GNU bug report logs - #23708
[PATCH] package: don’t hard code port number in test

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 23708 in the body.
You can then email your comments to 23708 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Michal Nazarewicz <mina86 <at> mina86.com>
To: bug-gnu-emacs <at> gnu.org
Subject: [PATCH] package: don’t hard code port number in test
Date: Mon,  6 Jun 2016 21:53:33 +0200

Hard coding port 8000 in package-tests.el causes the test to fail if
something is already listening on that port.  Change code so it uses
any available port.

* tests/lisp/emacs-lisp/package-tests.el
  (package-test-update-archives-async): Do not hard code port number
  and instead parse it from server’s standard output.

* tests/lisp/emacs-lisp/package-resources/package-test-server.py:
  Do not hard code port number.  Use any available one instead.
---
 .../package-resources/package-test-server.py       | 15 ++----
 test/lisp/emacs-lisp/package-tests.el              | 54 ++++++++++++----------
 2 files changed, 34 insertions(+), 35 deletions(-)

diff --git a/test/lisp/emacs-lisp/package-resources/package-test-server.py b/test/lisp/emacs-lisp/package-resources/package-test-server.py
index 35ca820..78e40b3 100644
--- a/test/lisp/emacs-lisp/package-resources/package-test-server.py
+++ b/test/lisp/emacs-lisp/package-resources/package-test-server.py
@@ -1,21 +1,16 @@
-import sys
+import os
 import BaseHTTPServer
 from SimpleHTTPServer import SimpleHTTPRequestHandler
 
-
 HandlerClass = SimpleHTTPRequestHandler
 ServerClass  = BaseHTTPServer.HTTPServer
 Protocol     = "HTTP/1.0"
 
-if sys.argv[1:]:
-    port = int(sys.argv[1])
-else:
-    port = 8000
-    server_address = ('127.0.0.1', port)
+os.chdir(os.path.dirname(__file__))
 
 HandlerClass.protocol_version = Protocol
-httpd = ServerClass(server_address, HandlerClass)
+httpd = ServerClass(('127.0.0.1', 0), HandlerClass)
 
-sa = httpd.socket.getsockname()
-print "Serving HTTP on", sa[0], "port", sa[1], "..."
+# This printed line is parsed by test code, don't change its format.
+print "Serving on http://%s:%s/" % httpd.socket.getsockname()
 httpd.serve_forever()
diff --git a/test/lisp/emacs-lisp/package-tests.el b/test/lisp/emacs-lisp/package-tests.el
index c7a5cc7..4ffb173 100644
--- a/test/lisp/emacs-lisp/package-tests.el
+++ b/test/lisp/emacs-lisp/package-tests.el
@@ -372,31 +372,35 @@ package-test-desc-version-string
   (skip-unless (executable-find "python2"))
   ;; For some reason this test doesn't work reliably on hydra.nixos.org.
   (skip-unless (not (getenv "NIX_STORE")))
-  (with-package-test (:basedir
-                      package-test-data-dir
-                      :location "http://0.0.0.0:8000/")
-    (let* ((package-menu-async t)
-           (process (start-process
-                     "package-server" "package-server-buffer"
-                     (executable-find "python2")
-                     (expand-file-name "package-test-server.py"))))
-      (unwind-protect
-          (progn
-            (list-packages)
-            (should package--downloads-in-progress)
-            (should mode-line-process)
-            (should-not
-             (with-timeout (10 'timeout)
-               (while package--downloads-in-progress
-                 (accept-process-output nil 1))
-               nil))
-            ;; If the server process died, there's some non-Emacs problem.
-            ;; Eg maybe the port was already in use.
-            (skip-unless (process-live-p process))
-            (goto-char (point-min))
-            (should
-             (search-forward-regexp "^ +simple-single" nil t)))
-        (if (process-live-p process) (kill-process process))))))
+
+  (with-temp-buffer
+    (cd package-test-data-dir)
+    (let ((package-menu-async t)
+          (process (start-process "package-server" (current-buffer)
+                                  (executable-find "python2")
+                                  (expand-file-name "package-test-server.py"))))
+      ;; Killing temp buffer will kill the process.
+      (set-process-query-on-exit-flag process nil)
+
+      (with-package-test
+          (:location (with-timeout (5 (should-not 'timeout))
+                       (while (and (goto-char (point-min))
+                                   (not (re-search-forward
+                                         "^Serving on \\(http://.*/\\)" nil t)))
+                         (accept-process-output process 1))
+                       (match-string 1)))
+        (list-packages)
+        (should package--downloads-in-progress)
+        (should mode-line-process)
+        (should-not (with-timeout (10 'timeout)
+                      (while package--downloads-in-progress
+                        (accept-process-output nil 1))
+                      nil))
+        ;; If the server process died, there's some non-Emacs problem.
+        ;; Eg maybe the port was already in use.
+        (skip-unless (process-live-p process))
+        (goto-char (point-min))
+        (should (search-forward-regexp "^ +simple-single" nil t))))))
 
 (ert-deftest package-test-describe-package ()
   "Test displaying help for a package."
-- 
2.8.0.rc3.226.g39d4020

Message #10 received at 23708-done <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: 23708-done <at> debbugs.gnu.org
Subject: Re: bug#23708: [PATCH] package: don’t hard code
 port number in test
Date: Tue, 07 Jun 2016 21:26:25 -0400

Version: 25.2

Thanks; I installed something similar.

Message #13 received at 23708 <at> debbugs.gnu.org (full text, mbox):

From: Michal Nazarewicz <mina86 <at> mina86.com>
To: 23708 <at> debbugs.gnu.org
Subject: Re: bug#23708: closed ( bug#23708: [PATCH] package: don’t hard code port number in test)
Date: Wed, 08 Jun 2016 14:34:19 +0200

Glenn Morris <rgm <at> gnu.org> wrote:
> Thanks; I installed something similar.

Note that currently argument parsing in package-test-server.py is
dysfunctional (passing an argument causes server_address to never be
set) so while my patch would fix that (by removing that unused feature).

-- 
Best regards
ミハウ “𝓶𝓲𝓷𝓪86” ナザレヴイツ
«If at first you don’t succeed, give up skydiving»

Message #16 received at 23708 <at> debbugs.gnu.org (full text, mbox):

From: Glenn Morris <rgm <at> gnu.org>
To: Michal Nazarewicz <mina86 <at> mina86.com>
Cc: 23708 <at> debbugs.gnu.org
Subject: Re: bug#23708: closed ( bug#23708: [PATCH] package: don’t hard code port number in test)
Date: Wed, 08 Jun 2016 12:02:33 -0400

Michal Nazarewicz wrote:

> Note that currently argument parsing in package-test-server.py is
> dysfunctional (passing an argument causes server_address to never be
> set) so while my patch would fix that (by removing that unused feature).

(Sorry, I did not notice that you have write access.) Feel free to fix
that if it bothers you.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.