For those who have not followed the developments of Microsoft Bridges announced at Build 2015, Project Centennial enables turning Win32 apps into Appx apps that can natively run in Windows 8.x/10, and most importantly, go into the Windows Store.
John Sheehan who presented it last year is back with more news and a demo with a working app converter :
Basically, the Centennial Bridge Converter works a bit like Wine bottles on Mac : it runs the Win32 installer (whichever one, “including a batch file”), and creates a container into which the Desktop app finds all its staples such as the Registry, and stuff typical of desktop like AppData. So for all intents and purposes, the (Xojo) Win32 app finds all its familiar environment. But to Windows 10, it is a New API UWP app. Very similar to a Mac Wine bottle, the app still has access to Documents, Pictures, and other user folders.
They had a booth at the Build conference where they were converting Win32 apps to appx for developers to see if it worked with their app. Which means it actually works now.
Centennial converted apps are not going to the Windows Store until later, when the team that manages the store has figured a way to onboard the apps. So for the moment it looks as if access to the Windows store for Xojo apps is still some ways away.
Centennial Bridge appx will not work on current version of Windows 10, but on what they call the anniversary edition, which will contain what they call the “bits”, which are necessary for Centennial converted apps.
Also, Win32 apps converted to UWA will be able to access the new API, which means for instance being able to create live tiles or call contracts like system dialogs (print, files, etc.) from the Win32 app.
Microsoft announced that they are going to make the bridge “more mainstream”. See here
This would make it easier to sell Xojo apps on the MS store.
Flexera already has a feature to create directly AppX, but it is only with the top of the line, $4900 or so installer maker.
I hope the new scheme will be easier.
Given the number of developers here who do well in the MAS, it would be nice to port the apps to the Windows Store.
I am actively working on a WS version of Check Writer. Most of the work consist in adapting it to tablet mode, with the virtual keyboard and 1360 x 768 screen.
The most interesting seems to be Advanced Installer. Their Professional edition at $399 apparently can directly produce an AppX, which alleviates the need for a second step with the converter. I downloaded the evaluation version. If I understand right, it will support AppX generation during the trial period.
Compared to the $4999 of Flexera’s InstallShield, that is much more affordable.
I plan on experimenting with the converter this week end for my first app. I just downloaded it from the Windows Store.
Fwiw, I’ve had an Advanced Installer license for the last 10 years and it’s been awesome. They’re very responsive to bug reports, keep their product very current and make it very easy to make good installers.
Excellent. I have downloaded the evaluation, and will try to play with it this week end. Sure, $399 is not quite trivial, but if sales in the Windows Store have any similarities with the MAS, it will break even in less than a month.
It’s been years since I’ve released anything public for Windows. I’d really appreciate you keeping us updated on your journey with this, Michel, as it could open up a new market for us more Apple-centric devs.
Getting a new outlet is exactly the idea. And I am sure there are opportunities for programs that already do well in the MAS. For anybody who is familiar with the MAS, the Windows Store is very easy to use.
Which is not exactly the case for the Desktop App Converter. I naively thought after downloading it from the Windows Store, it was ready. How delusional was I.
Now, after several hours, I hope I am OK to resume work tomorrow and try my first app conversion.
I will not start tonight, I would probably loose sleep.
All in all, though, it is not immensely more difficult than my first Mac app in the MAS. Just the same kind of tedious dealing with unfamiliar environment, a lot of cryptic command lines.
If I am successful, I will then proceed to upload to the Windows Store, and see how it goes.
I spent yesterday the whole afternoon and stayed on it until after midnight, trying to follow all the steps described at https://msdn.microsoft.com/en-us/windows/uwp/porting/desktop-to-uwp-run-desktop-app-converter to convert a silent installer of my latest app.
I still get nowhere, as the converter complains that it does not have a base image to match the host system. I have asked questions in the insider forums, hope to have an answer soon.
Base image is a separate download, you need one that match the DAC and the Windows you running on. If you using the Windows Store version then you need Window 10 Anniversary v14393 (the host) and base image 14393.
Desktop App Converter Base Images
That is what I found out the hard way. I can no longer be part of the Insider Preview because of that. I am in the process of reinstalling the regular Windows 10 so it will update to Anniversary Update.
Well ; I am keeping a log, and when I finally succeed converting an app, I will post here so others can benefit from the experience.
The most interesting spec I found was that some things did not work if I kept European time. After changing time zone to US Eastern time, it worked.
For testing, I have a Mac mini with 2 HDs uncluttered. The 1st HD for the latest official Windows 10 and the 2nd HD for the Insider Preview. Windows automatically create the Boot Manager menu to let me choose which to boot. This way I still stay with the Insider Preview.
I have another Mac mini set up similarly for macOS, although I seldom use it.
My work computer used to be Windows 7 (this is very cluttered) but just few days ago I updated it to Windows 10. I want to go into true UWA so this force me to have the tools right in front of me whenever I sit at the work computer. I have set up an older 2009 Mac mini with Windows 7 for whenever I need to test.
Here is a good guide along the way:
The path from a desktop app to a Universal Windows Platform (UWP) app
To make it into “a full Universal Windows Platform (UWP) app that can run on all Windows 10 devicessuch as Xbox One, HoloLens, phones, et cetera”, need to complete step 5 = EXPAND. Otherwise converted app will still be only a Desktop app that can only run on Wintel Windows. Step 2 to 4 = Enhance Extend Migrate requires the ability within Xojo to call the UWP APIs. I don’t know how to do this, or if it is possible.
Thank you. The three links I just posted show in detail what can be done. Yet, after downgrading my PC to 14393 I could not get the Desktop App Converter installed from the Windows Store as before. This extremely frustrating.
I tried to install the zip version and got stuck as well.
Anyway, have you already succeeded converting a Xojo app ?
Actually, I have stopped exploring further a while back after I realized that the only advantage of going through the DAC is to make the app discoverable in the Windows Store. App-wise, they are technically the same app to users. But converted desktop apps can only run on Windows 10 whereas unconverted desktop apps run on all Windows where supported.
I see the DAC as a scheme by MS to “quietly” move desktop developers to UWP, by providing a means to initially move their existing desktop app into the Windows Store. If you delve into more of their documents, you will see that the converted desktop app eventually need to be completely re-written for UWP.
With Windows 8/8.1, when they sort of try to abandon desktop apps and embrace Metro apps, they failed miserable when the Windows 8 Store was as good as empty. Now they are trying to correct their mistake, also having learned from their mistake of the half-baked VB6-to-VB.NET converter that never worked.
From what I see, they will make DAC work then they will not go much further. I see it as a dead end. So I decide to just go straight into real UWP and leave my existing desktop apps as desktop apps that will run on all Windows back to Windows XP SP2 because of side-by-side manifest implementation. I do not know whether some of the things I do qualify the apps for conversion.
The problem I see at present is that UWP apps (and converted desktop apps) can only target the 26% of total Windows installed base while unconverted desktop apps target the usual 100% of total Windows installed base.
I see what you mean.
However, for me, the equation is somewhat different. I mainly serve the general public.
I sell direct Windows Win32 apps on my sites and that is fine.
But I feel the Windows Store is picking up. Actually I verify it through the half dozen apps I made in VS a few years ago that are currently sold in the Windows Store.
Chances are the first correctly adapted Win32 apps converted and placed in the Windows Store will probably do good business, let alone because it is better to be leader in a small court than elbowing against 1.5 million other apps in the MAS.
Check Writer III today fully supports touch, down to a tablet mode with an on screen keyboard and AirPrint. I feel that once converted, it has chances serious enough to bring in good business.
Of course it would probably be better to have a native UWP app, but rewriting entirely Check Writer would take several months. I rather taste the waters with a converted app, before investing that amount of work.