Odd buildbot behaviour

Blair Zajac blair at orcaware.com
Thu Apr 12 08:56:21 PDT 2012


On 04/12/2012 08:27 AM, William Siegrist wrote:
>
> On Apr 11, 2012, at 8:12 PM, Blair Zajac<blair at orcaware.com>  wrote:
>
>> On 04/11/2012 08:08 AM, Joshua Root wrote:
>>> On 2012-4-12 01:04 , Blair Zajac wrote:
>>>> On 04/11/2012 07:47 AM, Joshua Root wrote:
>>>>> On 2012-4-10 11:30 , Joshua Root wrote:
>>>>> Seems like most of the problems were caused by ice33-cpp. Still not sure
>>>>> how the "not modified since last failure" check could have been
>>>>> triggered, but we'll see how it goes now that ice33-cpp is fixed.
>>>>
>>>> Did ice-cpp not have the same issue?  It would be good to keep the two
>>>> in sync if possible.
>>>
>>> It apparently built fine. It comes earlier in the list.
>>
>> Odd, I just did a build of ice33-cpp with the old Portfile and looked for the code that would be activated by create_runpath_symlink and it didn't trigger.  I didn't see any of the "Removed symbolic link..." or "Created symbolic link..." messages.
>>
>> On the buildbot, is ${prefix} a symlink to somewhere?
>
>
> Yes.

That explains the behavior in ice33-cpp, as Ice 3.3.x defines 
create_runpath_symlink as 'yes', which has the Makefile delete the 
symlink and then recreate it.  In 3.4.x, create_runpath_symlink is 'no' 
by default, so that code doesn't run.

Blair


More information about the macports-dev mailing list