${OS}-${ARCH}"` both match all three shards of a slot.
Placement stays on the `rpkgs-${ARCH}` group label, so shards queue against available capacity rather than oversubscribing it.
### 2. Extract the loop into `local/rebuild-missing.R`
The build is currently a single ~1 500-character `R -q -e` argument.
Shard arithmetic and resume logic do not belong in a YAML string, and none of it is testable there.
The loop moves to `local/rebuild-missing.R`, invoked as `Rscript local/rebuild-missing.R $SPLIT_INTO $SPLIT_INDEX`, mirroring `local/build-all.R`.
Its body is unchanged in substance: read `/tmp/rebuild_pkgs.txt`, subtract `local/excluded-packages.json`, loop with `tryCatch` around `bincraft::build_binary_package()`.
The slice is **interleaved**, not contiguous:
```r
# the list is alphabetical and build cost clusters by name (Rcpp*, Bioc*,
# rstan*), so contiguous thirds would be badly unbalanced
mine <- pkgs[seq(split_index, length(pkgs), by = split_into)]
```
`local/build-all.R:64` uses contiguous chunks via `cut()`.
That is fine there because its list is every CRAN package and version, so the chunks average out.
Here the list is a filtered backlog in which expensive families sit adjacent, so interleaving is the better default.
Interleaving also makes each shard's `[i/n]` progress representative of the slot as a whole.
### 3. Resume by re-deriving state from the bucket
Before the loop, the shard performs one `s3fs::s3_dir_info()` on `devxy-rpkgs-binaries/<arch>/<codename>/latest/src/contrib` and reads the `etag` column.
It fetches CRAN's `PACKAGES` once for the latest version and published `MD5sum` of every package.
A package is still outstanding if and only if the object at `<pkg>_<version>.tar.gz` has an ETag equal to CRAN's `MD5sum` for that version, which is the definition `check_s3_root_package()` already applies one package at a time.
```r
# one paginated listing instead of ~2900 sequential HEAD requests per shard
info <- s3fs::s3_dir_info(slot_dir)
etag <- setNames(gsub('^"|"$', "", info$etag), basename(info$uri))
key <- sprintf("%s_%s.tar.gz", mine, cran_version[mine])
# keep a package when no object exists yet, or when the object is still
# byte-identical to CRAN's source; drop it once a real binary is published
mine <- mine[is.na(etag[key]) | etag[key] == cran_md5[key]]
```
This is the whole resume mechanism.
There is no progress file, no volume, and no database cursor.
A restarted shard recomputes ground truth and continues where it stopped, and it is correct even when a sibling shard, a `process-updates` cron, or a manual `just rebuild` completed something in the meantime.
Three properties make this the right source of truth:
- **It is what the build itself checks.** Any other store can disagree with the bucket; this one cannot.
- **It is agent-independent.** `.crow/weekly-rebuild-missing.yaml` mounts no `volumes:`, unlike `.crow/build-all-versions.yaml:132-133`, so `/mnt/cache` is per-job and cannot carry progress anyway.
- **It costs one listing.** `cranlike`'s `s3` fork already does exactly this call against this bucket at ~24 000 objects, so the approach is proven at the required scale.
It must read ETags rather than the slot index's `Built` field, which is how `local/packages-to-build.R:104-130` answers the same question.
Under this design the index is not rewritten until the dependent re-index pipeline runs (section 5), so mid-run it cannot reflect the current run's progress.
Packages that genuinely fail to build re-publish their CRAN source, so they stay outstanding and would be retried on every restart.
That is already handled upstream: `bincraft::filter_packages_with_errors()` (`R/build_binaries.R:1018`, `:1143`) drops anything with `error_occurred = TRUE`, and `store_build_metadata = TRUE` is passed on every call.
No additional poison-pill filter is needed here.
Only the flat `src/contrib` path is considered.
The rebuild call passes no `is_r_minor_sensitive`, so it defaults to `FALSE` and only ever targets the flat path; the resume filter matches that scope deliberately.
### 4. Give each shard a wall-clock budget
`local/rebuild-missing.R` takes a budget, defaulting to 20 hours, and breaks out of the loop once it is exceeded:
```r
# exit cleanly rather than being killed, so the dependent re-index still runs
if (difftime(Sys.time(), started, units = "hours") > budget_hours) {
cat(sprintf("Budget of %sh reached after %d/%d packages; stopping cleanly\n", budget_hours, i, n))
break
}
```
It exits 0 and reports how much of the slice it covered.
Every run then has a terminating condition, the re-index and purge always fire, and the remainder is picked up by the next run with no bookkeeping, because section 3 recomputes the outstanding set from scratch.
### 5. Move the re-index and purge into `.crow/weekly-rebuild-reindex.yaml`
Three shards per slot means three concurrent `upload_package_index()` calls on the same S3 prefix.
`cranlike::update_PACKAGES()` lists the live bucket, so an early lister that uploads last publishes an index missing its siblings' work.
The re-index steps (`.crow/weekly-rebuild-missing.yaml:171-176`) and the purge step (`:187-207`) therefore leave that file entirely.
The new file carries:
```yaml
depends_on:
- weekly-rebuild-missing
runs_on: [success, failure]
```
`runs_on: [success, failure]` validates as a workflow-level key under `crow lint`, so a failing shard no longer withholds the re-index.
The file uses the same 18-row matrix and the same `when:` gating as `weekly-rebuild-missing`, so it only re-indexes slots that actually ran.
Each row re-indexes the flat slot and every per-minor slot.
`scripts/purge_cdn_zone.sh` runs once on a single row, because all hostnames share pull zone `3857050` and 18 identical zone purges would be waste.
## Failure behaviour
| case | today | after |
| -------------------------- | ----------------------------------- | ----------------------------------------------------- |
| one package errors | `tryCatch` logs, loop continues | unchanged |
| a shard fails outright | purge runs, re-index does not | re-index and purge run via `runs_on` |
| a shard exceeds its budget | cannot happen, runs until killed | exits 0, re-index and purge run |
| a shard is killed | nothing runs | still nothing; trigger the re-index pipeline alone |
| a shard restarts | re-walks the list, HEAD per package | one listing, resumes at the first outstanding package |
The known cost of `depends_on` being file-level rather than row-level: on the weekly cron no slot is re-indexed until the slowest of all 54 jobs finishes.
The 20-hour budget bounds that at roughly one day.
## Out of scope
- `build-all-versions` still cannot rebuild source fallbacks, because `local/build-all.R:113-122` drops every version with any `single_builds` row for the platform and arch, which is precisely the source-fallback set. That is a separate change.
- Bunny Perma-Cache eviction. `scripts/purge_cdn_zone.sh` purges the regular edge cache only; see the note in `CLAUDE.md` and issue history.
- The audit that produces the rebuild list is unchanged.
## Verification
- `crow lint .crow/` passes for both pipeline files.
- `local/rebuild-missing.R` gets unit coverage in `local/tests/` for the two pure pieces: the interleaved slice (disjoint, covering, deterministic) and the outstanding-set filter (source-served ETag kept, binary ETag dropped, absent object kept).
- A single-slot manual run of `alpine-324-amd64` shard 1 confirms the listing shortcut against the live bucket, and that the reported outstanding count is close to the 8 917 measured above divided by three.
- Restarting that shard mid-run confirms it resumes rather than replaying, by comparing the outstanding count it reports on the second start.