[Pharo-project] [update 3.0] #30016
30016 ----- 10168 Spec new layout management https://pharo.fogbugz.com/f/cases/10168 Diff information: http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Layout-StephaneDucasse.3... Stef ;)
why is the new update already in http://pharo.gforge.inria.fr/updates/pharo3.0/updates30.list when it is still present in the staged file? http://pharo.gforge.inria.fr/updates/pharo3.0/updates30.staged does that mean that the staged update process isn't used properly yet? On 2013-03-29, at 15:33, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
30016 -----
10168 Spec new layout management https://pharo.fogbugz.com/f/cases/10168
Diff information: http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Layout-StephaneDucasse.3...
Stef ;)
no, it is... or it should be :) On Mar 29, 2013, at 3:46 PM, Camillo Bruni <camillobruni@gmail.com> wrote:
why is the new update already in http://pharo.gforge.inria.fr/updates/pharo3.0/updates30.list
when it is still present in the staged file? http://pharo.gforge.inria.fr/updates/pharo3.0/updates30.staged
does that mean that the staged update process isn't used properly yet?
On 2013-03-29, at 15:33, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
30016 -----
10168 Spec new layout management https://pharo.fogbugz.com/f/cases/10168
Diff information: http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Layout-StephaneDucasse.3...
Stef ;)
camillo I do not announce the update before the 4 stages are run and green. This slows us down so we code in between updates but this is using the stagged process. I do not even know how you can determine that we are not using it :). Stef On Mar 29, 2013, at 3:46 PM, Camillo Bruni <camillobruni@gmail.com> wrote:
why is the new update already in http://pharo.gforge.inria.fr/updates/pharo3.0/updates30.list
when it is still present in the staged file? http://pharo.gforge.inria.fr/updates/pharo3.0/updates30.staged
does that mean that the staged update process isn't used properly yet?
On 2013-03-29, at 15:33, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
30016 -----
10168 Spec new layout management https://pharo.fogbugz.com/f/cases/10168
Diff information: http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Layout-StephaneDucasse.3...
Stef ;)
On 2013-03-31, at 11:08, stephane ducasse <stephane.ducasse@free.fr> wrote:
camillo
I do not announce the update before the 4 stages are run and green. This slows us down so we code in between updates but this is using the stagged process. I do not even know how you can determine that we are not using it :).
I just check when you announced the update and I see that the last update was present in both lists (updates30.list and updates30.staged) while jenkins was still validating the update. This makes a bit suspicious :) since technically this state is impossible (well there are bugs of course ;)) 1. new updates prepared 2. updates are put into updates30.staged 3. change triggers jenkins 4. changes are validated 5. updates30.staged is copied into updates30.list, only here at this point can we have two updates present in both files 6. updates30.staged is deleted
Stef
On Mar 29, 2013, at 3:46 PM, Camillo Bruni <camillobruni@gmail.com> wrote:
why is the new update already in http://pharo.gforge.inria.fr/updates/pharo3.0/updates30.list
when it is still present in the staged file? http://pharo.gforge.inria.fr/updates/pharo3.0/updates30.staged
does that mean that the staged update process isn't used properly yet?
On 2013-03-29, at 15:33, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:
30016 -----
10168 Spec new layout management https://pharo.fogbugz.com/f/cases/10168
Diff information: http://smalltalkhub.com/#!/~Pharo/Pharo30/diff/Spec-Layout-StephaneDucasse.3...
Stef ;)
On 2013-03-31, at 11:12, Camillo Bruni <camillobruni@gmail.com> wrote:
On 2013-03-31, at 11:08, stephane ducasse <stephane.ducasse@free.fr> wrote:
camillo
I do not announce the update before the 4 stages are run and green. This slows us down so we code in between updates but this is using the stagged process. I do not even know how you can determine that we are not using it :).
I just check when you announced the update and I see that the last update was present in both lists (updates30.list and updates30.staged) while jenkins was still validating the update. This makes a bit suspicious :) since technically this state is impossible (well there are bugs of course ;))
1. new updates prepared 2. updates are put into updates30.staged 3. change triggers jenkins 4. changes are validated 5. updates30.staged is copied into updates30.list, only here at this point can we have two updates present in both files 6. updates30.staged is deleted
and this is where the bug happened ;) udpates30.staged was never properly deleted. sorry for the noise :P
I do not announce the update before the 4 stages are run and green. This slows us down so we code in between updates but this is using the stagged process. I do not even know how you can determine that we are not using it :).
I just check when you announced the update and I see that the last update was present in both lists (updates30.list and updates30.staged) while jenkins was still validating the update. This makes a bit suspicious :) since technically this state is impossible (well there are bugs of course ;))
1. new updates prepared 2. updates are put into updates30.staged 3. change triggers jenkins 4. changes are validated 5. updates30.staged is copied into updates30.list, only here at this point can we have two updates present in both files 6. updates30.staged is deleted
and this is where the bug happened ;) udpates30.staged was never properly deleted. sorry for the noise :P
no this is good that you checked. :) because we want to improve on our process. Stef
participants (4)
-
Camillo Bruni -
Esteban Lorenzano -
stephane ducasse -
Stéphane Ducasse