January 2014 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] unload all

by Pavel Krivanek

Ben, can you take care of your child? ;-) I'm not against making UserManager part of the kernel packages, or better only Monticello prerequisity. But we need to separate the UI part. -- Pavel 2014/1/18 Benjamin <benjamin.vanryseghem.pharo(a)gmail.com> > COOOOOOOLLLLLLL !!!!!! > > Ben > > On 18 Jan 2014, at 16:37, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: > > Hi, > > I finally successfully created a script that is able to unload everything > *by Monticello* except kernel packages, network, Monticello, Gofer and > related packages. I will setup a job when Jenkins will be healthy again. > It uses several temporary patches that we need to solve properly: > - make NativeBoost optional > - make UserManager optional > - small super-easy issue https://pharo.fogbugz > .com/f/cases/12679/change-package-of-ClassTrait-soleInstance > > The size of resultant image is about 6.6 MB, 4.9 MB without MC data and > caches. BTW in the full image Monticello data have about 5MB. The image > has no Undeclared nor obsolete classes and is really able to load something > using Gofer. > > With some care we can make it usable base for the next remodularizationeffort. > > Cheers, > -- Pavel > > >

Jan. 19, 2014

Re: [Pharo-dev] ConfigurationOfGetText broken?

by HernÃ¡n Morales Durand

Hi Alex, Gettext is a software for i18n (http://en.wikipedia.org/wiki/Gettext) HernÃ¡n 2014/1/18 Alexandre Bergel <alexandre.bergel(a)me.com> > What GetText is all about? > http://smalltalkhub.com/#!/~PharoExtras/Gettext does not help much > > Alexandre > > > On Jan 17, 2014, at 7:42 AM, Usman Bhatti <usman.bhatti(a)gmail.com> wrote: > > > Hello, > > > > It seems that some recent changes in the ConfigurationOfText have broken > the loading in the Pharo 3.0. > > > > I'm loading GetText with the following script in my config. > > spec > > project: 'GetText' > > with: [ > > spec > > className: 'ConfigurationOfGettext'; > > file: 'ConfigurationOfGettext'; > > loads: #('Core' ); > > version: '1.3'; "tried #development too" > > repository: ' > http://smalltalkhub.com/mc/PharoExtras/Gettext/main']. > > > > I get this error whatever version I put. > > > > 'The version ''1.3'' is not defined in ConfigurationOfGettext for the > current platform, because an exception occurred while creating the > version:. Error: The imported version:''1.1-baseline'' for version: 1.3 > referenced from the method: #version13: in configuration > ConfigurationOfGettext has not been defined.. Evaluate the following to see > the error: ''[ConfigurationOfGettext project ] > > on: MetacelloErrorInProjectConstructionNotification > > do: [:ex | ex resume: true ].'' > > Possible versions include: #(#bleedingEdge #development #stable > ''1.0-baseline'' ''1.0'' ''1.1'' ''1.2'')' > > > > I tried to debug and compared with the changes from the previous version > but didn't find any workable solution. > > > > Usman > > -- > _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: > Alexandre Bergel http://www.bergel.eu > ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > > > > >

Jan. 19, 2014

[ANN] MessageTally in GTInspector

by Tudor Girba

Hi, The GTInspector just became a performance analysis tool, too. You can simply inspect a MessageTally and you get several useful views that help you identify performance problems. You can read a more detailed description here: http://www.humane-assessment.com/blog/assessing-pharo-performance-with-gtin… As an appetizer, I attached a screenshot with a Graph-ET chart (thank you Daniel Aviv for developing this engine). [image: Inline image 1] To play with the code, you can just work in the latest Moose 5.0 image: https://ci.inria.fr/moose/job/moose-5.0/lastSuccessfulBuild/artifact/moose-… Please let me know what you think. Cheers, Doru -- www.tudorgirba.com "Every thing has its own flow"

Jan. 18, 2014

Re: [Pharo-dev] unload all

by Sven Van Caekenberghe

On 18 Jan 2014, at 20:53, Benjamin <benjamin.vanryseghem.pharo(a)gmail.com> wrote: > COOOOOOOLLLLLLL !!!!!! +1 And interesting as well (I want to understand how to unload NB, for example). > Ben > > On 18 Jan 2014, at 16:37, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: > >> Hi, >> >> I finally successfully created a script that is able to unload everything *by Monticello* except kernel packages, network, Monticello, Gofer and related packages. I will setup a job when Jenkins will be healthy again. >> It uses several temporary patches that we need to solve properly: >> - make NativeBoost optional >> - make UserManager optional >> - small super-easy issue https://pharo.fogbugz.com/f/cases/12679/change-package-of-ClassTrait-soleIn… >> >> The size of resultant image is about 6.6 MB, 4.9 MB without MC data and caches. BTW in the full image Monticello data have about 5MB. The image has no Undeclared nor obsolete classes and is really able to load something using Gofer. >> >> With some care we can make it usable base for the next remodularization effort. >> >> Cheers, >> -- Pavel >> >

Jan. 18, 2014

Re: [Pharo-dev] ConfigurationOfGetText broken?

by Alexandre Bergel

What GetText is all about? http://smalltalkhub.com/#!/~PharoExtras/Gettext does not help much Alexandre On Jan 17, 2014, at 7:42 AM, Usman Bhatti <usman.bhatti(a)gmail.com> wrote: > Hello, > > It seems that some recent changes in the ConfigurationOfText have broken the loading in the Pharo 3.0. > > I'm loading GetText with the following script in my config. > spec > project: 'GetText' > with: [ > spec > className: 'ConfigurationOfGettext'; > file: 'ConfigurationOfGettext'; > loads: #('Core' ); > version: '1.3'; "tried #development too" > repository: 'http://smalltalkhub.com/mc/PharoExtras/Gettext/main']. > > I get this error whatever version I put. > > 'The version ''1.3'' is not defined in ConfigurationOfGettext for the current platform, because an exception occurred while creating the version:. Error: The imported version:''1.1-baseline'' for version: 1.3 referenced from the method: #version13: in configuration ConfigurationOfGettext has not been defined.. Evaluate the following to see the error: ''[ConfigurationOfGettext project ] > on: MetacelloErrorInProjectConstructionNotification > do: [:ex | ex resume: true ].'' > Possible versions include: #(#bleedingEdge #development #stable ''1.0-baseline'' ''1.0'' ''1.1'' ''1.2'')' > > I tried to debug and compared with the changes from the previous version but didn't find any workable solution. > > Usman -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Jan. 18, 2014

Re: [Pharo-dev] unload all

by StÃ©phane Ducasse

Excellent and impressive. Pavel did you publish it in SystemConfiguration? I would really like to see if we can add a unload protocol to the package manifest or configuration because it would be good to have the knowlegde modular. Stef > Hi, > > I finally successfully created a script that is able to unload everything *by Monticello* except kernel packages, network, Monticello, Gofer and related packages. I will setup a job when Jenkins will be healthy again. > It uses several temporary patches that we need to solve properly: > - make NativeBoost optional > - make UserManager optional > - small super-easy issue https://pharo.fogbugz.com/f/cases/12679/change-package-of-ClassTrait-soleIn… > > The size of resultant image is about 6.6 MB, 4.9 MB without MC data and caches. BTW in the full image Monticello data have about 5MB. The image has no Undeclared nor obsolete classes and is really able to load something using Gofer. > > With some care we can make it usable base for the next remodularization effort. > > Cheers, > -- Pavel >

Jan. 18, 2014

Re: [Pharo-dev] unload all

by Hilaire Fernandes

FANTASTIC! We have another Pharo Hero! Le 18/01/2014 20:37, Pavel Krivanek a Ã©crit : > I finally successfully created a script that is able to unload > everything *by Monticello* except kernel packages, network, Monticello, > Gofer and related -- Dr. Geo http://drgeo.eu

Jan. 18, 2014

Re: [Pharo-dev] unload all

by Tudor Girba

This is beautiful! Doru On Sat, Jan 18, 2014 at 8:37 PM, Pavel Krivanek <pavel.krivanek(a)gmail.com>wrote: > Hi, > > I finally successfully created a script that is able to unload everything > *by Monticello* except kernel packages, network, Monticello, Gofer and > related packages. I will setup a job when Jenkins will be healthy again. > It uses several temporary patches that we need to solve properly: > - make NativeBoost optional > - make UserManager optional > - small super-easy issue https://pharo.fogbugz > .com/f/cases/12679/change-package-of-ClassTrait-soleInstance > > The size of resultant image is about 6.6 MB, 4.9 MB without MC data and > caches. BTW in the full image Monticello data have about 5MB. The image > has no Undeclared nor obsolete classes and is really able to load something > using Gofer. > > With some care we can make it usable base for the next remodularizationeffort. > > Cheers, > -- Pavel > > -- www.tudorgirba.com "Every thing has its own flow"

Jan. 18, 2014

Re: [Pharo-dev] Announcer registry behaviour with Spy

by Nicolai Hess

2014/1/14 <btc(a)openinworld.com> > Igor Stasenko wrote: > > > > > On 14 January 2014 14:04, <btc(a)openinworld.com> wrote: > >> While investigating >> https://pharo.fogbugz.com/f/cases/12640/PhLTitledTreeModel-refresh-called-t… >> I came across the following behaviour... >> >> 1. In Workspace do... "Announcer new inspect." >> >> 2. In that Announcer inspector do... "registry inspect." >> >> 3. In that SubscriptionRegistry inspector monitor the 'subscriptions' >> ivar for subsequent steps. >> * 'subscription' is initially an empty IdentitySet. >> >> 4. In the Announcer inspector do... "self open" >> * window "Spy: an Announcer" appears >> * 'subscriptions' now contains one AnnouncementSubscription >> >> 5. In the Announcer inspector again do... "self open" >> * another window "Spy: an Announcer" appears >> * 'subscriptions' now contains two AnnouncementSubscription >> >> 6. Close both Spy windows >> * 'subscriptions' still contains two AnnouncementSubscription >> >> Is the result of step 6 expected and/or desired? Intuitively I would >> expect 'subscriptions' to be empty. >> Doing "Smalltalk garbageCollect" has no effect. >> >> > maybe because AnnouncementSpy should use weak subscriptions or take care > about unsubscribing once it is closed? > it seems like AnnouncementSpy never receives #close message.. (because > if you do it should unsubscribe) > either way this is bug and should be fixed. > > > cheers -ben >> >> >> > > > -- > Best regards, > Igor Stasenko. > > Thanks Igor. Thats something I might have a go at. > > https://pharo.fogbugz.com/f/cases/12649/Closing-AnnouncementSpy-does-not-un… > I just implemented windowIsClosing, as this is called by closing the system window. It calls self close in AnnouncementSpy for unregistering from the Announcer.

Jan. 18, 2014

Re: [Pharo-dev] unload all

by Benjamin

COOOOOOOLLLLLLL !!!!!! Ben On 18 Jan 2014, at 16:37, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: > Hi, > > I finally successfully created a script that is able to unload everything *by Monticello* except kernel packages, network, Monticello, Gofer and related packages. I will setup a job when Jenkins will be healthy again. > It uses several temporary patches that we need to solve properly: > - make NativeBoost optional > - make UserManager optional > - small super-easy issue https://pharo.fogbugz.com/f/cases/12679/change-package-of-ClassTrait-soleIn… > > The size of resultant image is about 6.6 MB, 4.9 MB without MC data and caches. BTW in the full image Monticello data have about 5MB. The image has no Undeclared nor obsolete classes and is really able to load something using Gofer. > > With some care we can make it usable base for the next remodularization effort. > > Cheers, > -- Pavel >

Jan. 18, 2014