- Pharo-dev - lists.pharo.org

[Pharo-project] got a surprising behavior

by StÃ©phane Ducasse

Hi guys I got a really strange behavior: I integrated - Issue 4607: Refactor Debugger >> content:notifying:. Thanks Camillo Bruni. http://code.google.com/p/pharo/issues/detail?id=4607 - Issue 4606: Fix drawing shadow of morphs. Thanks Igor Stasenko. http://code.google.com/p/pharo/issues/detail?id=4606 and I could not type enter on workspace I could everywhere else. I could reopen a workspace and it worked and suddenly stopped again. stef

Aug. 10, 2011

Re: [Pharo-project] Pharo Kernel OSProcess evaluator

by Pavel Krivanek

link: https://ci.lille.inria.fr/pharo/view/Pharo-Kernel/job/Pharo%20Kernel%20OSPr… On Wed, Aug 10, 2011 at 11:27 PM, Pavel Krivanek <pavel.krivanek(a)gmail.com> wrote: > Hi, > > I created a new job on CI server that creates a Pharo Kernel image > with OSProcess loaded and simple console based evaluator. Of course > something like Coral would be better ;-) > > Cheers > -- Pavel >

Aug. 10, 2011

[Pharo-project] Pharo Kernel OSProcess evaluator

by Pavel Krivanek

Hi, I created a new job on CI server that creates a Pharo Kernel image with OSProcess loaded and simple console based evaluator. Of course something like Coral would be better ;-) Cheers -- Pavel

Aug. 10, 2011

Re: [Pharo-project] Strange bug in text selection

by StÃ©phane Ducasse

On Aug 10, 2011, at 5:21 PM, Igor Stasenko wrote: > which leads to #basicNew: primitive fail, because it passing a > negative number (no surprise, it fails). > > According to stack trace, > a > Text>> copyFrom:to: > is called with arguments: > > start = 100 > stop = 44 > (instead of having start = 44 , stop = 100) > > > it is called from MultiNewParagraph>>buildSelectionBlocksFrom:to: method. > I don't know whether we should fix this method, or maybe it is better > to fix #copyFrom:to: > to accept inverted ranges (a simple check if stop < start then swap > the end points) Igor I saw that someboyd (luarent may be) was proposing a solution > yet it hard to reproduce, it works ok most of the times :( > > To reproduce the bug, try to do following: > - open a new workspace window, type an arbitrary text there > - then drag a window to the bottom of desktop, so only title bar of > workspace remain visible > - release drag > - click on the title bar and start dragging it back to make text visible > > -- > Best regards, > Igor Stasenko AKA sig. >

Aug. 10, 2011

Re: [Pharo-project] Strange bug in text selection

by Lukas Renggli

Related, but likely a different bug: Text selections makes colors disappear http://code.google.com/p/pharo/issues/detail?id=4615 Lukas On 10 August 2011 17:21, Igor Stasenko <siguctua(a)gmail.com> wrote: > which leads to #basicNew: primitive fail, because it passing a > negative number (no surprise, it fails). > > According to stack trace, > a > Text>> copyFrom:to: > is called with arguments: > > start = 100 > stop = 44 > (instead of having start = 44 , stop = 100) > > > it is called from MultiNewParagraph>>buildSelectionBlocksFrom:to: method. > I don't know whether we should fix this method, or maybe it is better > to fix #copyFrom:to: > to accept inverted ranges (a simple check if stop < start then swap > the end points) > > > yet it hard to reproduce, it works ok most of the times :( > > To reproduce the bug, try to do following: > Â - open a new workspace window, type an arbitrary text there > Â - then drag a window to the bottom of desktop, so only title bar of > workspace remain visible > Â - release drag > Â - click on the title bar and start dragging it back to make text visible > > -- > Best regards, > Igor Stasenko AKA sig. > > -- Lukas Renggli www.lukas-renggli.ch

Aug. 10, 2011

[Pharo-project] Strange bug in text selection

by Igor Stasenko

which leads to #basicNew: primitive fail, because it passing a negative number (no surprise, it fails). According to stack trace, a Text>> copyFrom:to: is called with arguments: start = 100 stop = 44 (instead of having start = 44 , stop = 100) it is called from MultiNewParagraph>>buildSelectionBlocksFrom:to: method. I don't know whether we should fix this method, or maybe it is better to fix #copyFrom:to: to accept inverted ranges (a simple check if stop < start then swap the end points) yet it hard to reproduce, it works ok most of the times :( To reproduce the bug, try to do following: - open a new workspace window, type an arbitrary text there - then drag a window to the bottom of desktop, so only title bar of workspace remain visible - release drag - click on the title bar and start dragging it back to make text visible -- Best regards, Igor Stasenko AKA sig.

Aug. 10, 2011

[Pharo-project] We need to condense changes in 1.4

by Igor Stasenko

I think it is time to condense changes, because file is grown too large (28Mb) -- Best regards, Igor Stasenko AKA sig.

Aug. 10, 2011

Re: [Pharo-project] News about NativeBoost

by Igor Stasenko

On 10 August 2011 16:09, Igor Stasenko <siguctua(a)gmail.com> wrote: > On 10 August 2011 15:44, Henrik Johansen <henrik.s.johansen(a)veloxit.no> wrote: >> A couple of issues I've had: >> >> 1) Â Â Â handleFailureIn: aContext nativeCode: aBlock does not recover after enabling native code for first time (successfully), but rather goes on to signal last error. (ie there'sno ifTrue: branch for enableNativeCode) > > Yes. Fixed >> > >> 2) Trying to use the interpreterproxy directly in a method with Â NBNativeCodeGen methodAssembly: [:gen | ... ] gives an error. >> Could #utils and Â #stackAlignement be pushed up from NBFFICallout? >> > Yes, both in > NativeBoost-Core-IgorStasenko.35 > err.. take .36 , i found typo there :) >> Cheers, >> Henry >> >> On Aug 9, 2011, at 5:11 24PM, Igor Stasenko wrote: >> >>> Hello, >>> >>> Major news: NativeBoost now working on Macs! >>> >>> The main problem with Mac OS platform was that it requires a stack >>> alignment discipline, while other platforms don't. >>> So, i had to extend the assembler to make sure that all external calls >>> will be made with properly aligned stack. >>> >>> Also, i spent time to improve error handling. Since in Cog/Stack VMs >>> we having nice primitive error code, now it can report what caused a >>> failure in more detail, >>> than just failing a primitive and leaving a language side to guess >>> what gone wrong. >>> >>> There's a lot of other small changes, like new tests, more examples >>> etc.. But i don't want to describe them all (and frankly i don't >>> remember all of them ;) >>> >>> Things not yet working: >>> Â The callbacks to language side is not functional (i have to revise >>> this code to synchronize it with callback machinery in Cog VMs). >>> Â The callbacks to native code working well (of course, because it >>> don't needs to deal with VM-specific stuff ;) >>> >>> The prebuilt VMs, which contain a NativeBoost plugin are now on Jenkins server. >>> >>> These VMs will be built automatically each time there is code update >>> for Cog VMs (in same way as all continuous integration jobs). >>> The only difference between original Cog VMs and Cog+NativeBoost VMs >>> is that they contain NativeBoost plugin. >>> So, you can use these VMs instead of Cog VMs if you like. >>> >>> You can download VMs from here: >>> https://ci.lille.inria.fr/pharo/view/NativeBoost/ >>> >>> >>> To install NativeBoost into your image follow instructions on: >>> http://code.google.com/p/nativeboost/wiki/Installation >>> >>> >>> -- >>> Best regards, >>> Igor Stasenko AKA sig. >>> >> >> >> > > > > -- > Best regards, > Igor Stasenko AKA sig. > -- Best regards, Igor Stasenko AKA sig.

Aug. 10, 2011

Re: [Pharo-project] News about NativeBoost

by Igor Stasenko

On 10 August 2011 15:44, Henrik Johansen <henrik.s.johansen(a)veloxit.no> wrote: > A couple of issues I've had: > > 1) Â Â Â handleFailureIn: aContext nativeCode: aBlock does not recover after enabling native code for first time (successfully), but rather goes on to signal last error. (ie there'sno ifTrue: branch for enableNativeCode) Yes. Fixed > > 2) Trying to use the interpreterproxy directly in a method with Â NBNativeCodeGen methodAssembly: [:gen | ... ] gives an error. > Could #utils and Â #stackAlignement be pushed up from NBFFICallout? > Yes, both in NativeBoost-Core-IgorStasenko.35 > Cheers, > Henry > > On Aug 9, 2011, at 5:11 24PM, Igor Stasenko wrote: > >> Hello, >> >> Major news: NativeBoost now working on Macs! >> >> The main problem with Mac OS platform was that it requires a stack >> alignment discipline, while other platforms don't. >> So, i had to extend the assembler to make sure that all external calls >> will be made with properly aligned stack. >> >> Also, i spent time to improve error handling. Since in Cog/Stack VMs >> we having nice primitive error code, now it can report what caused a >> failure in more detail, >> than just failing a primitive and leaving a language side to guess >> what gone wrong. >> >> There's a lot of other small changes, like new tests, more examples >> etc.. But i don't want to describe them all (and frankly i don't >> remember all of them ;) >> >> Things not yet working: >> Â The callbacks to language side is not functional (i have to revise >> this code to synchronize it with callback machinery in Cog VMs). >> Â The callbacks to native code working well (of course, because it >> don't needs to deal with VM-specific stuff ;) >> >> The prebuilt VMs, which contain a NativeBoost plugin are now on Jenkins server. >> >> These VMs will be built automatically each time there is code update >> for Cog VMs (in same way as all continuous integration jobs). >> The only difference between original Cog VMs and Cog+NativeBoost VMs >> is that they contain NativeBoost plugin. >> So, you can use these VMs instead of Cog VMs if you like. >> >> You can download VMs from here: >> https://ci.lille.inria.fr/pharo/view/NativeBoost/ >> >> >> To install NativeBoost into your image follow instructions on: >> http://code.google.com/p/nativeboost/wiki/Installation >> >> >> -- >> Best regards, >> Igor Stasenko AKA sig. >> > > > -- Best regards, Igor Stasenko AKA sig.

Aug. 10, 2011

Re: [Pharo-project] News about NativeBoost

by Henrik Johansen

A couple of issues I've had: 1) handleFailureIn: aContext nativeCode: aBlock does not recover after enabling native code for first time (successfully), but rather goes on to signal last error. (ie there'sno ifTrue: branch for enableNativeCode) 2) Trying to use the interpreterproxy directly in a method with NBNativeCodeGen methodAssembly: [:gen | ... ] gives an error. Could #utils and #stackAlignement be pushed up from NBFFICallout? Cheers, Henry On Aug 9, 2011, at 5:11 24PM, Igor Stasenko wrote: > Hello, > > Major news: NativeBoost now working on Macs! > > The main problem with Mac OS platform was that it requires a stack > alignment discipline, while other platforms don't. > So, i had to extend the assembler to make sure that all external calls > will be made with properly aligned stack. > > Also, i spent time to improve error handling. Since in Cog/Stack VMs > we having nice primitive error code, now it can report what caused a > failure in more detail, > than just failing a primitive and leaving a language side to guess > what gone wrong. > > There's a lot of other small changes, like new tests, more examples > etc.. But i don't want to describe them all (and frankly i don't > remember all of them ;) > > Things not yet working: > The callbacks to language side is not functional (i have to revise > this code to synchronize it with callback machinery in Cog VMs). > The callbacks to native code working well (of course, because it > don't needs to deal with VM-specific stuff ;) > > The prebuilt VMs, which contain a NativeBoost plugin are now on Jenkins server. > > These VMs will be built automatically each time there is code update > for Cog VMs (in same way as all continuous integration jobs). > The only difference between original Cog VMs and Cog+NativeBoost VMs > is that they contain NativeBoost plugin. > So, you can use these VMs instead of Cog VMs if you like. > > You can download VMs from here: > https://ci.lille.inria.fr/pharo/view/NativeBoost/ > > > To install NativeBoost into your image follow instructions on: > http://code.google.com/p/nativeboost/wiki/Installation > > > -- > Best regards, > Igor Stasenko AKA sig. >

Aug. 10, 2011