Rollo62

Yes, when usually toggling the insert key, to change between INSERT and OVERWRITE mode, this is ignored and stays at INSERT ( only in Delphi behaves like that, other apps do change the mode ). The solution is to change above combobox selection.

I still have hope that we will see VmWare or the like on M1 Macos, which is able to support x86 Win10/11. That would leave the problem to VmWare to make it run. I haven't thought too deeply about what that conversion would mean on the CPU virtualization side of an entire OS, but why not its also simply a "program" Yes, maybe a missing hypervisor emulation would be a problem.

That is a very good point, maybe that is not so good news if Wintel is needed, like for Delphi. It would be worth checking that with Apple, and try to get some clear guarantees. On the other hand, this time they are not moving from IBM to Intel, but from x86, which means they will have the chance to catch Wintel users in the long run. Why should they kill a well working emulator solution, which can attract all the Windows users to M1, I think cancelling that would be a stupid marketing decision. At least they had added some more interfaces, like HDMI, etc. this is what their advertising said, not sure in detail how many USB sockets they have. But I'm used to have a short USB-Hub dongle plugged in anyway, so not that big deal for me.

Me neither. Sometimes I recognize that INSERT mode had changed, and this won't get back by normal key, only by changing the mode here. Maybe there is also a hidden catch like that for Ctrl-F, somewhere hidden in the options ?

I only can say: Try to never use XCode "Betas", better always wait until the version is stable. They can cause a lot of issues, and can also make your whole environment unusable, by possibly enforcing further (beta) updates on iOS.

I want to re-activate this old thread, but its still a current topic. Apple launched their new M1 Pro and M1 chips and Max MacBook Pro's https://www.apple.com/apple-events/october-2021/?cid=CDM-USA-DM-P0021510-488761 Well, I'm not an Apple salesman, but I better should be. This really blows all other laptops away, if it works as smooth as desired. It has an tremendous "wanna-have" factor, and the prices are not even that dramatic as I expected, like for example the Intel i9 versions. I think Intel ( Wintel ) will loose even more ground on the chip business, because there are no big reasons to buy large desktop workstations any longer, with such power in a small box. By Rosetta this supports ARM M1 and Intel x86 anyway, with good compatibility and at high performance, so I think Wintel / AMD will see a hard future. Of course you have to move from soft Win-Wonderland to the aluminium-hard Mac-World, which is a little disturbing as a user/developer for some weeks or months. But as always, some things I find annoying everyday, some things I love, but there is nothing which you couldn't get close to. Who's next PC will be a MacBook Pro Max, I'm curious ?

I was also sure that they have been "included", and also I got some for testing. My goal was to get started with 2 or 3 API's, to see how its going. Since I could not find much time for testing, I had to realize one day that my unused licenses had been expired. I've tried to check, to re-acivate the 1-year licenses, but they told me they were intended only for testing only, and once expired cannot prolonged for free. The license terms were always very fuzzy to me ( as usual with Embarcadero stuff ). If I wanted to prolong, I could get a 14-day trial, but thats maximum I could get, after that I should buy them. So I feel screwed a lot, and also I can only say: "Dont tocuh it". Better grab your REST-APIs directly, CDATA is not worth it under such "all-or-nothing" business model.

Why removing BPL parts which maybe cause unknown sideeffects, and risking an unstable IDE as a newly installed tool, while of course still complaining about any instabilities after installation, when closing or ignoring of the Welcome page is just one click in the IDE ( so a no-brainer ) ? 🤔 I would keep IDE as "natural" as is, while exploring its behaviour and stability over a certain time, before I try to remove or install anything harmful with yet unknown results.

Yeeees, HTML/CSS/JS has a really terrible High DPI support

This has changed completely. https://www.ideasawakened.com/post/customize-the-rad-studio-11-welcome-page-for-single-click-activations

Thanks for the insights. I'm wondering if this might cause any problem with memory leaks or AV, when creating/destroying frames at runtime, in changing, random order. If the Application->MainForm is used over several units, inside specific components only, this always leaves a bad taste. This leads me to the question what is the right, intended, official way to use the Application->MainForm and the AOwner of Frames/Components. Shall it be in a strict tree-structure, or does it not matter, to mix tree and direct owenerships as you like ? Does it matter if the relations were mixed by designtime and runtime ? Does FMX behaves differently than VCL in that regard ? Thanks god, the termination sequence on mobile is not that critical, since this is killed by phone mostly

Good question which is using what. https://docwiki.embarcadero.com/RADStudio/Alexandria/en/LLVM-based_Delphi_Compilers Seems that Macos is not yet LLVM, according to above list. But it is acc. to this info https://docwiki.embarcadero.com/RADStudio/Alexandria/en/Delphi_Toolchains https://docwiki.embarcadero.com/RADStudio/Alexandria/en/DCCOSX64.EXE,_the_Delphi_Compiler_for_macOS_64-bit https://docwiki.embarcadero.com/RADStudio/Alexandria/en/DCCOSX64 https://static.webshopapp.com/shops/081736/files/088721693/embacadero-rad-studio-feature-matrixpd.pdf While I thought a long while that only D-Win32 is properitary compiler, and all the Rest is LLVM. Maybe it depends on language Delphi or C++ too. However, it compiles well, so it does its job for me

But you would be in "Delphi Heaven", not "DLL Hell".

Include files I like to use separation in clear defined units, something like this. That needs some kind of discipline, I'm not always that strict, but I would prefer such separation over cluttering one unit with a lot of includes Just roughly scribbled the code like that: unit MyUnit.Intf; interface Uses {$if DEFINED ( windows )} MyUnit.Intf.Win; type TMyUnit = MyUnit.Intf.Win.TMyUnit; {$ELSEIF DEFINED( linux ) } MyUnit.Intf.Linux; type TMyUnit = MyUnit.Intf.Linux.TMyUnit; {$ELSEIF DEFINED( macos ) } MyUnit.Intf.Macos; type TMyUnit = MyUnit.Intf.MAcos.TMyUnit; {$else} {$MESSAGE ERROR 'MyUnit.Intf: Unsupported platform' } {$endif} .... That results in two main advantages: - NO INCLUDE's needed in the underlying sub-units - a clear error message when unknown conditionals were used

1 click is fine, but I like at least to define a folder where it shall go. Yes EULA before install, but still I'm not sure about its purpose in Europe or other countries. I think this is mainly an US law thing, and everybody worldwide just copied the same behaviour, as it looks cool. I'm no lawyer, but I think in EU this doesn't make much sense anyway. If a user installs something on purpose, he is responsible, no matter if texts showed up before using it. Of course if the software formats the disk, than the manufacturer should be in focus, but I assume we talk about normal, non-critical apps for the moment, that could be installed and uninstalled with no harm. Maybe somebody has experiences from real cases, and could share this info.

Yes we all have such EULA pages. But to be honest, I don't know if this makes any legal difference, if we have it or not. Is there any known legal case, that was won or lost because of EULA or missing EULA ? Maybe in USA, such "overinformation" is needed, I don't know, but in good old Europe I doubt that. To the contrary, in EU new laws tend to ban such "read and click in a hurry" contracts, as not legal in some other cases ( not EULA ), but at least for web-pages contracts. I would assume the legal power of such EULA page is quite low (at least in EU). Instead of the EULA, I prefer to have a special, nice "welcome" page(s) in the app itself. This can be used for EULA, but also for all other setup requirements. I think this is much better than InnoSetup ever could be, so why not using such an app self-setup welcome page, and leave the setup minimal ?

Yes include files will work. But be aware of that precompiled units will break easily, when making changes the include file, which will lead to nasty errors. So this kind of unit + "include" should be always re-compiled.

The marketing of D11 claims that it has the latest billing service, if that is what you need, but I haven't checked if its working.

Thanks for the update, very good to hear that FMX cross platform support is coming closer. Do you have plans/roadmap to support all available platforms Android, iOS, Macos 32, Macos 64, Macos ARM, Win32, Win64, Linux ? In the current version, why is there no Macos 64 support, are there technical problems that prevent to go there ? Does the Macos 32 use maybe OS drivers, which are available on the platform, or do we have to add extra static/dynamic libraries for that ?

Ok, thats an explanation. BTW: I have found this nice article from Paolo, maybe thats helpful too http://blog.paolorossi.net/2017/04/27/jwt-authentication-with-delphi/ I think it was planned to part 3-of-3, and part-4 came later: http://blog.paolorossi.net/2019/07/15/jwt-authentication-with-delphi-part-4/ I've noticed this only later, when appearing in DelphiMagazine, so maybe I should monitor Paolo's blog more often 🙂 http://www.delphimagazine.com/2019/07/16/jwt-authentication-with-delphi-part-4/

Hi there, so far I have not considered GitHub as an OAuth service provider, also usable for external, productive apps, but this rough idea came always into my mind since some time. So I wanted to check now if and how that could be possible. The info about terms and conditions is a little unclear to me, about that OAuth Application from GitHub: https://docs.github.com/en/developers/apps/building-oauth-apps/creating-an-oauth-app https://membrane-soa.org/service-proxy-doc/4.2/oauth2-github.htm https://docs.readme.com/docs/setting-up-github-oauth I assume that a GitHub account >= Pro, can be used as a normal OAuth-Service provider, also for external apps not hosted on GitHub itself. Since I cannot find a lot information about these terms, I am very unsure if this is a good idea, or recommended only for development purposes. Who is using such already in productive applications ? What I can see that GitHub allows up to 100 OAuth services, and I see no difference especially for development or production. Are there special hidden costs, when using this productive, or is this all in the flat cost of the account ( which would starting from a Pro account at 4 USD/month, including up to 100 OAuth applications ) ? Are there special terms that forbid such kind of usage, outside the scope of GitHub projects ? Are there any technical restrictions that may bite me in a productive system ( any resources/packages occupied per app, performance, access limits, transfer limits, or the like ) ? If that is all OK, I think this would mean that GitHub is a very cost effective OAuth provider, only I cannot really believe that. Maybe somebody here has experiences with using GitHub as service provider, and would like to give some tips and and share recommendations ?

Ok, I see. Yes, it should be it, but stange that the issuer name is in the list below, but not the issuer key. I thought that several Issuers could be possible, it looks like there can be only one, or is it different for the currently logged in user ? I would have expected this issuer key near to the key's issuer name in the list below.

Yes, I think its just kind of access-token, and Apple recommends also a quite short lifetime < 20min for the token, to be accepted. By the way, I think all this is still beta probably, and I know Apple for changing a lot when its about improving security 🙂

Yes, this was also my first thought, but I have checked that the Jose JWT library adds this "JWT" (uppercase) automatically. But maybe worth to doublecheck, if this stays like that also after creation. The notes about URL are interesting, but I think Apple uses this in its own way. I also thought some times to use JWT Token, to create some unique, secured data, but without URL, that should make sense too IMHO. Did you get if from the AppStoreConnect, user section ? I see all the user/roles, etc. but no such ID anywhere. The problem I see is that I have several entries there anyway, under the same "Name", probably even with different Apple-Ids. If this is from the MemberCenter/Certification section, I cannot spot it either. I would assume that this is maybe taken from a different item, not the one that Apple wants to see here. Nevertheless, Apple seems not havin explained that origin very well.

Hi Dave, how does the AIssuer looks like ? It should be like 57246542-96fe-1a63-e053-0824d011072a , but I cannot find this from the Users table or anywhere else. Where did you get this from ?