Stocki,

Since your report was closed and I didn't fully (the closing) agree with that without research, here's a part of my research:

I found this reply by Danny Thorpe on stackoverflow: http://stackoverflow.com/questions/7759081/is-the-compiler-treatment-of-implicit-interface-variables-documented/19914385#19914385

To quote:
"I recall there have been a few bugs in this area over the years. Long ago (D3? D4?), the compiler didn't reference count the intermediate worked most of the time, but got into trouble in
parameter alias situations. Once that was addressed there was a follow up regarding const params, I believe./*There was always a desire to move disposal of the intermediate value interface up to as soon as possible after the statement in which it was needed, but I don't think that ever got implemented in the Win32 optimizer because the compiler just wasn't set up for handling disposal at statement or block granularity*/."

As you will see, the original intention by the Delphi architects was to release an interface as soon as possible.
The problem was**and is** that their compiler infrastructure doesn't allow that.
So the discussion on the buglist should be read in that context.

Even the designer of Delphi's interface implementation wished they were able to have a better implementation.
CUDO's to the Freepascal compiler guys for more or less achieving that ;)

You have to read it very carefully. It is not about the question, but about the *last part of the answer by Danny.**
*
I hope this sheds some new light on the matter that bothers you so much.

Regards,

Thaddy

[...]

I hope this sheds some new light on the matter that bothers you so much.

Addendum:
http://wiert.me/2014/08/06/delphi-using-iinterface-to-restore-cursor-at-end-of-mehod-prelude-to-a-memento-that-executes-any-code-at-end-of-method/ (look at the comments by me)

And unless we have it somewhere black on white in the official documentation we won't change anything there. And even then we'd likely only do it for mode Delphi...

Note: It would have been nice if you mentioned the issue which you are refering to. Not everyone has an eye on the bug tracker...

Regards,
Sven

Sven,
I made some mistakes here as well. About 20 years ago I knew what was going on.
I am sorry I didn't put the mail in the proper context.

Regards,

Thaddy

Sven, I made some mistakes here as well. About 20 years ago I knew what was going on. I am sorry I didn't put the mail in the proper context.

Then why haven't you make up for it and put up the link? ;)

For the record, we are talking about this issue:
http://bugs.freepascal.org/view.php?id=26602

Regards,
Sven

Oh, I see. Oeps.

In our previous episode, Sven Barth said:

And unless we have it somewhere black on white in the official documentation we won't change anything there. And even then we'd likely only do it for mode Delphi...

Make it a modeswitch then :-)

Mm,

Marco, since you were the first one to put me right (and I was referring to block rsolution, which was false) I do not think - in the light of Danny's writings - that a {$modeswitch delphibugs} is appropriate. ;)

Anyway, I hope that this reconciles Stocki's thinking about the subject.

Delphi got it wrong FPC got it better (maybe not right, but close and better than delphi) conceptually.