Attila Kovacs

44 minutes ago, Mark Williams said:

I set MaxConnection to 1000 I assume that will not of itself degrade performance

you are right, it wont.

this is all what happens:

  if (FMaxConnections > 0) and (FAddingActiveModules >= FMaxConnections) then
    raise EWebBrokerException.CreateRes(@sTooManyActiveConnections);

Calculating the max allowed concurrent user depends on many things and there is no exact formula.

You could do some stress-test to see how the server works under load.

This would also test how your dll and connection pool are performing, are there deadlocks, unexpected behavior, etc...

Well, I'd clone/backup the project and start removing all 3rd party units, including firedac.

- did you check the system/application events?

- if you kill all the w3wp processes in the taskmanager, will then IIS stop faster?

Ok.

- I'll assume then that you are aware that in case of TFDManager.Create(SomeComponent), including placing the component on a datamodule, FDManager will be free'd if SomeComponent is free'd.

- I'm sure, that you have tried commenting out the whole OnTerminate, and IIS still stalls on stop.

Let me know if you find something. I'm really curious.

Ok. How do you create FDManager (Create(nil)?, not that you free it twice) and where is the var declared?

5 minutes ago, Mark Williams said:

application pool is still taking an age to close down

How can you see that? Hourglass, or the time until the dll is released?

sidenote: in "TWebApplication.Create" there is already a "if IsLibrary then IsMultiThread := True;" 

ok

-what if you remove the .free? Will it leak memory or does it get freed when IIS unloads the dll?

-what if you create your FDManager with "TISAPIApplication(Application)" as owner? It's a TComponent descendant so it will free FDManager.

I'm not sure why do you need a critical session for freeing something? Do you think other threads would still using it? And you just freeing it?

Just asking.

It shows the values as octal up to 31 except TAB. Nice.

I have no clue what is this for but 0x8 is octal 10 and 0x9 is TAB which is "\t".

We still can't see the most important part of the code, how do you evaluate the "output".

Anyway, no matter what it is, this is for sure not the way this should be done.

Edit: Also, the output is suspicious.

"This code has been working on windows for years."

How many years? 20? Was it compiled with an AnsiString Delphi compiler?

why don't you hit 'Break' and look into the call stack?

you have to pass it one by one, like:

SELECT * FROM tTableName WHERE Status IN (:ST0,:ST1)

thus, you have to build your query with

there is no "IN (:ONEPARAM)"

I don't get it, if its a remote server, how comes that there is no REST between the client and the server?

https://www.php.net/manual/en/function.exif-read-data.php

You can't just register a clipboard format and expect that every control starting using it. That's not how windows clipboard works.

My implementation was attached 11 hours ago, need further testing. It looks good so far.

By the way, with the very same unit it would be possible to encrypt the clipboard data too. But I see no use for that, just saying.

Ok, here is my prototype.

- The goal was to keep application data safe from the clipboard.

- It tries only CF_UNICODETEXT to keep safe, every other formats are free to come and go.  (for now?)

- RichEdit does not pass the necessary events, so it needs further client code to protect

- no extended testing yet, but I can assure that this behavior will drive the user crazy 😉

- no performance penalty tests

- no clue what Antivirus programs says for modifying the import table

It does capture clipboard events and keeps copied unicode text in a variable instead of the clipboard.

On paste command, it pastes from this variable. (as far my basic paste algo works)

On OS Clipboard change, if it's CF_UNICODETEXT, overwrites the In-App clipboard variable to be able to paste into the app from the outside.

I have no clue yet if there are any flaws, but it was lot of fun to make it.

InAppClipBrd.pas

so why do you comment on something you don't understand?:)

Especially with ctrl-c/ctrl-v. Is that all how clipboard works in your opinion?:)

Hm, thanks Lars, this will be a good start. Hope I can isolate, not just watch. Lets see!

7 minutes ago, Ugochukwu Mmaduekwe said:

you might have to end up writing your own app specific clipboard

Thanks @Ugochukwu Mmaduekwe, no problem, I just wanted to inform if there is already something for that.

I'll check hooking up some windows messages.

@Fr0sT.Brutal  Not quite. I think my description is also crystal clear. 😉

Hi all,

Is there any existing unit/module whatever for handling In-App clipboard functionality?

App should not leak data over clipboard, but I would like to keep the feature inside the app.

It's shouldn't be too hard, I'm just curious if there is already any solution.

Handling text is sufficient.

(VCL/Windows)

it's a recurring problem

most likely no ODBC driver installed

High five!

@Kryvich thx, works as you described. (db login-error-form, prior to any other, including main-form created)

What if you create a TBaseBase with protected overloaded methods and public in TBase(TBaseBase) TChild(TBaseBase)

or TBase with protecdted overloaded methods and public in TBaseExposed(TBase) and TChild(TBase) ?

Or you go from New() to .FromJSON(JSON: string): TMyClass and .FromInteger(I: integer): TMyClass?

Or creating records with Implicit operators for string and integer/whatever, then you could write Child := '{somejson}'; or Child := 1; (did I go too far?:)