January 2011 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] Deprecations and Hudson

by Mariano Martinez Peck

On Sun, Jan 23, 2011 at 12:05 PM, Marcus Denker <marcus.denker(a)inria.fr>wrote: > > On Jan 23, 2011, at 11:27 AM, Marcus Denker wrote: > > > ((Smalltalk at: #ConfigurationOfPharo) project lastVersion) perform: > #silently: with: true; perform: #load.Deprecation raiseWarning: true. > > > This way, the build image can be always build and then at runtime, the user > will see the popups. > > Marcus, please, let me know if you change this becasue I need it to fix a > problem with Metacello. > > > done. > > > Ok, so it breaks the build when running tests. I removed turning on > Deprecation warnings > Ok BUT, the final image, (the one that the user can download), must have the flag in true. can you do that ? > (last line) > lets see. > > > > -- > Marcus Denker -- http://www.marcusdenker.de > INRIA Lille -- Nord Europe. Team RMoD. > >

Jan. 23, 2011

Re: [Pharo-project] Deprecations and Hudson

by Henrik Sperre Johansen

On 23.01.2011 12:05, Marcus Denker wrote: > > On Jan 23, 2011, at 11:27 AM, Marcus Denker wrote: >>> >>> ((Smalltalk at: #ConfigurationOfPharo) project lastVersion) perform: >>> #silently: with: true; perform: #load.Deprecation raiseWarning: true. >>> >>> >>> This way, the build image can be always build and then at runtime, >>> the user will see the popups. >>> >>> Marcus, please, let me know if you change this becasue I need it to >>> fix a problem with Metacello. >> >> done. > > Ok, so it breaks the build when running tests. I removed turning on > Deprecation warnings > (last line) > lets see. > > > > -- > Marcus Denker -- http://www.marcusdenker.de <http://www.marcusdenker.de/> > INRIA Lille -- Nord Europe. Team RMoD. > I'd suggest using Deprecation>>#deprecationsWhile: That way you can report them in Hudson, so they are easier to fix. Cheers, Henry

Jan. 23, 2011

Re: [Pharo-project] Deprecations and Hudson

by Marcus Denker

On Jan 23, 2011, at 11:27 AM, Marcus Denker wrote: >> >> ((Smalltalk at: #ConfigurationOfPharo) project lastVersion) perform: #silently: with: true; perform: #load. >> Deprecation raiseWarning: true. >> >> >> This way, the build image can be always build and then at runtime, the user will see the popups. >> >> Marcus, please, let me know if you change this becasue I need it to fix a problem with Metacello. > > > done. Ok, so it breaks the build when running tests. I removed turning on Deprecation warnings (last line) lets see. -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

Jan. 23, 2011

Re: [Pharo-project] Deprecations and Hudson

by Marcus Denker

On Jan 23, 2011, at 10:57 AM, Mariano Martinez Peck wrote: > Hi. With Hudson we want to automatically build images. It can happens that external packages did not yet update their code and that are still using depreacted methods. So, by default we raise an exception. This is good because when the developer see an annoying popup, maybe he fix it ;) > > Now, the problem is that Hudson FAILS if there is a deprecated. So, I suggest changing the hudson dev build script to: > > > Deprecation raiseWarning: false. > Gofer new > squeaksource: 'MetacelloRepository'; > package: 'ConfigurationOfPharo'; > load. > > ((Smalltalk at: #ConfigurationOfPharo) project lastVersion) perform: #silently: with: true; perform: #load. > Deprecation raiseWarning: true. > > > This way, the build image can be always build and then at runtime, the user will see the popups. > > Marcus, please, let me know if you change this becasue I need it to fix a problem with Metacello. done. -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

Jan. 23, 2011

Re: [Pharo-project] default styler is shout

by Mariano Martinez Peck

I think Tudor is right. We cannot have shout everywhere by default. Take an example. Today morning, I opened the latest PharoDev and I couldn't understand why my nice welcome workspace was converted to an ugly red text workspace, like if it was code, which was not, of course. Now I read this email and I think it is related. Take a look to the PharoDev image from hudson and you will see what I am talking about. +1 not having Shout for everybody and instead, have the previous behavior. Of course I have no idea how to do it because otherwise I would have done it. Now, if we change to what Doru says, so I think some senders have to be updated if they would expect to have shout as the default. For example, FinderUI. I guess it would like to use the default/null styler in a PharoCore image, but in presence of Shout, it would probably like to use shout styler. But in such case, maybe he can use #styler: rather that #setDefaultStyler ? cheers mariano On Sun, Jan 23, 2011 at 8:47 AM, Marcus Denker <marcus.denker(a)inria.fr>wrote: > Hi, > > We need the imperfect, intermediate solution that is good enough for people > to live with > in 1.2. > > On Jan 22, 2011, at 4:25 PM, Tudor Girba wrote: > > > Hi, > > > > In the latest Morphic version, the default styler of the > PluggableTextMorph is given by: > > useDefaultStyler > > "This should be changed to a proper registry but as long as there > is only shout this will do" > > > > Smalltalk globals > > at: #SHTextStylerST80 > > ifPresent: [ :stylerClass | self styler: (stylerClass new > view: self) ] > > ifAbsent: [ self styler: ( NullTextStyler new view: self) > ] > > > > Thus, if Shout is loaded, you will get it in all text morphs by default. > Given that TextMorph should be general and be used in all sorts of contexts, > the default behavior should be the original: > > > > useDefaultStyler > > "This should be changed to a proper registry but as long as there > is only shout this will do" > > > > self styler: (NullTextStyler new view: self; yourself). > > > > > > What do you say? > > > > Cheers, > > Doru > > > > -- > > www.tudorgirba.com > > > > "We are all great at making mistakes." > > > > > > > > > > > > > > > > > > -- > Marcus Denker -- http://www.marcusdenker.de > INRIA Lille -- Nord Europe. Team RMoD. > > >

Jan. 23, 2011

Re: [Pharo-project] Issue 3579 in pharo: "declare temp" adds the variable twice

by pharo＠googlecode.com

Comment #1 on issue 3579 by marianopeck: "declare temp" adds the variable twice http://code.google.com/p/pharo/issues/detail?id=3579 naaaaa what a bug! I have just tested and I can reproduce it not only in OBSystemBrowserAdaptor but also with the default browser (the one it is in pharocore). However, in the PharoCore I cannot reproduce it. Maybe it is related to shout and its smart characters??

Jan. 23, 2011

[Pharo-project] Deprecations and Hudson

by Mariano Martinez Peck

Hi. With Hudson we want to automatically build images. It can happens that external packages did not yet update their code and that are still using depreacted methods. So, by default we raise an exception. This is good because when the developer see an annoying popup, maybe he fix it ;) Now, the problem is that Hudson FAILS if there is a deprecated. So, I suggest changing the hudson dev build script to: Deprecation raiseWarning: false. Gofer new squeaksource: 'MetacelloRepository'; package: 'ConfigurationOfPharo'; load. ((Smalltalk at: #ConfigurationOfPharo) project lastVersion) perform: #silently: with: true; perform: #load.Deprecation raiseWarning: true. This way, the build image can be always build and then at runtime, the user will see the popups. Marcus, please, let me know if you change this becasue I need it to fix a problem with Metacello. mariano

Jan. 23, 2011

Re: [Pharo-project] Issue 3575 in pharo: [Failing Tests] Tests.Release.ReleaseTest.testObsoleteClasses

by pharo＠googlecode.com

Comment #1 on issue 3575 by marcus.d...(a)gmail.com: [Failing Tests] Tests.Release.ReleaseTest.testObsoleteClasses http://code.google.com/p/pharo/issues/detail?id=3575 AnObsoleteMCMockClassG

Jan. 23, 2011

[Pharo-project] Issue 3581 in pharo: [Failing test] 1.1.1 Full: MethodWrappers.Tests.ObjectAsOneTimeMethodWrapperTest.testInstallOnClassCategory

by pharo＠googlecode.com

Status: Accepted Owner: marcus.d...(a)gmail.com Labels: Milestone-1.1.2 New issue 3581 by marcus.d...(a)gmail.com: [Failing test] 1.1.1 Full: MethodWrappers.Tests.ObjectAsOneTimeMethodWrapperTest.testInstallOnClassCategory http://code.google.com/p/pharo/issues/detail?id=3581 MethodWrappers.Tests.ObjectAsOneTimeMethodWrapperTest.testInstallOnClassCategory fails sometimes.

Jan. 23, 2011

Re: [Pharo-project] Issue 3576 in pharo: [Failing Tests] Hudson Full 1.1 Tests.Release.ReleaseTest.testUndeclared

by pharo＠googlecode.com

Comment #2 on issue 3576 by marcus.d...(a)gmail.com: [Failing Tests] Hudson Full 1.1 Tests.Release.ReleaseTest.testUndeclared http://code.google.com/p/pharo/issues/detail?id=3576 #(#MetacelloSymbolicVersionMethodSpec #MetacelloMCVersionValidator #MetacelloCleanLoadAndTestsNotification #MetacelloCannotUpdateReleasedVersionError #MetacelloSymbolicVersionDoesNotExistError #MetacelloCleanNotification #MetacelloToolBoxConstructor #MetacelloVersionMethodSpec #MetacelloVersionDoesNotExistError #MetacelloValidationWarning #MetacelloCleanLoadNotification)

Jan. 23, 2011