Why rename it? I'll take a look at updating the conf files soon.

Thanks for getting this upgraded. What did you do to update it? (so the info 
isn't lost forever, heh)

-Colin (via thumbs)

On Jul 24, 2011, at 8:21 AM, Adrian Godoroja <robot...@me.com> wrote:

> Hi,
> 
> i've managed to upgrade buildbot master to the latest and greatest,
> and also muscovy has been upgraded to the latest version of buildbot and 
> Xcode.
> 
> So, it's up and running. 
> 
> Now the next logical step is to finish what Colin started, i.e. multiple 
> nighties, etc. etc. (read below), which i'm trying to do atm.
> 
> I'd also propose to rename Lumberjack, to Wall-E for example :)
> 
> Thanks Eric, thanks Colin for the provided support!
> 
> Adrian.
> 
> 
> On Jul 19, 2011, at 22:04 , Adrian Godoroja wrote:
> 
>> Colin,
>> 
>> can you give me access?
>> 
>> Thanks.
>> 
>> On Jul 19, 2011, at 01:04 , Adrian Godoroja wrote:
>> 
>>> Hi guys,
>>> 
>>> me and Patrick (pks) wanted to help. If someone could give us access, it 
>>> would be great.
>>> 
>>> Thanks, Adrian.
>>> 
>>> On Jul 19, 2011, at 00:56 , Colin Barrett wrote:
>>> 
>>>> Evan (and list),
>>>> 
>>>> I would love to. I just need to find time to get buildbot upgraded -- 
>>>> either by my own hand or at the instruction of someone else. The current 
>>>> issue is that we need to upgrade Twisted to install the required version 
>>>> of buildbot (we need some newer features). My plan was to do that by 
>>>> installing buildbot in a virtualenv[1], but if we could just upgrade 
>>>> buildbot and twisted on that machine, that'd work too.
>>>> 
>>>> In the end, the story is that not only am I an awful sysadmin, but I 
>>>> dislike doing it as well, which is the main reason why this stuff hasn't 
>>>> gotten done. I certainly SHOULD be able to make time for this, but the 
>>>> reality is that I haven't :(
>>>> 
>>>> -Colin
>>>> 
>>>> [1] an isolated python install; this poses some complications with our 
>>>> init.d scripts though.
>>>> 
>>>> On Jul 16, 2011, at 1:50 PM, Evan Schoenberg, M.D. wrote:
>>>> 
>>>>> Colin,
>>>>> 
>>>>> Still planning on making these changes? Seems like you've done most of 
>>>>> the heavy lifting already :)
>>>>> 
>>>>> -Evan
>>>>> 
>>>>> On May 7, 2011, at 7:45 PM, Zachary West wrote:
>>>>> 
>>>>>> 
>>>>>> 
>>>>>> On Mon, Apr 4, 2011 at 15:40, Colin Barrett <co...@springsandstruts.com> 
>>>>>> wrote:
>>>>>> How it works:
>>>>>> 
>>>>>> buildbot has an array of repos and branches (repos are paths on 
>>>>>> "hg.adium.im" so "adium" an "adium-1.4", branches are in repository 
>>>>>> branches). When it builds each one, it injects NIGHTLY_REPO and 
>>>>>> NIGHTLY_BRANCH into the environment. (This is new as of last night -- 
>>>>>> right now that list simply contains ("adium", "default"). All nightlies, 
>>>>>> branch or not, will be using the same infrastructure.) To add a new one, 
>>>>>> you'll need to modify master.cfg. (This could probably be automated if 
>>>>>> it happens enough.)
>>>>>> 
>>>>>> That information gets used by the release makefile to set some 
>>>>>> Info.plist keys. The parameters "repo" & "branch" are then passed along 
>>>>>> when we hit the Sparkle update URL.
>>>>>> 
>>>>>> When the nightly is uploaded, it's now stored in a directory on the 
>>>>>> server like "repo-branch". The two scripts that look for nightlies have 
>>>>>> been modified to look in the right places. When a branch is merged in to 
>>>>>> trunk, deleting its directory on the nightly server is enough to get it 
>>>>>> to fall back to the default branch of the repo, or to the default branch 
>>>>>> of adium as a last resort.
>>>>>> 
>>>>>> Pretty much the only outstanding issue is presenting a list of 
>>>>>> alternative nightlies for display on nightly.adium.im.
>>>>>> 
>>>>>> The adium-side changes seem solid. 
>>>>>> 
>>>>>> On the sparkle side of things, I think the appcast-nightly.php changes 
>>>>>> need a little modifications. With the changes, existing_branch() will be 
>>>>>> pinging the nightly server on each invocation, which means for every 
>>>>>> user update check; this is a bit too much for my blood.
>>>>>> 
>>>>>> How about this:
>>>>>> (1) Check and see if a cache file exists for the requested latest.info
>>>>>> (2) If the file exists, and isn't out of date, use it.
>>>>>> (3) If the file exists, but is out of date, update it.
>>>>>> (4) If the file does not exist, try and ping the nightly server to see 
>>>>>> if it's real.
>>>>>> 
>>>>>> Otherwise the changes seem good.
>>>>>> 
>>>>>> -- 
>>>>>> Zachary West
>>>>> 
>>>> 
>>>> 
>>> 
>>> 
>> 
>> 
> 
> 

Reply via email to