After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 623306 - GDBus: nonce-tcp reuses port numbers
GDBus: nonce-tcp reuses port numbers
Status: RESOLVED DUPLICATE of bug 631379
Product: glib
Classification: Platform
Component: gdbus
unspecified
Other All
: Normal normal
: ---
Assigned To: David Zeuthen (not reading bugmail)
gtkdev
Depends on:
Blocks:
 
 
Reported: 2010-07-01 16:34 UTC by Allison Karlitskaya (desrt)
Modified: 2010-10-19 11:40 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Allison Karlitskaya (desrt) 2010-07-01 16:34:13 UTC
TEST: gdbus-peer... (pid=20496)
  /gdbus/peer-to-peer:                                                 OK
  /gdbus/delayed-message-processing:                                   **
ERROR:gdbus-peer.c:846:dmp_thread_func: assertion failed (error == NULL): Error binding to address: Address already in use (g-io-error-quark, 33)
FAIL
GTester: last random seed: R02S45299181f7e64d94fa788e0ca0f7c56c

consistently.
Comment 1 Allison Karlitskaya (desrt) 2010-07-01 16:57:19 UTC
I check this in (at David's request) to work around the issue for now:

commit 93bf09a9e7a7d35d50802faf4b271ca4267e49cb
Author: Ryan Lortie <desrt@desrt.ca>
Date:   Thu Jul 1 12:55:04 2010 -0400

    Use "unix:" instead of tcp in test for now...
    
    Temporarily avoids the issue reported in bug #623306.

nonce-tcp is broken, it seems (or maybe GNIO?)
Comment 2 David Zeuthen (not reading bugmail) 2010-07-01 17:13:53 UTC
Thanks Ryan - the weird thing is that it works fine here if using nonce-tcp: ... (of course, wouldn't have checked in the code where 'make check' fails). You mentioned that we're trying to bind to the same port twice - copy/pasting from IRC:

<desrt> it's because you are trying to bind to port 56658 twice in the same process
 [pid 24477] bind(7, {sa_family=AF_INET, sin_port=htons(0), sin_addr=inet_addr("127.0.0.1")}, 16 <unfinished ...>
 [pid 24459] <... sched_yield resumed> ) = 0
 [pid 24477] <... bind resumed> )        = 0
--> jkprg (~jarda@r10ld239.net.upc.cz) has joined #gtk+
<desrt> [pid 24459] sched_yield()               = 0
 [pid 24459] sched_yield( <unfinished ...>
 [pid 24477] listen(7, 10 <unfinished ...>
 here
 then later:
 [pid 24477] bind(11, {sa_family=AF_INET, sin_port=htons(56658), sin_addr=inet_addr("127.0.0.1")}, 16 <unfinished ...>
 [pid 24459] sched_yield( <unfinished ...>
 [pid 24477] <... bind resumed> )        = -1 EADDRINUSE (Address already in use)
<davidz> desrt: doesn't happen here...
<desrt> without closing 7


(Related: Would be nice if we had a commit hook that ran 'make check'... e.g. you'd be free to commit to a branch called, say, staging and then a robot would merge to the master branch only if 'make check' passes. </dreaming>)
Comment 3 Allison Karlitskaya (desrt) 2010-10-19 11:40:45 UTC
i thought we had fixed this....

*** This bug has been marked as a duplicate of bug 631379 ***