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

Re: [Pharo-dev] keyboard events in the windows VM

by EuanM

I'm happy to test on Windows 7, and on Raspbian Linux (Although Raspbian Linux runs Pharo UI events slo-o-owly, so I am not the best person to do the Linux testing) On 12 November 2015 at 15:42, EuanM <euanmee(a)gmail.com> wrote: > I think anything that makes a consistent UI to Pharo for users on > different platforms is a good thing. > > However, I do not know what the downside-effects of this change are. > > The key thing is to capture all breaks caused by the changed, so they > can be coded through or around. And ensure that the test results are > viewed with at least the same importance as the automated Jenkins CI > test results. > > I do not know the extent that the Jenkins CI tests actually simulate > user keyboard input, so... (Pharo pseudocode for comic effect and > clarity) > > We can write Jenkins tests that simulate user input > ifTrue: [ we should do so ]. > ifFalse: [ > we should > write a canonical set of regression test scripts for human > users to follow . > we should > do: [:eachPlatformSupportedByPharo | > run the tests in which humans use keyboard tasks ] . > we should > capture the results ; > and feed them into automated the regression tests results database > ] > > > On 12 November 2015 at 08:35, Nicolai Hess <nicolaihess(a)gmail.com> wrote: >> A short overview: >> >> With Ctrl-Key pressed, >> >> some keyboard events don't generate smalltalk keystrokes (EventKeyChar) >> ctrl+tab, ctrl+1, ctrl+2, ..., ctrl+0, ctrl+Tab, ctrl+m >> >> for some keyboard events there are only key down and key up events (from >> windows) but the vm "generates" a keypress (EventKeyChar) event >> ctrl+Home, ctrl+End, ctrl+PageUp, ... >> >> But the key value , char code and ctrl flag for this generated events, make >> them undistinguishable from other ctrl+key shortcuts >> ctrl+a = ctrl+Home, ctrl+d = ctrl+End, ctrl+k=ctrl+PageUp,... >> >> And one key (ctrl+Enter) creates two EventKeyChar events one "normal" and >> one that is generated from the vm >> one with keyvalue 10 and one with 13. >> >> And the keycode for char like "a","b", ..., are different if you they are >> typed with or without ctrl-key >> >> just key "a": >> keycode: >> 65 (keydown event) >> 97 (keyvalue/charcode event) (65 if you press shift key) >> 65 (keyup event) >> >> with ctrl key pressed: >> keycode >> 65 (keydown event) >> 1 (keyvalue/charcode event) >> 65 (keyup event) >> >> >> I would like to change this, that means: >> >> generate EventKeyChar events for >> ctrl+1, ctrl+2, ..., ctrl+Tab, ctrl+m >> >> remove the double key event for ctrl+Enter >> >> and adopt the behavior of the linux vm, for keyValue, charCode values for >> events like >> ctrl+a, ctrl+home, ... >> >> for example, the linux vm generates event buffer values: >> #(2 7329656 1 0 2 65 0 1) for ctrl+a (keyValue 1/charCode 65) >> but >> #(2 7329656 1 0 2 1 0 1) for ctrl+home (keyValue 1/charCode 1) >> >> this may need some changes on the image side. For example, in Pharo we >> have some special handling that adds "96" to the keyValue for keystrokes >> with ctrl-key). >> >> What do you think? >> >> >> >> nicolai >>

Nov. 12, 2015

Re: [Pharo-dev] keyboard events in the windows VM

by EuanM

I think anything that makes a consistent UI to Pharo for users on different platforms is a good thing. However, I do not know what the downside-effects of this change are. The key thing is to capture all breaks caused by the changed, so they can be coded through or around. And ensure that the test results are viewed with at least the same importance as the automated Jenkins CI test results. I do not know the extent that the Jenkins CI tests actually simulate user keyboard input, so... (Pharo pseudocode for comic effect and clarity) We can write Jenkins tests that simulate user input ifTrue: [ we should do so ]. ifFalse: [ we should write a canonical set of regression test scripts for human users to follow . we should do: [:eachPlatformSupportedByPharo | run the tests in which humans use keyboard tasks ] . we should capture the results ; and feed them into automated the regression tests results database ] On 12 November 2015 at 08:35, Nicolai Hess <nicolaihess(a)gmail.com> wrote: > A short overview: > > With Ctrl-Key pressed, > > some keyboard events don't generate smalltalk keystrokes (EventKeyChar) > ctrl+tab, ctrl+1, ctrl+2, ..., ctrl+0, ctrl+Tab, ctrl+m > > for some keyboard events there are only key down and key up events (from > windows) but the vm "generates" a keypress (EventKeyChar) event > ctrl+Home, ctrl+End, ctrl+PageUp, ... > > But the key value , char code and ctrl flag for this generated events, make > them undistinguishable from other ctrl+key shortcuts > ctrl+a = ctrl+Home, ctrl+d = ctrl+End, ctrl+k=ctrl+PageUp,... > > And one key (ctrl+Enter) creates two EventKeyChar events one "normal" and > one that is generated from the vm > one with keyvalue 10 and one with 13. > > And the keycode for char like "a","b", ..., are different if you they are > typed with or without ctrl-key > > just key "a": > keycode: > 65 (keydown event) > 97 (keyvalue/charcode event) (65 if you press shift key) > 65 (keyup event) > > with ctrl key pressed: > keycode > 65 (keydown event) > 1 (keyvalue/charcode event) > 65 (keyup event) > > > I would like to change this, that means: > > generate EventKeyChar events for > ctrl+1, ctrl+2, ..., ctrl+Tab, ctrl+m > > remove the double key event for ctrl+Enter > > and adopt the behavior of the linux vm, for keyValue, charCode values for > events like > ctrl+a, ctrl+home, ... > > for example, the linux vm generates event buffer values: > #(2 7329656 1 0 2 65 0 1) for ctrl+a (keyValue 1/charCode 65) > but > #(2 7329656 1 0 2 1 0 1) for ctrl+home (keyValue 1/charCode 1) > > this may need some changes on the image side. For example, in Pharo we > have some special handling that adds "96" to the keyValue for keystrokes > with ctrl-key). > > What do you think? > > > > nicolai >

Nov. 12, 2015

Using a newer vm with PharoLauncher, platform differences

by Stephan Eggermont

I wanted to use a newer vm with PharoLauncher, so I added a vms directory in /Users/stephan/Library/Application Support/Pharo/ and used curl get.pharo.org/vmLatest50 | bash to install it there. In PharoLauncher I then set the vm full path to either /Users/stephan/Library/Application Support/Pharo/vms/pharo-vm/Pharo.app/Contents/MacOS/pharo or /Users/stephan/Library/Application Support/Pharo/vms/pharo-vm/Pharo.app Neither works to launch an image (Mac 10.10.5). The similar procedure on Ubuntu 15.04 works (though the apt dependencies of SDL2/Cairo/Freetype:i386 are somewhat broken). What is the vm path supposed to be? Stephan

Nov. 12, 2015

[pharo-project/pharo-core]

by GitHub

Branch: refs/tags/50439 Home: https://github.com/pharo-project/pharo-core

Nov. 12, 2015

[pharo-project/pharo-core] 56b5ef: 50439

by GitHub

Branch: refs/heads/5.0 Home: https://github.com/pharo-project/pharo-core Commit: 56b5ef44e07c2b445b71e979444fe8b2da4dc3b8 https://github.com/pharo-project/pharo-core/commit/56b5ef44e07c2b445b71e979… Author: Jenkins Build Server <board(a)pharo-project.org> Date: 2015-11-12 (Thu, 12 Nov 2015) Changed paths: M ConfigurationOfVersionner.package/ConfigurationOfVersionner.class/instance/symbolic versions/stable_.st A ConfigurationOfVersionner.package/ConfigurationOfVersionner.class/instance/versions/version2137_.st A Deprecated50.package/extension/TBehavior/instance/methodDictAddSelectorSilently_withMethod_.st R ScriptLoader50.package/ScriptLoader.class/instance/pharo - scripts/script50438.st A ScriptLoader50.package/ScriptLoader.class/instance/pharo - scripts/script50439.st R ScriptLoader50.package/ScriptLoader.class/instance/pharo - updates/update50438.st A ScriptLoader50.package/ScriptLoader.class/instance/pharo - updates/update50439.st M ScriptLoader50.package/ScriptLoader.class/instance/public/commentForCurrentUpdate.st R Traits.package/TBehavior.class/instance/adding%2Fremoving methods/methodDictAddSelectorSilently_withMethod_.st A Traits.package/TClass.class/instance/subclass creation/subclass_uses_instanceVariableNames_classVariableNames_package_.st M Versionner-Core-DependenciesModel.package/MTDependency.class/instance/comparing/=.st M Versionner-Core-Model.package/MBGroupInfo.class/instance/accessing/name.st Log Message: ----------- 50439 16992 deprecate methodDictAddSelectorSilently:withMethod: https://pharo.fogbugz.com/f/cases/16992 16952 Change class template creation https://pharo.fogbugz.com/f/cases/16952 16975 versionner (Object>>#name) https://pharo.fogbugz.com/f/cases/16975 http://files.pharo.org/image/50/50439.zip

Nov. 12, 2015

Re: [Pharo-dev] [Pharo-project] [Vm-dev] Cog issue? Re: StandardFileStream size limit?

by Nicolai Hess

2015-11-12 14:10 GMT+01:00 Dimitris Chloupis <kilon.alios(a)gmail.com>: > latest vm has issues with OSProcess stdout the last time I checked on > Yosemite. I would not call it stable unless we talk only crashes here. > Ah, I didn't know this was a vm issue. Does this means it is working with the "official" stable vm ? > > On Thu, Nov 12, 2015 at 2:55 PM Nicolai Hess <nicolaihess(a)gmail.com> > wrote: > >> 2015-11-12 6:14 GMT+01:00 Jimmie Houchin <jlhouchin(a)gmail.com>: >> >>> This still fails three and a half years later with a default Pharo 4 vm >>> and image. >>> :( >>> >>> I spent a few hours trying to diagnose why Pharo was telling me that I >>> file I knew existed didn't exist. Very frustrating trying to debug this and >>> then to find it is a vm and not an image issue. I then to search the >>> archives to find this is a known issue for this long and still broken. >>> >>> The latest linux vm works fine. >>> http://files.pharo.org/vm/pharo/linux/Pharo-VM-linux-latest.zip >>> >> >> Yes, the latest vm is pretty stable. But I think the problem is, there >> are to many changes for to many >> different projects: >> - I added a change to the windows keyevent handling and created a small >> regression (can not use ctrl+space anymore) >> - changes for OSWindow /SDL >> - spur (?) or any other changes from the main squeak vm >> - ... >> >> I think we should have create release candidate branch and make sure more >> people testing that version and all >> newer changes go to the master/development branch only. >> >> nicolai >> >> >>> >>> Is there a reason why the "latest" vm isn't the stable vm? Or at least >>> this bugfix applied to the stable vm? >>> >>> Thanks. >>> >>> Jimmie >>> >>> >>> >>> On 06/13/2012 05:00 PM, eliot.miranda at gmail.com (Eliot Miranda) >>> wrote: >>> >>>> On Wed, Jun 13, 2012 at 7:40 AM, David T. Lewis <lewis at mail.msen.com> >>>> wrote: >>>> >>>> On Wed, Jun 13, 2012 at 04:08:14PM +0200, G?ran Krampe wrote: >>>>> >>>>>> Hey! >>>>>> >>>>>> Ok, so the plot thickens: >>>>>> >>>>>> If I run this in a "oneclick 1.4" I will get a file write error on >>>>>> 2Gb: >>>>>> >>>>>> | f b | >>>>>> f := StandardFileStream newFileNamed: 'test'. >>>>>> b := ByteArray new: 1024*1024*100. "100Mb" >>>>>> [30 timesRepeat: [f nextPutAll: b]] ensure: [f close] "3Gb" >>>>>> >>>>>> >>>>>> ...but it worked fine using a self built "standard" VM from >>>>>> squeakvm.org >>>>>> (4.4.7.2357)! Same image btw. >>>>>> >>>>>> Now, does it work with the bleeding edge Cog? Tried r2556 and nope, >>>>>> same >>>>>> problem. >>>>>> >>>>>> So I "guess" it is a Cog thing? I haven't tried building Cog from >>>>>> source. >>>>>> >>>>>> regards, G?ran >>>>>> >>>>> Background on large file support, see especially Bert's summary: >>>>> >>>>> >>>>> >>>>> http://forum.world.st/Re-squeak-dev-filesize-reporting-0-for-very-large-fil… >>>>> http://bugs.squeak.org/view.php?id=7522 >>>>> http://www.suse.de/~aj/linux_lfs.html >>>>> >>>>> I'm not sure if the Cog VMs are being compiled with the the LFS option, >>>>> although >>>>> I expect that if you compile it yourself with the build options that >>>>> Bert >>>>> explains, >>>>> then it should start working. >>>>> >>>>> I just committed the necessary changes for my branch. Will rebuild >>>> soon. >>>> G?ran, if you're in a hurry build your own? You'd need to be in >>>> http://www.squeakvm.org/svn/squeak/branches/Cog/unixbuild/bld and run >>>> ./mvm. >>>> >>> >>> >>>

Nov. 12, 2015

Re: [Pharo-dev] [Pharo-project] [Vm-dev] Cog issue? Re: StandardFileStream size limit?

by Dimitris Chloupis

latest vm has issues with OSProcess stdout the last time I checked on Yosemite. I would not call it stable unless we talk only crashes here. On Thu, Nov 12, 2015 at 2:55 PM Nicolai Hess <nicolaihess(a)gmail.com> wrote: > 2015-11-12 6:14 GMT+01:00 Jimmie Houchin <jlhouchin(a)gmail.com>: > >> This still fails three and a half years later with a default Pharo 4 vm >> and image. >> :( >> >> I spent a few hours trying to diagnose why Pharo was telling me that I >> file I knew existed didn't exist. Very frustrating trying to debug this and >> then to find it is a vm and not an image issue. I then to search the >> archives to find this is a known issue for this long and still broken. >> >> The latest linux vm works fine. >> http://files.pharo.org/vm/pharo/linux/Pharo-VM-linux-latest.zip >> > > Yes, the latest vm is pretty stable. But I think the problem is, there are > to many changes for to many > different projects: > - I added a change to the windows keyevent handling and created a small > regression (can not use ctrl+space anymore) > - changes for OSWindow /SDL > - spur (?) or any other changes from the main squeak vm > - ... > > I think we should have create release candidate branch and make sure more > people testing that version and all > newer changes go to the master/development branch only. > > nicolai > > >> >> Is there a reason why the "latest" vm isn't the stable vm? Or at least >> this bugfix applied to the stable vm? >> >> Thanks. >> >> Jimmie >> >> >> >> On 06/13/2012 05:00 PM, eliot.miranda at gmail.com (Eliot Miranda) wrote: >> >>> On Wed, Jun 13, 2012 at 7:40 AM, David T. Lewis <lewis at mail.msen.com> >>> wrote: >>> >>> On Wed, Jun 13, 2012 at 04:08:14PM +0200, G?ran Krampe wrote: >>>> >>>>> Hey! >>>>> >>>>> Ok, so the plot thickens: >>>>> >>>>> If I run this in a "oneclick 1.4" I will get a file write error on 2Gb: >>>>> >>>>> | f b | >>>>> f := StandardFileStream newFileNamed: 'test'. >>>>> b := ByteArray new: 1024*1024*100. "100Mb" >>>>> [30 timesRepeat: [f nextPutAll: b]] ensure: [f close] "3Gb" >>>>> >>>>> >>>>> ...but it worked fine using a self built "standard" VM from >>>>> squeakvm.org >>>>> (4.4.7.2357)! Same image btw. >>>>> >>>>> Now, does it work with the bleeding edge Cog? Tried r2556 and nope, >>>>> same >>>>> problem. >>>>> >>>>> So I "guess" it is a Cog thing? I haven't tried building Cog from >>>>> source. >>>>> >>>>> regards, G?ran >>>>> >>>> Background on large file support, see especially Bert's summary: >>>> >>>> >>>> >>>> http://forum.world.st/Re-squeak-dev-filesize-reporting-0-for-very-large-fil… >>>> http://bugs.squeak.org/view.php?id=7522 >>>> http://www.suse.de/~aj/linux_lfs.html >>>> >>>> I'm not sure if the Cog VMs are being compiled with the the LFS option, >>>> although >>>> I expect that if you compile it yourself with the build options that >>>> Bert >>>> explains, >>>> then it should start working. >>>> >>>> I just committed the necessary changes for my branch. Will rebuild >>> soon. >>> G?ran, if you're in a hurry build your own? You'd need to be in >>> http://www.squeakvm.org/svn/squeak/branches/Cog/unixbuild/bld and run >>> ./mvm. >>> >> >> >>

Nov. 12, 2015

Re: [Pharo-dev] [Pharo-project] [Vm-dev] Cog issue? Re: StandardFileStream size limit?

by Nicolai Hess

2015-11-12 6:14 GMT+01:00 Jimmie Houchin <jlhouchin(a)gmail.com>: > This still fails three and a half years later with a default Pharo 4 vm > and image. > :( > > I spent a few hours trying to diagnose why Pharo was telling me that I > file I knew existed didn't exist. Very frustrating trying to debug this and > then to find it is a vm and not an image issue. I then to search the > archives to find this is a known issue for this long and still broken. > > The latest linux vm works fine. > http://files.pharo.org/vm/pharo/linux/Pharo-VM-linux-latest.zip > Yes, the latest vm is pretty stable. But I think the problem is, there are to many changes for to many different projects: - I added a change to the windows keyevent handling and created a small regression (can not use ctrl+space anymore) - changes for OSWindow /SDL - spur (?) or any other changes from the main squeak vm - ... I think we should have create release candidate branch and make sure more people testing that version and all newer changes go to the master/development branch only. nicolai > > Is there a reason why the "latest" vm isn't the stable vm? Or at least > this bugfix applied to the stable vm? > > Thanks. > > Jimmie > > > > On 06/13/2012 05:00 PM, eliot.miranda at gmail.com (Eliot Miranda) wrote: > >> On Wed, Jun 13, 2012 at 7:40 AM, David T. Lewis <lewis at mail.msen.com> >> wrote: >> >> On Wed, Jun 13, 2012 at 04:08:14PM +0200, G?ran Krampe wrote: >>> >>>> Hey! >>>> >>>> Ok, so the plot thickens: >>>> >>>> If I run this in a "oneclick 1.4" I will get a file write error on 2Gb: >>>> >>>> | f b | >>>> f := StandardFileStream newFileNamed: 'test'. >>>> b := ByteArray new: 1024*1024*100. "100Mb" >>>> [30 timesRepeat: [f nextPutAll: b]] ensure: [f close] "3Gb" >>>> >>>> >>>> ...but it worked fine using a self built "standard" VM from >>>> squeakvm.org >>>> (4.4.7.2357)! Same image btw. >>>> >>>> Now, does it work with the bleeding edge Cog? Tried r2556 and nope, same >>>> problem. >>>> >>>> So I "guess" it is a Cog thing? I haven't tried building Cog from >>>> source. >>>> >>>> regards, G?ran >>>> >>> Background on large file support, see especially Bert's summary: >>> >>> >>> >>> http://forum.world.st/Re-squeak-dev-filesize-reporting-0-for-very-large-fil… >>> http://bugs.squeak.org/view.php?id=7522 >>> http://www.suse.de/~aj/linux_lfs.html >>> >>> I'm not sure if the Cog VMs are being compiled with the the LFS option, >>> although >>> I expect that if you compile it yourself with the build options that Bert >>> explains, >>> then it should start working. >>> >>> I just committed the necessary changes for my branch. Will rebuild soon. >> G?ran, if you're in a hurry build your own? You'd need to be in >> http://www.squeakvm.org/svn/squeak/branches/Cog/unixbuild/bld and run >> ./mvm. >> > > >

Nov. 12, 2015

Re: [Pharo-dev] Smalltalkhub returning 500 response for a public repository

by Denis Kudriashov

Same for me now. I try commit to CAR/Ghost repository 2015-11-11 20:30 GMT+01:00 Euan Mee <euanmee(a)gmail.com>: > I've been trying unsuccessfully to write to the > SandstoneDb public repository at smalltalkhub.com > > http://smalltalkhub.com/#!/~gnaritas/SandstoneDb > shows the green tickmark and PUBLIC > > Writing to a package within it receives a 500 error. > > The specific package I am writing to is: > SandstoneDbTests > which is marked in Monticello (after I have downloaded it) as belonging in > http://smalltalkhub.com/mc/gnaritas/SandstoneDb/main > > as do all the other listed packages, i.e. > ConfigurationOfSandstoneDb > SandstoneDb > SandstoneDbFileSystemStore > SandstoneDbFuel > SandstoneDbFileDirectoryStore > SandstoneDbSqueak > SandstoneGOODS > SandstondDbGOODSTests > > (Thank you to Max Leske for confirming it was a 500 error at the > smalltalkhub.com end that was causing the debugger to respond to my > attempted commit). > >

Nov. 12, 2015

Re: [Pharo-dev] Does anybody know if Sport is available in Pharo?

by Christophe Demarey

Le 12 nov. 2015 Ã 10:15, Frank Shearar a Ã©crit : > On 12 November 2015 at 09:04, Christophe Demarey > <Christophe.Demarey(a)inria.fr> wrote: >> >> Le 10 nov. 2015 Ã 21:39, Sven Van Caekenberghe a Ã©crit : >> >>> << Sport is a Smalltalk portability layer. Sport presents a consistent interface and implemented in a number of Smalltalk dialects. Smalltalk code written to the Sport interfaces can be run in any Smalltalk dialect without alteration. >> >>> >>> Anyway, that was the idea some years ago. It is somewhat comparable to Grease (used by Seaside). >> >> Any reason to prefer one or the other? > > While they're both compatibility layers, Grease is a portability layer > _for Seaside_, while Sport is a portability layer for time, sockets, > and files. ok, thanks

Nov. 12, 2015