No, I mean I want the build to fail when the image can't update.
Hopefully last night's changes will address the problem. I now force
updating to happen with a Cog VM (so running "rake release" on an OS
where Cog doesn't work, will fail instantly with an appropriate error
message), and use Smalltalk snapshot: false andQuitWithErrorCode: 1.
(On an Interpreter VM that just means it quits with 0, but see
previous "update using Cog" point.) After the update, I force the
image format to an Interpreter friendly one.
Bear in mind that SqueakTrunk's build times should progressively
increase, because it takes longer & longer to update the image. Every
now and then we'll need to update the base image CI uses just for this
reason.
frank
On 22 January 2014 22:41, Chris Muller <
[hidden email]> wrote:
> I assume you mean you want to fix the timeout, not the conflict (since
> it looks like you have the correct response for the conflict).
>
> Hopefully it just needs to be increased. Trunk has been feeling sluggish.
>
>
> On Wed, Jan 22, 2014 at 3:35 PM, Frank Shearar <
[hidden email]> wrote:
>> On 22 January 2014 20:27, Chris Muller <
[hidden email]> wrote:
>>> Hi Frank, when I downloaded the Squeak4.5.zip located at
>>>
>>>
http://build.squeak.org/job/ReleaseSqueakTrunk/lastSuccessfulBuild/>>>
>>> it is still -13352 image, but that page says it was built yesterday.
>>>
>>> I thought I would find the latest trunk updates there?
>>
>> It's because the build doesn't blow up properly when things go wrong.
>> If you look at
http://build.squeak.org/job/SqueakTrunk/773/console>> you'll see this:
>>
>> 2014-01-22T17:01:33.898+01:00: Updating
http://source.squeak.org/trunk>> 2014-01-22T17:01:40.724+01:00: Checking
http://source.squeak.org/trunk>> 2014-01-22T17:02:53.152+01:00: Conflict; choosing remote:
>> 2014-01-22T17:02:53.176+01:00: Utilities class>>updateFromServer
>> 2014-01-22T17:02:53.192+01:00: Resolved? true
>>
>> SocketStream>>signalTimeout
>> [] in SocketStream>>waitForData
>> Socket>>waitForDataFor:ifClosed:ifTimedOut:
>> SocketStream>>waitForData
>>
>> I flagged this up, er, day before yesterday?, but haven't had a chance
>> to fix the problem. It's a very serious problem: until it's fixed, you
>> have no idea what kind of image will come out the SqueakTrunk job -
>> either 4.5-13352, or a fully updated one.
>>
>> frank
Locked
