September 2016 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] FFI callout arguments - references to internal objects

by Esteban Lorenzano

Hi Ben, I will take a better look tomorrow morning, but as first sight, seems to be that you want to declare SomeObject and receive it as argument to the callback, while you are receiving just an ExternalAddress. If is that the problem what happens is that callbacks do not handle *any type* as the rest of UFFIâ¦ they just handle standard C types, so you will receive an ExternalAddress isnât? You need to obtain the actual object as you need it. I see two strategies: 1) just create the object using SomeObject>>#fromHandle: This work fine if object is stateless and just want to access itâs defined vocabulary. 2) orâ¦ you can just cachÃ© the data someplace and then find it using the ExternalAddress as index. (I hope I interpreted the problem correctlyâ¦ Iâm a bit sleep here :P) cheers, Esteban > On 24 Sep 2016, at 19:40, Ben Coman <btc(a)openInWorld.com> wrote: > > I'm seeking some support for Part 5 of my FFI tutorial > http://blog.openinworld.com/2016/09/pharo-libclang-ffi-part-5-client-data-a… > > where I'm trying to replicate some C code where: > * client_data is a void pointer, and an argument to a callout function > * a reference to some data (an int, or a struct) is passed to the > callout function as the client_data > * the callout function passes that same client_data to a callback > function which casts the client-data to the required type > > How can I replicate that with UFFI? The key thing is that *any* type > can be passed via client_data. So for a normal Pharo object defined > like... > > Object subclass: MyObject > instanceVariables: 'x y' > > myobject := MyObject new x:3 y:4. > > I want to pass a reference to myobject to the callout, and in the > callback dereference client_data to get myobject that I can operate on > normally. > > Is this possible with current infrastructure, or would it need > something new like FFIInternalPinnedReference to wrap myobject for the > callout and unwrap it in the callback? > > cheers -ben >

Sept. 25, 2016

[pharo-project/pharo-core]

by GitHub

Branch: refs/tags/60240 Home: https://github.com/pharo-project/pharo-core

Sept. 25, 2016

[pharo-project/pharo-core] d5075d: 60240

by GitHub

Branch: refs/heads/6.0 Home: https://github.com/pharo-project/pharo-core Commit: d5075d407114ba1d3b529fb2740cbb34a4eaa50b https://github.com/pharo-project/pharo-core/commit/d5075d407114ba1d3b529fb2… Author: Jenkins Build Server <board(a)pharo-project.org> Date: 2016-09-25 (Sun, 25 Sep 2016) Changed paths: M Kernel.package/Object.class/instance/accessing/yourself.st R ScriptLoader60.package/ScriptLoader.class/instance/pharo - scripts/script60239.st A ScriptLoader60.package/ScriptLoader.class/instance/pharo - scripts/script60240.st R ScriptLoader60.package/ScriptLoader.class/instance/pharo - updates/update60239.st A ScriptLoader60.package/ScriptLoader.class/instance/pharo - updates/update60240.st M ScriptLoader60.package/ScriptLoader.class/instance/public/commentForCurrentUpdate.st A ScriptingExtensions.package/extension/String/instance/asClass.st A ScriptingExtensions.package/extension/String/instance/asClassIfAbsent_.st A ScriptingExtensions.package/extension/String/instance/asClassIfPresent_.st A ScriptingExtensions.package/extension/String/instance/asClassInEnvironment_.st A ScriptingExtensions.package/extension/String/instance/asClassInEnvironment_ifAbsent_.st R System-Support.package/extension/String/instance/asClass.st R System-Support.package/extension/String/instance/asClassIfAbsent_.st R System-Support.package/extension/String/instance/asClassIfPresent_.st R System-Support.package/extension/String/instance/asClassInEnvironment_.st R System-Support.package/extension/String/instance/asClassInEnvironment_ifAbsent_.st Log Message: ----------- 60240 19140 asClass should be packaged in ScriptingExtensions https://pharo.fogbugz.com/f/cases/19140 19139 Method comment of yourself should be really improved https://pharo.fogbugz.com/f/cases/19139 http://files.pharo.org/image/60/60240.zip

Sept. 25, 2016

Re: [Pharo-dev] glitches in text selection

by stepharo

Hi sean What I mentioned what really worse. You would select a piece of text and you would lose some characters at the end. So I do not think that it is link with your report. Did you try on latest Pharo 60 to know if your problem is still there? Stef Le 25/9/16 Ã 05:50, Sean P. DeNigris a Ã©crit : > Nicolai Hess-3-2 wrote >> Yes, see case 19113, fixed in 60233 > Is this related to > https://pharo.fogbugz.com/f/cases/19036/Bizarre-Text-Selection-in-Pharo-5-0 > ? If so, I strongly suggest we backport it. The inaccurate text selection in > 5.0 is a workflow killer!! > > > > ----- > Cheers, > Sean > -- > View this message in context: http://forum.world.st/glitches-in-text-selection-tp4916366p4917003.html > Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com. > >

Sept. 25, 2016

Re: [Pharo-dev] glitches in text selection

by Sean P. DeNigris

Nicolai Hess-3-2 wrote > Yes, see case 19113, fixed in 60233 Is this related to https://pharo.fogbugz.com/f/cases/19036/Bizarre-Text-Selection-in-Pharo-5-0 ? If so, I strongly suggest we backport it. The inaccurate text selection in 5.0 is a workflow killer!! ----- Cheers, Sean -- View this message in context: http://forum.world.st/glitches-in-text-selection-tp4916366p4917003.html Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.

Sept. 25, 2016

Re: [Pharo-dev] [ANN] FFI Tutorial, Pharo5 / Libclang

by Ben Coman

The immediate driver is that I want to parse the VM platform sources and use Roassal to arrange them graphically to look for cross-platform patterns. I started doing this manually but its a error prone and tedious. http://forum.world.st/code-chart-of-time-inside-the-VM-td4916951.html I am replicating a bit the TalkFFI project [1] which provides Pharo 4 NativeBoost bindings to libclang but that needs to be ported to Pharo 5 UFFI and a secondary goal was for me to learn about FFI, so I'm proceeding from scratch. What I learn will help in porting TalkFFI to Pharo 5. [1] http://smalltalkhub.com/#!/~CipT/TalkFFI In the back of my mind I consider a possibility of storing C code in-Image, and calling out to libclang to compile it, dynamically l ink and run it without ever leaving Pharo. Or imagine for the VM when saving a modified Slang primitive it generated the C code just for that primitive, via FFI got libclang to compile and link it to the platform library, then started off the new VM running tests all in one go. But thats a pie in the sky idea probably a long way off. cheers -ben On Sun, Sep 25, 2016 at 4:56 AM, Jan Vrany <jan.vrany(a)fit.cvut.cz> wrote: > Hi, > > I'm quite interested in this, though not in UFFI bit, rather > in bindings as such. > > What is the goal of the project? > > Thanks! > > Best, Jan > > > > On Sun, 2016-09-25 at 01:25 +0800, Ben Coman wrote: >> hi all, >> >> Just announcing that I'm writing a series of posts on using FFI in >> Pharo 5 to interface to libclang, the interface library for the LLVM >> C >> compiler. >> >> http://blog.openinworld.com/2016/09/pharo-libclang-ffi-part-1-preambl >> e/ >> >> I'm writing this from the perspective of a FFI newbie progressively >> learning the system. I've left in a few mis-steps since I think it >> can be useful seeing what didn't work. >> >> cheers -ben >> >

Sept. 25, 2016

Re: [Pharo-dev] object pinning versus garbage collection

by Ben Coman

On Sun, Sep 25, 2016 at 7:14 AM, Eliot Miranda <eliot.miranda(a)gmail.com> wrote: > Hi Ben, > >> On Sep 23, 2016, at 8:17 PM, Ben Coman <btc(a)openinworld.com> wrote: >> >> How does object pinning interact with garbage collection? Does it >> block garbage collection until it is manually unpinned? Or does it >> garbage collection proceed anyway? Intuitively I'd guess the former?? > > The latter. The properties are orthogonal. Unreachable pinned objects are still unreachable. Therefore they are garbage collected. A use case that wishes to maintain a pinned object over some time period must arrange that there is a string reference to the object to define that time span. Thanks Eliot. So we need to be careful/aware of the corner case where an FFI callout function may store a reference to a pinned object past the return from the callout, and arrange things as you say. cheers -ben

Sept. 25, 2016

Re: [Pharo-dev] object pinning versus garbage collection

by Eliot Miranda

Hi Ben, > On Sep 23, 2016, at 8:17 PM, Ben Coman <btc(a)openinworld.com> wrote: > > How does object pinning interact with garbage collection? Does it > block garbage collection until it is manually unpinned? Or does it > garbage collection proceed anyway? Intuitively I'd guess the former?? The latter. The properties are orthogonal. Unreachable pinned objects are still unreachable. Therefore they are garbage collected. A use case that wishes to maintain a pinned object over some time period must arrange that there is a string reference to the object to define that time span. > > cheers -ben >

Sept. 24, 2016

Re: [Pharo-dev] [ANN] FFI Tutorial, Pharo5 / Libclang

by Jan Vrany

Hi, I'm quite interested in this, though not in UFFI bit, rather in bindings as such.Â What is the goal of the project?Â Thanks!Â Best, Jan On Sun, 2016-09-25 at 01:25 +0800, Ben Coman wrote: > hi all, > > Just announcing that I'm writing a series of posts on using FFI in > Pharo 5 to interface to libclang, the interface library for the LLVM > C > compiler. > > http://blog.openinworld.com/2016/09/pharo-libclang-ffi-part-1-preambl > e/ > > I'm writing this from the perspective of a FFI newbie progressively > learning the system.Â Â I've left in a few mis-steps since I think it > can be useful seeing what didn't work. > > cheers -ben >

Sept. 24, 2016

FFI callout arguments - references to internal objects

by Ben Coman

I'm seeking some support for Part 5 of my FFI tutorial http://blog.openinworld.com/2016/09/pharo-libclang-ffi-part-5-client-data-a… where I'm trying to replicate some C code where: * client_data is a void pointer, and an argument to a callout function * a reference to some data (an int, or a struct) is passed to the callout function as the client_data * the callout function passes that same client_data to a callback function which casts the client-data to the required type How can I replicate that with UFFI? The key thing is that *any* type can be passed via client_data. So for a normal Pharo object defined like... Object subclass: MyObject instanceVariables: 'x y' myobject := MyObject new x:3 y:4. I want to pass a reference to myobject to the callout, and in the callback dereference client_data to get myobject that I can operate on normally. Is this possible with current infrastructure, or would it need something new like FFIInternalPinnedReference to wrap myobject for the callout and unwrap it in the callback? cheers -ben

Sept. 24, 2016