[squeak-dev] trunk

Chris Muller ma.chris.m at gmail.com
Mon Jan 27 04:52:36 UTC 2014


Thanks Frank.  Hey, as you know from our prior release experience, I
want our release process to allow the Release "Candidate" to live up
to its designation by being THE actual binary image file that
_everyone_ bangs on.  I don't want community to verify one binary, but
then we release a different binary even if it "should" be basically
the same.

I'm using ReleaseBuilder's new #prepareNewRelease to produce THE
Release Candidate image which will point to the releaseRepository
squeak45 instead of #buildRepository trunk.  You'll be glad to know I
intend to use tomorrow's successful build (with these last trunk
commits) from the latest CI build artifact.  No more "rolling our own"
all the way from the prior alpha, cool!

Only after the release-image is made, zipped up, and uploaded to
ftp.squeak.org, will I consider it a "Release Candidate".  That way
we're sure to all be testing exactly the same binary.

 - Chris


On Sun, Jan 26, 2014 at 3:23 PM, Frank Shearar <frank.shearar at gmail.com> wrote:
> Right. http://build.squeak.org/job/ReleaseSqueakTrunk/758/console has
> a bright, shiny green light. That artifact _should_ be good to go.
> Chris, would you mind banging on the image a bit?
>
> frank
>
> On 26 January 2014 20:05, Frank Shearar <frank.shearar at gmail.com> wrote:
>>  http://build.squeak.org/job/ReleaseSqueakTrunk/754/console shows the
>> desired failure.
>> http://build.squeak.org/job/ReleaseSqueakTrunk/755/console shows the
>> Warning being resumed (before another error breaks things).
>>
>> Ah, progress.
>>
>> frank
>>
>> On 26 January 2014 17:07, Frank Shearar <frank.shearar at gmail.com> wrote:
>>> It does now :) Sorry, I meant to mention that a few hours ago. Except
>>> I've just learned to my amazement that #messageTest != #messageText.
>>> So give me a moment and I'll have two fixes: one to make CI blow up
>>> properly when a process fails, and one to dump the Warning's
>>> _messageText_. Sigh!
>>>
>>> frank
>>>
>>> On 26 January 2014 16:42, Chris Muller <ma.chris.m at gmail.com> wrote:
>>>> That's just a Warning, I thought release.st caught and resumed Warnings?
>>>>
>>>> On Sat, Jan 25, 2014 at 8:47 AM, Frank Shearar <frank.shearar at gmail.com> wrote:
>>>>> On 25 January 2014 09:56, Frank Shearar <frank.shearar at gmail.com> wrote:
>>>>>> On 25 January 2014 09:48, Frank Shearar <frank.shearar at gmail.com> wrote:
>>>>>>> On 24 January 2014 14:20, Frank Shearar <frank.shearar at gmail.com> wrote:
>>>>>>>> On 24 January 2014 14:09, Tobias Pape <Das.Linux at gmx.de> wrote:
>>>>>>>>> Hi:
>>>>>>>>> the just downloaded trunk (build 774)
>>>>>>>>> cannot be opened in Squeak 5.8b12.app
>>>>>>>>> Just fyi :)
>>>>>>>>
>>>>>>>> I can't look at it now (but will this evening), but I think I broke
>>>>>>>> the job from #775 onwards.
>>>>>>>
>>>>>>> http://build.squeak.org/job/ReleaseSqueakTrunk/748/console has a
>>>>>>> 4.5-13655. Took me long enough. Anyway, please take a look and see if
>>>>>>> this is a suitable release candidate.
>>>>>>>
>>>>>>> Thanks!
>>>>>>>
>>>>>>> (SqueakTrunk (a) takes a base image and updates it, and (b) tests it.
>>>>>>> ReleaseSqueakTrunk (a) takes a base image and updates it, (b) runs the
>>>>>>> release process, and (c) tests it.)
>>>>>>
>>>>>> I'll have a look myself later today, but one thing worries me about
>>>>>> the artifact mentioned above:
>>>>>>
>>>>>> Releasing based off Squeak4.5-13655
>>>>>> /var/lib/jenkins/workspace/ReleaseSqueakTrunk/release.st
>>>>>> spawning command 3 with timeout 900 seconds:
>>>>>> /var/lib/jenkins/workspace/ReleaseSqueakTrunk/target/cog.r2776/coglinux/bin/squeak
>>>>>> -vm-sound-null -vm-display-null
>>>>>> "/var/lib/jenkins/workspace/ReleaseSqueakTrunk/target/ReleaseCandidate.image"
>>>>>> ../release.st
>>>>>> release.st
>>>>>> (Command started with PID 1906)
>>>>>> !!! Killing command 3 for exceeding allotted time:
>>>>>> /var/lib/jenkins/workspace/ReleaseSqueakTrunk/target/cog.r2776/coglinux/bin/squeak
>>>>>> -vm-sound-null -vm-display-null
>>>>>> "/var/lib/jenkins/workspace/ReleaseSqueakTrunk/target/ReleaseCandidate.image"
>>>>>> ../release.st.
>>>>>>
>>>>>> That job has a 15 minute timeout attached to it. Preparing a new build
>>>>>> does take an age, so it might be that we need a longer timeout... or
>>>>>> there's a debugger popped up.
>>>>>
>>>>> "Display extent not set to 800 at 600
>>>>>
>>>>> Select Proceed to continue, or close this window to cancel the operation."
>>>>>
>>>>> That's why the command to release times out.
>>>>>
>>>>> frank


More information about the Squeak-dev mailing list