digitalmars.D.learn - Why this code can't take advantage from CTFE?
- Andrea Fontana (19/19) Feb 03 2016 This code:
- anonymous (8/20) Feb 03 2016 The compiler doesn't try to apply CTFE everywhere, because it could take...
- Adam D. Ruppe (6/9) Feb 03 2016 You never asked for CTFE.
- Messenger (5/8) Feb 03 2016 What is a good way to try to force it? Using enum? Then
- Adam D. Ruppe (2/5) Feb 03 2016 Yes, or static local variables are ctfe initialized too.
- Andrea Fontana (4/9) Feb 03 2016 But if I want a normal var, not static or enum?
- Messenger (3/5) Feb 03 2016 Just make a mutable copy of it, I think.
- Steven Schveighoffer (4/13) Feb 03 2016 Hm... trick you could do is:
- Marc =?UTF-8?B?U2Now7x0eg==?= (5/8) Feb 03 2016 template forceCTFE(alias expr) {
- Andrea Fontana (3/11) Feb 03 2016 Interesting :)
- Timon Gehr (4/17) Feb 03 2016 I would use enum forceCTFE(alias expr)=expr; though. With alias it won't...
- Andrea Fontana (31/34) Feb 05 2016 I've a code that build a JSON object using a wrapper over
- Luis (3/23) Feb 05 2016 Reading/parsing a JSON or a XML using std.json / std.xml could be
- Andrea Fontana (9/11) Feb 05 2016 parseJSON() from std.json doesn't work with CTFE.
This code:
import std.stdio;
int very_very_long_function(in int k)
{
if (!__ctfe) writeln("Can't use ctfe!");
return k/2;
}
void main()
{
enum first = very_very_long_function(10);
writeln("First is ", first);
auto second = very_very_long_function(12);
writeln("Second is ", second);
auto third = first;
third += 1;
writeln("Third is ", third);
}
Why second init doesn't work with CTFE? It could be something
like third, but with one less copy. What am I missing?
Feb 03 2016
anonymous <anonymous example.com>
On 03.02.2016 16:34, Andrea Fontana wrote:
void main()
{
enum first = very_very_long_function(10);
writeln("First is ", first);
auto second = very_very_long_function(12);
writeln("Second is ", second);
auto third = first;
third += 1;
writeln("Third is ", third);
}
Why second init doesn't work with CTFE? It could be something like
third, but with one less copy. What am I missing?
The compiler doesn't try to apply CTFE everywhere, because it could take
forever. So CTFE is only done when the programmer requests it by using a
function in a static context.
The compiler is still free to precompute the value at compile time, if
it thinks that's a good idea, but that's just an optimization. And while
it's evaluation of a function at compile time, it's not CTFE in the
sense of __ctfe.
Feb 03 2016
On Wednesday, 3 February 2016 at 15:34:51 UTC, Andrea Fontana
wrote:
enum first = very_very_long_function(10);
auto second = very_very_long_function(12);
Why second init doesn't work with CTFE?
You never asked for CTFE.
CTFE only happens when it *has* to - when you write code that
specifically asks for or requires it.
Ordinary variables don't require ctfe, so they don't get it.
Feb 03 2016
On Wednesday, 3 February 2016 at 15:59:43 UTC, Adam D. Ruppe wrote:You never asked for CTFE. CTFE only happens when it *has* to - when you write code that specifically asks for or requires it.What is a good way to try to force it? Using enum? Then optionally copying the value once to avoid the "manifest constant" copy/paste behaviour, where applicable?
Feb 03 2016
On Wednesday, 3 February 2016 at 16:07:59 UTC, Messenger wrote:What is a good way to try to force it? Using enum? Then optionally copying the value once to avoid the "manifest constant" copy/paste behaviour, where applicable?Yes, or static local variables are ctfe initialized too.
Feb 03 2016
On Wednesday, 3 February 2016 at 16:24:19 UTC, Adam D. Ruppe wrote:On Wednesday, 3 February 2016 at 16:07:59 UTC, Messenger wrote:But if I want a normal var, not static or enum? There's no way to force it?What is a good way to try to force it? Using enum? Then optionally copying the value once to avoid the "manifest constant" copy/paste behaviour, where applicable?Yes, or static local variables are ctfe initialized too.
Feb 03 2016
On Wednesday, 3 February 2016 at 17:16:30 UTC, Andrea Fontana wrote:But if I want a normal var, not static or enum? There's no way to force it?Just make a mutable copy of it, I think.
Feb 03 2016
On 2/3/16 12:16 PM, Andrea Fontana wrote:On Wednesday, 3 February 2016 at 16:24:19 UTC, Adam D. Ruppe wrote:Hm... trick you could do is: auto myvar = AliasSeq!(someCTFEFunc())[0]; -SteveOn Wednesday, 3 February 2016 at 16:07:59 UTC, Messenger wrote:But if I want a normal var, not static or enum? There's no way to force it?What is a good way to try to force it? Using enum? Then optionally copying the value once to avoid the "manifest constant" copy/paste behaviour, where applicable?Yes, or static local variables are ctfe initialized too.
Feb 03 2016
On Wednesday, 3 February 2016 at 16:07:59 UTC, Messenger wrote:What is a good way to try to force it? Using enum? Then optionally copying the value once to avoid the "manifest constant" copy/paste behaviour, where applicable?template forceCTFE(alias expr) { alias forceCTFE = expr; } auto value = forceCTFE!(some_expression());
Feb 03 2016
On Wednesday, 3 February 2016 at 17:49:39 UTC, Marc Schütz wrote:On Wednesday, 3 February 2016 at 16:07:59 UTC, Messenger wrote:Interesting :) Thank you!What is a good way to try to force it? Using enum? Then optionally copying the value once to avoid the "manifest constant" copy/paste behaviour, where applicable?template forceCTFE(alias expr) { alias forceCTFE = expr; } auto value = forceCTFE!(some_expression());
Feb 03 2016
On 02/03/2016 11:39 PM, Andrea Fontana wrote:On Wednesday, 3 February 2016 at 17:49:39 UTC, Marc Schütz wrote:I would use enum forceCTFE(alias expr)=expr; though. With alias it won't force compile-time evaluation of expressions that can be interpreted as symbols.On Wednesday, 3 February 2016 at 16:07:59 UTC, Messenger wrote:Interesting :) Thank you!What is a good way to try to force it? Using enum? Then optionally copying the value once to avoid the "manifest constant" copy/paste behaviour, where applicable?template forceCTFE(alias expr) { alias forceCTFE = expr; } auto value = forceCTFE!(some_expression());
Feb 03 2016
On Wednesday, 3 February 2016 at 22:45:47 UTC, Timon Gehr wrote:I would use enum forceCTFE(alias expr)=expr; though. With alias it won't force compile-time evaluation of expressions that can be interpreted as symbols.I've a code that build a JSON object using a wrapper over std.json. This code: ---- enum forceCTFE(alias expr)=expr; auto j = forceCTFE!( JSOB ( "hello", JSOB("world", 3), "arr", JSAB("hello", JSOB("world", 1)) ) ); j.put("/hello/world", "!"); j.put("/hello/blah", 42); // Here j == {"arr":["hello",{"world":1}],"hello":{"blah":42,"world":"!"}} ---- shows a 10% performance gain over this: ---- auto j = JSOB ( "hello", JSOB("world", 3), "arr", JSAB("hello", JSOB("world", 1)) ); j.put("/hello/world", "!"); j.put("/hello/blah", 42); ---- (inside a loop). And it's just a small json in this case. That's pretty good IMHO :)
Feb 05 2016
On Friday, 5 February 2016 at 09:36:37 UTC, Andrea Fontana wrote:On Wednesday, 3 February 2016 at 22:45:47 UTC, Timon Gehr wrote:Reading/parsing a JSON or a XML using std.json / std.xml could be done on CTFE ?I would use enum forceCTFE(alias expr)=expr; though. With alias it won't force compile-time evaluation of expressions that can be interpreted as symbols.I've a code that build a JSON object using a wrapper over std.json. This code: ---- enum forceCTFE(alias expr)=expr; auto j = forceCTFE!( JSOB ( "hello", JSOB("world", 3), "arr", JSAB("hello", JSOB("world", 1)) ) ); j.put("/hello/world", "!"); j.put("/hello/blah", 42); // Here j == {"arr":["hello",{"world":1}],"hello":{"blah":42,"world":"!"}}
Feb 05 2016
On Friday, 5 February 2016 at 09:49:38 UTC, Luis wrote:Reading/parsing a JSON or a XML using std.json / std.xml could be done on CTFE ?parseJSON() from std.json doesn't work with CTFE. But I can build objects with with my code that works over std.json. So if you convert (with mixins) { "hello" : "world" } to JSOB("hello", "world) You can parse it ar compile time. :)
Feb 05 2016