digitalmars.D - Paging Walter Bright: Contextual Keywords
- Meta (16/16) Nov 06 2016 Hello Walter,
- deadalnix (3/13) Nov 06 2016 Why not make it a plain identifier ? We give identifier special
- Stefan Koch (2/17) Nov 06 2016 If I am not mistaken we do not need body ay at all.
- Dicebot (17/21) Nov 07 2016 protected-headers="v1"
- Timon Gehr (2/12) Nov 07 2016 What's the new syntax for contracts if body is removed?
Hello Walter, I'm currently working on a DIP on the topic of either removing or replacing the `body` keyword in D's contract programming syntax. I've found that the best possible way to do this, without any breaking changes to user code, would be to turn `body` into a contextual keyword. I know that you've expressed your opposition to contextual keywords in the past, but it's hard to find any concrete (and more important, recent) posts stating it clearly (as well as your reasons), other than this one: http://forum.dlang.org/post/npsp8a$mv4$1 digitalmars.com Can I ask you to give a definitive statement on contextual keywords in D (whether support, oppose, or are neutral to their addition), as well as a bit of rationale? This will then give me something official to link to in the DIP (and we will able to update the D FAQ with it as well). Link to DIP: https://github.com/dlang/DIPs/pull/48
Nov 06 2016
On Monday, 7 November 2016 at 03:33:47 UTC, Meta wrote:Hello Walter, I'm currently working on a DIP on the topic of either removing or replacing the `body` keyword in D's contract programming syntax. I've found that the best possible way to do this, without any breaking changes to user code, would be to turn `body` into a contextual keyword. I know that you've expressed your opposition to contextual keywords in the past, but it's hard to find any concrete (and more important, recent) posts stating it clearly (as well as your reasons), other than this one: http://forum.dlang.org/post/npsp8a$mv4$1 digitalmars.comWhy not make it a plain identifier ? We give identifier special meaning in various places, like scope(exit).
Nov 06 2016
On Monday, 7 November 2016 at 04:32:52 UTC, deadalnix wrote:On Monday, 7 November 2016 at 03:33:47 UTC, Meta wrote:If I am not mistaken we do not need body ay at all.Hello Walter, I'm currently working on a DIP on the topic of either removing or replacing the `body` keyword in D's contract programming syntax. I've found that the best possible way to do this, without any breaking changes to user code, would be to turn `body` into a contextual keyword. I know that you've expressed your opposition to contextual keywords in the past, but it's hard to find any concrete (and more important, recent) posts stating it clearly (as well as your reasons), other than this one: http://forum.dlang.org/post/npsp8a$mv4$1 digitalmars.comWhy not make it a plain identifier ? We give identifier special meaning in various places, like scope(exit).
Nov 06 2016
protected-headers="v1" From: Dicebot <public dicebot.lv> Newsgroups: d,i,g,i,t,a,l,m,a,r,s,.,D Subject: Re: Paging Walter Bright: Contextual Keywords References: <ckgmowedszvevzhvsnih forum.dlang.org> <nbsocfsjfcsnkxrvyavh forum.dlang.org> <ayffqkcdpyzlkbmdtahk forum.dlang.org> In-Reply-To: <ayffqkcdpyzlkbmdtahk forum.dlang.org> --c42qSUQScaJkCMOXXNlKTbvm637SxDH9a Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable On 11/07/2016 08:40 AM, Stefan Koch wrote:It is matter of how smooth deprecation phase is. Completely removal of `body` is OK long-term but that would mean one have to wait several years of deprecation phase before it can be used as identifier in user code. Making it deprecated _and_ contextual results in immediate availability. --c42qSUQScaJkCMOXXNlKTbvm637SxDH9a--Why not make it a plain identifier ? We give identifier special meaning in various places, like scope(exit).=20 If I am not mistaken we do not need body ay at all.
Nov 07 2016
On 07.11.2016 12:38, Dicebot wrote:On 11/07/2016 08:40 AM, Stefan Koch wrote:What's the new syntax for contracts if body is removed?It is matter of how smooth deprecation phase is. Completely removal of `body` is OK long-term but that would mean one have to wait several years of deprecation phase before it can be used as identifier in user code. Making it deprecated _and_ contextual results in immediate availability.Why not make it a plain identifier ? We give identifier special meaning in various places, like scope(exit).If I am not mistaken we do not need body ay at all.
Nov 07 2016
On Monday, 7 November 2016 at 13:24:20 UTC, Timon Gehr wrote:On 07.11.2016 12:38, Dicebot wrote:Currently https://github.com/dlang/DIPs/pull/48 (https://github.com/MetaLang/DIPs/blob/8f3ac3a133cd1f9095dd992f38ae377f5987c a4/DIPs/DIP1003.md) proposes either use `function` keyword instead of `body` or simply omit it with two blocks following each other.On 11/07/2016 08:40 AM, Stefan Koch wrote:What's the new syntax for contracts if body is removed?It is matter of how smooth deprecation phase is. Completely removal of `body` is OK long-term but that would mean one have to wait several years of deprecation phase before it can be used as identifier in user code. Making it deprecated _and_ contextual results in immediate availability.Why not make it a plain identifier ? We give identifier special meaning in various places, like scope(exit).If I am not mistaken we do not need body ay at all.
Nov 07 2016
I'm not sure if there's a point as Walter exclusively uses a mail client, but I'm bumping this back to the first page anyway.
Nov 09 2016