spec: class: add ctor, dtor and __xdtor aliases #3125

ljmf00 · 2021-11-22T16:16:43Z

Signed-off-by: Luís Ferreira [email protected]

dlang-bot · 2021-11-22T16:16:45Z

Thanks for your pull request, @ljmf00!

Bugzilla references

Your PR doesn't reference any Bugzilla issue.

If your PR contains non-trivial changes, please reference a Bugzilla issue or create a manual changelog.

ljmf00 · 2021-11-22T16:17:14Z

Should we also refer to this on structs ?

ljmf00 · 2021-11-22T16:53:26Z

spec/class.dd

@@ -371,6 +371,9 @@ $(GNAME Constructor):
        }
        ------

+        $(P The compiler may generate a $(D __ctor) alias that matches with


I'm not sure if the right wording is may, should or omit it. Should we enforce this to be implemented? I think so

Is the compiler required to generate these aliases to comply to the spec? Or are they just an optional thing?

Is the compiler required to generate these aliases to comply to the spec? Or are they just an optional thing?

Technically yes, since e.g. __ctor is mentioned in several specification examples but this can be considered an implementation detail. You, as a user, probably don't require this, unless doing e.g. custom class allocators, custom traits or other advanced code. This is mostly used internally by the D runtime hooks and standard library traits, although I think it is useful to make it not implementation-defined to ease the user for the reasons I stated above.

maxhaton · 2021-11-23T20:44:45Z

spec/class.dd

@@ -729,6 +732,11 @@ $(GNAME Destructor):
        by the gc.
        )

+        $(P The compiler may generate $(D __dtor) and $(D __xdtor) alias that
+        matches the class destructor definition and inclusive destructors,


is the term "inclusive destructor" defined elsewhere in the spec? If not I think it should have a definition somewhere.

I note that __xpostblit()'s description in the struct page of the spec mentions the creation of aliases.

void __xpostblit(). The field and aggregated postblits, although generated for a struct, are not actual struct members. In order to be able to call them, the compiler internally creates an alias, called __xpostblit which is a member of the struct and which points to the generated postblit that is the most inclusive.

However, I note that this wording seems to be descriptive rather than prescriptive, which isn't ideal for the language specification.

I added the inclusive destructor concept to the glossary. Please review it.

About the __xpostblit(), I think it is a bit out of scope of this PR, although I can write a separate one fixing that wording.

dkorpel · 2021-11-23T20:51:01Z

There's no linked bugzilla issue, so what's the idea behind this PR? It's already specified that identifiers starting with a double underscore are reserved.

ljmf00 · 2021-11-23T23:26:32Z

There's no linked bugzilla issue, so what's the idea behind this PR? It's already specified that identifiers starting with a double underscore are reserved.

Yes, but this has a special purpose, and as I said above, it is noted in some specification examples such as traits and structs. We should also follow the same struct section, which specifies the behaviour of __postblit and __xpostblit.

Signed-off-by: Luís Ferreira <[email protected]>

RazvanN7 · 2021-12-09T14:25:37Z

spec/class.dd

@@ -371,6 +371,9 @@ $(GNAME Constructor):
        }
        ------

+        $(P The compiler generate a $(D __ctor) alias that matches with


Suggested change

$(P The compiler generate a $(D __ctor) alias that matches with

$(P The compiler generates a $(D __ctor) alias that matches with

RazvanN7 · 2021-12-09T14:26:00Z

spec/class.dd

@@ -371,6 +371,9 @@ $(GNAME Constructor):
        }
        ------

+        $(P The compiler generate a $(D __ctor) alias that matches with
+        constructors definitions.)


Suggested change

constructors definitions.)

constructor definitions.)

I don't really understand what you mean by this sentence. Could you please elaborate?

PetarKirov · 2023-05-22T14:39:08Z

@ljmf00 can you please rebase your PR?

ljmf00 · 2023-05-22T15:23:40Z

@ljmf00 can you please rebase your PR?

Yes, I'll add to my TODO list.

ntrel · 2023-05-29T15:43:50Z

glossary.dd

@@ -258,6 +258,11 @@ void test()
        syntax) refers to the ability to call a function as if it were a method of
        its first argument. For example `funct(arg)` may be written as
        `arg.funct()`.)
+
+        $(DT $(LNAME2 inclusive_destructor, inclusive-destructor, Inclusive


Suggested change

$(DT $(LNAME2 inclusive_destructor, inclusive-destructor, Inclusive

$(DT $(LNAME2 inclusive-destructor, Inclusive

pbackus · 2023-11-08T14:43:46Z

glossary.dd

+        $(DT $(LNAME2 inclusive_destructor, inclusive-destructor, Inclusive
+        Destructor))
+        $(DD An inclusive destructor is a destructor that includes destruction
+        of the object itself and its fields, if reaching its lifetime.)


Suggested change

of the object itself and its fields, if reaching its lifetime.)

of the object itself and its fields.)

Destructors can be called early (for example, by destroy), and this definition applies regardless of when it is called.

RazvanN7 · 2024-03-06T10:45:53Z

ping @ljmf00

ljmf00 requested review from andralex and WalterBright as code owners November 22, 2021 16:16

ljmf00 changed the title ~~spec: class: add __ctor() and __dtor() aliases~~ spec: class: add __ctor(), __dtor() and __xdtor() aliases Nov 22, 2021

ljmf00 changed the title ~~spec: class: add __ctor(), __dtor() and __xdtor() aliases~~ spec: class: add __ctor, __dtor and __xdtor aliases Nov 22, 2021

ljmf00 force-pushed the ctor-dtor branch from b4db5f2 to d0f806a Compare November 22, 2021 16:18

ljmf00 commented Nov 22, 2021

View reviewed changes

maxhaton reviewed Nov 23, 2021

View reviewed changes

spec: class: add __ctor, __dtor and __xdtor aliases

33428bf

Signed-off-by: Luís Ferreira <[email protected]>

ljmf00 force-pushed the ctor-dtor branch from d0f806a to 33428bf Compare November 28, 2021 22:33

ljmf00 requested a review from maxhaton November 28, 2021 22:35

RazvanN7 reviewed Dec 9, 2021

View reviewed changes

dlang-bot added the stalled label Mar 9, 2022

PetarKirov requested review from RazvanN7 and dkorpel May 22, 2023 14:38

ntrel requested changes May 29, 2023

View reviewed changes

pbackus reviewed Nov 8, 2023

View reviewed changes

	$(P The compiler generate a $(D __ctor) alias that matches with
	$(P The compiler generates a $(D __ctor) alias that matches with

	$(DT $(LNAME2 inclusive_destructor, inclusive-destructor, Inclusive
	$(DT $(LNAME2 inclusive-destructor, Inclusive

	of the object itself and its fields, if reaching its lifetime.)
	of the object itself and its fields.)

Uh oh!

spec: class: add __ctor, __dtor and __xdtor aliases #3125

Are you sure you want to change the base?

spec: class: add __ctor, __dtor and __xdtor aliases #3125

Uh oh!

Conversation

ljmf00 commented Nov 22, 2021

Uh oh!

dlang-bot commented Nov 22, 2021

Bugzilla references

Uh oh!

ljmf00 commented Nov 22, 2021

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ljmf00 Nov 23, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dkorpel commented Nov 23, 2021

Uh oh!

ljmf00 commented Nov 23, 2021

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

PetarKirov commented May 22, 2023

Uh oh!

ljmf00 commented May 22, 2023

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

RazvanN7 commented Mar 6, 2024

Uh oh!

Uh oh!

spec: class: add ctor, dtor and __xdtor aliases #3125

spec: class: add ctor, dtor and __xdtor aliases #3125

ljmf00 Nov 23, 2021 •

edited

Loading