From 1d2edf1051e5ece91cd0ec237863821930125f1d Mon Sep 17 00:00:00 2001
From: Marc Abramowitz <marc@marc-abramowitz.com>
Date: Tue, 26 Mar 2019 09:10:41 -0700
Subject: [PATCH] mkpkg: Grab sdist instead of first URL from PyPI (#356)

The first URL from PyPI is often a platform-specific wheel, so it's
better if we prefer to grab a source distribution that we can build with
emscripten.

Before:

```
$ bin/pyodide mkpkg lxml
$ grep 'url: ' packages/lxml/meta.yaml
  url: https://files.pythonhosted.org/packages/28/32/951337aa3d14e58cae3f585e468e402141f5ac3728589eccca765c4e17ce/lxml-4.3.2-cp27-cp27m-macosx_10_6_intel.macosx_10_9_intel.macosx_10_9_x86_64.macosx_10_10_intel.macosx_10_10_x86_64.whl
```

After:

```
$ bin/pyodide mkpkg lxml
$ grep 'url: ' packages/lxml/meta.yaml
  url: https://files.pythonhosted.org/packages/65/6f/d070609b415f4a24bf36b7deb94bfd791e8d023c0dbf4233659af53fe0ab/lxml-4.3.2.tar.gz
```

Closes: #363.
---
 pyodide_build/mkpkg.py | 32 +++++++++++++++++++++++++++++---
 1 file changed, 29 insertions(+), 3 deletions(-)

diff --git a/pyodide_build/mkpkg.py b/pyodide_build/mkpkg.py
index 873721c42..a88e4a71e 100755
--- a/pyodide_build/mkpkg.py
+++ b/pyodide_build/mkpkg.py
@@ -3,12 +3,38 @@
 import argparse
 import json
 import os
-from pathlib import Path
+import shutil
 import urllib.request
-
+from pathlib import Path
 
 PACKAGES_ROOT = Path(__file__).parent.parent / 'packages'
 
+SDIST_EXTENSIONS = []
+
+
+def get_sdist_extensions():
+    if SDIST_EXTENSIONS:
+        return SDIST_EXTENSIONS
+
+    for format in shutil.get_unpack_formats():
+        for ext in format[1]:
+            SDIST_EXTENSIONS.append(ext)
+
+    return SDIST_EXTENSIONS
+
+
+def get_sdist_url_entry(json_content):
+    sdist_extensions_tuple = tuple(get_sdist_extensions())
+
+    for entry in json_content['urls']:
+        if entry['filename'].endswith(sdist_extensions_tuple):
+            return entry
+
+    raise Exception('No sdist URL found for package %s (%s)' % (
+        json_content['info'].get('name'),
+        json_content['info'].get('package_url'),
+    ))
+
 
 def make_package(package):
     import yaml
@@ -18,7 +44,7 @@ def make_package(package):
     with urllib.request.urlopen(url) as fd:
         json_content = json.load(fd)
 
-    entry = json_content['urls'][0]
+    entry = get_sdist_url_entry(json_content)
     download_url = entry['url']
     sha256 = entry['digests']['sha256']
     version = json_content['info']['version']