February 2017 - Pharo-dev - lists.pharo.org

Re: [Pharo-dev] Spotter - menu items

by Tudor Girba

Hi, > On Feb 22, 2017, at 5:47 AM, Ben Coman <btc(a)openInWorld.com> wrote: > > I happened to be Spotting for image saveXXX methods and at the top of > results is > #Menu > + Save > + Save as... > + Save and quit > > Now what would be really nice is if the code pane showed the > implementation of the menu item. Usually I resort to using Halos to > inspect the menu item morph to view its implementation. > > > On a slightly separate point, it feels awkward clicking Save in the > Spotter search, > since there is no visual feedback anything happened. I had to go back > and experiment by choosing [Save as] to understand that clicking it > actually executed it. > > Perhaps executing a menu item from Spotter could #inform: us > that a menu item was executed. This is in spite that choosing [Save] > from the World menu doesn't provide any feedback. > I think that conforms to expected behaviour, but most other Spotter results > only browse, not execute. This is an interesting idea. Could you open an issue for it? > Actually how many other Spotter extensions execute something? > It is maybe potentially dangerous that it behaves differently to other results, > and clicking on it (e.g. trying to expand it like other results) > causes side effects. Clicking triggers an action. What the action does is not controlled and nor should it. Cheers, Doru > Perhaps the code pane showing the menu item implementation could have > <Execute> and <Debug> buttons above it. > > cheers -ben > -- www.tudorgirba.com www.feenk.com âThe smaller and more pervasive the hardware becomes, the more physical the software gets."

Feb. 26, 2017

Re: [Pharo-dev] Why do we have doublequoting for "?

by Tudor Girba

Hi Stef, Just a clarification. The mechanism you mention is obtained through Cmd+p, Enter and this embeds the printout as a comment in the existing editor. This is particularly useful when you want to keep a trace of multiple executions in the same editor (typically a Playground). In order to make any string a valid Pharo comment we use String>>asComment, and this escapes â (double-quote) inside a comment. Without it, the result would not be a valid comment and the syntax highlighting would be broken. For your goal of having the result in the debugger this is obviously not ideal. This is why we ended up with Cmd+p,Cmd+p, but this is not available in Pharo 5. Cheers, Doru > On Feb 24, 2017, at 8:30 AM, stepharong <stepharong(a)free.fr> wrote: > > Yuriy > > > Why does it make sense to have "" "" inside strings? > > > So if you paste this as a comment somewhere you donât have to manually add double double-quotes. > > Did you read my example? > I do not need any double double quotes. > > > I donât use the result of print it at all, but I guess there are two scenarios for itâs usage. One as you described: copy the result and use it in tests or in some further computation. > > But this is exactly what I cannot do with this stupid doubling of double quotes > > (but if you use TDD test are already written ;P). > > well... > > Another one is to copy the result and paste it somewhere as a comment in your code, to access it easily afterwards. If you have a double quote in your result it will break the comment, so the thing is escaped automatically. > > Probably but here this is inside a string. > I think that this heuristic is totally bogus. > > > Uko > >> On 23 Feb 2017, at 22:25, stepharong <stepharong(a)free.fr> wrote: >> >> Thnkas gabriel >> I'm hacking in an old Pharo 50 image and it was terrible because printit even a copyReplaceAll: '""' byt: '"' >> did not work. >> I do not understand what is the scenario to double quote comment character " >> >> I understand single quotes but not double quotes. >> stef >> >> Try doing Ctrl+P twice. It will not comment the printed string. >> >> On Thu, Feb 23, 2017 at 4:17 PM, stepharong <stepharong(a)free.fr> wrote: >> about quoting ;( >> >> Hi >> >> I have a string >> >> '<li> >> Stéphane Ducasse and Damien Pollet, Fingerprints, Journal of Information System, 2010. >> </li> >> ' >> >> and when I do printIt in the debugger I get the following >> >> >> "'<li> >> Stéphane Ducasse and Damien Pollet, Fingerprints, Journal of Information System, 2010. >> </li> >> '" >> >> I do not get why we doublequote the character " >> inside string. >> >> It means that I have to remove all the " inside to string to be able to express tests. >> >> Stef >> >> >> >> >> >> -- >> Using Opera's mail client: http://www.opera.com/mail/ > > > > > -- > Using Opera's mail client: http://www.opera.com/mail/ -- www.tudorgirba.com www.feenk.com "Innovation comes in the least expected form. That is, if it is expected, it already happened."

Feb. 26, 2017

refresh in EyeInspector disabled?

by Andrei Chis

Hi, In latest Pharo 6 version (60411) I see that EyeInspector does not do automatic refresh. Doing 'Morph new basicInspect' selecting the 'color' attribute and then executing in the code pane 'self color: Color red' has no effect. Was this disabled explicitly or is it an error? Cheers, Andrei

Feb. 26, 2017

Re: [Pharo-dev] [ANN] Woden 2 first release for OS X and Linux

by Alexandre Bergel

This is great Ronie! I tried and it works like a charm. Impressive work! This will revive the Roassal3D project! Alexandre > On Feb 23, 2017, at 10:02 PM, Ronie Salgado <roniesalg(a)gmail.com> wrote: > > Hello, > > Finally I am glad to announce a first release of Woden 2. I made a hopefully simplified process for loading Woden 2 out of the box. > > For loading and testing Woden 2, follow the instructions at: > > https://github.com/ronsaldo/woden2 > > In this release, I am not loading the level editor by default because it depends on an outdated version of Bloc. Until I fix the dependency problem of the Level editor, it will remain unusable. > > Because Woden 2 requires Metal on OS X, it requires 64 bits on OS X, which can be unstable. This is tested on OS X El Capitan. I do not know whether it will work or not on OS X Sierra. > > Best regards, > Ronie > <Captura de pantalla 2017-02-23 a las 10.00.45 p.m..png> -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.

Feb. 26, 2017

Re: [Pharo-dev] Any idea for a cool name for the remote tool suite?

by Pierce Ng

On Fri, Feb 24, 2017 at 04:42:03PM -0800, Sean P. DeNigris wrote: > Andreas Wacknitz wrote > > Everything that is used outside Pharo could have a fancy name. > > For internal tools and frameworks this is... like talking in argot and > > creates high walls around Pharo. > > +100 Just call it Bike Teleporter already. :-P Pierce

Feb. 26, 2017

Re: [Pharo-dev] Build Pharo VM

by Pierce Ng

On Thu, Feb 23, 2017 at 03:51:07PM +0100, Esteban Lorenzano wrote: > for linux you can: > $ cd pharo-vm/opensmalltalk-vm//build.linux32x86/pharo.cog.spur/build > $ ./mvm On a fairly fresh 32-bit Linux Mint 17 (based on Ubuntu 14.04), first apt-get install curl, libssl-dev, cmake, uuid-dev, and libtool which are the dependencies. On libtool, the (generated?) source contains a version which doesn't understand a command line argument that it was fed. I changed that libtool file to a symlink to the system-installed one. The HT version that I built segfaults now and then on my 64-bit Linux Mint 17. I'm using the ITHB one. Pierce

Feb. 26, 2017

Re: [Pharo-dev] [Moose-dev] Re: GTInspector duplicates overriden extensions

by Peter Uhnak

On Sat, Feb 25, 2017 at 04:44:15PM +0100, Andrei Chis wrote: > Hi, > > Usually when we want to specialize a view in a subclass we override the > method and do not add the gtInspectorPresentationOrder: annotation. But > also removing duplicationd would make sense. Ah, I didn't think of just removing the pragma. :) Thanks! Peter

Feb. 25, 2017

Re: [Pharo-dev] [Moose-dev] GTInspector duplicates overriden extensions

by Andrei Chis

Hi, Usually when we want to specialize a view in a subclass we override the method and do not add the gtInspectorPresentationOrder: annotation. But also removing duplicationd would make sense. Cheers, Andrei On Feb 25, 2017 4:36 PM, "Peter Uhnak" <i.uhnak(a)gmail.com> wrote: Hi, when I override a method with a GTInspector (because I need to change it), then inspector contains the extension twice. It seems Inspector scans the whole hierarchy of the inspected object for methods, but doesn't deduplicate them. Peter _______________________________________________ Moose-dev mailing list Moose-dev(a)list.inf.unibe.ch https://www.list.inf.unibe.ch/listinfo/moose-dev

Feb. 25, 2017

GTInspector duplicates overriden extensions

by Peter Uhnak

Hi, when I override a method with a GTInspector (because I need to change it), then inspector contains the extension twice. It seems Inspector scans the whole hierarchy of the inspected object for methods, but doesn't deduplicate them. Peter

Feb. 25, 2017

Re: [Pharo-dev] [ANN] Woden 2 first release for OS X and Linux

by Ronie Salgado

2017-02-25 2:08 GMT-03:00 Ben Coman <btc(a)openinworld.com>: > What happens if the wrong type object is passed to a lowcode method? You should get an error. A Lowcode method checks the type of its argument when called from a normal Pharo method. When calling a Lowcode method from another Lowcode method, the type check happens in compilation time. The type check of the arguments is only removed when a Lowcode method is inlined. By default, each variable has the object type, which corresponds with any normal Pharo Object. > Unfortunately, when we did this work the 64 bits version of Spur was > incomplete and in an experimental state [and we were] not able to compare > the performance difference for floating point arithmetics against a > Smalltalk with support for immediate floating point objects > > Have you been able to do this performance comparison since the paper > was written? > I have not tried it. But, we now have the 64 bits Lowcode VM, which is used by Woden 2 by default on OS X. Currently I am just too lazy for doing this comparison, because there lot of more interesting things that can be done with Lowcode. Anyway, in theory most of the performance gain could be obtained from loops that need to manipulate memory pointers without bounds checks, instead of the small linear algebra that I am testing in that. I have not finished the support for the pointers in the Lowcode Opal Compiler, so I am quite sure we still are not getting all of the performance improvement that can be obtained with Lowcode. I am seriously thinking on remaking the Lowcode Opal Compiler, but using Slovim for generating more optimized code, and doing more aggressive inlining. Slovim is a LLVM style SSA intermediate representation that I am currently using for generating all of the shaders used by Woden 2. In Slovim I already have backends for SpirV, GLSL, the Metal Shader language, a x86 (and x86_64) machine code generator that I am refactoring. Now, I am only missing Lowcode backend, and we could be compiling actual C code alongside Pharo code in the near future. The only bad think about integrating C is its syntax. Because in C not everything is a expression, I do not believe it would be behave well in playground. Best regards, Ronie

Feb. 25, 2017