build-cran-binaries/local/fetch-rebuild-packages-from-issue.R

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")
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")
}

# Prefer the audit's freshly-written RDS. The audit overwrites it each run
# (saveRDS), so unlike the Forgejo issue body it is never subject to the
# duplicate-subsection accumulation bug. Fall back to parsing the issue when the
# RDS is absent (e.g. a fresh runner with no shared cache).
rds_file <- file.path(
  Sys.getenv("REBUILD_PKG_RDS_DIR", "/mnt/cache/packages"),
  sprintf("weekly_rebuild_%s_%s.rds", platform, arch)
)
if (file.exists(rds_file)) {
  pkgs <- tryCatch(as.character(readRDS(rds_file)), error = function(e) NULL)
  if (!is.null(pkgs) && length(pkgs) > 0) {
    cat(sprintf(
      "Using audit RDS %s: %d rebuildable packages for %s/%s\n",
      rds_file,
      length(pkgs),
      platform,
      arch
    ))
    writeLines(pkgs, output_file)
    cat(sprintf("Wrote package list to %s\n", output_file))
    q("no")
  }
  cat(sprintf(
    "RDS %s present but empty/unreadable -- falling back to issue\n",
    rds_file
  ))
}

token <- Sys.getenv("FORGEJO_TOKEN")
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))