www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - `in` no longer same as `const ref`

reply =?UTF-8?B?Tm9yZGzDtnc=?= <per.nordlow gmail.com> writes:
https://github.com/dlang/druntime/pull/1748/files

Why is `in` no longer the same as `const ref`?
Jan 29 2017
parent reply =?UTF-8?B?Tm9yZGzDtnc=?= <per.nordlow gmail.com> writes:
On Sunday, 29 January 2017 at 11:49:19 UTC, Nordlöw wrote:

 https://github.com/dlang/druntime/pull/1748/files

 Why is `in` no longer the same as `const ref`?


Correction:

Why is `in` no longer the same as `const scope`?
Jan 29 2017
parent reply Walter Bright <newshound2 digitalmars.com> writes:
On 1/29/2017 3:50 AM, Nordlöw wrote:

 Why is `in` no longer the same as `const scope`?


Because it was unchecked and largely unimplemented. I was afraid that by 
checking it, too much code would break.
Jan 29 2017
next sibling parent reply =?UTF-8?B?Tm9yZGzDtnc=?= <per.nordlow gmail.com> writes:
On Monday, 30 January 2017 at 00:26:27 UTC, Walter Bright wrote:

 On 1/29/2017 3:50 AM, Nordlöw wrote:

 Why is `in` no longer the same as `const scope`?


 Because it was unchecked and largely unimplemented. I was 
 afraid that by checking it, too much code would break.


Will we make them equal in the long run? So much shorter to read.
Jan 29 2017
parent reply Walter Bright <newshound2 digitalmars.com> writes:
On 1/29/2017 9:50 PM, Nordlöw wrote:

 Will we make them equal in the long run? So much shorter to read.


I know it's shorter. I'd like to do it too, we'll have to see.
Jan 30 2017
parent reply Radu <radu void.null> writes:
On Monday, 30 January 2017 at 08:02:12 UTC, Walter Bright wrote:

 On 1/29/2017 9:50 PM, Nordlöw wrote:

 Will we make them equal in the long run? So much shorter to 
 read.


 I know it's shorter. I'd like to do it too, we'll have to see.


Can it be enabled for -dip1000 flag?
Doc was pretty clear on what it was suppose to do and people 
kinda knew that scope was not finished.

Having it turned on for dip flag will allow some testing of the 
existing code. I would like to have in's intended semantic 
finally implemented.
Jan 30 2017
parent reply Walter Bright <newshound2 digitalmars.com> writes:
On 1/30/2017 12:42 AM, Radu wrote:

 Can it be enabled for -dip1000 flag?
 Doc was pretty clear on what it was suppose to do and people kinda knew that
 scope was not finished.

 Having it turned on for dip flag will allow some testing of the existing code.
I
 would like to have in's intended semantic finally implemented.


I'd rather do one thing at a time.
Jan 30 2017
parent reply =?UTF-8?B?Tm9yZGzDtnc=?= <per.nordlow gmail.com> writes:
On Monday, 30 January 2017 at 11:32:11 UTC, Walter Bright wrote:

 I'd rather do one thing at a time.


+1

And it won't be too difficult to search-and-replace `const scope` 
with `in` once things has stabilized and all the corner cases 
have been sorted out.

Thanks for making D more safe, Walter. D really needs it in the 
competition with Rust.

Once things have stabilized we really need to announce this work 
and highlight its developer productivity compared to Rust.
Jan 30 2017
parent Walter Bright <newshound2 digitalmars.com> writes:
On 1/30/2017 7:06 AM, Nordlöw wrote:

 And it won't be too difficult to search-and-replace `const scope` with `in`
once
 things has stabilized and all the corner cases have been sorted out.

 Thanks for making D more safe, Walter. D really needs it in the competition
with
 Rust.


You're welcome.



 Once things have stabilized we really need to announce this work and highlight
 its developer productivity compared to Rust.


It's what I plan to talk about at DConf.
Jan 30 2017
prev sibling next sibling parent reply Jonathan M Davis via Digitalmars-d <digitalmars-d puremagic.com> writes:
On Sunday, January 29, 2017 16:26:27 Walter Bright via Digitalmars-d wrote:

 On 1/29/2017 3:50 AM, Nordlöw wrote:

 Why is `in` no longer the same as `const scope`?


 Because it was unchecked and largely unimplemented. I was afraid that by
 checking it, too much code would break.


Which is why I've been arguing against anyone using in for years, but some
folks really liked how it was a parallel of out and used it anyway. Some
understood what scope was at least theoretically supposed to mean and were
willing to deal with the consequences if and when scope were actually
implemented beyond delegates, but most just used it without that
understanding.

Personally, I think that effectively having an alias for two attributes in a
single attribute is a confusing design decision anyway and think that it was
a mistake, but we've had folks slapping in on stuff for years with no
enforcement, and flipping the switch on that would likely not be pretty.

- Jonathan M Davis
Jan 29 2017
parent reply Olivier FAURE <olivier.faure epitech.eu> writes:
On Monday, 30 January 2017 at 06:38:11 UTC, Jonathan M Davis 
wrote:

 Personally, I think that effectively having an alias for two 
 attributes in a single attribute is a confusing design decision 
 anyway and think that it was a mistake, but we've had folks 
 slapping in on stuff for years with no enforcement, and 
 flipping the switch on that would likely not be pretty.

 - Jonathan M Davis


I've always thought of 'in' as a visual shorthand for "this 
parameter doesn't care whether you give it a deep copy or a 
shallow reference", personally.
Jan 30 2017
parent reply Q. Schroll <qs.il.paperinik gmail.com> writes:
On Monday, 30 January 2017 at 12:08:06 UTC, Olivier FAURE wrote:

 On Monday, 30 January 2017 at 06:38:11 UTC, Jonathan M Davis 
 wrote:

 Personally, I think that effectively having an alias for two 
 attributes in a single attribute is a confusing design 
 decision anyway and think that it was a mistake, but we've had 
 folks slapping in on stuff for years with no enforcement, and 
 flipping the switch on that would likely not be pretty.

 - Jonathan M Davis


 I've always thought of 'in' as a visual shorthand for "this 
 parameter doesn't care whether you give it a deep copy or a 
 shallow reference", personally.


Would have been a far better definition. Why does anyone really 
need a shorthand attribute for two attributes that could be 
easily spelled out? You can type anything for "const  scope" 
while programming and then do search-and-replace. That's even 
trivial.

Can't we make "in" mean "const scope ref", that binds on 
r-values, too? Effectively, that's (similar to) what "const T&" 
in C++ means. It's a non-copying const view on the object. We 
have the longstanding problem, one must overload a function to 
effectively bind both l- and r-values. That is what I'd suppose 
to be the dual to "out".
Jan 30 2017
next sibling parent bitwise <bitwise.pvt gmail.com> writes:
On Monday, 30 January 2017 at 19:05:33 UTC, Q. Schroll wrote:

 Can't we make "in" mean "const scope ref", that binds on 
 r-values, too? Effectively, that's (similar to) what "const T&" 
 in C++ means. It's a non-copying const view on the object.


'ref' being separate from 'in' allows you to use both:

foo(in ref T x);
foo(in T* x);

However, I agree that 'in' should be the opposite of 'out' (const 
scope ref). Making the 'in' and 'out' keywords semantically 
asymmetrical just to save a few key strokes is dumb.


 We have the longstanding problem, one must overload a function 
 to effectively bind both l- and r-values. That is what I'd 
 suppose to be the dual to "out".


+1 on allowing rvalue to bind to 'const scope ref'...or 'in' done 
right.
Jan 30 2017
prev sibling next sibling parent kinke <noone nowhere.com> writes:
On Monday, 30 January 2017 at 19:05:33 UTC, Q. Schroll wrote:

 Can't we make "in" mean "const scope ref", that binds on 
 r-values, too? Effectively, that's (similar to) what "const T&" 
 in C++ means. It's a non-copying const view on the object. We 
 have the longstanding problem, one must overload a function to 
 effectively bind both l- and r-values. That is what I'd suppose 
 to be the dual to "out".


+1000, I really love this proposal; I *hate* that this point is 
still missing in D in 2017.
It shouldn't even break existing code using `in` (but obviously 
the ABI) just because now a pointer to a const instance is passed 
instead of making a const copy.

Generic stuff:          void foo(in T arg)    // const scope ref 
=> safe `const T&`
By-value optimizations: void foo(const T arg)

I love it.
Jan 30 2017
prev sibling parent Chris Wright <dhasenan gmail.com> writes:
On Mon, 30 Jan 2017 19:05:33 +0000, Q. Schroll wrote:

 Would have been a far better definition. Why does anyone really need a
 shorthand attribute for two attributes that could be easily spelled out?


Because there was existing code thath used `in` and `scope const` was the 
nearest equivalent.
Jan 30 2017
prev sibling next sibling parent reply Adam D. Ruppe <destructionator gmail.com> writes:
On Monday, 30 January 2017 at 00:26:27 UTC, Walter Bright wrote:

 I was afraid that by checking it, too much code would break.


Code that was using it improperly was *already* broken. Now, the 
compiler will simply tell them, at compile time, why instead of 
letting it silently accept undefined behavior.
Jan 30 2017
parent Olivier FAURE <olivier.faure epitech.eu> writes:
On Monday, 30 January 2017 at 13:57:10 UTC, Adam D. Ruppe wrote:

 On Monday, 30 January 2017 at 00:26:27 UTC, Walter Bright wrote:

 I was afraid that by checking it, too much code would break.


 Code that was using it improperly was *already* broken. Now, 
 the compiler will simply tell them, at compile time, why 
 instead of letting it silently accept undefined behavior.


Well, it's a trade-off. Some people would rather their project 
with potentially broken code does not stop compiling because they 
upgraded their compiler.

Although I guess you could solve this by having -dip1000 emit 
only warnings and no error until the adaptation period had passed.
Jan 30 2017
prev sibling parent Soulsbane <paul acheronsoft.com> writes:
On Monday, 30 January 2017 at 00:26:27 UTC, Walter Bright wrote:

 On 1/29/2017 3:50 AM, Nordlöw wrote:

 Why is `in` no longer the same as `const scope`?


 Because it was unchecked and largely unimplemented. I was 
 afraid that by checking it, too much code would break.


Unless I'm completely missing something the documentation still 
says:

http://dlang.org/spec/function.html#parameters
in	equivalent to const scope
Jan 30 2017