feat(local): aggregate blocked-on-dependency reporting by dependency (#131)
Some checks failed
ci/crow/cron/process-updates/1 Pipeline was successful
ci/crow/cron/process-updates/2 Pipeline failed
ci/crow/cron/process-updates/7 Pipeline was successful
ci/crow/cron/process-updates/8 Pipeline was successful
ci/crow/cron/process-updates/3 Pipeline was successful
ci/crow/cron/process-updates/4 Pipeline was successful
ci/crow/manual/weekly-patch-proposals Pipeline was successful
ci/crow/cron/process-updates/9 Pipeline was successful
ci/crow/cron/process-updates/10 Pipeline was successful
ci/crow/manual/trial-build-registry/7 Pipeline failed
ci/crow/manual/trial-build-registry/5 Pipeline failed
ci/crow/manual/trial-build-registry/3 Pipeline failed
ci/crow/manual/trial-build-registry/13 Pipeline failed
ci/crow/manual/trial-build-registry/1 Pipeline failed
ci/crow/manual/trial-build-registry/11 Pipeline failed
ci/crow/manual/trial-build-registry/4 Pipeline failed
ci/crow/manual/trial-build-registry/2 Pipeline failed
ci/crow/manual/trial-build-registry/6 Pipeline failed
ci/crow/manual/trial-build-registry/9 Pipeline failed
ci/crow/manual/trial-build-registry/17 Pipeline failed
ci/crow/manual/trial-build-registry/15 Pipeline failed
ci/crow/manual/trial-build-registry/10 Pipeline failed
ci/crow/manual/trial-build-registry/14 Pipeline failed
ci/crow/manual/trial-build-registry/16 Pipeline failed
ci/crow/manual/trial-build-registry/12 Pipeline failed
ci/crow/manual/trial-build-registry/8 Pipeline failed
ci/crow/manual/trial-build-registry/18 Pipeline failed
ci/crow/cron/process-updates/13 Pipeline was successful
ci/crow/cron/process-updates/14 Pipeline was successful
ci/crow/cron/process-updates/15 Pipeline was successful
ci/crow/cron/process-updates/16 Pipeline was successful

## Why

With cascade detection (#128) live, the latest `auto-apply-patches` run did exactly the right thing — **proposed nothing** (`No auto-proposable candidates`) because every failure is a dependency cascade, and it surfaced the ~30 root-cause dependencies to fix.

But the "Blocked on a dependency" list printed **one line per fingerprint group**, so the same dependency repeated (rstan ×4, lpsymphony ×4, salso ×2, BH ×2, GO.db ×2, RcppCWB ×2, …), burying the priority.

## What

Aggregate blocked packages across all groups **by the dependency they wait on**:

- Expose `blocked_map` (package → dependency) from `build_triage_report()`.
- Add `blocked_by_dependency()` — dedupes dependents (a package in two groups counts once) and ranks dependencies by how many distinct dependents they block.
- Proposer and tracker (log + issue) now print one line per dependency, sorted by impact. Replaces the per-group `blocked_summary`.

## Result (same data, aggregated)

```
Blocked on a dependency (3 dependencies block 6 dependents; fix the dependency, not each dependent):
  RcppParallel             3 dependent(s)
  rstan                    2 dependent(s)
  sf                       1 dependent(s)
```

So the real run becomes a crisp, ranked worklist: RcppParallel (894), sf (128), rstan (~96), Rfast (33), clarabel/DescTools (26), Rglpk (22), xgboost (18), …

## Verified

New test covers cross-group aggregation, dedup (a dependent in two groups counted once), the example cap, and ranking. 112 tests pass; hooks pass.

Reviewed-on: #131
This commit is contained in:
Patrick Schratz 2026-07-16 21:24:03 +00:00 committed by Patrick Schratz
commit 6c03f278ec

View file

@ -160,24 +160,38 @@ unclassified_summary <- function(report, max_groups = 30L, max_pkgs = 15L) {
)
}
# Groups with packages blocked on a dependency (a package-specific fix pinned
# via `applies_to`, or a data-driven cascade where the package fails building a
# dependency): report the dependency + how many dependents wait on it, so fixing
# it once is recognised as clearing the batch.
blocked_summary <- function(report, max_pkgs = 15L) {
bl <- Filter(
function(g) length(g$blocked_packages %||% character(0L)) > 0L,
report
)
lapply(bl, function(g) {
pkgs <- g$blocked_packages
# Aggregate blocked packages across ALL groups by the dependency they wait on,
# so one dependency (RcppParallel, rstan, sf, ...) is a single line -- deduped
# and ranked by how many distinct dependents it blocks -- instead of repeating
# once per fingerprint group. Reads each group's `blocked_map` (package -> the
# dependency it is blocked on). Returns records sorted by dependent count desc,
# each with up to `max_pkgs` example dependents.
blocked_by_dependency <- function(report, max_pkgs = 15L) {
acc <- list() # dependency -> character vector of dependent packages
for (g in report) {
bm <- g$blocked_map
if (is.null(bm) || length(bm) == 0L) {
next
}
for (pkg in names(bm)) {
for (dep in as.character(unlist(bm[[pkg]]))) {
acc[[dep]] <- unique(c(acc[[dep]], pkg))
}
}
}
if (length(acc) == 0L) {
return(list())
}
out <- lapply(names(acc), function(dep) {
pkgs <- acc[[dep]]
list(
blocked_on = g$blocked_on,
dependency = dep,
n_packages = length(pkgs),
packages = utils::head(pkgs, max_pkgs),
packages_truncated = length(pkgs) > max_pkgs
)
})
out[order(-vapply(out, function(x) x$n_packages, integer(1L)))]
}
# Does a registry entry's `platforms` apply to a build on `os` (e.g.