julia.withPackages: handle non-archive artifacts (fixes #277410)

1 files changed, 63 insertions, 27 deletions

diff --git a/pkgs/development/julia-modules/python/extract_artifacts.py b/pkgs/development/julia-modules/python/extract_artifacts.py
index ecbdf10ed7148..f811c6624e851 100755
--- a/pkgs/development/julia-modules/python/extract_artifacts.py
+++ b/pkgs/development/julia-modules/python/extract_artifacts.py
@@ -5,10 +5,33 @@ import multiprocessing
 import subprocess
 import sys
 import toml
+from urllib.parse import urlparse
 import yaml
 
 import dag
 
+# This should match the behavior of the default unpackPhase.
+# See https://github.com/NixOS/nixpkgs/blob/59fa082abdbf462515facc8800d517f5728c909d/pkgs/stdenv/generic/setup.sh#L1044
+archive_extensions = [
+  # xz extensions
+  ".tar.xz",
+  ".tar.lzma",
+  ".txz",
+
+  # *.tar or *.tar.*
+  ".tar",
+  ".tar.Z",
+  ".tar.bz2",
+  ".tar.gz",
+
+  # Other tar extensions
+  ".tgz",
+  ".tbz2",
+  ".tbz",
+
+  ".zip"
+  ]
+
 dependencies_path = Path(sys.argv[1])
 closure_yaml_path = Path(sys.argv[2])
 julia_path = Path(sys.argv[3])
@@ -33,33 +56,14 @@ with open(closure_yaml_path, "r") as f:
     if contents.get("depends_on"):
       closure_dependencies_dag.add_node(uuid, dependencies=contents["depends_on"].values())
 
-with open(out_path, "w") as f:
-  f.write("{ lib, fetchurl, glibc, pkgs, stdenv }:\n\n")
-  f.write("rec {\n")
-
-  def process_item(item):
-    uuid, src = item
-    lines = []
-    artifacts = toml.loads(subprocess.check_output([julia_path, extract_artifacts_script, uuid, src]).decode())
-    if not artifacts: return f'  uuid-{uuid} = {{}};\n'
-
-    lines.append(f'  uuid-{uuid} = {{')
-
-    for artifact_name, details in artifacts.items():
-      if len(details["download"]) == 0: continue
-      download = details["download"][0]
-      url = download["url"]
-      sha256 = download["sha256"]
-
-      git_tree_sha1 = details["git-tree-sha1"]
-
-      depends_on = set()
-      if closure_dependencies_dag.has_node(uuid):
-        depends_on = set(closure_dependencies_dag.get_dependencies(uuid)).intersection(dependency_uuids)
+def get_archive_derivation(uuid, artifact_name, url, sha256):
+  depends_on = set()
+  if closure_dependencies_dag.has_node(uuid):
+    depends_on = set(closure_dependencies_dag.get_dependencies(uuid)).intersection(dependency_uuids)
 
-      other_libs = extra_libs.get(uuid, [])
+  other_libs = extra_libs.get(uuid, [])
 
-      fixup = f"""fixupPhase = let
+  fixup = f"""fixupPhase = let
           libs = lib.concatMap (lib.mapAttrsToList (k: v: v.path))
                                [{" ".join(["uuid-" + x for x in depends_on])}];
           in ''
@@ -69,7 +73,7 @@ with open(out_path, "w") as f:
               patchelf --set-interpreter ${{glibc}}/lib/ld-linux-x86-64.so.2 {{}} \;
           ''"""
 
-      derivation = f"""{{
+  return f"""stdenv.mkDerivation {{
         name = "{artifact_name}";
         src = fetchurl {{
           url = "{url}";
@@ -82,9 +86,41 @@ with open(out_path, "w") as f:
         {fixup};
       }}"""
 
+def get_plain_derivation(url, sha256):
+  return f"""fetchurl {{
+        url = "{url}";
+        sha256 = "{sha256}";
+      }}"""
+
+with open(out_path, "w") as f:
+  f.write("{ lib, fetchurl, glibc, pkgs, stdenv }:\n\n")
+  f.write("rec {\n")
+
+  def process_item(item):
+    uuid, src = item
+    lines = []
+    artifacts = toml.loads(subprocess.check_output([julia_path, extract_artifacts_script, uuid, src]).decode())
+    if not artifacts: return f'  uuid-{uuid} = {{}};\n'
+
+    lines.append(f'  uuid-{uuid} = {{')
+
+    for artifact_name, details in artifacts.items():
+      if len(details["download"]) == 0: continue
+      download = details["download"][0]
+      url = download["url"]
+      sha256 = download["sha256"]
+
+      git_tree_sha1 = details["git-tree-sha1"]
+
+      parsed_url = urlparse(url)
+      if any(parsed_url.path.endswith(x) for x in archive_extensions):
+        derivation = get_archive_derivation(uuid, artifact_name, url, sha256)
+      else:
+        derivation = get_plain_derivation(url, sha256)
+
       lines.append(f"""    "{artifact_name}" = {{
       sha1 = "{git_tree_sha1}";
-      path = stdenv.mkDerivation {derivation};
+      path = {derivation};
     }};\n""")
 
     lines.append('  };\n')