November 2008 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] Issues

by StÃ©phane Ducasse

Thanks we need to find a way to order a bit the list. I will harvest from times to times items from the list, my goal right now is really etoy curving. Stef On Nov 7, 2008, at 8:10 AM, Torsten Bergmann wrote: >> this is important when you open an issue to tell it in the mailing- >> list > > Hi Stef, > > http://code.google.com/p/pharo/issues/detail?id=312 can be harvested > savely. > It's a typo in a symbol > > There is also > > 314 ShoutWorkspace / Support for different Workspace types > (adaptions if shout will make it into the std. image) shout will be in the default image but not the core image. > 232 Recreate Special Objects Array should preserve external > semaphores (the loading after FFI install problem) is the solution ready to be integrated? > > 231 Image does not open full screen anymore (outdated Win32 VM) > > Last one is obsolete when Markus uses newer sources for building the > one click Win32 VM > > Bye > Torsten > -- > GMX Download-Spiele: Preizsturz! Alle Puzzle-Spiele Deluxe Ã¼ber 60% > billiger. > http://games.entertainment.gmx.net/de/entertainment/games/download/puzzle/i… > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Nov. 7, 2008

Fwd: [Pharo-project] ArrayedCollection, Boolean SmallInteger>>ccg: cg generateCoerceToOopFrom: aNode on: aStream

by StÃ©phane Ducasse

Begin forwarded message: > From: "David T. Lewis" <lewis(a)mail.msen.com> > Date: November 7, 2008 1:08:04 AM CEST > To: John M McIntosh <johnmci(a)smalltalkconsulting.com> > Cc: stephane.ducasse(a)inria.fr > Subject: Re: Fwd: [Pharo-project] ArrayedCollection, Boolean > SmallInteger>>ccg: cg generateCoerceToOopFrom: aNode on: aStream > > Hi John, > > It is safe to remove all the 'ccg:*' methods from Pharo, and no > other updates are required for VMMaker. > > All of the methods in question are already included in the VMMaker > package. The methods in the base image are exact duplicates of the > corresponding methods in the VMMaker package (except for method > category of course). I would guess that Tim may have just forgotten > to delete them. > > I have entered Mantis issue #7221 to track this, with a change set > to remove the duplicate methods from a base 3.10 image. I also > confirmed > that after removing the methods from the base image, I was able to > load > VMMaker and build a VM. > > - Dave > > > On Thu, Nov 06, 2008 at 12:34:11PM -0800, John M McIntosh wrote: >> David if you want to look into this that would be good. I think there >> are a few other methods (4 message sends I think) >> >> ccg:emitLoadFor:from:on: >> ccg:generateCoerceToOopFrom:on: >> ccg:generateCoerceToValueFrom:on: >> ccg:prolog:expr:index: >> >> on Array, Boolean, ByteArray, Float, FloatArray, IntegerArry, Object, >> and WordArray >> >> As the Pharo folks noticed these really should be part of VMMaker and >> not in the base squeak distribution. >> >> >> Begin forwarded message: >> >>> From: St?phane Ducasse <stephane.ducasse(a)inria.fr> >>> Date: November 5, 2008 12:22:57 PM PST (CA) >>> To: Pharo Development <Pharo-project(a)lists.gforge.inria.fr> >>> Subject: [Pharo-project] ArrayedCollection, Boolean >>> SmallInteger>>ccg: cg generateCoerceToOopFrom: aNode on: aStream >>> Reply-To: Pharo Development <pharo-project(a)lists.gforge.inria.fr> >>> >>> I have the impression that we could remove these methods from the >>> images. >>> It seems that this is related to the VM generation? >>> Does someone know more about them? >>> Stef >>> >>> ccg: cg generateCoerceToOopFrom: aNode on: aStream >>> >>> self instSize > 0 ifTrue: >>> [self error: 'cannot auto-coerce arrays with named instance >>> variables']. >>> cg generateCoerceToObjectFromPtr: aNode on: aStream >>> >

Nov. 7, 2008

[Pharo-project] Issues

by Torsten Bergmann

>this is important when you open an issue to tell it in the mailing-list Hi Stef, http://code.google.com/p/pharo/issues/detail?id=312 can be harvested savely. It's a typo in a symbol There is also 314 ShoutWorkspace / Support for different Workspace types (adaptions if shout will make it into the std. image) 232 Recreate Special Objects Array should preserve external semaphores (the loading after FFI install problem) 231 Image does not open full screen anymore (outdated Win32 VM) Last one is obsolete when Markus uses newer sources for building the one click Win32 VM Bye Torsten -- GMX Download-Spiele: Preizsturz! Alle Puzzle-Spiele Deluxe Ã¼ber 60% billiger. http://games.entertainment.gmx.net/de/entertainment/games/download/puzzle/i…

Nov. 7, 2008

Re: [Pharo-project] Strange bugs?

by Yoshiki Ohshima

At Sun, 2 Nov 2008 16:09:21 +0100, Marcus Denker wrote: > > > > > > - hex removal (I do not remember why we removed it). > > The only sender of hex: was the method asHTMLColor, in Pharo, this was > changed (bf) to not call hex: anymore. > I guess this change removed the hex:, as there where no senders. But during a debugging session, don't you type blah blah blah hex to see the values in hexadecimal quite often? I really miss it in images without #hex for Numbers and Characters. -- Yoshiki

Nov. 7, 2008

[Pharo-project] [ANN] 10129 and 10130

by StÃ©phane Ducasse

Hi 10130 - I cleaned a lot of etoys related things based on pavel submissions. Now it will take long since a lot of packages got published but empty because of a bug in the process I used. Sorry. 10131 313: FileList2 crashes if some subdirectory present Stef

Nov. 6, 2008

Re: [Pharo-project] latest update #10129 breaks FileList2 (NonBooleanReceiver: proceed for truth).

by StÃ©phane Ducasse

Hi guys this is important when you open an issue to tell it in the mailing-list. I will harvest it now. Stef On Nov 6, 2008, at 9:26 PM, Pavel Krivanek wrote: > see http://code.google.com/p/pharo/issues/detail?id=313&colspec=ID%20Type%20Sta… > > On Thu, Nov 6, 2008 at 9:18 PM, John M McIntosh <johnmci(a)smalltalkconsulting.com > > wrote: > After updating my test image and then attempting to open a file list > browser I get a walkback > This seems to be related to a change in > FileList2>>initialize > > where > showDirsInFileList _ false. > was dropped > by lr 10/29/2008 21:28 > > > Unfortunately showDirsInFileList is not initialized anywhere else so > the block then which uses the variable > evaluates to nil... > > -- > = > = > = > = > = > ====================================================================== > John M. McIntosh <johnmci(a)smalltalkconsulting.com> > Corporate Smalltalk Consulting Ltd. http:// > www.smalltalkconsulting.com > = > = > = > = > = > ====================================================================== > > > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Nov. 6, 2008

[Pharo-project] Debugger problem

by Michael Roberts

Before I go digging too much, has anyone seen these symptoms? In executing some sunits I hit an error, so go poking around up the stack and try and restart a method or block to walk through the sequence of sends again. If I am inside a block I don't get prompted for the "restart the method or block?" type question. Squeak used to do this, no? I can't exactly work out what the exact sequence of events is, but I end up with multiple "unwind error during termination" notifiers popping up which I can't kill the process. This used to really wind me up in squeak so it might not be a new problem. any thoughts? Mike

Nov. 6, 2008

Re: [Pharo-project] Too much updates without an image

by Adrian Lienhard

I've fixed the problem. Adrian On Nov 6, 2008, at 13:07 , StÃ©phane Ducasse wrote: > Amazing :) > I started to migrate part of the update to the installer but never > finished. > > Stef > > >> I found the problem. Its a quite complex bug, actually, as it is >> not related to the update #120 but to #116. I inadvertently >> introduced the problem with the memory clean up script. It resets >> the update downloader, and this makes the update queue to get into >> an inconsistent state after consuming 5 more updates. This is why >> we did not see the problem earlier but only now that we update from >> 116 through to 120. >> >> I don't have time to fix that now, but I will try this evening. >> >> Adrian >> >> >> On Nov 6, 2008, at 11:11 , StÃ©phane Ducasse wrote: >> >>> I will try too. >>> I retried from 10095 during the sprint at my place. >>> but now I will retry from 107 >>> >>> This was when we got the problem with lukas removal. >>> and as you see I did not update the log >>> >>> This is the last time I do update from bern (because of this >>> caching proxy >>> I could not test directly). >>> >>> ------------------------------------------------------ >>> update10118 >>> - fix script loading broken. >>> Issue 294: VM does not send script files to image >>> ------------------------------------------------------ >>> update10119 >>> - fix script loading broken. >>> Issue 294: VM does not send script files to image >>> ------------------------------------------------------ >>> update10119 >>> - fix script loading broken. >>> Issue 294: VM does not send script files to image >>> ------------------------------------------------------ >>> update10120 >>> - Etoy removal >>> TextMorph >>> ThumbnailMorph >>> TileMorphTest >>> TraitDescription >>> UndefinedObjectTest >>> UpdatingStringMorph >>> Vocabulary >>> ------------------------------------------------------ >>> >>> >>> On Nov 6, 2008, at 10:56 AM, Marcus Denker wrote: >>> >>>> >>>> On 04.11.2008, at 10:30, Damien Cassou wrote: >>>> >>>>> Hi, >>>>> >>>>> could you please release an image with the latest changes so >>>>> that I >>>>> can base the new pharo-dev on them? >>>>> >>>> >>>> >>>> So... I tried to update the last core image (107) from the >>>> gforge. But that one stops after >>>> update 120. (blocked in a semaphore). Very strange. >>>> >>>> I have an image that I updated often, I never had problems >>>> updating that one. >>>> >>>> Very strange. >>>> >>>> Marcus >>>> >>>> >>>> -- >>>> Marcus Denker -- denker(a)iam.unibe.ch >>>> http://www.iam.unibe.ch/~denker >>>> >>>> >>>> _______________________________________________ >>>> Pharo-project mailing list >>>> Pharo-project(a)lists.gforge.inria.fr >>>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >>>> >>> >>> >>> _______________________________________________ >>> Pharo-project mailing list >>> Pharo-project(a)lists.gforge.inria.fr >>> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> >> >> _______________________________________________ >> Pharo-project mailing list >> Pharo-project(a)lists.gforge.inria.fr >> http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >> > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project

Nov. 6, 2008

Re: [Pharo-project] ArrayedCollection, Boolean SmallInteger>>ccg: cg generateCoerceToOopFrom: aNode on: aStream

by John M McIntosh

I asked David T Lewis to look into the issue, you are correct they should be part of vmmaker and not in the base distribution. I think you'll find there are: ccg:emitLoadFor:from:on: ccg:generateCoerceToOopFrom:on: ccg:generateCoerceToValueFrom:on: ccg:prolog:expr:index: on Array, Boolean, ByteArray, Float, FloatArray, IntegerArry, Object, and WordArray Fortunately they all start with ccg: so easy to find On Nov 6, 2008, at 2:00 AM, StÃ©phane Ducasse wrote: > excellent then these methods could be pushed there. > John it looked to me that you have access to VMMaker so could you > add these methods > there and I would remove them from the core. > > Stef -- = = = ======================================================================== John M. McIntosh <johnmci(a)smalltalkconsulting.com> Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com = = = ========================================================================

Nov. 6, 2008

Re: [Pharo-project] latest update #10129 breaks FileList2 (NonBooleanReceiver: proceed for truth).

by StÃ©phane Ducasse

thanks john. Stef On Nov 6, 2008, at 9:18 PM, John M McIntosh wrote: > After updating my test image and then attempting to open a file list > browser I get a walkback > This seems to be related to a change in > FileList2>>initialize > > where > showDirsInFileList _ false. > was dropped > by lr 10/29/2008 21:28 > > > Unfortunately showDirsInFileList is not initialized anywhere else so > the block then which uses the variable > evaluates to nil... > > -- > = > = > = > = > = > ====================================================================== > John M. McIntosh <johnmci(a)smalltalkconsulting.com> > Corporate Smalltalk Consulting Ltd. http:// > www.smalltalkconsulting.com > = > = > = > = > = > ====================================================================== > > > > > _______________________________________________ > Pharo-project mailing list > Pharo-project(a)lists.gforge.inria.fr > http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project >

Nov. 6, 2008