R: improve support for CRAN and Bioconductor package sets

 - Update the instructions for re-generating each of the package set files.
 - Provide test-evaluation.nix expression to verify that the package sets evaluates.
 - Update list of known broken packages.

1 files changed, 24 insertions, 17 deletions

diff --git a/pkgs/development/r-modules/generate-r-packages.R b/pkgs/development/r-modules/generate-r-packages.R
index f9346ddc127c8..d788457963755 100755
--- a/pkgs/development/r-modules/generate-r-packages.R
+++ b/pkgs/development/r-modules/generate-r-packages.R
@@ -5,47 +5,54 @@ library(parallel)
 cl <- makeCluster(10)
 
 mirrorType <- commandArgs(trailingOnly=TRUE)[1]
-packagesFile <- paste(mirrorType, 'packages.nix', sep='-')
-mirrorUrls <- list(
-  bioc="http://bioconductor.statistik.tu-dortmund.de/packages/3.0/bioc",
-  cran="http://cran.r-project.org"
-)
-mirrorUrl <- paste(mirrorUrls[mirrorType], "/src/contrib/", sep="")
+stopifnot(mirrorType %in% c("bioc","cran"))
 
+packagesFile <- paste(mirrorType, 'packages.nix', sep='-')
 readFormatted <- as.data.table(read.table(skip=6, sep='"', text=head(readLines(packagesFile), -1)))
 
+mirrorUrls <- list( bioc="http://bioconductor.statistik.tu-dortmund.de/packages/3.2/bioc/src/contrib/"
+                  , cran="http://cran.r-project.org/src/contrib/"
+                  )
+mirrorUrl <- mirrorUrls[mirrorType][[1]]
+knownPackages <- lapply(mirrorUrls, function(url) as.data.table(available.packages(url, filters=c("R_version", "OS_type", "duplicates"))))
+pkgs <- knownPackages[mirrorType][[1]]
+setkey(pkgs, Package)
+knownPackages <- c(unique(do.call("rbind", knownPackages)$Package))
+knownPackages <- sapply(knownPackages, gsub, pattern=".", replacement="_", fixed=TRUE)
+
 nixPrefetch <- function(name, version) {
   prevV <- readFormatted$V2 == name & readFormatted$V4 == version
   if (sum(prevV) == 1) as.character(readFormatted$V6[ prevV ]) else
     system(paste0("nix-prefetch-url --type sha256 ", mirrorUrl, name, "_", version, ".tar.gz"), intern=TRUE)
 }
 
-formatPackage <- function(name, version, sha256, depends, imports, linkingTo, knownPackages) {
+formatPackage <- function(name, version, sha256, depends, imports, linkingTo) {
     attr <- gsub(".", "_", name, fixed=TRUE)
-    if (is.na(depends)) depends <- "";
-    depends <- unlist(strsplit(depends, split="[ \t\n]*,[ \t\n]*", fixed=FALSE))
-    depends <- c(depends, unlist(strsplit(imports, split="[ \t\n]*,[ \t\n]*", fixed=FALSE)))
-    depends <- c(depends, unlist(strsplit(linkingTo, split="[ \t\n]*,[ \t\n]*", fixed=FALSE)))
+    depends <- paste( if (is.na(depends)) "" else gsub("[ \t\n]+", "", depends)
+                    , if (is.na(imports)) "" else gsub("[ \t\n]+", "", imports)
+                    , if (is.na(linkingTo)) "" else gsub("[ \t\n]+", "", linkingTo)
+                    , sep=","
+                    )
+    depends <- unlist(strsplit(depends, split=",", fixed=TRUE))
     depends <- sapply(depends, gsub, pattern="([^ \t\n(]+).*", replacement="\\1")
-    depends <- depends[depends %in% knownPackages]
     depends <- sapply(depends, gsub, pattern=".", replacement="_", fixed=TRUE)
-    depends <- paste(depends, collapse=" ")
+    depends <- depends[depends %in% knownPackages]
+    depends <- paste(sort(unique(depends)), collapse=" ")
     paste0(attr, " = derive { name=\"", name, "\"; version=\"", version, "\"; sha256=\"", sha256, "\"; depends=[", depends, "]; };")
 }
 
-clusterExport(cl, c("nixPrefetch","readFormatted", "mirrorUrl"))
+clusterExport(cl, c("nixPrefetch","readFormatted", "mirrorUrl", "knownPackages"))
 
 pkgs <- as.data.table(available.packages(mirrorUrl, filters=c("R_version", "OS_type", "duplicates")))
 pkgs <- pkgs[order(Package)]
 pkgs$sha256 <- parApply(cl, pkgs, 1, function(p) nixPrefetch(p[1], p[2]))
-knownPackages <- unique(pkgs$Package)
 
-nix <- apply(pkgs, 1, function(p) formatPackage(p[1], p[2], p[18], p[4], p[5], p[6], knownPackages))
+nix <- apply(pkgs, 1, function(p) formatPackage(p[1], p[2], p[18], p[4], p[5], p[6]))
 
 cat("# This file is generated from generate-r-packages.R. DO NOT EDIT.\n")
 cat("# Execute the following command to update the file.\n")
 cat("#\n")
-cat(paste("# Rscript generate-r-packages.R", mirrorType, ">", packagesFile))
+cat(paste("# Rscript generate-r-packages.R", mirrorType, ">new && mv new", packagesFile))
 cat("\n\n")
 cat("{ self, derive }: with self; {\n")
 cat(paste(nix, collapse="\n"), "\n")