Split the private webstore install API into two parts....

chrome/test/data/extensions/api_test/extension_gallery_install/invalid_begin.html

Index: chrome/test/data/extensions/api_test/extension_gallery_install/invalid_begin.html
===================================================================
--- chrome/test/data/extensions/api_test/extension_gallery_install/invalid_begin.html	(revision 0)
+++ chrome/test/data/extensions/api_test/extension_gallery_install/invalid_begin.html	(revision 0)
@@ -0,0 +1,15 @@
+<script src="common.js"></script>
+<script>
+// This tests that an invalid id passed to beginInstall generates an error, and
+// that completeInstall raises an error in this case as well.
+chrome.webstorePrivate.beginInstall("some invalid id", function() {
+  assertEq(chrome.extension.lastError.message, "Invalid id");
+
+  chrome.webstorePrivate.completeInstall(extension_id, function() {

[Erik does not do reviews, 2010/11/10 16:47:45]

this seems redundant since you already have the separate complete without begin
test.  if you want to test that complete doesn't trigger just because someone
called begin before, then it seems like a better test would be to try a
different valid id for begin.  I'm still not sure that's very interesting
though.

[asargent_no_longer_on_chrome, 2010/11/10 22:58:23]

Agreed - I've removed this part. 

+    var expected_error = extension_id +
+                         " does not match a previous call to beginInstall";
+    assertEq(chrome.extension.lastError.message, expected_error);
+    succeed();
+  });
+});
+</script>
Property changes on: chrome/test/data/extensions/api_test/extension_gallery_install/invalid_begin.html
___________________________________________________________________
Added: svn:eol-style
   + LF