fix: audit workflows

local/fetch-rebuild-packages-from-issue.R

@@ -0,0 +1,87 @@
+library(httr2, quietly = TRUE)
+
+forgejo_base <- "https://git.devxy.io/api/v1"
+repo         <- "devxy/build-cran-binaries"
+platform     <- Sys.getenv("PLATFORM")
+arch         <- Sys.getenv("ARCH")
+token        <- Sys.getenv("FORGEJO_TOKEN")
+output_file  <- Sys.getenv("REBUILD_PKG_LIST", "/tmp/rebuild_pkgs.txt")
+
+if (nchar(platform) == 0) stop("PLATFORM env var is not set")
+if (nchar(arch)     == 0) stop("ARCH env var is not set")
+if (nchar(token)    == 0) stop("FORGEJO_TOKEN env var is not set")
+
+os_family <- if (grepl("^ubuntu", platform)) {
+  "Ubuntu"
+} else if (grepl("^alpine", platform)) {
+  "Alpine"
+} else if (grepl("^redhat", platform)) {
+  "Red Hat"
+} else {
+  platform
+}
+
+issue_title <- sprintf("Missing package binaries for latest version (%s)", os_family)
+cat(sprintf("Searching for issue: %s\n", issue_title))
+
+search_url <- sprintf(
+  "%s/repos/%s/issues?type=issues&state=open&q=%s&limit=50",
+  forgejo_base, repo,
+  utils::URLencode(issue_title, reserved = TRUE)
+)
+search_resp <- request(search_url) |>
+  req_headers(Authorization = paste("token", token)) |>
+  req_perform()
+
+issues <- resp_body_json(search_resp, simplifyVector = FALSE)
+match_idx <- which(vapply(issues, function(x) x$title, character(1)) == issue_title)
+
+if (length(match_idx) == 0) {
+  cat("No matching issue found - nothing to rebuild\n")
+  writeLines(character(0), output_file)
+  q("no")
+}
+
+body <- issues[[match_idx[1]]]$body
+if (is.null(body) || nchar(body) == 0) {
+  cat("Issue body is empty - nothing to rebuild\n")
+  writeLines(character(0), output_file)
+  q("no")
+}
+
+lines <- strsplit(body, "\n", fixed = TRUE)[[1]]
+
+# Find the platform section (## alpine-322)
+plat_header <- sprintf("## %s", platform)
+plat_idx <- which(lines == plat_header)
+if (length(plat_idx) == 0) {
+  cat(sprintf("No section found for platform %s - nothing to rebuild\n", platform))
+  writeLines(character(0), output_file)
+  q("no")
+}
+
+# Find the arch subsection (### arm64 (...))
+arch_pattern <- sprintf("^### %s ", arch)
+arch_idx <- which(grepl(arch_pattern, lines) & seq_along(lines) > plat_idx[1])
+if (length(arch_idx) == 0) {
+  cat(sprintf("No section found for arch %s under %s - nothing to rebuild\n", arch, platform))
+  writeLines(character(0), output_file)
+  q("no")
+}
+
+ai <- arch_idx[1]
+
+# Find the end of the rebuildable section: stop at ####, ###, ##, ---, or EOF
+end_idx <- which(grepl("^####|^###|^##|^---", lines) & seq_along(lines) > ai)
+end_idx <- if (length(end_idx) > 0) end_idx[1] - 1 else length(lines)
+
+section_lines <- lines[seq(ai + 1, end_idx)]
+
+# Extract package names from "- PackageName (Version)" lines
+pkg_lines <- section_lines[grepl("^- ", section_lines)]
+pkgs <- sub("^- ([^ ]+) \\(.*\\)$", "\\1", pkg_lines)
+pkgs <- pkgs[nchar(pkgs) > 0 & pkgs != "_None_"]
+
+cat(sprintf("Found %d rebuildable packages for %s/%s\n", length(pkgs), platform, arch))
+writeLines(pkgs, output_file)
+cat(sprintf("Wrote package list to %s\n", output_file))

local/weekly-missing-binaries-audit.R

@@ -20,8 +20,11 @@ arch     <- Sys.getenv("ARCH")
 if (nchar(platform) == 0) stop("PLATFORM env var is not set")
 if (nchar(arch)     == 0) stop("ARCH env var is not set")
 
-# e.g. "ubuntu-2204" -> "ubuntu2204"
+# Map platform names to S3 codenames
 s3_codename <- gsub("-", "", platform)
+s3_codename <- sub("^redhat", "rhel", s3_codename)
+s3_codename <- sub("^ubuntu2204$", "jammy", s3_codename)
+s3_codename <- sub("^ubuntu2404$", "noble", s3_codename)
 
 # Derive OS family for the Forgejo issue title
 os_family <- if (grepl("^ubuntu", platform)) {
@@ -38,7 +41,38 @@ cat(sprintf("Platform: %s | Arch: %s | S3 codename: %s | OS family: %s\n",
             platform, arch, s3_codename, os_family))
 
 # ---------------------------------------------------------------------------
-# 1. CRAN release packages
+# 1. Query PostgreSQL for known build failures (before s3fs init to avoid
+#    C++ pointer conflicts between s3fs/curl and RPostgres/libpq)
+# ---------------------------------------------------------------------------
+cat("Connecting to PostgreSQL...\n")
+con <- DBI::dbConnect(
+  RPostgres::Postgres(),
+  dbname   = "build_metadata",
+  host     = "r-binaries.devxy.io",
+  port     = 15432,
+  user     = "rpkgs",
+  password = Sys.getenv("PGPASS"),
+  sslmode  = "require"
+)
+
+errored_pkgs <- DBI::dbGetQuery(
+  con,
+  sprintf(
+    "SELECT name, tag FROM single_builds WHERE error_occurred = TRUE AND platform = '%s' AND arch = '%s'",
+    platform, arch
+  )
+)
+DBI::dbDisconnect(con)
+
+errored_dt <- as.data.table(errored_pkgs)
+if (nrow(errored_dt) > 0) {
+  setnames(errored_dt, c("Package", "Version"))
+  setkey(errored_dt, Package, Version)
+}
+cat(sprintf("Found %d known build failures for %s/%s\n", nrow(errored_dt), platform, arch))
+
+# ---------------------------------------------------------------------------
+# 2. CRAN release packages (uses curl internally - must come after PG)
 # ---------------------------------------------------------------------------
 cat("Fetching CRAN package database...\n")
 cran_release <- tools::CRAN_package_db()
@@ -48,7 +82,7 @@ cran_dt <- data.table(
 )
 
 # ---------------------------------------------------------------------------
-# 2. S3 tarballs
+# 3. S3 tarballs
 # ---------------------------------------------------------------------------
 cat("Connecting to S3...\n")
 s3fs::s3_file_system(
@@ -74,15 +108,19 @@ file_names <- basename(s3_pkgs)
 matches    <- regexec("^([A-Za-z0-9.]+)_([0-9][^/]*)\\.tar\\.gz$", file_names)
 parts      <- regmatches(file_names, matches)
 parts      <- parts[sapply(parts, length) == 3]
-s3_dt <- data.table(
-  Package = sapply(parts, `[`, 2),
-  Version = sapply(parts, `[`, 3)
-)
+if (length(parts) == 0) {
+  s3_dt <- data.table(Package = character(0), Version = character(0))
+} else {
+  s3_dt <- data.table(
+    Package = sapply(parts, `[`, 2),
+    Version = sapply(parts, `[`, 3)
+  )
+}
 
 cat(sprintf("S3 contains %d tarballs for %s/%s\n", nrow(s3_dt), arch, s3_codename))
 
 # ---------------------------------------------------------------------------
-# 3. Find missing packages (CRAN release version not in S3)
+# 4. Find missing packages (CRAN release version not in S3)
 # ---------------------------------------------------------------------------
 setkey(cran_dt, Package, Version)
 setkey(s3_dt,   Package, Version)
@@ -90,7 +128,7 @@ missing_dt <- cran_dt[!s3_dt]
 cat(sprintf("%d CRAN release packages missing from S3\n", nrow(missing_dt)))
 
 # ---------------------------------------------------------------------------
-# 4. Remove excluded packages
+# 5. Remove excluded packages
 # ---------------------------------------------------------------------------
 script_dir <- tryCatch(
   dirname(normalizePath(
@@ -114,36 +152,6 @@ if (file.exists(excluded_path)) {
   cat(sprintf("No excluded-packages.json found at %s -- skipping exclusion step\n", excluded_path))
 }
 
-# ---------------------------------------------------------------------------
-# 5. Query PostgreSQL for known build failures
-# ---------------------------------------------------------------------------
-cat("Connecting to PostgreSQL...\n")
-con <- DBI::dbConnect(
-  RPostgres::Postgres(),
-  dbname   = "build_metadata",
-  host     = "r-binaries.devxy.io",
-  port     = 15432,
-  user     = "rpkgs",
-  password = Sys.getenv("PGPASS"),
-  sslmode  = "require"
-)
-on.exit(DBI::dbDisconnect(con), add = TRUE)
-
-errored_pkgs <- DBI::dbGetQuery(
-  con,
-  sprintf(
-    "SELECT name, tag FROM single_builds WHERE error_occurred = TRUE AND platform = '%s' AND arch = '%s'",
-    platform, arch
-  )
-)
-errored_dt <- as.data.table(errored_pkgs)
-if (nrow(errored_dt) > 0) {
-  setnames(errored_dt, c("Package", "Version"))
-  setkey(errored_dt, Package, Version)
-}
-
-cat(sprintf("Found %d known build failures for %s/%s\n", nrow(errored_dt), platform, arch))
-
 # ---------------------------------------------------------------------------
 # 6. Split into rebuildable vs known failures
 # ---------------------------------------------------------------------------