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

[Pharo-project] Issue 3911 in pharo: Change updater to allow updating a release image

by pharo＠googlecode.com

Status: Accepted Owner: marcus.d...(a)gmail.com Labels: Milestone-1.3 New issue 3911 by marcus.d...(a)gmail.com: Change updater to allow updating a release image http://code.google.com/p/pharo/issues/detail?id=3911 right now (1.2 and 1.3), the update logic checks for some character in the version number to allow updating and image. Now the fact is that people only look at released images. Leading to lots of suggestions of fixes directly after a release, but no way to update that release. Allowing the update process to update release images will remove a *significant* amount of work from the release manager who now needs to do a complete release cycle even for point releases (1.2.1a followed by 1.2.1 folowed by 1.2.2a followed by 1.2.2... with lots of manual intervention).

March 29, 2011

Re: [Pharo-project] Migration morph objects from 1.1 to 1.2

by Denis Kudriashov

2011/3/30 Marcus Denker <marcus.denker(a)inria.fr> > > No, I don't think so... keeping the system compatibel at that level is > *really* difficult... simple refactoring can be done > (e.g. etoys did that for years), but complex changes in the image will be > nearly impossible if you require serializability compatibility > at that level. (especially when using image segment style low-level > serialization). > > Being compatible on that level was religion for Squeak (and it's even > today). I personally don't think that what we have is > valuable enough to define compatibility on that level to be more impotant > than being able to evolve the system. > > The system as it is now is just not good enough to have any form of future > with it, so trading "improve it" vs. "be backward compatible > on a binary level" is selection between having the past and having a > future... > > I agree with you absolutelly. And I solve my problem :) With smalltalk compatibility is not problem to my mind. Just open debugger and you can do everything :) But I really want ask where I can read about morphic changes between 1.1 and 1.2 pharo. With this my migration should be easy

March 29, 2011

Re: [Pharo-project] [ANN 1.2.1Core] pre-build 1.2.1 Core image

by Marcus Denker

On Mar 29, 2011, at 10:18 PM, StÃ©phane Ducasse wrote: >> Pharo 1.2.1 looks good so far, except for the GLMOrangeUITheme that is >> still in a circular dependency. >> >> Lukas > > Can you open a ticket? We can do a 1.2.2... and add the TextContstants at:put: fix, too. Marcus -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

March 29, 2011

Re: [Pharo-project] Migration morph objects from 1.1 to 1.2

by StÃ©phane Ducasse

Hi denis Check the archive of the list since we already explained it at least two times :) Let us know if you have questions after. Basically turning the last dictionary into a SharedPool. We should finish to clean the sharedPool now. Stef On Mar 29, 2011, at 10:12 PM, Denis Kudriashov wrote: > Ok, I load my morph. > > I just implement "TextConstants class>>at: aSymbol" with: "^classPool at: aSymbol". > Now all work good. But I want to now what really changed in TextConstants stuff? > > > 2011/3/29 Denis Kudriashov <dionisiydk(a)gmail.com> > Hello > > I have many morphs saved in files by smartReferenceStream. I doing it in pharo1.1 image. > Now I want to load this objects to pharo1.2. > When I doing this I have exception about absent TextConstants stuff. My saved Morph containes textMorph. > I know pharo1.2 has textMorph refactored and TextConstants was removed. > > How old object can be migrated to new stuff? Is some rules exists? > For example, when I import morph from file (created in pharo1.1) I see in debugger "TextConstants class at: index" where "index=#DefaultTabsArray" > > Best regards, > Denis > >

March 29, 2011

[Pharo-project] Issue 3910 in pharo: -0.01s2 printString is loosing its sign

by pharo＠googlecode.com

Status: Accepted Owner: stephane...(a)gmail.com CC: nicolas....(a)gmail.com New issue 3910 by stephane...(a)gmail.com: -0.01s2 printString is loosing its sign http://code.google.com/p/pharo/issues/detail?id=3910 -0.01s2 printString loosing its sign

March 29, 2011

Re: [Pharo-project] [ANN 1.2.1Core] pre-build 1.2.1 Core image

by StÃ©phane Ducasse

> Pharo 1.2.1 looks good so far, except for the GLMOrangeUITheme that is > still in a circular dependency. > > Lukas Can you open a ticket? I want to clean all the themes soon in 1.3 Stef

March 29, 2011

Re: [Pharo-project] [ANN 1.2.1Core] pre-build 1.2.1 Core image

by Marcus Denker

On Mar 29, 2011, at 10:14 PM, Lukas Renggli wrote: > Pharo 1.2.1 looks good so far, except for the GLMOrangeUITheme that is > still in a circular dependency. > There was no bug report filed. I nearly asked when you sent the mail if this was meant to be a bug report... but I thought you would do it if you would think that it's a show stopper. Marcus -- Marcus Denker -- http://www.marcusdenker.de INRIA Lille -- Nord Europe. Team RMoD.

March 29, 2011

[Pharo-project] [update 1.3] #13118

by StÃ©phane Ducasse

13118 ----- - Issue 3429: Better Number performance. Thanks Levente and Nicolas Cellier. - Issue 3845: Polymorph extension shortSummary is not programmed the object way. Thanks Nicolas Cellier. - Issue 3902: Text field converter refactoring. Thanks Gary Chambers. - Issue 3713: Additional obsolete methods. Thanks Edmond - Issue 3878: #enableTextChange selector name is misleading, i propose changing the name to #canChangeText. Thanks Fernando Olivero. Thanks guys..... More tomorrow

March 29, 2011

Re: [Pharo-project] Issue 3429 in pharo: Better Number performance

by StÃ©phane Ducasse

Thanks. No stress :). even if the minus can make all the difference :) Stef > Sorry, it seems I let another bug, like -0.01s2 printString loosing its sign. > There might be some tests missing in KernelTests-Numbers > > I will open a new issue and give a new patch from trunk. > > Nicolas > > 2011/3/29 <pharo(a)googlecode.com>: >> >> Comment #4 on issue 3429 by stephane...(a)gmail.com: Better Number performance >> http://code.google.com/p/pharo/issues/detail?id=3429 >> >> in 13118 >> >> >> >

March 29, 2011

Re: [Pharo-project] [squeak-dev] Immutability, read-only objects

by Eliot Miranda

Hi Janko, 2011/3/27 Janko MivÅ¡ek <janko.mivsek(a)eranova.si> > Dear all, > > What is current status of immutable/read-only support in VM? Or more > broader, object instvar change catching support (for which I mostly need > immutability anyway)? > My understanding is that this exists only in the Newspeak VM sources. It is now reasonably easy to port to the standard and Cog VMs and the necessary primitive error code support is in the VMs and the image. I'm currently working in Newspeak at Cadence expect that this summer I will have the opportunity to port the Cog VM under Newspeak. At that point I'll be able to add the support to Cog and the standard VM. There are now mechanisms in the VM which would allow one to include the immutability support conditionally, in which case the source could be included in the base VMs, and not have to live in a fork, which would be my preference. e.g. Interpreter methods for stack bytecodes storeAndPopReceiverVariableBytecode "Note: This code uses storePointerUnchecked:ofObject:withValue: and does the store check explicitly in order to help the translator produce better code." | rcvr top | <expandInCase: 'currentBytecode'> "this bytecode will be expanded so that refs to currentBytecode below will be constant" rcvr := receiver. top := self internalStackTop. self cppIf: Immutability ifTrue: [(self internalIsImmutable: rcvr) ifTrue: [self internalPop: 1. self internalCannotAssign: top to: receiver withIndex: (currentBytecode bitAnd: 7)]. "cannot fetch next bytecode until after immutability check so pc is correct (set to following bytecode, not the bytecode after that) after attemptToAssign:withIndex:"]. self fetchNextBytecode. rcvr < youngStart ifTrue: [self possibleRootStoreInto: rcvr value: top]. self storePointerUnchecked: (currentBytecode bitAnd: 7) ofObject: rcvr withValue: top. self internalPop: 1 best, Eliot > > Best regards > Janko > > -- > Janko MivÅ¡ek > Aida/Web > Smalltalk Web Application Server > http://www.aidaweb.si > >

March 29, 2011