Thread problem causing some calls to hang for 25

Ralf Habacker ralf.habacker at freenet.de
Thu Jun 10 08:28:24 PDT 2010


Thiago Macieira schrieb:
> Em Quinta-feira 10 Junho 2010, às 11:02:54, Ralf Habacker escreveu:
>   
>> Thiago Macieira schrieb:
>>     
>>> Em Quinta-feira 10 Junho 2010, às 10:26:38, Thiago Macieira escreveu:
>>>       
>>>> Em Terça-feira 08 Junho 2010, às 08:06:49, Carlsson, Johannes escreveu:
>>>>         
>>>>> Hi
>>>>>
>>>>> I am not sure if you mean me, but I dont have commit rights, so if
>>>>> anyone else could commit the patches it would be great, especially
>>>>> since there is a new release coming up.
>>>>>           
>>>> Done.
>>>>
>>>> Sorry for the delay, I was away for the past two days on a workshop.
>>>>         
>>> Actually, Ralf did it.
>>>
>>> But in doing so, he also merged D-Bus 1.3 into the dbus-1.2 branch.
>>>       
>> you are refering to the following patches ?
>>
>> 3 days    Fixed compiler warnings.
>> 3 days    Minor include file fix.
>> 3 days    Keep in sync with autotools.
>> 3 days    Keep in sync with autotools.
>>
>> The should do not have any bad side effect
>>
>> git is sometime hard to understand - I applied the patch to the master
>> branch first, then switched to the dbus-1.2 branch with  'git branch
>> dbus-1.2', run 'git checkout -f' to be sure that there are no local
>> changes and checked the log with 'git log', which shown nothing wrong.
>> Then I applied the patch with git am <patchname> and did 'git push' - is
>> there something wrong with this sequence ?
>>     
>
> Yes.
>
> You apply the patch only to the lowest branch where it makes sense. That 
> branch is merged to the next branch, then sequentially until master is 
> reached,
>
> However, your description doesn't show the error.
>   
Yeah, I don't believe that the problem was initiated by my commits.
> Yet the commit cd815c347a279f324e5025fcda71a16ad9b08c7d (Fixed compiler 
> warnings.) is in both master and dbus-1.2.
>   
The graph on the left side of the image I sent to the list let me assume 
that master and dbus-1.2 are merged together into one remaining branch - 
which would explain this behavior
> From what I can tell, 01ae2cb79376acb65b876c6620477fde86e6ef12 is the first bad 
> commit. It should have been in master only, but it exists in both branches.
>   
If you download the dbus-1.2 branch tar ball from this commit see 
http://cgit.freedesktop.org/dbus/dbus/commit/?h=dbus-1.2&id=01ae2cb79376acb65b876c6620477fde86e6ef12
you will see that is contains the source from the master branch - 
another indication that dbus-1.2 and master are merged together with 
this commit.

Ralf




More information about the dbus mailing list