April 2015 - Pharo-dev - lists.pharo.org

Pharo 4.0 - display

by recursive68＠gmail.com

Hi, This is probably already a potential known solution to the display issue on MAC Retina displays but I've just tried the latest CogVM with the Pharo 4.0 image and using Cog fixes the issue, it no longer leaves outlines of windows or tracers. Thanks

April 25, 2015

Re: [Pharo-dev] SmalltalkImage>>#imageName

by Sean P. DeNigris

Given that there is consensus, I created: Issue 15399 SmalltalkImage>>#imageName is redundant But, thinking more deeply, maybe the cleaner approach would be to remove it completely. Since #imageFile returns a FS object, one already has easy access to whatever variant of the name one wants: #fullName #basename #base Why pollute the protocol? I don't think it's worth it to save from `imageFile base` to `imageName`... And this would apply to #changesName as well... ----- Cheers, Sean -- View this message in context: http://forum.world.st/SmalltalkImage-imageName-tp4821681p4821788.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

April 25, 2015

Re: [Pharo-dev] [pharo-project/pharo-core] de6e0a: 50003

by Ben Coman

While the fix is simple, its not a minor thing just to change a preferences file external to the image. This is going to bite *everyone* who moves to 5.0-development, and when 5.0 is released its going to bite lots of general public. A debugger can't be presented at startup with any reasonable expectation that users think Pharo is a quality product. First impressions count. This should be handled more gracefully by presenting instructions on how to migrate to StartupPreferencesLoader using version specific folders." https://pharo.fogbugz.com/default.asp?15398 Yuriy, Did you log an issue already to cleanup the StartupLoader references? cheers -ben On Wed, Apr 22, 2015 at 2:34 PM, Yuriy Tymchuk <yuriy.tymchuk(a)me.com> wrote: > Yes, I know. But while I was trying to understand the problem, Iâve found > places in StartupPreferencesLoader where code is generated with > StartupLoader. Itâs a minor thing, but do we really need a place with > potential problems in our system? > > Uko > > > On 22 Apr 2015, at 08:13, Esteban Lorenzano <estebanlm(a)gmail.com> wrote: > > > > > >> On 21 Apr 2015, at 19:27, Yuriy Tymchuk <yuriy.tymchuk(a)me.com> wrote: > >> > >> Oh, it should be replaced by StartupPreferencesLoader. Iâll submit a > slice > > > > a slice for what? > > the changes you need to do are in your preferences file, not inside the > image. > > > > Esteban > > > > >>> On 21 Apr 2015, at 19:20, Yuriy Tymchuk <yuriy.tymchuk(a)me.com> wrote: > >>> > >>> So now startup actions are not working because StartupLoader was > removed. What are the plans about this? > >>> > >>>> On 21 Apr 2015, at 17:40, GitHub <noreply(a)github.com> wrote: > >>>> 15368 unload Deprecate40 in Pharo5 > >>>> https://pharo.fogbugz.com/f/cases/15368 >

April 25, 2015

Re: [Pharo-dev] Simple benchmark RBParser vs. Shout

by Tudor Girba

Nice experiment. Very interesting. At this speed, we can consider not using Shout at all. Doru On Sat, Apr 25, 2015 at 12:00 PM, Marcus Denker <marcus.denker(a)inria.fr> wrote: > > > On 25 Apr 2015, at 09:35, Marcus Denker <marcus.denker(a)inria.fr> wrote: > > > > Hi, > > > > I want to get a feeling of how much faster Shout is compared to > RBParser. This is a very simplistic > > bench: > > > > > > [Morph methods do: [:method | RBParser parseFaultyMethod: method > sourceCode ]] bench. > > "'3.197 per second'" > > > > [Morph methods do: [ :method | > > SHTextStylerST80 new rangesIn: method sourceCode setWorkspace: false.]] > > bench > > > > "'2.797 per secondâ" > > > > This means that parsing performance alone is actually not bad with RBâ¦ > > (it would be interesting to look at edge casesâ¦) > > > > And then of course the next thing to do is the highlighting itself. > > With highlighting AST vs. Shout, Shout is a bit faster: > > [Morph methods do: [:method | SHTextStylerST80 new classOrMetaClass: > method methodClass; styledTextFor: method sourceCode asText]] bench. > "'2.301 per second'" > > [Morph methods do: [:method | > SHRBTextStyler new classOrMetaClass: method methodClass; styledTextFor: > method sourceCode asText]] bench > > "'1.935 per second'" > > Marcus > -- www.tudorgirba.com "Every thing has its own flow"

April 25, 2015

Re: [Pharo-dev] SmalltalkImage>>#imageName

by Ben Coman

On Sat, Apr 25, 2015 at 12:23 AM, Sean P. DeNigris <sean(a)clipperadams.com> wrote: > After a brief IRC chat, it seems like returning just the file base (i.e. > without the extension). Any objections before I open an issue? > > SeanDeNigris_: I found it surprising that `Smalltalk imageName` returns the > full path. Just me? > [11:34am] maxleske: ++ > [11:43am] SeanDeNigris_: Worth an issue? Only sender seems to be Spotter > [11:49am] DamienCassou: all these little things are annoying. So, +1 to > create an issue > > +1. We have... Smalltalk>>imageName "Answer the full path name for the current image." ^ FilePathEncoder decode: self primImageName "<primitive: 121>" Smalltalk>>imagePath "Answer the full path name for the current image." ^ FilePathEncoder decode: self primImagePath "<primitive: 121>" ...same duplicate comment, same duplicate result. cheers -ben

April 25, 2015

Re: [Pharo-dev] Simple benchmark RBParser vs. Shout

by Marcus Denker

> On 25 Apr 2015, at 09:35, Marcus Denker <marcus.denker(a)inria.fr> wrote: > > Hi, > > I want to get a feeling of how much faster Shout is compared to RBParser. This is a very simplistic > bench: > > > [Morph methods do: [:method | RBParser parseFaultyMethod: method sourceCode ]] bench. > "'3.197 per second'" > > [Morph methods do: [ :method | > SHTextStylerST80 new rangesIn: method sourceCode setWorkspace: false.]] > bench > > "'2.797 per secondâ" > > This means that parsing performance alone is actually not bad with RBâ¦ > (it would be interesting to look at edge casesâ¦) > > And then of course the next thing to do is the highlighting itself. With highlighting AST vs. Shout, Shout is a bit faster: [Morph methods do: [:method | SHTextStylerST80 new classOrMetaClass: method methodClass; styledTextFor: method sourceCode asText]] bench. "'2.301 per second'" [Morph methods do: [:method | SHRBTextStyler new classOrMetaClass: method methodClass; styledTextFor: method sourceCode asText]] bench "'1.935 per second'" Marcus

April 25, 2015

Re: [Pharo-dev] changing nautilus method icons

by Peter UhnÃ¡k

Perfect! Thanks! Peter On Sat, Apr 25, 2015 at 11:22 AM, Nicolai Hess <nicolaihess(a)web.de> wrote: > > > 2015-04-25 11:16 GMT+02:00 Nicolai Hess <nicolaihess(a)web.de>: > >> >> >> 2015-04-25 9:00 GMT+02:00 Peter UhnÃ¡k <i.uhnak(a)gmail.com>: >> >>> Hi, >>> >>> is there some (API) way to control the icons for methods in Nautilus? >>> I would like to show an icon for methods that contains "self >>> shouldBeImplemented". >>> Something like what is shown when there is "self halt" in the method. >>> Is this possible without overriding something inside Nautilus? >>> >> >> Just implement a new subclass of >> AbstractMethodIconAction >> and implement >> #privateActionIcon #actionStateToCache #isActionHandled #actionOrder >> > > (and of course, once you will have to call "AbstractNautilusUI > resetIconCaches" ) > > > >> >> >>> >>> (Possibly through Nautilus Plugin Manager, but I have no idea how that >>> works.) >>> >>> Thanks, >>> Peter >>> >> >> >

April 25, 2015

Re: [Pharo-dev] changing nautilus method icons

by Nicolai Hess

2015-04-25 11:16 GMT+02:00 Nicolai Hess <nicolaihess(a)web.de>: > > > 2015-04-25 9:00 GMT+02:00 Peter UhnÃ¡k <i.uhnak(a)gmail.com>: > >> Hi, >> >> is there some (API) way to control the icons for methods in Nautilus? >> I would like to show an icon for methods that contains "self >> shouldBeImplemented". >> Something like what is shown when there is "self halt" in the method. >> Is this possible without overriding something inside Nautilus? >> > > Just implement a new subclass of > AbstractMethodIconAction > and implement > #privateActionIcon #actionStateToCache #isActionHandled #actionOrder > (and of course, once you will have to call "AbstractNautilusUI resetIconCaches" ) > > >> >> (Possibly through Nautilus Plugin Manager, but I have no idea how that >> works.) >> >> Thanks, >> Peter >> > >

April 25, 2015

Re: [Pharo-dev] changing nautilus method icons

by Nicolai Hess

2015-04-25 9:00 GMT+02:00 Peter UhnÃ¡k <i.uhnak(a)gmail.com>: > Hi, > > is there some (API) way to control the icons for methods in Nautilus? > I would like to show an icon for methods that contains "self > shouldBeImplemented". > Something like what is shown when there is "self halt" in the method. > Is this possible without overriding something inside Nautilus? > Just implement a new subclass of AbstractMethodIconAction and implement #privateActionIcon #actionStateToCache #isActionHandled #actionOrder > > (Possibly through Nautilus Plugin Manager, but I have no idea how that > works.) > > Thanks, > Peter >

April 25, 2015

Simple benchmark RBParser vs. Shout

by Marcus Denker

Hi, I want to get a feeling of how much faster Shout is compared to RBParser. This is a very simplistic bench: [Morph methods do: [:method | RBParser parseFaultyMethod: method sourceCode ]] bench. "'3.197 per second'" [Morph methods do: [ :method | SHTextStylerST80 new rangesIn: method sourceCode setWorkspace: false.]] bench "'2.797 per secondâ" This means that parsing performance alone is actually not bad with RBâ¦ (it would be interesting to look at edge casesâ¦) And then of course the next thing to do is the highlighting itself. Marcus

April 25, 2015