On Wed, 21 Nov 2012, Pascal Stumpf wrote:

> On Tue, 20 Nov 2012 17:14:19 -0700 (MST), Jon Trulson wrote:
>> On Tue, 20 Nov 2012, Pascal Stumpf wrote:
>>
>>> On Mon, 19 Nov 2012 19:37:53 -0700 (MST), Jon Trulson wrote:
>>>> On Sun, 18 Nov 2012, Pascal Stumpf wrote:
>>>>
>>>>> On Mon, 15 Oct 2012 18:10:26 -0600 (MDT), Jon Trulson wrote:
>>>>>> On Mon, 15 Oct 2012, Pascal Stumpf wrote:
>>>>>>
>>>>>>> Something has seriously broken ttsession on OpenBSD (and anything using
>>>>>>> it).  Either segfaulting and leaving zombie processes or immediately
>>>>>>> locking up on startup.  I probably won't have time to look into it until
>>>>>>> next weekend, but if anyone has an idea or wants to help debugging ...
>>>>>>>
>>>>>>
>>>>>> There have been a few changes there - cd into lib/tt and do a:
>>>>>>
>>>>>> git log -p .
>>>>>>
>>>>>> To see relevant commits.  Not sure what specifically might have been
>>>>>> broken on CDE for OpenBSD...
>>>>>
>>>>> I *think* this was it.
>>>>>
>>>>
>>>> I'm not sure what you mean here... If you are saying that
>>>> OPT_UNIX_SOCKET_RPC needs to be enabled for OpenBSD, then this should
>>>> be done in tt/lib/tt_options.h (define it for OpenBSD)...
>>>>
>>>> Or am I missing something?
>>>
>>> No, I'm saying the svcfd_create() is still needed in the
>>> !OPT_UNIX_SOCKET_RPC case.  I don't really understand how this works on
>>> other OSes, but well ...
>>
>> Hmm... it should be all or nothing WRT OPT_UNIX_SOCKET_RPC... Could it
>> have been this commit: 406fa95994e68bf8bf3004930d0498ce892412a3 ?
>
> Nope, RPC_ANYSOCK is defined to -1 on OpenBSD, so that commit doesn't
> change anything functionality-wise.
>
>> Have you bisected?
>
> Yes, it's definitely commit 44e384aedb3b8aabbd3c556af28848fc58c09c60.
> So it seems the commit message is wrong and the code is indeed called
> even in the !OPT_UNIX_SOCKET_RPC case.
>
> Since about when does Linux have svcfd_create()?
>

Hmm.  I think that patch (44e384aedb3b8aabbd3c556af28848fc58c09c60) is
wrong... The comment too.  I will revert it.

svcfd_create() seems to exist on my system - don't know how long it's
been around.  I do know there is a bug report for debian where it is
not specified in any header files of manpages, but does exist:

https://lsbbugs.linuxfoundation.org/show_bug.cgi?id=3404

Wonder if that's what he ran into.  At any rate, it should be fixed
properly on those systems where it is a problem.

-- 
Jon Trulson

"Oh the land of the free, and the home of the brave.
  Are you heaven on Earth, or the gloom of the grave."
                 -- Iron Sky



------------------------------------------------------------------------------
Monitor your physical, virtual and cloud infrastructure from a single
web console. Get in-depth insight into apps, servers, databases, vmware,
SAP, cloud infrastructure, etc. Download 30-day Free Trial.
Pricing starts from $795 for 25 servers or applications!
http://p.sf.net/sfu/zoho_dev2dev_nov
_______________________________________________
cdesktopenv-devel mailing list
cdesktopenv-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/cdesktopenv-devel

Reply via email to