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

[Pharo-project] [update 1.4] #14077

by StÃ©phane Ducasse

14077 ----- - Issue 4598: Improve MCBrowser search. Thanks Camillo. http://code.google.com/p/pharo/issues/detail?id=4598 - Issue 4600: Add Search Morph http://code.google.com/p/pharo/issues/detail?id=4600. Thanks Camillo. - Issue 4533: Transcript patch for the Kernel image http://code.google.com/p/pharo/issues/detail?id=4533. Thanks Pavel Krivanek

Aug. 4, 2011

[Pharo-project] classDescription change

by StÃ©phane Ducasse

Hi veronica I was reading your change in ClassDescription>>definition aStream cr; tab; nextPutAll: 'category: '; store: self category asString. replaced by store: (self environment organization categoryOfElement: self name) asString. and I'm wondering why this is needed? Because I was thinking that I would prefer to have category fixed. Now category Class>>category "Answer the system organization category for the receiver. First check whether the category name stored in the ivar is still correct and only if this fails look it up (latter is much more expensive)" | result | self basicCategory ifNotNil: [ :symbol | ((self environment organization listAtCategoryNamed: symbol) includes: self name) ifTrue: [ ^symbol ] ]. self basicCategory: (result := self environment organization categoryOfElement: self name). ^result

Aug. 4, 2011

Re: [Pharo-project] The shiftEnclose: horror show

by StÃ©phane Ducasse

so far I have no idea :) But I will learn. Stef On Aug 4, 2011, at 3:07 PM, Nicolas Cellier wrote: > as long as the keyMapper only map to #enclose: and not #shiftEnclose:, > no problem. > > Nicolas > > 2011/8/4 StÃ©phane Ducasse <stephane.ducasse(a)inria.fr>: >> What is on our todo is to introduce keymapper or another solution to have no hardcoded keys in the complete system. >> Now my time is short... sadly. But indeed this is good if we can clean that part. >> Stef >> >> >>> http://code.google.com/p/pharo/issues/detail?id=4599 >>> >>> 2011/8/4 Nicolas Cellier <nicolas.cellier.aka.nice(a)gmail.com>: >>>> I'm typing this message on a mac French keyboard. >>>> That means I can obtain the holy brackets [ | ] only through weird >>>> keystroke combinations >>>> | = shit+option+L >>>> [ = shift+option+( where shift+5 is ( >>>> >>>> Enclosing a text in square brackets would require some finger torture >>>> both in Squeak/Pharo >>>> shift+option+command+( >>>> But this does not work because of #shiftEnclose: rules encountered in >>>> #initializeShiftCmdKeyShortcuts >>>> (found both ParagraphEditor class and TextEditor class) >>>> >>>> "Note: Command key overrides shift key, so, for example, cmd-shift-9 >>>> produces $9 not $(" >>>> '9[,''' do: [ :char | cmdMap at: (char asciiValue + 1) put: >>>> #shiftEnclose: ]. "({< and double-quote" >>>> "Note: Must use cmd-9 or ctrl-9 to get '()' since cmd-shift-9 is a >>>> Mac FKey command." >>>> >>>> shiftEnclose: is badly designed because it does hardcode the keyboard >>>> layout (see below). >>>> This is not compatible with modern VMs, at least the mac ones, because >>>> they deliver a unicode codePoint for $[ or $|, not a raw keycode. >>>> It's easy to remove this anachronism and correct the mapping: >>>> >>>> "On some keyboards, these characters require a shift" >>>> '([<{|"''' do: [:char | cmdMap at: char asciiValue + 1 put: #enclose:]. >>>> >>>> To avoid pushing a mac-centric change in trunk, I need to know if the >>>> Linux/Windows VM would support above modification. >>>> Can anyone check for me ? >>>> >>>> Nicolas >>>> >>>> >>>> >>>> >>>> >>>> Example of hardcoded keyboard layout: >>>> TextEditor>>shiftEnclose: aKeyboardEvent >>>> "Insert or remove bracket characters around the current selection. >>>> Flushes typeahead." >>>> >>>> | char left right startIndex stopIndex oldSelection which text | >>>> char := aKeyboardEvent keyCharacter. >>>> char = $9 ifTrue: [ char := $( ]. >>>> char = $, ifTrue: [ char := $< ]. >>>> char = $[ ifTrue: [ char := ${ ]. >>>> char = $' ifTrue: [ char := $" ]. >>>> char asciiValue = 27 ifTrue: [ char := ${ ]. "ctrl-[" >>>> snip... >>>> >>> >> >> >> >

Aug. 4, 2011

Re: [Pharo-project] Lazy initialization [was Help with a WeakOrderedCollection and objects not being GCed]

by StÃ©phane Ducasse

Nicolas do you have a pointer on In Squeak trunk it would work because removals are performed AFTER >>> methodAdditions/Changes. >>> In Pharo, removals are performed BEFORE, so yes, the SLICE needs to be >>> split in two parts... >>> Or you can just manually merge it (apply the removal after...). Stef

Aug. 4, 2011

Re: [Pharo-project] Pharo 1.4 #14076 update failed.

by Marcus Denker

On Aug 4, 2011, at 3:05 PM, Mariano Martinez Peck wrote: > It seems you simply have a timeout. What I don't understand is how that can crash the VM. Do you have a dump file as well ? > > It did not chrash the VM, it's simply a time out while reading from SqueakSource. I can update, so it's not a problem of Squeaksource. It would be interesting to know from *which* image version the update was tried. 140? but even that should not be relevant. The best is to just try again. -- Marcus Denker -- http://marcusdenker.de

Aug. 4, 2011

Re: [Pharo-project] The shiftEnclose: horror show

by Nicolas Cellier

as long as the keyMapper only map to #enclose: and not #shiftEnclose:, no problem. Nicolas 2011/8/4 StÃ©phane Ducasse <stephane.ducasse(a)inria.fr>: > What is on our todo is to introduce keymapper or another solution to have no hardcoded keys in the complete system. > Now my time is short... sadly. But indeed this is good if we can clean that part. > Stef > > >> http://code.google.com/p/pharo/issues/detail?id=4599 >> >> 2011/8/4 Nicolas Cellier <nicolas.cellier.aka.nice(a)gmail.com>: >>> I'm typing this message on a mac French keyboard. >>> That means I can obtain the holy brackets [ | ] only through weird >>> keystroke combinations >>> | = shit+option+L >>> [ = shift+option+( where shift+5 is ( >>> >>> Enclosing a text in square brackets would require some finger torture >>> both in Squeak/Pharo >>> shift+option+command+( >>> But this does not work because of #shiftEnclose: rules encountered in >>> #initializeShiftCmdKeyShortcuts >>> (found both ParagraphEditor class and TextEditor class) >>> >>> Â Â Â Â "Note: Command key overrides shift key, so, for example, cmd-shift-9 >>> produces $9 not $(" >>> Â Â Â Â '9[,''' do: [ :char | cmdMap at: (char asciiValue + 1) put: >>> #shiftEnclose: ]. Â Â Â "({< and double-quote" >>> Â Â Â Â "Note: Must use cmd-9 or ctrl-9 to get '()' since cmd-shift-9 is a >>> Mac FKey command." >>> >>> shiftEnclose: is badly designed because it does hardcode the keyboard >>> layout (see below). >>> This is not compatible with modern VMs, at least the mac ones, because >>> they deliver a unicode codePoint for $[ or $|, not a raw keycode. >>> It's easy to remove this anachronism and correct the mapping: >>> >>> Â Â Â Â "On some keyboards, these characters require a shift" >>> Â Â Â Â '([<{|"''' do: [:char | cmdMap at: char asciiValue + 1 put: #enclose:]. >>> >>> To avoid pushing a mac-centric change in trunk, I need to know if the >>> Linux/Windows VM would support above modification. >>> Can anyone check for me ? >>> >>> Nicolas >>> >>> >>> >>> >>> >>> Example of hardcoded keyboard layout: >>> TextEditor>>shiftEnclose: aKeyboardEvent >>> Â Â Â Â "Insert or remove bracket characters around the current selection. >>> Â Â Â Â Flushes typeahead." >>> >>> Â Â Â Â | char left right startIndex stopIndex oldSelection which text | >>> Â Â Â Â char := aKeyboardEvent keyCharacter. >>> Â Â Â Â char = $9 ifTrue: [ char := $( ]. >>> Â Â Â Â char = $, ifTrue: [ char := $< ]. >>> Â Â Â Â char = $[ ifTrue: [ char := ${ ]. >>> Â Â Â Â char = $' ifTrue: [ char := $" ]. >>> Â Â Â Â char asciiValue = 27 ifTrue: [ char := ${ ]. Â Â "ctrl-[" >>> snip... >>> >> > > >

Aug. 4, 2011

Re: [Pharo-project] Pharo 1.4 #14076 update failed.

by Mariano Martinez Peck

It seems you simply have a timeout. What I don't understand is how that can crash the VM. Do you have a dump file as well ? [ConnectionTimedOut signal: 'Data receive timed out.'] in Socket>>waitForDataFor: Receiver: a Socket[connected] Arguments and temporary variables: Receiver's instance variables: semaphore: a Semaphore() socketHandle: #[173 40 138 8 0 0 0 0 160 126 6 0] readSemaphore: a Semaphore() writeSemaphore: a Semaphore() Socket>>waitForDataFor:ifClosed:ifTimedOut: Receiver: a Socket[connected] Arguments and temporary variables: timeout: 30 closedBlock: [ConnectionClosed signal: 'Connection closed while waiting for dat...etc... timedOutBlock: [ConnectionTimedOut signal: 'Data receive timed out.'] startTime: 32994 msecsDelta: 30000 Receiver's instance variables: semaphore: a Semaphore() socketHandle: #[173 40 138 8 0 0 0 0 160 126 6 0] readSemaphore: a Semaphore() writeSemaphore: a Semaphore() 30 seconds is ok for the timeout? sven ? On Thu, Aug 4, 2011 at 2:57 PM, Laszlo Zsolt Kiss <lzskiss(a)dublin.ie> wrote: > > here is. > thx. > best reagards > laszlo zsolt > > > ----- Original Message ----- > From: > "Mariano Martinez Peck" <marianopeck(a)gmail.com> > > To: > <Pharo-project(a)lists.gforge.inria.fr> > Cc: > "Laszlo Zsolt Kiss" <lzskiss(a)dublin.ie> > Sent: > Thu, 4 Aug 2011 14:19:14 +0200 > Subject: > Re: [Pharo-project] Pharo 1.4 #14076 update failed. > > > > Can you please attach PharoDebug.log? <http://PharoDebug.log?%C2%A0> it > should be present in the same directory where .image is. > > Cheers > > On Thu, Aug 4, 2011 at 1:52 PM, Marcus Denker <marcus.denker(a)inria.fr>wrote: > >> >> On Aug 4, 2011, at 1:46 PM, Laszlo Zsolt Kiss wrote: >> >> > >> > aloha pharoers! >> > >> > i don' know what is the problem >> > My image is clean 1.4. >> >> which version? >> >> > But I geted the attached message. >> > Please help. >> > >> >> Works fine here... >> >> >> >> >> -- >> Marcus Denker -- http://marcusdenker.de >> >> >> > > > -- > Mariano > http://marianopeck.wordpress.com > > -- Mariano http://marianopeck.wordpress.com

Aug. 4, 2011

[Pharo-project] Lazy initialization [was Help with a WeakOrderedCollection and objects not being GCed]

by Nicolas Cellier

2011/8/3 StÃ©phane Ducasse <stephane.ducasse(a)inria.fr>: > > On Aug 3, 2011, at 1:49 PM, Nicolas Cellier wrote: > >> Oh I see, for the SLICE, this is again the Pharo's version of >> MCPackageLoader>>basicLoad which is striking. >> In Squeak trunk it would work because removals are performed AFTER >> methodAdditions/Changes. >> In Pharo, removals are performed BEFORE, so yes, the SLICE needs to be >> split in two parts... >> Or you can just manually merge it (apply the removal after...). > > oops we should fix that > > >> I'm eager to see a true atomic load with a single massive becomeForward: :) > > Yes me too > >> IMHO, this should be coupled with lazy initializations mechanism as I >> proposed once (based on pragmas). > > Can you point us to that? > > I cannot find my original posts, maybe they remained a draft, so here it is. http://smallissimo.blogspot.com/2011/08/lazy-initialization-of-shared-varia… Nicolas

Aug. 4, 2011

Re: [Pharo-project] Pharo 1.4 #14076 update failed.

by Laszlo Zsolt Kiss

Â here is.thx.best reagardslaszlo zsolt ----- Original Message ----- From: "Mariano Martinez Peck" To: Cc:"Laszlo Zsolt Kiss" Sent:Thu, 4 Aug 2011 14:19:14 +0200 Subject:Re: [Pharo-project] Pharo 1.4 #14076 update failed. Can you please attach PharoDebug.log?Â [1] it should be present in the same directory where .image is Cheers On Thu, Aug 4, 2011 at 1:52 PM, Marcus Denker wrote: On Aug 4, 2011, at 1:46 PM, Laszlo Zsolt Kiss wrote: > > aloha pharoers! > > i don' know what is the problem > My image is clean 1.4. which version? > But I geted the attached message. > Please help. > Works fine here... -- Marcus Denker -- http://marcusdenker.de [3] -- Mariano http://marianopeck.wordpress.com [4] Links: ------ [1] http://PharoDebug.log?%C2%A0 [2] mailto:marcus.denker@inria.fr [3] http://marcusdenker.de [4] http://marianopeck.wordpress.com

Aug. 4, 2011

Re: [Pharo-project] Pharo 1.4 #14076 update failed.

by Mariano Martinez Peck

Can you please attach PharoDebug.log? it should be present in the same directory where .image is. Cheers On Thu, Aug 4, 2011 at 1:52 PM, Marcus Denker <marcus.denker(a)inria.fr>wrote: > > On Aug 4, 2011, at 1:46 PM, Laszlo Zsolt Kiss wrote: > > > > > aloha pharoers! > > > > i don' know what is the problem > > My image is clean 1.4. > > which version? > > > But I geted the attached message. > > Please help. > > > > Works fine here... > > > > > -- > Marcus Denker -- http://marcusdenker.de > > > -- Mariano http://marianopeck.wordpress.com

Aug. 4, 2011