September 2012 - Pharo-dev - lists.pharo.org

[Pharo-project] Pharo 1.4 -- debugger initialExtent is too wide, centered

by Martin McClure

In Pharo 1.4, if your Display is not tiny, the initial extent of the debugger is 2/3 the width and 4/5 the height of Display. For large screens, this is just too big, especially too wide. My current Display is 2396x1741x32, and I have to immediately resize any debugger smaller in order to make it usable. I've experimented with leaving the height as is and limiting the width to 900 in Debugger>>initialExtent; this produces a much more usable result. ((w - (w // 3)) min: 900) @ (h - (h // 5)) The debugger also always centers its window in Display. Since my Display is spread across two monitors, this means it always must be moved to be readable. If I comment out the line window center: Display center. in MorphicUIManager>>openDebugger:fullMorphicLabel:, I'm much happier. I'd think these changes might be desirable for everyone -- comments? Regards, -Martin

Sept. 22, 2012

[Pharo-project] The roadmap aka fight plan to make image control host window

by Igor Stasenko

First, we need to make one thing: Change the VM to not call ioProcessEvents function directly, but rather use a variable which will hold a function pointer. Initially, when VM starts, that function pointer is NULL, and so, "process events" is NOP. Then we need to implement a module which will manage host windows, and once image tells it to create a window, it can change the function pointer to ioProcessEvents to own function. This change, would allow us to temporary disable any external event processing (like in case of callbacks). Second, we should move all event-related stuff to new plugin (for instance, all definitions in sq.h like sqInputEvent , ioSetInputSemaphore() ioGetNextEvent() etc). For all UI-related stuff the image should communicate with a new plugin, while VM should get rid of all knowledge of any existence of UI events, like input semaphores, primitives for fetching them etc. Third, of course, all screen update/refresh primitives should be deprecated, and image should use new primitives to update the contents of host window(s). -- Best regards, Igor Stasenko.

Sept. 22, 2012

Re: [Pharo-project] headless CogVM on Windows

by Igor Stasenko

Pavel, as you may know, we wanting to make image be in control whether it wants to display any window or not, i.e. it should tell VM to create host window if it wants to, and VM should not provide any window by default (or simulate that it is there). And we want it for a years, still nobody spent time implementing this. As to me, it worth putting effort to make this happen , instead of fixing broken headless mode. On 22 September 2012 18:39, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: > Well, on Mac OS X it works too but the Pharo window is always displayed. > > -- Pavel > > On Fri, Sep 21, 2012 at 7:23 PM, Pavel Krivanek > <pavel.krivanek(a)gmail.com> wrote: >> This is the reason why I need the headless mode on Windows: >> http://youtu.be/L6P6VuqDmaM >> >> "This video shows a very simple Pharo application based on XULRunner. >> We firstly start XULRunner that displays a splashscreen. During the >> splash screen the XULRunner starts a new process with Pharo and starts >> Zinc server in it. As soon as the server is started, it starts to >> communicate to it using a WebSocket and opens the main window. When >> the window is closed, the Pharo image is closed immediately too." >> >> ...so this is the way we can create seamless applications with native >> GUI in Pharo without need of special VM extensions. Currently on >> Windows it works the same way however the Pharo window is displayed. I >> will test the Mac later. >> >> -- Pavel >> >> On Fri, Sep 21, 2012 at 5:07 PM, Igor Stasenko <siguctua(a)gmail.com> wrote: >>> On 21 September 2012 16:28, Torsten Bergmann <astares(a)gmx.de> wrote: >>>>>yeah... we need a real headless mode (no systry, nothing... just a deamon >running) >>>> >>>> The regulary Squeak VM for Windows is able to do just that. >>>> It runs Smalltalk as an NT service by using the option "-service" >>>> See >>>> >>>> http://wiki.squeak.org/squeak/105 >>>> >>>> Dont know if this also works on Cog/Pharo standard VM. >>>> >>> >>> yeah.. but this is something different.. it registers itself as a >>> service in system. >>> Which is not the same as just running it headless. >>> >>>> Bye >>>> T. >>>> >>> >>> >>> >>> -- >>> Best regards, >>> Igor Stasenko. >>> -- Best regards, Igor Stasenko.

Sept. 22, 2012

Re: [Pharo-project] headless CogVM on Windows

by Pavel Krivanek

Well, on Mac OS X it works too but the Pharo window is always displayed. -- Pavel On Fri, Sep 21, 2012 at 7:23 PM, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: > This is the reason why I need the headless mode on Windows: > http://youtu.be/L6P6VuqDmaM > > "This video shows a very simple Pharo application based on XULRunner. > We firstly start XULRunner that displays a splashscreen. During the > splash screen the XULRunner starts a new process with Pharo and starts > Zinc server in it. As soon as the server is started, it starts to > communicate to it using a WebSocket and opens the main window. When > the window is closed, the Pharo image is closed immediately too." > > ...so this is the way we can create seamless applications with native > GUI in Pharo without need of special VM extensions. Currently on > Windows it works the same way however the Pharo window is displayed. I > will test the Mac later. > > -- Pavel > > On Fri, Sep 21, 2012 at 5:07 PM, Igor Stasenko <siguctua(a)gmail.com> wrote: >> On 21 September 2012 16:28, Torsten Bergmann <astares(a)gmx.de> wrote: >>>>yeah... we need a real headless mode (no systry, nothing... just a deamon >running) >>> >>> The regulary Squeak VM for Windows is able to do just that. >>> It runs Smalltalk as an NT service by using the option "-service" >>> See >>> >>> http://wiki.squeak.org/squeak/105 >>> >>> Dont know if this also works on Cog/Pharo standard VM. >>> >> >> yeah.. but this is something different.. it registers itself as a >> service in system. >> Which is not the same as just running it headless. >> >>> Bye >>> T. >>> >> >> >> >> -- >> Best regards, >> Igor Stasenko. >>

Sept. 22, 2012

Re: [Pharo-project] headless CogVM on Windows

by Sven Van Caekenberghe

Cool stuff ! -- Sven Van Caekenberghe http://stfx.eu Smalltalk is the Red Pill On 21 Sep 2012, at 19:23, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: > This is the reason why I need the headless mode on Windows: > http://youtu.be/L6P6VuqDmaM > > "This video shows a very simple Pharo application based on XULRunner. > We firstly start XULRunner that displays a splashscreen. During the > splash screen the XULRunner starts a new process with Pharo and starts > Zinc server in it. As soon as the server is started, it starts to > communicate to it using a WebSocket and opens the main window. When > the window is closed, the Pharo image is closed immediately too." > > ...so this is the way we can create seamless applications with native > GUI in Pharo without need of special VM extensions. Currently on > Windows it works the same way however the Pharo window is displayed. I > will test the Mac later. > > -- Pavel > > On Fri, Sep 21, 2012 at 5:07 PM, Igor Stasenko <siguctua(a)gmail.com> wrote: >> On 21 September 2012 16:28, Torsten Bergmann <astares(a)gmx.de> wrote: >>>> yeah... we need a real headless mode (no systry, nothing... just a deamon >running) >>> >>> The regulary Squeak VM for Windows is able to do just that. >>> It runs Smalltalk as an NT service by using the option "-service" >>> See >>> >>> http://wiki.squeak.org/squeak/105 >>> >>> Dont know if this also works on Cog/Pharo standard VM. >>> >> >> yeah.. but this is something different.. it registers itself as a >> service in system. >> Which is not the same as just running it headless. >> >>> Bye >>> T. >>> >> >> >> >> -- >> Best regards, >> Igor Stasenko. >> >

Sept. 21, 2012

Re: [Pharo-project] [update 2.0] #20301

by Sven Van Caekenberghe

On 21 Sep 2012, at 13:52, Marcus Denker <marcus.denker(a)inria.fr> wrote: > 20301 > ----- > > Issue 6697: Zn+Zdc Update 2012-09-19 > http://code.google.com/p/pharo/issues/detail?id=6697 Thanks! Zinc-HTTP-Deprecated is an empty category now and can be removed. > Issue 6699: Share binding of metaclass methods > http://code.google.com/p/pharo/issues/detail?id=6699 > > -- > Marcus Denker -- http://marcusdenker.de -- Sven Van Caekenberghe http://stfx.eu Smalltalk is the Red Pill

Sept. 21, 2012

Re: [Pharo-project] [Vm-dev] [NB] NativeBoost meets JIT

by Eliot Miranda

Hi Igor, great news! On Fri, Sep 21, 2012 at 7:59 AM, Igor Stasenko <siguctua(a)gmail.com> wrote: > > Hello there, > > so, we're entered a new area, where native code, generated from image > side can be run directly by JIT. > This feature was one of the first things which i wanted to try, once > Eliot released Cog :) > > The way how we do that, is when VM decides to JIT a specific method, > we copying the native code (from method trailer) > directly into the method's code. > All you need to do is to use special primitive for that 220 ( > #primitiveVoltage) > > So, a first question, which we wanted to be answered is how faster to > run native code by JIT, > comparing to running native code via NativeBoost primitive , which is > #primitiveNativeCall.. > > For here are methods, which just answer 42: > > This one using #primitiveNativeCall > > nbFoo2 > <primitive: #primitiveNativeCall module: #NativeBoostPlugin error: > errorCode> > > ^ NBNativeCodeGen methodAssembly: [:gen :proxy :asm | > asm noStackFrame. > asm > mov: (42 << 1) + 1 to: asm EAX; > ret. > ] > > And this one uses JIT: > > nbFoo > <primitive: 220 error: errorCode> > > [ errorCode = ErrRunningViaInterpreter ] whileTrue: [ ^ self > nbFoo ]. > > ^ NBNativeCodeGen jitMethodAssembly: [:gen :proxy :asm | > asm noStackFrame. > asm > mov: (42 << 1) + 1 to: asm EDX; > ret: 4 asUImm. > ] > > And this one is code which JIT can do: > > nbFoo42 > ^ 42 > > So, here the numbers: > > Time to run via #primitiveNativeCall : > > [100000000 timesRepeat: [ MyClass nbFoo2 ] ] timeToRun > 6995 > > Time to run via JIT: > > [100000000 timesRepeat: [ MyClass nbFoo ] ] timeToRun > 897 > > Time to run JITed method: > > [100000000 timesRepeat: [ MyClass nbFoo42 ] ] timeToRun > 899 > > so, as you can see, the JITed method and our custom generated code is > on par (which is logical ;). > > Time to run an empty loop: > > [100000000 timesRepeat: [ ] ] timeToRun 679 > > > So, here the result, if we extract the loop overhead, we can see the > difference in > calling our native code when it uses JIT vs using #primitiveNativeCall : > > (6995 - 679 ) / (897- 679) asFloat 28.972477064220183 > > 28 times faster!!!! > > So, with this new feature, we now can make our generated code to run > with unmatched speed, > without overhead related to #primitiveNativeCall. > This is especially useful for implementing primives which involving > heavy numeric crunching. > > I would release this code to public, but there's one little > discrepancy i need to deal with first: > > (one little problem, which i hope Eliot can help to solve) > > it looks like primitivePerform: never enters the JIT mode, but always > executing the method via interpreter. > I'll take a look. This is all very detailed so I'll need a little time. This is why you see this code: > [ errorCode = ErrRunningViaInterpreter ] whileTrue: [ ^ self > nbFoo ]. > > because if i do it inside of NBNativeCodeGen>>jitMethodAssembly:, > which checks for same error and retries the send using perform > primitive, it never enters the JIT mode, > resulting in endless loop :( > > This is despite the fact that method is JITed, because we enforce the > JITing of that method during error handling: > > lastError = ErrRunningViaInterpreter ifTrue: [ > "a method contains native code, but executed by > interpreter " > method forceJIT ifFalse: [ self error: 'Failed to JIT the > compiled > method. Try reducing it''s size ' ]. > ^ self retrySend: aContext > ]. > > The #forceJit is the primitive which i implemented like following: > > primitiveForceJIT > > <export: true > > > | val result | > > val := self stackTop. > > (self isIntegerObject: val) ifTrue: [ ^ self primitiveFail ]. > (self isCompiledMethod: val) ifFalse: [ ^ self primitiveFail ]. > > (self methodHasCogMethod: val) ifFalse: [ > cogit cog: val selector: objectMemory nilObject ]. > > result := (self methodHasCogMethod: val ) ifTrue: [ objectMemory > trueObject ] ifFalse: [ objectMemory falseObject ]. > > ^ self pop: 1 thenPush: result. > > As you can see from its usage, if VM, for some reason will fail to jit > the method, the primitive will answer false, > and we will stop with an error.. Which apparently never happens. > Still, a #primitivePerform seems like ignoring that the method > contains machine code an always runs it interpreted :( > > I do not like the idea, that users will be forced to manually put such > loops in every method they will write.. > any ideas/suggestions how to overcome that? > Yes. The JIT should be told that methods that have NB code should be jitted. But right now I don't understand enough of how NB code is generated and methods marked that they have NB code etc to know exactly how to do this. I need to play around a bit. -- best, Eliot

Sept. 21, 2012

Re: [Pharo-project] headless CogVM on Windows

by Pavel Krivanek

This is the reason why I need the headless mode on Windows: http://youtu.be/L6P6VuqDmaM "This video shows a very simple Pharo application based on XULRunner. We firstly start XULRunner that displays a splashscreen. During the splash screen the XULRunner starts a new process with Pharo and starts Zinc server in it. As soon as the server is started, it starts to communicate to it using a WebSocket and opens the main window. When the window is closed, the Pharo image is closed immediately too." ...so this is the way we can create seamless applications with native GUI in Pharo without need of special VM extensions. Currently on Windows it works the same way however the Pharo window is displayed. I will test the Mac later. -- Pavel On Fri, Sep 21, 2012 at 5:07 PM, Igor Stasenko <siguctua(a)gmail.com> wrote: > On 21 September 2012 16:28, Torsten Bergmann <astares(a)gmx.de> wrote: >>>yeah... we need a real headless mode (no systry, nothing... just a deamon >running) >> >> The regulary Squeak VM for Windows is able to do just that. >> It runs Smalltalk as an NT service by using the option "-service" >> See >> >> http://wiki.squeak.org/squeak/105 >> >> Dont know if this also works on Cog/Pharo standard VM. >> > > yeah.. but this is something different.. it registers itself as a > service in system. > Which is not the same as just running it headless. > >> Bye >> T. >> > > > > -- > Best regards, > Igor Stasenko. >

Sept. 21, 2012

Re: [Pharo-project] headless CogVM on Windows

by Igor Stasenko

On 21 September 2012 16:28, Torsten Bergmann <astares(a)gmx.de> wrote: >>yeah... we need a real headless mode (no systry, nothing... just a deamon >running) > > The regulary Squeak VM for Windows is able to do just that. > It runs Smalltalk as an NT service by using the option "-service" > See > > http://wiki.squeak.org/squeak/105 > > Dont know if this also works on Cog/Pharo standard VM. > yeah.. but this is something different.. it registers itself as a service in system. Which is not the same as just running it headless. > Bye > T. > -- Best regards, Igor Stasenko.

Sept. 21, 2012

[Pharo-project] [NB] NativeBoost meets JIT

by Igor Stasenko

Hello there, so, we're entered a new area, where native code, generated from image side can be run directly by JIT. This feature was one of the first things which i wanted to try, once Eliot released Cog :) The way how we do that, is when VM decides to JIT a specific method, we copying the native code (from method trailer) directly into the method's code. All you need to do is to use special primitive for that 220 ( #primitiveVoltage) So, a first question, which we wanted to be answered is how faster to run native code by JIT, comparing to running native code via NativeBoost primitive , which is #primitiveNativeCall.. For here are methods, which just answer 42: This one using #primitiveNativeCall nbFoo2 <primitive: #primitiveNativeCall module: #NativeBoostPlugin error: errorCode> ^ NBNativeCodeGen methodAssembly: [:gen :proxy :asm | asm noStackFrame. asm mov: (42 << 1) + 1 to: asm EAX; ret. ] And this one uses JIT: nbFoo <primitive: 220 error: errorCode> [ errorCode = ErrRunningViaInterpreter ] whileTrue: [ ^ self nbFoo ]. ^ NBNativeCodeGen jitMethodAssembly: [:gen :proxy :asm | asm noStackFrame. asm mov: (42 << 1) + 1 to: asm EDX; ret: 4 asUImm. ] And this one is code which JIT can do: nbFoo42 ^ 42 So, here the numbers: Time to run via #primitiveNativeCall : [100000000 timesRepeat: [ MyClass nbFoo2 ] ] timeToRun 6995 Time to run via JIT: [100000000 timesRepeat: [ MyClass nbFoo ] ] timeToRun 897 Time to run JITed method: [100000000 timesRepeat: [ MyClass nbFoo42 ] ] timeToRun 899 so, as you can see, the JITed method and our custom generated code is on par (which is logical ;). Time to run an empty loop: [100000000 timesRepeat: [ ] ] timeToRun 679 So, here the result, if we extract the loop overhead, we can see the difference in calling our native code when it uses JIT vs using #primitiveNativeCall : (6995 - 679 ) / (897- 679) asFloat 28.972477064220183 28 times faster!!!! So, with this new feature, we now can make our generated code to run with unmatched speed, without overhead related to #primitiveNativeCall. This is especially useful for implementing primives which involving heavy numeric crunching. I would release this code to public, but there's one little discrepancy i need to deal with first: (one little problem, which i hope Eliot can help to solve) it looks like primitivePerform: never enters the JIT mode, but always executing the method via interpreter. This is why you see this code: [ errorCode = ErrRunningViaInterpreter ] whileTrue: [ ^ self nbFoo ]. because if i do it inside of NBNativeCodeGen>>jitMethodAssembly:, which checks for same error and retries the send using perform primitive, it never enters the JIT mode, resulting in endless loop :( This is despite the fact that method is JITed, because we enforce the JITing of that method during error handling: lastError = ErrRunningViaInterpreter ifTrue: [ "a method contains native code, but executed by interpreter " method forceJIT ifFalse: [ self error: 'Failed to JIT the compiled method. Try reducing it''s size ' ]. ^ self retrySend: aContext ]. The #forceJit is the primitive which i implemented like following: primitiveForceJIT <export: true > | val result | val := self stackTop. (self isIntegerObject: val) ifTrue: [ ^ self primitiveFail ]. (self isCompiledMethod: val) ifFalse: [ ^ self primitiveFail ]. (self methodHasCogMethod: val) ifFalse: [ cogit cog: val selector: objectMemory nilObject ]. result := (self methodHasCogMethod: val ) ifTrue: [ objectMemory trueObject ] ifFalse: [ objectMemory falseObject ]. ^ self pop: 1 thenPush: result. As you can see from its usage, if VM, for some reason will fail to jit the method, the primitive will answer false, and we will stop with an error.. Which apparently never happens. Still, a #primitivePerform seems like ignoring that the method contains machine code an always runs it interpreted :( I do not like the idea, that users will be forced to manually put such loops in every method they will write.. any ideas/suggestions how to overcome that? -- Best regards, Igor Stasenko.

Sept. 21, 2012