Re: Current state of GUI's
On 2015-09-06 01:43, motaito wrote: Bummer... However, I appreciate that you are giving it a shot! But you see my point. An out of the box solution like QT would be worth gold. That's the one thing Microsoft got right. They have awesome developer tools. Or in the words of Steve Ballmer: "Developers! Developers! Developers!" :) Instead of loosing time you could just focus on building some software. Imagine D would have a kick ass IDE like visual studio but cross platform like QT. Productivity, efficiency and performance in one. Too bad D doesn't have the option for paid developers to create something like this. It's just too much work for your spare time. So, it's unlikely to happen any time soon :( Yeah, I completely agree. Still, I very much appreciate people like Rikki Cattermole who builds a base to work of for others. The problem is for a few individuals that just takes a long time to complete. That's also why I work on DWT and not a custom library. It's a port of SWT to D. -- /Jacob Carlborg
Re: Current state of GUI's
On Sunday, 6 September 2015 at 02:27:42 UTC, Rikki Cattermole wrote: I would very much appreciate people helping out. E.g. Contributing manipulation functions, image loader / exporter. Most importantly unittests. Fix bugs. I'll most likely do the PNG test case infrastructure, but it would make things go a few months faster if somebody took ownership of testing in general. I'll also be creating a TODO list for what needs to happen with Phobos. E.g. @nogc version of std.string:indexOf. So plenty of places anybody can jump in and help :) I can imagine that you could need help. With work and private projects I am pretty much max out already for the foreseeable future. That's also why I was hoping that D already has the tool, to speed up the process. I hope you find someone though!
Re: Current state of GUI's
On Friday, 4 September 2015 at 14:47:47 UTC, Jacob Carlborg wrote: On 2015-09-04 15:23, motaito wrote: BSD is certainly the most difficult to find a solution. I tried compiling DWT on PC-BSD and it compile just if I remove one check for Linux. Although it doesn't link because I haven't managed to install the necessary libraries yet. Glad to see someone giving an effort for BSD. It would be awesome to have more developers keeping it in mind.
Re: Current state of GUI's
On Friday, 4 September 2015 at 15:03:59 UTC, Chris wrote: I see what you mean. As far as I know, the libraries (PyD, LuaD etc.) would be in the static build of the executable, so you wouldn't have to "distribute" them, only Python or Lua. But that's no problem either, just include lua5.x.so/dll. Please guys, correct me, if I'm wrong here, I don't wanna give motaito any wrong information. I'll just try it and see if it works. If I can simply include the .so/dll I guess it would work well enough. It's more of a hustle if e.g. python has to be installed by the user.
Re: Current state of GUI's
On 2015-09-05 13:41, motaito wrote: Glad to see someone giving an effort for BSD. It would be awesome to have more developers keeping it in mind. It links now when I manged to install all the dependencies, but it doesn't run :(. Segfaults for some reason, I haven't looked in to it. -- /Jacob Carlborg
Re: Current state of GUI's
On 06/09/15 11:43 AM, motaito wrote: On Saturday, 5 September 2015 at 19:13:12 UTC, Jacob Carlborg wrote: It links now when I manged to install all the dependencies, but it doesn't run :(. Segfaults for some reason, I haven't looked in to it. Bummer... However, I appreciate that you are giving it a shot! But you see my point. An out of the box solution like QT would be worth gold. That's the one thing Microsoft got right. They have awesome developer tools. Or in the words of Steve Ballmer: "Developers! Developers! Developers!" :) Instead of loosing time you could just focus on building some software. Imagine D would have a kick ass IDE like visual studio but cross platform like QT. Productivity, efficiency and performance in one. Too bad D doesn't have the option for paid developers to create something like this. It's just too much work for your spare time. So, it's unlikely to happen any time soon :( Still, I very much appreciate people like Rikki Cattermole who builds a base to work of for others. The problem is for a few individuals that just takes a long time to complete. I would very much appreciate people helping out. E.g. Contributing manipulation functions, image loader / exporter. Most importantly unittests. Fix bugs. I'll most likely do the PNG test case infrastructure, but it would make things go a few months faster if somebody took ownership of testing in general. I'll also be creating a TODO list for what needs to happen with Phobos. E.g. @nogc version of std.string:indexOf. So plenty of places anybody can jump in and help :)
Re: Current state of GUI's
On Saturday, 5 September 2015 at 19:13:12 UTC, Jacob Carlborg wrote: It links now when I manged to install all the dependencies, but it doesn't run :(. Segfaults for some reason, I haven't looked in to it. Bummer... However, I appreciate that you are giving it a shot! But you see my point. An out of the box solution like QT would be worth gold. That's the one thing Microsoft got right. They have awesome developer tools. Or in the words of Steve Ballmer: "Developers! Developers! Developers!" :) Instead of loosing time you could just focus on building some software. Imagine D would have a kick ass IDE like visual studio but cross platform like QT. Productivity, efficiency and performance in one. Too bad D doesn't have the option for paid developers to create something like this. It's just too much work for your spare time. So, it's unlikely to happen any time soon :( Still, I very much appreciate people like Rikki Cattermole who builds a base to work of for others. The problem is for a few individuals that just takes a long time to complete.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 15:15:47 UTC, motaito wrote: Thanks for the details! I was afraid, of something like this. To me D doesn't seam very competitive without a complete toolset :( Given that D has been around for a while now I am somewhat surprised it's not further along. I know it's a lot of work and not easy to do but clearly there has been a need for this for many years now. The resources I found were several years old (not including that they must have been working on it for a while before abandoning the project). It will be hard to compete with something like QT. I was hoping for a cross platform solution that would deliver good performance without the complexity of c/c++. But if there is no built in solution QT seams a better option. I wanted to avoid QT because of the added complexity and the QT-windows tend to lag on a windows platform. This may go away with updates and bugfixes though. Anyway, I hope you will be successful with alphaPhobos. I think D needs it, so don't abandon it! I will keep an eye out and check again in about a year or so. But for now I have to rethink whether to use D or not. Out of curiosity, why do you think that I wouldn't want a Phobos solution? Wouldn't an integrated solution be better? I would not have to maintain a build of an external third party project, that may or may not be abandoned in the future, not knowing whether there will be updates, bugfixes or missing features to be added. It seams like quite a bit extra management for my project if I have to keep a third party project up to date and may need to adjust my code due to an update in an external project. Don't give up yet. I've learned that a D GUI is not so important. There is a plethora of platforms and devices (mobile phones, tablets, PCs etc etc), so it's better not to get married to one particular UI. Keep your program GUI-agnostic and then you can connect it to any GUI you want. And D is easy to connect to: C(++), Lua (LuaD) and Python (PyD). With D you can concentrate on the program, the GUI could be anything. E.g. look at how the DCD/Dscanner plug-in works with Textadept and other editors. https://github.com/Hackerpilot/DCD/wiki/IDEs-and-Editors-with-DCD-support
Re: Current state of GUI's
On Friday, 4 September 2015 at 09:12:23 UTC, Chris wrote: Don't give up yet. I've learned that a D GUI is not so important. There is a plethora of platforms and devices (mobile phones, tablets, PCs etc etc), so it's better not to get married to one particular UI. Keep your program GUI-agnostic and then you can connect it to any GUI you want. And D is easy to connect to: C(++), Lua (LuaD) and Python (PyD). With D you can concentrate on the program, the GUI could be anything. E.g. look at how the DCD/Dscanner plug-in works with Textadept and other editors. https://github.com/Hackerpilot/DCD/wiki/IDEs-and-Editors-with-DCD-support At first I thought, yes your right. But on second thought that kind of brings me back to the initial question. I would prefer an integrated solution, so I don't have the hustle with dependencies. With your approach, I also have to maintain a dependency. E.g. if I choose PyD I also need to distribute it to the end user. I am not even sure, if I would also need to have python as separate dependency or if it's included in PyD. Either way while python may be common on Linux, it is not a standard on Windows. I have the same problem with something like GtkD. The Gtk Framework is common on Linux but not on windows. I don't have such issues with QT. There I have all in one package. All I have to do is include the required dll's. QT is also very common in many Linux distributions. So, I can just compile and give it to anyone who is interested without the hustle of more dependencies. BSD is certainly the most difficult to find a solution. Such a great OS so little support. But they too have QT. I was hoping to basically use QT with D instead of C++. Shannon made a post with an interesting approach. I think I will look more into that first. But thanks for the tip!
Re: Current state of GUI's
On 2015-09-04 15:23, motaito wrote: BSD is certainly the most difficult to find a solution. I tried compiling DWT on PC-BSD and it compile just if I remove one check for Linux. Although it doesn't link because I haven't managed to install the necessary libraries yet. -- /Jacob Carlborg
Re: Current state of GUI's
On Friday, 4 September 2015 at 13:23:47 UTC, motaito wrote: On Friday, 4 September 2015 at 09:12:23 UTC, Chris wrote: Don't give up yet. I've learned that a D GUI is not so important. There is a plethora of platforms and devices (mobile phones, tablets, PCs etc etc), so it's better not to get married to one particular UI. Keep your program GUI-agnostic and then you can connect it to any GUI you want. And D is easy to connect to: C(++), Lua (LuaD) and Python (PyD). With D you can concentrate on the program, the GUI could be anything. E.g. look at how the DCD/Dscanner plug-in works with Textadept and other editors. https://github.com/Hackerpilot/DCD/wiki/IDEs-and-Editors-with-DCD-support At first I thought, yes your right. But on second thought that kind of brings me back to the initial question. I would prefer an integrated solution, so I don't have the hustle with dependencies. With your approach, I also have to maintain a dependency. E.g. if I choose PyD I also need to distribute it to the end user. I am not even sure, if I would also need to have python as separate dependency or if it's included in PyD. Either way while python may be common on Linux, it is not a standard on Windows. I have the same problem with something like GtkD. The Gtk Framework is common on Linux but not on windows. I don't have such issues with QT. There I have all in one package. All I have to do is include the required dll's. QT is also very common in many Linux distributions. So, I can just compile and give it to anyone who is interested without the hustle of more dependencies. BSD is certainly the most difficult to find a solution. Such a great OS so little support. But they too have QT. I was hoping to basically use QT with D instead of C++. Shannon made a post with an interesting approach. I think I will look more into that first. But thanks for the tip! I see what you mean. As far as I know, the libraries (PyD, LuaD etc.) would be in the static build of the executable, so you wouldn't have to "distribute" them, only Python or Lua. But that's no problem either, just include lua5.x.so/dll. Please guys, correct me, if I'm wrong here, I don't wanna give motaito any wrong information. This said, the GUI should only be a very very thin layer on top of your actual program, regardless. What if your users say that they want a different GUI (Mac users for example, they don't accept anything that is a bit different), or the D-GUI doesn't work on a given OS for some reason, or a new GUI is introduced with a new OS. That's why I said that your program should never get married to a specific GUI.
Re: Current state of GUI's
On Friday, 4 September 2015 at 00:11:02 UTC, Nick Sabalausky wrote: On 09/03/2015 01:51 PM, motaito wrote: Yeah, it's a shame that the mono framework does not support WPF. I think WPF and MEF are the two most powerful features from .NET (for me anyway, that may be subjective). And working in visual studio is really nice too. I wish there was a cross platform IDE like that. I don't quite like the text editor from eclipse, but I haven't used it in a while. Maybe I need to give it another try. I thought Mono did support that. Like in GitExtentions which runs fine on Linux (well, aside from frequent crashing). Or is WPF different from "win forms"? Afaik, WPF is not supported nor are there any plans to do so. http://www.mono-project.com/docs/gui/wpf/ MEF may be supported, I haven't found any info. It's possible that it will be supported with with the .NET 5 framework coming up.
Re: Current state of GUI's
On Friday, 4 September 2015 at 00:11:02 UTC, Nick Sabalausky wrote: On 09/03/2015 01:51 PM, motaito wrote: Yeah, it's a shame that the mono framework does not support WPF. I thought Mono did support that. Like in GitExtentions which runs fine on Linux (well, aside from frequent crashing). Or is WPF different from "win forms"? Yes, it's totally different. WinForms is based on good old WinAPI with classical windows, dialogs, controls and GDI for drawing. WPF is based on a new engine over DirectX with lots of vector graphics, transformations and effects, everything GPU-accelerated. They remade everything - windows, controls and layouts are described declaratively in XAML format (XML-based), then everything is rendered with that new engine, even text is drawn differently (and in first versions looked differently). Mono never had and probably will never have WPF because this DirectX-based rendering engine is too hard to port.
Re: Current state of GUI's
On Friday, 4 September 2015 at 03:36:37 UTC, Shannon wrote: Anyway, though I have so enjoyed the huge part of Qt that is non-gui, and especially their model/view components where I can do all my model manipulation behind the scenes and practically hot-swap or connect multiple views to a model. I have rarely a need for non-gui apps myself. Maybe I have been looking at D with wrong expectations to begin with. QT has some nifty tools :) I have to admit that I became lazi with gui modeling, after having spent lots of time using c#. With WPF you practically get sorting and filtering of tables for free. I have come to really like the MVVM pattern with WPF. It is very convenient. And I've found a novel way to relatively easily hook up thin Qt graphical interfaces on top of the commandline app that interacts with JSON via stdin/stdout as it does on servers. That sound very interesting! So, best of both worlds. The D implementation is generally easier/more enjoyable for the actual guts of the tool, and I get a perfectly slick Qt user interface. That's exactly was I was hoping for with D. Maybe I should try to give your approach a try. sorry, this is already too long. I mean to open up and example repo/wiki with that pattern. RSN (Real Soon Now) If you have the time, please to so!
Re: Current state of GUI's
On 04/09/15 1:21 AM, motaito wrote: On Thursday, 3 September 2015 at 12:51:18 UTC, CraigDillabaugh wrote: Not sure of its current state, but you should check out DOtherSide: https://github.com/filcuc/DOtherSide Thanks for the link. I will take a closer look at it. But does this also mean, that there is no "built in" GUI solution for D? Short answer: Don't expect anything in Phobos short term. In the mean time you'll probably want a c/c++ library such as QT. Long answer: We're probably about 2 years away before anything useful and even then it will have a huge push back unless e.g. Andrei says we must have a GUI toolkit! We've got money to make it happen. That's if my work gets into Phobos when I am planning it to (unlikely). I know its grim, but keep in mind we need a platform abstraction that supports things like an event loop, window creation independent of what the system actually supports. An image abstraction with common manipulation/formats. Then on top of a platform abstraction some people also need other forms of event handling and the list goes on. But I am actively working on an image library and a window + context creation library for Phobos. It's just a year away. https://github.com/rikkimax/alphaPhobos Not to scare you aware or anything, but you probably don't want a pure Phobos solution anyway.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 14:35:44 UTC, Jacob Carlborg wrote: On 2015-09-03 14:19, motaito wrote: I am looking for a cross platform (Windows, Linux, BSD) solution. Others have already mentioned DWT. It officially only support Windows and Linux. It might be easy to get it to work on BSD since the platform is similar to Linux, assuming GTK is used. Thanks for the tip. It's always hard to find out of the box support for BSD. That request was probably a bit of a stretch :)
Re: Current state of GUI's
On 04/09/15 3:15 AM, motaito wrote: Anyway, I hope you will be successful with alphaPhobos. I think D needs it, so don't abandon it! I will keep an eye out and check again in about a year or so. But for now I have to rethink whether to use D or not. Thanks, this is a long term thing. I've been planning this for ~2 years. Previous was Devisualization (Github org of mine). So a bunch of code has actually originated here. Before that DOOGLE. Even some of that came from OOGL (ported it from c++ to D as D_OOGL). Out of curiosity, why do you think that I wouldn't want a Phobos solution? Wouldn't an integrated solution be better? I would not have to maintain a build of an external third party project, that may or may not be abandoned in the future, not knowing whether there will be updates, bugfixes or missing features to be added. It seams like quite a bit extra management for my project if I have to keep a third party project up to date and may need to adjust my code due to an update in an external project. Because, at best case the graphics could be very ugly and lacking of features in a GUI toolkit. We're just too far away from it. We need to focus on getting the core problems solved. So that people can flourish with building e.g. GUI toolkits unimpeded by what we support. Also helps interop between them. After we have a decent one, then we can consider it for Phobos. I know the current situation is unfortunate. We've had some pretty major changes in the last few years and that is ok. I hope we do get to a point where it is easy to get started in almost any problem domain :)
Re: Current state of GUI's
On Thursday, 3 September 2015 at 13:21:45 UTC, motaito wrote: On Thursday, 3 September 2015 at 12:51:18 UTC, CraigDillabaugh wrote: Not sure of its current state, but you should check out DOtherSide: https://github.com/filcuc/DOtherSide Thanks for the link. I will take a closer look at it. But does this also mean, that there is no "built in" GUI solution for D? No there is no standard D GUI. There are lots of efforts by different community members like DWT (https://github.com/d-widget-toolkit/dwt). I pointed you to DOtherSide because of all the GUI's I've heard of it seemed the closest to what you are looking for. I think DWT is the most complete of the D GUI libraries available though. I think part of the issues is that there is no strong consensus in the community as to what form a D GUI would take, through there is certainly a lot of discussion.
Re: Current state of GUI's
On 2015-09-03 14:19, motaito wrote: I am looking for a cross platform (Windows, Linux, BSD) solution. Others have already mentioned DWT. It officially only support Windows and Linux. It might be easy to get it to work on BSD since the platform is similar to Linux, assuming GTK is used. -- /Jacob Carlborg
Re: Current state of GUI's
On Thursday, 3 September 2015 at 14:16:19 UTC, CraigDillabaugh wrote: No there is no standard D GUI. There are lots of efforts by different community members like DWT (https://github.com/d-widget-toolkit/dwt). I pointed you to DOtherSide because of all the GUI's I've heard of it seemed the closest to what you are looking for. I think DWT is the most complete of the D GUI libraries available though. I think part of the issues is that there is no strong consensus in the community as to what form a D GUI would take, through there is certainly a lot of discussion. Thanks for the input! I was hoping for a standard D Gui though. I guess I have to give it a shot, before making a final decision.
Re: Current state of GUI's
On 04/09/15 3:37 AM, motaito wrote: On Thursday, 3 September 2015 at 15:26:45 UTC, Rikki Cattermole wrote: On 04/09/15 3:15 AM, motaito wrote: Anyway, I hope you will be successful with alphaPhobos. I think D needs it, so don't abandon it! I will keep an eye out and check again in about a year or so. But for now I have to rethink whether to use D or not. Thanks, this is a long term thing. I've been planning this for ~2 years. Previous was Devisualization (Github org of mine). So a bunch of code has actually originated here. Before that DOOGLE. Even some of that came from OOGL (ported it from c++ to D as D_OOGL). Out of curiosity, why do you think that I wouldn't want a Phobos solution? Wouldn't an integrated solution be better? I would not have to maintain a build of an external third party project, that may or may not be abandoned in the future, not knowing whether there will be updates, bugfixes or missing features to be added. It seams like quite a bit extra management for my project if I have to keep a third party project up to date and may need to adjust my code due to an update in an external project. Because, at best case the graphics could be very ugly and lacking of features in a GUI toolkit. We're just too far away from it. We need to focus on getting the core problems solved. So that people can flourish with building e.g. GUI toolkits unimpeded by what we support. Also helps interop between them. After we have a decent one, then we can consider it for Phobos. I know the current situation is unfortunate. We've had some pretty major changes in the last few years and that is ok. I hope we do get to a point where it is easy to get started in almost any problem domain :) Again Thanks for the details. It's very much appreciated. From a usability stand point for alphaPhobos. Are you more going into the direction of WPF or QML, using a declaration file or more along the lines of a QWidget, building the GUI through instantiating objects inside a D-module? Neither, I'm not touching GUI's itself. Only the creation of a window and a context. GUI's are far too controversial and for good reason. They are next to impossible to do properly. But that is not a popular opinion :p Here is more or less what I'm planning on supporting: https://github.com/rikkimax/alphaPhobos/blob/master/source/std/experimental/ui/window/defs.d There is a bunch of optional features you will notice in other files. Such as a menu, notifications and even screenshots.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 14:13:26 UTC, Rikki Cattermole wrote: Short answer: Don't expect anything in Phobos short term. In the mean time you'll probably want a c/c++ library such as QT. Long answer: We're probably about 2 years away before anything useful and even then it will have a huge push back unless e.g. Andrei says we must have a GUI toolkit! We've got money to make it happen. That's if my work gets into Phobos when I am planning it to (unlikely). I know its grim, but keep in mind we need a platform abstraction that supports things like an event loop, window creation independent of what the system actually supports. An image abstraction with common manipulation/formats. Then on top of a platform abstraction some people also need other forms of event handling and the list goes on. But I am actively working on an image library and a window + context creation library for Phobos. It's just a year away. https://github.com/rikkimax/alphaPhobos Not to scare you aware or anything, but you probably don't want a pure Phobos solution anyway. Thanks for the details! I was afraid, of something like this. To me D doesn't seam very competitive without a complete toolset :( Given that D has been around for a while now I am somewhat surprised it's not further along. I know it's a lot of work and not easy to do but clearly there has been a need for this for many years now. The resources I found were several years old (not including that they must have been working on it for a while before abandoning the project). It will be hard to compete with something like QT. I was hoping for a cross platform solution that would deliver good performance without the complexity of c/c++. But if there is no built in solution QT seams a better option. I wanted to avoid QT because of the added complexity and the QT-windows tend to lag on a windows platform. This may go away with updates and bugfixes though. Anyway, I hope you will be successful with alphaPhobos. I think D needs it, so don't abandon it! I will keep an eye out and check again in about a year or so. But for now I have to rethink whether to use D or not. Out of curiosity, why do you think that I wouldn't want a Phobos solution? Wouldn't an integrated solution be better? I would not have to maintain a build of an external third party project, that may or may not be abandoned in the future, not knowing whether there will be updates, bugfixes or missing features to be added. It seams like quite a bit extra management for my project if I have to keep a third party project up to date and may need to adjust my code due to an update in an external project.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 15:26:45 UTC, Rikki Cattermole wrote: On 04/09/15 3:15 AM, motaito wrote: Anyway, I hope you will be successful with alphaPhobos. I think D needs it, so don't abandon it! I will keep an eye out and check again in about a year or so. But for now I have to rethink whether to use D or not. Thanks, this is a long term thing. I've been planning this for ~2 years. Previous was Devisualization (Github org of mine). So a bunch of code has actually originated here. Before that DOOGLE. Even some of that came from OOGL (ported it from c++ to D as D_OOGL). Out of curiosity, why do you think that I wouldn't want a Phobos solution? Wouldn't an integrated solution be better? I would not have to maintain a build of an external third party project, that may or may not be abandoned in the future, not knowing whether there will be updates, bugfixes or missing features to be added. It seams like quite a bit extra management for my project if I have to keep a third party project up to date and may need to adjust my code due to an update in an external project. Because, at best case the graphics could be very ugly and lacking of features in a GUI toolkit. We're just too far away from it. We need to focus on getting the core problems solved. So that people can flourish with building e.g. GUI toolkits unimpeded by what we support. Also helps interop between them. After we have a decent one, then we can consider it for Phobos. I know the current situation is unfortunate. We've had some pretty major changes in the last few years and that is ok. I hope we do get to a point where it is easy to get started in almost any problem domain :) Again Thanks for the details. It's very much appreciated. From a usability stand point for alphaPhobos. Are you more going into the direction of WPF or QML, using a declaration file or more along the lines of a QWidget, building the GUI through instantiating objects inside a D-module?
Re: Current state of GUI's
On Thursday, 3 September 2015 at 12:19:25 UTC, motaito wrote: Hi, I haven't used D before and wanted to take a closer look at it. However, there are a few things that keep me from doing so. Most notably I am somewhat confused about the state of GUI's. I have looked at a couple projects and online resources, but the resources I found seam to be quite outdated. Sorry in advance, if I have missed some update on the topic. I wanted to ask, if there is something in the workings (or already done) to provide a standard way to create GUI's. Something supported by D, not an independent project. Independent projects tend to scare me away, as they almost always get abandoned at some point. I am looking for a cross platform (Windows, Linux, BSD) solution. [...] Have a look at GtkD and DWT - they are the most complete. I've written small apps using both and they work :-) With GtkD you can use the Glade Builder to simplify design of the GUI elements.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 14:24:29 UTC, Mike James wrote: Have a look at GtkD and DWT - they are the most complete. I've written small apps using both and they work :-) With GtkD you can use the Glade Builder to simplify design of the GUI elements. Thanks for the reply. A graphical tool to build the GUI might come in handy :) I will remember it, in case I decide to use D. Right now I am on the fence... (I pointed this already out in an other reply to rikki, but I will like wait before using D).
Re: Current state of GUI's
On Thursday, 3 September 2015 at 15:41:34 UTC, Rikki Cattermole wrote: Neither, I'm not touching GUI's itself. Only the creation of a window and a context. GUI's are far too controversial and for good reason. They are next to impossible to do properly. But that is not a popular opinion :p Here is more or less what I'm planning on supporting: https://github.com/rikkimax/alphaPhobos/blob/master/source/std/experimental/ui/window/defs.d There is a bunch of optional features you will notice in other files. Such as a menu, notifications and even screenshots. I see, more like a unified groundwork for someone to build a GUI on top. Looks like an interesting project. Good luck with alphaPhobos! It will be interesting to see it move forward.
Re: Current state of GUI's
On 9/3/2015 10:51 AM, motaito wrote: And working in visual studio is really nice too. I wish there was a cross platform IDE like that. I don't quite like the text editor from eclipse, but I haven't used it in a while. Maybe I need to give it another try. Just in case you hadn't seen it, there is now Visual Studio Code: https://code.visualstudio.com/ Although I haven't really looked at it myself. I don't think it can accept plug-ins (yet) such as VisualD so that may rule it out.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 18:12:38 UTC, Jim Hewes wrote: On 9/3/2015 10:51 AM, motaito wrote: And working in visual studio is really nice too. I wish there was a cross platform IDE like that. I don't quite like the text editor from eclipse, but I haven't used it in a while. Maybe I need to give it another try. Just in case you hadn't seen it, there is now Visual Studio Code: https://code.visualstudio.com/ Although I haven't really looked at it myself. I don't think it can accept plug-ins (yet) such as VisualD so that may rule it out. I forgot about visual studio code. Right now I am trying to move away from windows products and only support the windows platform for cross platform reasons. It's a shame though, I used visual studio for years and it's my favorite IDE. However, I have a moral issue with the data collection philosophy that windows is currently trying to enforce. Just like visual studio code, among the first things you have to agree to is the terms and privacy policy, which will try to collect as much data as they can. I think something like visual studio code is their way to get data from other platforms as well... Otherwise it would look interesting and I would certainly give it a shot but I try to support this behavior as little as possible. Thanks for the tip though!
Re: Current state of GUI's
On Thursday, 3 September 2015 at 16:36:31 UTC, thedeemon wrote: On Thursday, 3 September 2015 at 16:31:33 UTC, Zekereth wrote: https://github.com/buggins/dlangui is fairly complete and contains DML - DlangUI Markup Language - similar to QML. +1, came here to mention dlangui. It already works and contains a lot of good stuff, including layout engine and markup language. @Zekereth and @thedeemon I guess I will stick to QT for my current project, but I will build some smaller learning projects in D, to try out various GUI libraries. DlangUI seams promising. I probably will either use DlangUI or GtkD.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 16:18:34 UTC, motaito wrote: On Thursday, 3 September 2015 at 15:41:34 UTC, Rikki Cattermole wrote: Neither, I'm not touching GUI's itself. Only the creation of a window and a context. GUI's are far too controversial and for good reason. They are next to impossible to do properly. But that is not a popular opinion :p Here is more or less what I'm planning on supporting: https://github.com/rikkimax/alphaPhobos/blob/master/source/std/experimental/ui/window/defs.d There is a bunch of optional features you will notice in other files. Such as a menu, notifications and even screenshots. I see, more like a unified groundwork for someone to build a GUI on top. Looks like an interesting project. Good luck with alphaPhobos! It will be interesting to see it move forward. https://github.com/buggins/dlangui is fairly complete and contains DML - DlangUI Markup Language - similar to QML.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 16:40:56 UTC, Jim Hewes wrote: I'm in a similar place as you are. While D is better than C++, right now C++ plus JUCE is better for me overall than D plus . C# plus WPF seems attractive too except that WPF isn't cross platform. Anyway, I still always keep an eye on D. Yeah, it's a shame that the mono framework does not support WPF. I think WPF and MEF are the two most powerful features from .NET (for me anyway, that may be subjective). And working in visual studio is really nice too. I wish there was a cross platform IDE like that. I don't quite like the text editor from eclipse, but I haven't used it in a while. Maybe I need to give it another try.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 16:31:33 UTC, Zekereth wrote: https://github.com/buggins/dlangui is fairly complete and contains DML - DlangUI Markup Language - similar to QML. +1, came here to mention dlangui. It already works and contains a lot of good stuff, including layout engine and markup language.
Re: Current state of GUI's
On 9/3/2015 5:19 AM, motaito wrote: Hi, I haven't used D before and wanted to take a closer look at it. However, there are a few things that keep me from doing so. I'm in a similar place as you are. While D is better than C++, right now C++ plus JUCE is better for me overall than D plus . C# plus WPF seems attractive too except that WPF isn't cross platform. Anyway, I still always keep an eye on D.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 17:58:21 UTC, motaito wrote: @Zekereth and @thedeemon I guess I will stick to QT for my current project, but I will build some smaller learning projects in D, to try out various GUI libraries. DlangUI seams promising. I probably will either use DlangUI or GtkD. Hopefully you will help by creating documentation (writing a blog post about your adventure would be a valuable exercise) and filing bug reports/feature requests. Even if you can't use any of them now you can push the process along.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 19:59:24 UTC, bachmeier wrote: On Thursday, 3 September 2015 at 17:58:21 UTC, motaito wrote: @Zekereth and @thedeemon I guess I will stick to QT for my current project, but I will build some smaller learning projects in D, to try out various GUI libraries. DlangUI seams promising. I probably will either use DlangUI or GtkD. Hopefully you will help by creating documentation (writing a blog post about your adventure would be a valuable exercise) and filing bug reports/feature requests. Even if you can't use any of them now you can push the process along. I can't make any promises. It will depend on how busy I will be. But I keep it in mind. If I write something, you could find it on www.motaito.com. I have only a few things there (installing arch and gentoo for beginners). I could make an additional post in this forum (provide a link), if I find the time to do so. Right now I would say it's somewhat unlikely, but if the interest is there, I can try to find the time.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 12:19:25 UTC, motaito wrote: ... Please see: http://wiki.dlang.org/GUI_Libraries
Current state of GUI's
Hi,
I haven't used D before and wanted to take a closer look at it.
However, there are a few things that keep me from doing so. Most
notably I am somewhat confused about the state of GUI's. I have
looked at a couple projects and online resources, but the
resources I found seam to be quite outdated. Sorry in advance, if
I have missed some update on the topic. I wanted to ask, if there
is something in the workings (or already done) to provide a
standard way to create GUI's. Something supported by D, not an
independent project. Independent projects tend to scare me away,
as they almost always get abandoned at some point. I am looking
for a cross platform (Windows, Linux, BSD) solution.
I very much enjoyed working with QT before, especially QML and
would like to use something, that essentially replaces C++ with D
for QML. I only found QtD which also seams to be abandoned. Or
(maybe even preferable) something similar to .NET and WPF as this
makes it very easy and flexible to create modern GUI's. I would
like to install a D environment and get working without having to
build other projects for GUI's.
e.g.
import std.GUI;
void main(string[] args)
{
auto win = loadGUIFromXML("Path/to/gui.xml");
win.show();
}
The XML could then have bindings to D-Modules to handle events
and perform the actual tasks.
e.g.
// some component to be loaded into the window above
orientation="LeftToRight">
Does D offer something like this by now? What is the current
state/standard for GUI creation?
Re: Current state of GUI's
On 09/03/2015 01:51 PM, motaito wrote: Yeah, it's a shame that the mono framework does not support WPF. I think WPF and MEF are the two most powerful features from .NET (for me anyway, that may be subjective). And working in visual studio is really nice too. I wish there was a cross platform IDE like that. I don't quite like the text editor from eclipse, but I haven't used it in a while. Maybe I need to give it another try. I thought Mono did support that. Like in GitExtentions which runs fine on Linux (well, aside from frequent crashing). Or is WPF different from "win forms"?
Re: Current state of GUI's
On Thursday, 3 September 2015 at 15:15:47 UTC, motaito wrote: Thanks for the details! I was afraid, of something like this. To me D doesn't seam very competitive without a complete toolset :( Given that D has been around for a while now I am somewhat surprised it's not further along. I know it's a lot of work and not easy to do but clearly there has been a need for this for many years now. The resources I found were several years old (not including that they must have been working on it for a while before abandoning the project). It will be hard to compete with something like QT. I was hoping for a cross platform solution that would deliver good performance without the complexity of c/c++. But if there is no built in solution QT seams a better option. I wanted to avoid QT because of the added complexity and the QT-windows tend to lag on a windows platform. This may go away with updates and bugfixes though. With D's relatively easy ability to link C libs, gtkD is the way I would've gone, IF I wasn't such a satisfied and long time user of Qt across BeOS-OS/2-BSD-Windows-OSX-Android-iOS-embedded Linux (did I miss any?). The monumental amount of continual work involved in keeping a consistent professional interface blended into the "major" operating systems and adjusted to make the native user feel at home seems folly to attempt - though I do keep my fingers crossed for dlangui for a gui lib done the "D" way. Anyway, though I have so enjoyed the huge part of Qt that is non-gui, and especially their model/view components where I can do all my model manipulation behind the scenes and practically hot-swap or connect multiple views to a model. I've spent much of the last year building commandline applications with D running on the desktop and behind production web applications on aws servers. And I've found a novel way to relatively easily hook up thin Qt graphical interfaces on top of the commandline app that interacts with JSON via stdin/stdout as it does on servers. In the Qt graphical interface I put, for example, a QTableView which I connect to a thin class that inherits QAbstractTableModel and implements four methods, three of which are 1 liners. The other method simply connects to stdin and converts the JSON that it receives, from the D commandline app, into data for each (row,column) that the View calls for. I actually embed the D executable into the Qt application, then extract and run it when the Qt app starts up. I'm always amazed how un-noticably fast that is. So, best of both worlds. The D implementation is generally easier/more enjoyable for the actual guts of the tool, and I get a perfectly slick Qt user interface. sorry, this is already too long. I mean to open up and example repo/wiki with that pattern. RSN (Real Soon Now) good luck.
Re: Current state of GUI's
On Friday, 4 September 2015 at 03:36:37 UTC, Shannon wrote: On Thursday, 3 September 2015 at 15:15:47 UTC, motaito wrote: oh, DOtherSide was mentioned. the first time I looked at it. it implements QAbstractListModel .. very familiar. I should study that some more.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 14:13:26 UTC, Rikki Cattermole wrote: Short answer: Don't expect anything in Phobos short term. In the mean time you'll probably want a c/c++ library such as QT. I really don't know much about GUI programming. I was looking at some basic info about GTK+, which led me to GObject, which led me to the Vala language. My understanding is that Vala is basically an OO version of C (i.e. it works well with GObject without a lot of the complexity of writing GObject code). People seem to be using it with GNOME. Maybe to your point about not having a pure D solution, it might be useful to investigate this as well.
Re: Current state of GUI's
On Thursday, 3 September 2015 at 12:51:18 UTC, CraigDillabaugh wrote: Not sure of its current state, but you should check out DOtherSide: https://github.com/filcuc/DOtherSide Thanks for the link. I will take a closer look at it. But does this also mean, that there is no "built in" GUI solution for D?
Re: Current state of GUI's
On Thursday, 3 September 2015 at 12:19:25 UTC, motaito wrote: Hi, I haven't used D before and wanted to take a closer look at it. However, there are a few things that keep me from doing so. Most notably I am somewhat confused about the state of GUI's. I have looked at a couple projects and online resources, but the resources I found seam to be quite outdated. Sorry in advance, if I have missed some update on the topic. I wanted to ask, if there is something in the workings (or already done) to provide a standard way to create GUI's. Something supported by D, not an independent project. Independent projects tend to scare me away, as they almost always get abandoned at some point. I am looking for a cross platform (Windows, Linux, BSD) solution. [...] Not sure of its current state, but you should check out DOtherSide: https://github.com/filcuc/DOtherSide
