Anders Melander

I got that one just last week. AFAIR, it too had to do with a array[0..0] of stuff. I can't really remember much about it because I was in a rush and just worked around it

Supposedly each request is handled by a thread - Like in every other middleware. If your experience is that an exception in the request handling of the middle-tier blocks the whole server then there's either something seriously wrong with the middleware or the architecture of the middle-tier. The problem is probably that the main thread is somehow involved in the request handling. Not that it's related to your problem, but if you want middle-tier exceptions passed back to the client, as exceptions, you can look in the DataSnap source code to see how it's done.

Impressive dedication 🙂 See if you can locate the Embarcadero Licence Manager somewhere on your system: LicenseManager.exe If you can run that, then you should be able to delete any prior license that might be in conflict with the new one. If everything else fails then you can try to delete all the license info manually. I believe it's in C:\ProgramData\Embarcadero\.licenses Note that this contains the license data for all Embarcadero products on your system so be careful. See also: https://docwiki.embarcadero.com/Support/en/Moving_from_Delphi_or_C%2B%2BBuilder_Community_Edition_to_another_edition

Well, you said The Unit Scope Names setting helps abstract away implementation details which was what I have been trying to get an explanation of - but never mind. The example I gave with PByteArray is real; I have a core unit which declares TByteArray (of which PByteArray is a pointer to) as array[0..0] while SysUtils declares it as array[0..32767]. Since these two are ambiguous I have to scope my use of PByteArray when both units are used. No problem, SysUtils.PByteArray(some_pointer) doesn't obfuscate the code much. System.SysUtils.PByteArray(some_pointer) however is bit too verbose for my taste. FWIW, I just resolved it with a local type declaration: type PByteArray = System.SysUtils.PByteArray;

You still have to explicitly declare, in the uses clause, where PByteArray comes from - regardless of which form you use. And in the case of identifier ambiguity there is no difference in how it is resolved, regardless of fully qualified namespace or implicit namespace. So how does that explain your statement: If they move an identifier from one unit to another, you still have to update your uses clause and any explicit namespacing, since Delphi doesn't really have namespaces. They can hide a move from Foo.SysUtils to Bar.SysUtils but not from System.SysUtils to System.FooUtils. I guess what you were referring to was a move from SysUtils to System.SysUtils.

Huh? How do they do that? Can you give an example?

I'm not sure I agree on that. I've just been through a refactoring session where SysUtils was replaced with System.Sysutils which in turn made it necessary to resolve a series of type references as System.Sysutils.PByteArray instead of Sysutils.PByteArray due to a duplicate type declaration. That certainly didn't improve readability.

Supposedly compiles faster...?

You are looking at the function declaration. As Christian implied, for API functions that exist in both an Ansi and a Unicode version, you need to look at the import declaration: function FindFirstFile; external kernelbase name 'FindFirstFileW'; and the declaration of TWin32FindData: TWin32FindData = TWin32FindDataW; I'm guessing it has been declared this way since the very first Unicode version of Delphi - as it should be. It's the same in C. In fact, the declarations were probably generated automatically from the C headers. I'm guessing the official Microsoft FindFirstFile documentation also mentions it.

Bike shedding For example like this: https://github.com/graphics32/graphics32/blob/b45d1108a8f57b66739731e952f81e2636c63abd/Source/GR32.ImageFormats.pas#L386 Example of use: const FileSignaturePNG: AnsiString = #$89#$50#$4e#$47#$0d#$0a#$1a#$0a; FileSignaturePNGMask: AnsiString = #$ff#$ff#$ff#$ff#$ff#$ff#$ff#$ff; ... function TImageFormatAdapterPNG.CanLoadFromStream(AStream: TStream): boolean; begin Result := CheckFileSignature(AStream, FileSignaturePNG, FileSignaturePNGMask); end;

Good luck and please don't keep us informed about how it goes.

I can't see that they do. Here's the TLDR as I read it: Allen is saying that you shouldn't (mindless or not) use FreeAndNil to solve logic/design bugs and AFAICT nobody is disagreeing with that. Primoz and David are saying that they are using it to find stale reference bugs. Allen's suggestion to use a debug memory manager instead isn't valid because you can't do that on a production system, as Primoz also points out.

Okay, now I understand what you meant. Yes, it's true that I'm trading the ability to catch one problem for another. However, the problem you are describing does not seem to be one that we are experiencing much if ever, so I'll take that trade. Regardless, we have made an active choice to use FreeAndNil this way. Not because we don't know any better but because we have a specific (recurring, I might add) problem that it helps us solve. I totally get the argument against just mindlessly using it CCP style but I can't really see how that applies to how we use it. I'd still like to know about these alternatives I should be using though... Well, that's nice but no that's not what we are discussing. I gave an example of how and why we used it and you said that that usage had been "debunked".

I have not claimed otherwise. It's not a magic "find all your bugs" tool. It's a "better than nothing" tool. Again, you are arguing against real world experience on actual code. I don't see the point. I understand it and no, (roses are red, violets are blue) it does not answer my question and neither did you. Such as? Debug memory managers are not suited for this use case; They can not be used in production code because of their overhead. FreeAndNil has practically no overhead. Again, I never claimed otherwise. Sometimes is better better than never. If there was a better alternative I would happily use that but I'm not hearing any. That's fine. In our code we make our intent clear by documenting it: With comments in the code where we rely on the order of nil and free. With a section in our development guidelines and style guide about all the other cases; When, why, and how and when not to use it.

That's a pretty unnuanced statement. One could just as easily, and just as unnuanced, argue that not using it is CCP. If you use something, with a specific purpose, how is that CCP?

Oh really? It appears to me that you are arguing a theoretical standpoint against the actual real world experience I have with finding stale object pointers using FreeAndNil. Yes, that's the point. Did anyone claim otherwise? It will succeed because it has been designed to do so. How is that relevant to FreeAndNil? What second time? What logic error? What are you even talking about? All this focus on free is pretty irrelevant to my use of FreeAndNil. I'm obviously not using it to guard against double frees because it doesn't do that. I'm using it to catch access though stale object pointers. End of story. I like to think that I actually know pretty well what I'm doing but I get the impression that you imagine I code by just throwing stuff against the wall to see what sticks.

Yes. We now know Marco can read the help. Beyond that I don't know what we learned. Maybe that they don't have any compiler engineers at hand since the product manager had to do first level support and guess about the work required. The issue was filed as a New Feature but treated as a Bug report. Not impressed.

Real life race condition :-)

Instead of FreeAndNil implying a certain pattern I think it would be better to explicitly state, in a comment, when and why a certain behavior is expected. The projects I'm working on right now has something in the neighborhood of 200 calls to FreeAndNil in it. All except two are there to catch stale pointers; FreeAndNil gives us a nice AV that we can easily debug. Free would most likely give us a sporadic random error somewhere else. The two remaining cases are in a framework where we need a container var to be nilled before the object is destroyed. I won't go into why it's necessary; I'm sure you know the pattern. Anyway, for these two cases, the comments in the code clearly document why the FreeAndNil pattern is necessary. This is an old project and the introduction of FreeAndNil, where it makes sense of course, has helped us catch and eliminate countless bugs. Before I took over as the lead the code was littered with hundreds of empty try..except blocks simply because they had given up on trying to find the cause of the exceptions.

Kudos to @Kas Ob. 🎖️

FWIW, here's an example of improper use of FreeAndNil causing a race condition because a context switch happened between between the nilling and the free: https://supportcenter.devexpress.com/ticket/details/t812550/an-av-occurs-if-the-spell-checker-whose-usethreadedload-property-is-set-to-true-reloads

Don't expect it to be "fixed". The best you can hope for is that it is documented. The current behavior is consistent with class initialization; If you have the same field in a class and create a new instance, then the ordinal value of the field will be zero - as expected.

Bummer. I've just checked both my desktop (which is a *ehem* 15 years old home build) and my laptop which is the top Lenovo X1 model, 2 years old, and neither of them support S0. This means that I unfortunately can't help debug the problem. If anyone else wants to have a go at it, the command to check for S0 support is powercfg /a. Note though that if Modern Standby isn't already enabled then it requires a complete OS reinstall to enable it. Yup, you read that right. Nice one, Microsoft.

You are looking at the symptoms at the end of a long chain of circumstances. You need to look at the other end, which supposedly (assuming WM_TIMER is the problem) is closer to the source of the problem. If standby killed WM_TIMER then millions of applications would break so I think it's very unlikely. The bug in PuTTY wasn't that the timer didn't fire but that they assumed that it would fire with a certain time interval and WM_TIMER is never guaranteed to do that. AFAIK WM_TIMER is a low priority message which is only synthesized when the message queue is polled and no higher priority messages are in the queue (or synthesized). I don't know your application but it could be that the queue is simply flooded with other stuff after resume from standby. So instead of assuming that WM_TIMER doesn't work I recommend that you simply verify with a simple application that does nothing but write a time stamp to a TMemo every time a 1 second TTimer fires. If that doesn't work then there's a problem with TTimer and we can start examining what that is.

I think the timer is still alive. It just takes a (unusually long) while to fire after resume from standby. See also: https://github.com/cyd01/KiTTY/issues/475