- Pharo-dev - lists.pharo.org

Re: [Pharo-dev] Moz2d Vs. Skia

by ducasse

Tx doru! > On 8 Nov 2019, at 11:41, Tudor Girba <tudor(a)tudorgirba.com> wrote: > > Hi > > Thanks for following the development of Bloc :). > > We invest in multiple backends because of several reasons. For example, for window hosting we have SDL2 (which relies on polling) and Glutin (which relies on a push model), but we also have Morphic. And now, when it comes to rendering we have both Moz2D and Skia. These have different properties. > > When we started to work with Moz2D, it was the only cross-platform vector graphical framework with rich support of text rendering. We knew that we eventually need to move away from it, but it allowed us to bootstrap nicely. At that time Skia was actually one of the backends of Moz2D. In the meantime, Skia evolved significantly. It hasnât quite reach the level of Moz2D when it comes to text yet, but for our typical use cases, it is already good enough. > > While Moz2D works fine today, we now want to move to Skia because of maintenance and support for newer technologies. For example, both Moz2D and Skia are written in C++ which mean we can not directly use both of them via FFI and have to implement C ABI bindings anyway. However, for Skia, there exists a project (https://github.com/rust-skia/rust-skia) that provides an up-to-date Skia bindings in Rust. This provides a quick way to build and to integrate into CI, and it also looks memory safe. Skia also has support for Vulkan. > > Now, while we target Skia in the short term, we also consider and play with other combinations. In particular, WebRenderer looks interesting. For text, we could use Skia, or we could also consider using something like font-kit. > > I wanted to detail a bit how we work to show how we exercise the underlying architecture continuously. A similar story happens on the host side. We think it is critical to have as much flexibility as possible at that level, while also exposing all underlying abilities in the image. For example, an important piece when dealing with text is to have access to accurate measurements. So, those are available through Sparta. > > Cheers, > Doru > > >> On Nov 7, 2019, at 6:52 AM, ducasse <stepharo(a)netcourrier.com> wrote: >> >> Hi doru and alex >> >> Can you tell me why you drop Moz2D and switch to Skia (which to us looks a much safer path starting from the packaging :)? >> >> S >> >> >> > > -- > feenk.com > > "Quality cannot be an afterthought." > >

Nov. 8, 2019

[Pharo 8.0] Build #961: Fix-bug-in-MCClassTraitDefinition

by ci-pharo-ci-jenkins2＠inria.fr

There is a new Pharo build available! The status of the build #961 was: FAILURE. The Pull Request #5099 was integrated: "Fix-bug-in-MCClassTraitDefinition" Pull request url: https://github.com/pharo-project/pharo/pull/5099 Issue Url: https://github.com/pharo-project/pharo/issues/FIx Build Url: https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2…

Nov. 8, 2019

[Pharo 8.0] Build #960: Compiler API: turn forSyntaxHighlighting into an option

by ci-pharo-ci-jenkins2＠inria.fr

There is a new Pharo build available! The status of the build #960 was: SUCCESS. The Pull Request #5082 was integrated: "Compiler API: turn forSyntaxHighlighting into an option" Pull request url: https://github.com/pharo-project/pharo/pull/5082 Issue Url: https://github.com/pharo-project/pharo/issues/CompilerAPICodeCompletion Build Url: https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2…

Nov. 8, 2019

Re: [Pharo-dev] Moz2d Vs. Skia

by Tudor Girba

Hi Thanks for following the development of Bloc :). We invest in multiple backends because of several reasons. For example, for window hosting we have SDL2 (which relies on polling) and Glutin (which relies on a push model), but we also have Morphic. And now, when it comes to rendering we have both Moz2D and Skia. These have different properties. When we started to work with Moz2D, it was the only cross-platform vector graphical framework with rich support of text rendering. We knew that we eventually need to move away from it, but it allowed us to bootstrap nicely. At that time Skia was actually one of the backends of Moz2D. In the meantime, Skia evolved significantly. It hasnât quite reach the level of Moz2D when it comes to text yet, but for our typical use cases, it is already good enough. While Moz2D works fine today, we now want to move to Skia because of maintenance and support for newer technologies. For example, both Moz2D and Skia are written in C++ which mean we can not directly use both of them via FFI and have to implement C ABI bindings anyway. However, for Skia, there exists a project (https://github.com/rust-skia/rust-skia) that provides an up-to-date Skia bindings in Rust. This provides a quick way to build and to integrate into CI, and it also looks memory safe. Skia also has support for Vulkan. Now, while we target Skia in the short term, we also consider and play with other combinations. In particular, WebRenderer looks interesting. For text, we could use Skia, or we could also consider using something like font-kit. I wanted to detail a bit how we work to show how we exercise the underlying architecture continuously. A similar story happens on the host side. We think it is critical to have as much flexibility as possible at that level, while also exposing all underlying abilities in the image. For example, an important piece when dealing with text is to have access to accurate measurements. So, those are available through Sparta. Cheers, Doru > On Nov 7, 2019, at 6:52 AM, ducasse <stepharo(a)netcourrier.com> wrote: > > Hi doru and alex > > Can you tell me why you drop Moz2D and switch to Skia (which to us looks a much safer path starting from the packaging :)? > > S > > > -- feenk.com "Quality cannot be an afterthought."

Nov. 8, 2019

[Pharo 8.0] Build #959: Clean Display packages group dependencies

by ci-pharo-ci-jenkins2＠inria.fr

There is a new Pharo build available! The status of the build #959 was: SUCCESS. The Pull Request #5092 was integrated: "Clean Display packages group dependencies" Pull request url: https://github.com/pharo-project/pharo/pull/5092 Issue Url: https://github.com/pharo-project/pharo/issues/displayDependencies1 Build Url: https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2…

Nov. 7, 2019

[Pharo 8.0] Build #958: 5017-should-have-a-setting-to-select-the-prettyprinter

by ci-pharo-ci-jenkins2＠inria.fr

There is a new Pharo build available! The status of the build #958 was: FAILURE. The Pull Request #5091 was integrated: "5017-should-have-a-setting-to-select-the-prettyprinter" Pull request url: https://github.com/pharo-project/pharo/pull/5091 Issue Url: https://github.com/pharo-project/pharo/issues/5017 Build Url: https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2…

Nov. 7, 2019

Re: [Pharo-dev] Horizontal scrolling in FastTable

by ducasse

thanks danil esteban is working on FastTable right now. So I imagine that he will have a look at your enh. Could you open a bug entry and issue a PR so that it does not get lost in email? Stef > On 7 Nov 2019, at 17:53, Danil Osipchuk <danil.osipchuk(a)gmail.com> wrote: > > As a followup, attached is a changeset for Pharo8 with a simple implementation of horizontal scrolling. I tried to keep it minimal. > It adds startColumnIndex var to FTTableContainerMorph which is being set when FTTableContainerMorph>>adjustToHorizontalScrollBarValue: is triggered. > The main change is in #calculateColumnWidths - it starts distributing space width from the startColumnIndex to ensure this area gets most visibility. If startColumnIndex is 0, it reverts to old (current) behavior - trying to draw columns starting from the first one as space permits. > > The scrolling thus jumps to a beginning of the column and I'm not particularly proud of the thing, but it is better then nothing and suits my immediate needs. So may be it is worth a look for those who are interested. > > To try one can use an example: > FTExamples exampleTableHorizontalScroll > > Or even set trialHSB := true in FTTableMorth>>initialize (like I have to enable it everywhere by default) > (as a side note the list is very quiet, compared to what it used to be years ago) > > Ð¿Ð½, 4 Ð½Ð¾ÑÐ±. 2019 Ð³. Ð² 14:14, Danil Osipchuk <danil.osipchuk(a)gmail.com <mailto:danil.osipchuk@gmail.com>>: > Hi all > > I wonder what is happening to the horizontal scrolling in FastTable, the stub is there for years (I've checked the latest 8.0). > I need it rather badly, could someone advise what is the hindrance there? I would try doing it myself, but I've never was doing much of UI programming and only recently returned to Pharo again, so have to catch up (a lot of things changed for good). > > How difficult would be to adopt vertical scrolling approach (which seems to be the same problem already solved to layperson ) > > > horizontalScrollBarValue: aNumber > trialHSB ifNotNil: [ > self inform: 'FTTableMorph>>#horizontalScrollBarValue: ', aNumber printString , String cr, > 'Does nothing yet.' , String cr, 'Need some advise here.' > ]. > > regards, > Danil > <FastTableHorizontalScrolling.cs>

Nov. 7, 2019

Re: [Pharo-dev] [ANN] Roadmap progress report October 2019 - Open meeting 14/11 16h EU

by Alexandre Bergel

Impressive work Esteban! Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. > On 07-11-2019, at 12:47, Sean P. DeNigris <sean(a)clipperadams.com> wrote: > > EstebanLM wrote >> This is the log of what we have been working on last month. > > Wow, that's quite a list! Some of those have been on the wishlist for a > loooong time. Thanks and keep it up :) > > > > ----- > Cheers, > Sean > -- > Sent from: http://forum.world.st/Pharo-Smalltalk-Developers-f1294837.html >

Nov. 7, 2019

[Pharo 8.0] Build #957: MCClassTraitDefinition-should-take-category-into-account-for-equality

by ci-pharo-ci-jenkins2＠inria.fr

There is a new Pharo build available! The status of the build #957 was: FAILURE. The Pull Request #5072 was integrated: "MCClassTraitDefinition-should-take-category-into-account-for-equality" Pull request url: https://github.com/pharo-project/pharo/pull/5072 Issue Url: https://github.com/pharo-project/pharo/issues/MCClassTraitDefinition Build Url: https://ci.inria.fr/pharo-ci-jenkins2/job/Test%20pending%20pull%20request%2…

Nov. 7, 2019

Re: [Pharo-dev] Horizontal scrolling in FastTable

by Danil Osipchuk

As a followup, attached is a changeset for Pharo8 with a simple implementation of horizontal scrolling. I tried to keep it minimal. It adds startColumnIndex var to FTTableContainerMorph which is being set when FTTableContainerMorph>>adjustToHorizontalScrollBarValue: is triggered. The main change is in #calculateColumnWidths - it starts distributing space width from the startColumnIndex to ensure this area gets most visibility. If startColumnIndex is 0, it reverts to old (current) behavior - trying to draw columns starting from the first one as space permits. The scrolling thus jumps to a beginning of the column and I'm not particularly proud of the thing, but it is better then nothing and suits my immediate needs. So may be it is worth a look for those who are interested. To try one can use an example: FTExamples exampleTableHorizontalScroll Or even set trialHSB := true in FTTableMorth>>initialize (like I have to enable it everywhere by default) (as a side note the list is very quiet, compared to what it used to be years ago) Ð¿Ð½, 4 Ð½Ð¾ÑÐ±. 2019 Ð³. Ð² 14:14, Danil Osipchuk <danil.osipchuk(a)gmail.com>: > Hi all > > I wonder what is happening to the horizontal scrolling in FastTable, the > stub is there for years (I've checked the latest 8.0). > I need it rather badly, could someone advise what is the hindrance there? > I would try doing it myself, but I've never was doing much of UI > programming and only recently returned to Pharo again, so have to catch up > (a lot of things changed for good). > > How difficult would be to adopt vertical scrolling approach (which seems > to be the same problem already solved to layperson ) > > > horizontalScrollBarValue: aNumber > trialHSB ifNotNil: [ > self inform: 'FTTableMorph>>#horizontalScrollBarValue: ', aNumber > printString , String cr, > 'Does nothing yet.' , String cr, 'Need some advise here.' > ]. > > regards, > Danil >

Nov. 7, 2019