From 7afba1087499d047cc86048e8d75c25506905c08 Mon Sep 17 00:00:00 2001
From: Tomas Kopecek <tkopecek@redhat.com>
Date: Jan 02 2020 09:45:56 +0000
Subject: PR#1829: Add a sanity check on remotely opened RPMs


Merges #1829
https://pagure.io/koji/pull-request/1829

Fixes: #290
https://pagure.io/koji/issue/290
consider some validation/retry/check for openRemoteFile

---

diff --git a/builder/kojid b/builder/kojid
index b0f031d..9cc959c 100755
--- a/builder/kojid
+++ b/builder/kojid
@@ -1106,6 +1106,7 @@ class BuildTask(BaseTaskHandler):
         opts = dict([(k, getattr(self.options, k)) for k in ('topurl','topdir')])
         opts['tempdir'] = self.workdir
         with koji.openRemoteFile(relpath, **opts) as fo:
+            koji.check_rpm_file(fo)
             h = koji.get_rpm_header(fo)
         if not koji.get_header_field(h, 'sourcepackage'):
             raise koji.BuildError("%s is not a source package" % srpm)
diff --git a/koji/__init__.py b/koji/__init__.py
index 69940cb..aa4371e 100644
--- a/koji/__init__.py
+++ b/koji/__init__.py
@@ -1691,6 +1691,38 @@ def openRemoteFile(relpath, topurl=None, topdir=None, tempdir=None):
     return fo
 
 
+def check_rpm_file(rpmfile):
+    """Do a initial sanity check on an RPM
+
+    rpmfile can either be a file name or a file object
+
+    This check is used to detect issues with RPMs before they break builds
+    See: https://pagure.io/koji/issue/290
+    """
+    if isinstance(rpmfile, six.string_types):
+        with open(rpmfile, 'rb') as fo:
+            return _check_rpm_file(fo)
+    else:
+        return _check_rpm_file(rpmfile)
+
+
+def _check_rpm_file(fo):
+    """Check that the open file appears to be an rpm"""
+    # TODO: trap exception and raise something with more infomation
+    ts = rpm.TransactionSet()
+    try:
+        hdr = ts.hdrFromFdno(fo.fileno())
+    except rpm.error as ex:
+        raise GenericError("rpm's header can't be extracted: %s (rpm error: %s)" % \
+                           (fo.name, ', '.join(ex.args)))
+    try:
+        rpm.TransactionSet().hdrCheck(hdr.unload())
+    except rpm.error as ex:
+        raise GenericError("rpm's header can't be checked: %s (rpm error: %s)" % \
+                           (fo.name, ', '.join(ex.args)))
+    fo.seek(0)
+
+
 def config_directory_contents(dir_name, strict=False):
     configs = []
     try: