I will tomorrow :) Ben On 04 Nov 2013, at 13:22, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:

Ben

this is strange when I look at these changes I do not really see that value is used instead of contents. Can you check to see if your changes got really integrated?

Stef On Nov 3, 2013, at 8:17 PM, GitHub <noreply@github.com> wrote:

...

Branch: refs/heads/3.0 Home: https://github.com/pharo-project/pharo-core Commit: a5bce43720062026ab415167af4db194d3958d62 https://github.com/pharo-project/pharo-core/commit/a5bce43720062026ab415167a... Author: Jenkins Build Server <board@pharo-project.org> Date: 2013-11-03 (Sun, 03 Nov 2013)

Changed paths: A ScriptLoader30.package/ScriptLoader.class/instance/pharo - scripts/script199.st A ScriptLoader30.package/ScriptLoader.class/instance/pharo - updates/update30544.st M ScriptLoader30.package/ScriptLoader.class/instance/public/commentForCurrentUpdate.st M Spec-Core.package/TextModel.class/definition.st M Spec-Core.package/TextModel.class/instance/initialization/initialize.st A Spec-Core.package/TextModel.class/instance/protocol-events/whenMenuHolderChanged_.st M Spec-Core.package/TextModel.class/instance/protocol/codePaneMenu_shifted_.st A Spec-Core.package/TextModel.class/instance/protocol/menuHolder.st A Spec-Core.package/TextModel.class/instance/protocol/menuHolder_.st A StartupPreferences.package/extension/SettingBrowser/instance/exportAllSettings_by_withBasename_.st M StartupPreferences.package/extension/SettingBrowser/instance/exportSettings.st A StartupPreferences.package/extension/SettingBrowser/instance/exportSettings_toFileNamed_.st

Log Message: ----------- 30544 12061 SettingBrowser>>exportSettings need refactoring https://pharo.fogbugz.com/f/cases/12061

11920 NewValueHolder should use value instead of contents https://pharo.fogbugz.com/f/cases/11920

11819 TextModel>>#codePaneMenu:shifted: should work with a value holder https://pharo.fogbugz.com/f/cases/11819

http://files.pharo.org/image/30/30544.zip

So there must be some serious MC issue. On 2013-11-05, at 11:15, Benjamin <benjamin.vanryseghem.pharo@gmail.com> wrote:

Indeed, it seems to not be in the system.

Most probably because afterward another spec related issue has been integrated, and that the diff were not properly detected. (Already happened)

Ben

On 04 Nov 2013, at 23:06, Benjamin <benjamin.vanryseghem.pharo@gmail.com> wrote:

...

I will tomorrow :)

Ben

On 04 Nov 2013, at 13:22, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:

...

Ben

this is strange when I look at these changes I do not really see that value is used instead of contents. Can you check to see if your changes got really integrated?

Stef On Nov 3, 2013, at 8:17 PM, GitHub <noreply@github.com> wrote:

...

Branch: refs/heads/3.0 Home: https://github.com/pharo-project/pharo-core Commit: a5bce43720062026ab415167af4db194d3958d62 https://github.com/pharo-project/pharo-core/commit/a5bce43720062026ab415167a... Author: Jenkins Build Server <board@pharo-project.org> Date: 2013-11-03 (Sun, 03 Nov 2013)

Changed paths: A ScriptLoader30.package/ScriptLoader.class/instance/pharo - scripts/script199.st A ScriptLoader30.package/ScriptLoader.class/instance/pharo - updates/update30544.st M ScriptLoader30.package/ScriptLoader.class/instance/public/commentForCurrentUpdate.st M Spec-Core.package/TextModel.class/definition.st M Spec-Core.package/TextModel.class/instance/initialization/initialize.st A Spec-Core.package/TextModel.class/instance/protocol-events/whenMenuHolderChanged_.st M Spec-Core.package/TextModel.class/instance/protocol/codePaneMenu_shifted_.st A Spec-Core.package/TextModel.class/instance/protocol/menuHolder.st A Spec-Core.package/TextModel.class/instance/protocol/menuHolder_.st A StartupPreferences.package/extension/SettingBrowser/instance/exportAllSettings_by_withBasename_.st M StartupPreferences.package/extension/SettingBrowser/instance/exportSettings.st A StartupPreferences.package/extension/SettingBrowser/instance/exportSettings_toFileNamed_.st

Log Message: ----------- 30544 12061 SettingBrowser>>exportSettings need refactoring https://pharo.fogbugz.com/f/cases/12061

11920 NewValueHolder should use value instead of contents https://pharo.fogbugz.com/f/cases/11920

11819 TextModel>>#codePaneMenu:shifted: should work with a value holder https://pharo.fogbugz.com/f/cases/11819

http://files.pharo.org/image/30/30544.zip

it shouldn't be an MC issue. the integrator is configured in "merge accepting upcoming changes". So if you integrate and accept without you picking the right version, when the automatic process run it blindly overrides the previous change. So I would call it "a human mistake", before an MC one. unless, of course, there is no mistake and MC failed to see the change, who would be too serious to notice it just right now, don't you think? Esteban On Tue, Nov 5, 2013 at 8:38 AM, Nicolas Cellier < nicolas.cellier.aka.nice@gmail.com> wrote:

That was my impression too. But the interaction with RPackage Annoucement etc.. makes it hard to isolate. I suspect these layers, because I use MC everyday in Squeak where it's much more robust. That means that tests are not currently sufficient and we should focus on this (MC used to have a fairly good coverage).

2013/11/5 Camillo Bruni <camillobruni@gmail.com>

...

So there must be some serious MC issue.

On 2013-11-05, at 11:15, Benjamin <benjamin.vanryseghem.pharo@gmail.com> wrote:

...

Indeed, it seems to not be in the system.

Most probably because afterward another spec related issue has been integrated, and that the diff were not properly detected. (Already happened)

Ben

On 04 Nov 2013, at 23:06, Benjamin < benjamin.vanryseghem.pharo@gmail.com> wrote:

...

I will tomorrow :)

Ben

On 04 Nov 2013, at 13:22, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:

...

Ben

this is strange when I look at these changes I do not really see that value is used instead of contents. Can you check to see if your changes got really integrated?

Stef On Nov 3, 2013, at 8:17 PM, GitHub <noreply@github.com> wrote:

...

Branch: refs/heads/3.0 Home: https://github.com/pharo-project/pharo-core Commit: a5bce43720062026ab415167af4db194d3958d62

https://github.com/pharo-project/pharo-core/commit/a5bce43720062026ab415167a...

...

Author: Jenkins Build Server <board@pharo-project.org> Date: 2013-11-03 (Sun, 03 Nov 2013)

Changed paths: A ScriptLoader30.package/ScriptLoader.class/instance/pharo - scripts/script199.st A ScriptLoader30.package/ScriptLoader.class/instance/pharo - updates/update30544.st M ScriptLoader30.package/ScriptLoader.class/instance/public/commentForCurrentUpdate.st M Spec-Core.package/TextModel.class/definition.st M Spec-Core.package/TextModel.class/instance/initialization/ initialize.st A Spec-Core.package/TextModel.class/instance/protocol-events/whenMenuHolderChanged_.st M Spec-Core.package/TextModel.class/instance/protocol/codePaneMenu_shifted_.st A Spec-Core.package/TextModel.class/instance/protocol/menuHolder.st A Spec-Core.package/TextModel.class/instance/protocol/menuHolder_.st A StartupPreferences.package/extension/SettingBrowser/instance/exportAllSettings_by_withBasename_.st M StartupPreferences.package/extension/SettingBrowser/instance/exportSettings.st A StartupPreferences.package/extension/SettingBrowser/instance/exportSettings_toFileNamed_.st

Log Message: ----------- 30544 12061 SettingBrowser>>exportSettings need refactoring https://pharo.fogbugz.com/f/cases/12061

11920 NewValueHolder should use value instead of contents https://pharo.fogbugz.com/f/cases/11920

11819 TextModel>>#codePaneMenu:shifted: should work with a value holder https://pharo.fogbugz.com/f/cases/11819

http://files.pharo.org/image/30/30544.zip

On 05 Nov 2013, at 13:51, Esteban Lorenzano <estebanlm@gmail.com> wrote:

it shouldn't be an MC issue. the integrator is configured in "merge accepting upcoming changes". So if you integrate and accept without you picking the right version, when the automatic process run it blindly overrides the previous change.

Then this is clearly a problem. It works in case fixes are made linearly one after the other. Which is far from being the case Ben

So I would call it "a human mistake", before an MC one.

unless, of course, there is no mistake and MC failed to see the change, who would be too serious to notice it just right now, don't you think?

Esteban

On Tue, Nov 5, 2013 at 8:38 AM, Nicolas Cellier <nicolas.cellier.aka.nice@gmail.com> wrote: That was my impression too. But the interaction with RPackage Annoucement etc.. makes it hard to isolate. I suspect these layers, because I use MC everyday in Squeak where it's much more robust. That means that tests are not currently sufficient and we should focus on this (MC used to have a fairly good coverage).

2013/11/5 Camillo Bruni <camillobruni@gmail.com> So there must be some serious MC issue.

On 2013-11-05, at 11:15, Benjamin <benjamin.vanryseghem.pharo@gmail.com> wrote:

...

Indeed, it seems to not be in the system.

Most probably because afterward another spec related issue has been integrated, and that the diff were not properly detected. (Already happened)

Ben

On 04 Nov 2013, at 23:06, Benjamin <benjamin.vanryseghem.pharo@gmail.com> wrote:

...

I will tomorrow :)

Ben

On 04 Nov 2013, at 13:22, Stéphane Ducasse <stephane.ducasse@inria.fr> wrote:

...

Ben

this is strange when I look at these changes I do not really see that value is used instead of contents. Can you check to see if your changes got really integrated?

Stef On Nov 3, 2013, at 8:17 PM, GitHub <noreply@github.com> wrote:

...

Branch: refs/heads/3.0 Home: https://github.com/pharo-project/pharo-core Commit: a5bce43720062026ab415167af4db194d3958d62 https://github.com/pharo-project/pharo-core/commit/a5bce43720062026ab415167a... Author: Jenkins Build Server <board@pharo-project.org> Date: 2013-11-03 (Sun, 03 Nov 2013)

Changed paths: A ScriptLoader30.package/ScriptLoader.class/instance/pharo - scripts/script199.st A ScriptLoader30.package/ScriptLoader.class/instance/pharo - updates/update30544.st M ScriptLoader30.package/ScriptLoader.class/instance/public/commentForCurrentUpdate.st M Spec-Core.package/TextModel.class/definition.st M Spec-Core.package/TextModel.class/instance/initialization/initialize.st A Spec-Core.package/TextModel.class/instance/protocol-events/whenMenuHolderChanged_.st M Spec-Core.package/TextModel.class/instance/protocol/codePaneMenu_shifted_.st A Spec-Core.package/TextModel.class/instance/protocol/menuHolder.st A Spec-Core.package/TextModel.class/instance/protocol/menuHolder_.st A StartupPreferences.package/extension/SettingBrowser/instance/exportAllSettings_by_withBasename_.st M StartupPreferences.package/extension/SettingBrowser/instance/exportSettings.st A StartupPreferences.package/extension/SettingBrowser/instance/exportSettings_toFileNamed_.st

Log Message: ----------- 30544 12061 SettingBrowser>>exportSettings need refactoring https://pharo.fogbugz.com/f/cases/12061

11920 NewValueHolder should use value instead of contents https://pharo.fogbugz.com/f/cases/11920

11819 TextModel>>#codePaneMenu:shifted: should work with a value holder https://pharo.fogbugz.com/f/cases/11819

http://files.pharo.org/image/30/30544.zip

On 05 Nov 2013, at 14:13, Camillo Bruni <camillobruni@gmail.com> wrote:

On 2013-11-05, at 14:09, Benjamin <benjamin.vanryseghem.pharo@gmail.com> wrote:

...

On 05 Nov 2013, at 13:51, Esteban Lorenzano <estebanlm@gmail.com> wrote:

...

it shouldn't be an MC issue. the integrator is configured in "merge accepting upcoming changes". So if you integrate and accept without you picking the right version, when the automatic process run it blindly overrides the previous change.

Then this is clearly a problem.

It works in case fixes are made linearly one after the other. Which is far from being the case

but we even proposed, several times, a remerged SLICE which would for instance produce again a merge conflict. So there is a strange problem around, which has something to do with MC.

This is for sure another issue :) But I guess the value holder fix got lost because I - produce the slice A for value holder - produce the slice B for something else spec related - A got integrated - B got integrated. And of course in this scenario, if conflicts are not resolved, of if merge is not done properly, it tools back A Ben