D issues are now tracked on GitHub. This Bugzilla instance remains as a read-only archive.
Issue 6447 - iota(BigInt) too
Summary: iota(BigInt) too
Status: REOPENED
Alias: None
Product: D
Classification: Unclassified
Component: phobos (show other issues)
Version: D2
Hardware: All All
: P4 enhancement
Assignee: No Owner
URL:
Keywords: pull
Depends on:
Blocks:
 
Reported: 2011-08-06 19:56 UTC by bearophile_hugs
Modified: 2024-12-01 16:14 UTC (History)
4 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this issue.
Description bearophile_hugs 2011-08-06 19:56:56 UTC
Sometimes I'd like to generate a range of bigints too:

Enhancement:

import std.bigint: BigInt;
import std.range: iota;
void main() {
    auto x = iota(BigInt(10));
}


Currently DMD gives:

...\std\range.d(4028): Error: template std.range.iota(B,E,S) if ((isIntegral!(CommonType!(B,E)) || isPointer!(CommonType!(B,E))) && isIntegral!(S)) does not match any function template declaration
...\std\range.d(4028): Error: template std.range.iota(B,E,S) if ((isIntegral!(CommonType!(B,E)) || isPointer!(CommonType!(B,E))) && isIntegral!(S)) cannot deduce template function from argument types !()(BigInt,BigInt)
test.d(4): Error: template instance std.range.iota!(BigInt) error instantiating
Comment 1 Jonathan M Davis 2011-08-06 20:05:50 UTC
BigInt isn't considered an integral by isIntegral, hence why it doesn't work. We should probably look at having a template in std.traits which covers both the built-in integral types and BigInt so that those functions which can work with both have an easy way to test for it. It's either that or every function that could use BigInt is going to have to test for it explicitly on top of testing for isIntegral.

Regardless, I don't see any reason why using iota with BigInt shouldn't be made to work.
Comment 2 Russel Winder 2013-05-28 03:37:57 UTC
I don't have the time, and likely not the knowledge, to fix this and provide a pull request. Is there anyone who does? This is a two year old problem that means D is not useful for applications using data types other than the hardware types, which is a lot of applications.

Is this a symptom of the fact that Phobos is really predicated on use of hardware types more generally?

(This comment is really to get me on the cc list as there seems to be a requirement to make a comment in order to add oneself to the cc list :-((
Comment 3 bearophile_hugs 2013-05-28 04:15:16 UTC
(In reply to comment #2)

> Is this a symptom of the fact that Phobos is really predicated on use of
> hardware types more generally?

I think it's mostly a symptom of D/Phobos not having many contributors :-)
Comment 4 hsteoh 2013-08-18 22:39:52 UTC
Related: issue #10762.
Comment 5 bearophile_hugs 2014-12-04 23:53:02 UTC
A workaround:

10.iota.map!BigInt
Comment 6 hsteoh 2014-12-05 00:12:34 UTC
That doesn't help when you actually *need* a BigInt, e.g., iota(BigInt(2)^^65536). Of course, you'd run into other problems with that (e.g., you'll be waiting a looong time for your program to finish), but the idea is that you want to iterate over numbers that only BigInt can represent, right?
Comment 7 bearophile_hugs 2014-12-05 00:23:49 UTC
(In reply to hsteoh from comment #6)
> That doesn't help when you actually *need* a BigInt, e.g.,
> iota(BigInt(2)^^65536). Of course, you'd run into other problems with that
> (e.g., you'll be waiting a looong time for your program to finish), but the
> idea is that you want to iterate over numbers that only BigInt can
> represent, right?

The idea is that I'd like iota(BigInt(10)) to work. In the meantime I use some workarounds.
Comment 8 hsteoh 2014-12-05 00:54:51 UTC
Yes, but if BigInt(10) should work, then BigInt(2)^^65536 should work too. I don't think you'd like it if the library imposed some arbitrary subrange on your numeric types that may be used with iota (what if iota(10) worked but iota(11) didn't?).
Comment 9 bearophile_hugs 2014-12-05 01:03:23 UTC
(In reply to hsteoh from comment #8)
> Yes, but if BigInt(10) should work, then BigInt(2)^^65536 should work too. I
> don't think you'd like it if the library imposed some arbitrary subrange on
> your numeric types that may be used with iota (what if iota(10) worked but
> iota(11) didn't?).

Yes, of course. I'd like iota to work with bigints in general :-)
Comment 10 hsteoh 2015-01-22 18:22:31 UTC
This seems to be a subset of https://issues.dlang.org/show_bug.cgi?id=10762
Comment 12 github-bugzilla 2015-01-26 09:27:04 UTC
Commit pushed to master at https://github.com/D-Programming-Language/phobos

https://github.com/D-Programming-Language/phobos/commit/b159a5bdc980abb90833b32fa04044a002dfc794
Merge pull request #2895 from quickfur/iota_bigint

Issue 6447 & 10762: support user-defined types in iota()
Comment 13 bearophile_hugs 2015-01-26 12:23:10 UTC
I keep this issue open because this still doesn't work:


void main() {
    import std.bigint: BigInt;
    import std.range: iota;
    iota(BigInt(1), BigInt(100), BigInt(5));
    iota(BigInt(1), BigInt(100), 5);
}



test.d(4,9): Error: template std.range.iota cannot deduce function from argument types !()(BigInt, BigInt, BigInt), candidates are:
...\dmd2\src\phobos\std\range\package.d(4008,6):        std.range.iota(B, E, S)(B begin, E end, S step) if ((isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E))) && isIntegral!S)
...\dmd2\src\phobos\std\range\package.d(4098,6):        std.range.iota(B, E)(B begin, E end) if (isFloatingPoint!(CommonType!(B, E)))
...\dmd2\src\phobos\std\range\package.d(4105,6):        std.range.iota(B, E)(B begin, E end) if (isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E)))
...\dmd2\src\phobos\std\range\package.d(4168,6):        std.range.iota(E)(E end)
...\dmd2\src\phobos\std\range\package.d(4176,6):        std.range.iota(B, E, S)(B begin, E end, S step) if (isFloatingPoint!(CommonType!(B, E, S)))
test.d(4,9):        ... (1 more, -v to show) ...
test.d(5,9): Error: template std.range.iota cannot deduce function from argument types !()(BigInt, BigInt, int), candidates are:
...\dmd2\src\phobos\std\range\package.d(4008,6):        std.range.iota(B, E, S)(B begin, E end, S step) if ((isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E))) && isIntegral!S)
...\dmd2\src\phobos\std\range\package.d(4098,6):        std.range.iota(B, E)(B begin, E end) if (isFloatingPoint!(CommonType!(B, E)))
...\dmd2\src\phobos\std\range\package.d(4105,6):        std.range.iota(B, E)(B begin, E end) if (isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E)))
...\dmd2\src\phobos\std\range\package.d(4168,6):        std.range.iota(E)(E end)
...\dmd2\src\phobos\std\range\package.d(4176,6):        std.range.iota(B, E, S)(B begin, E end, S step) if (isFloatingPoint!(CommonType!(B, E, S)))
test.d(5,9):        ... (1 more, -v to show) ...
Comment 14 hsteoh 2015-01-26 16:16:50 UTC
One thing at a time! The last PR implemented the iota(start,end) case. The next step is to extend it to handle the iota(start,end,step) case when += is supported.
Comment 15 github-bugzilla 2015-02-18 03:41:30 UTC
Commit pushed to 2.067 at https://github.com/D-Programming-Language/phobos

https://github.com/D-Programming-Language/phobos/commit/b159a5bdc980abb90833b32fa04044a002dfc794
Merge pull request #2895 from quickfur/iota_bigint
Comment 16 Dmitry Olshansky 2018-05-29 14:03:57 UTC
Works now.
Comment 17 Russel Winder 2018-06-05 10:22:47 UTC
Using LDC on Debian Sid, the code:

import std.bigint: BigInt;
import std.range: iota;
void main() {
    auto x = iota(BigInt(10));
}

compiles and executes fine, but the code:

void main() {
    import std.bigint: BigInt;
    import std.range: iota;
    iota(BigInt(1), BigInt(100), BigInt(5));
    iota(BigInt(1), BigInt(100), 5);
}

results in:

test_b.d(4): Error: template std.range.iota cannot deduce function from argument types !()(BigInt, BigInt, BigInt), candidates are:
/usr/lib/ldc/x86_64-linux-gnu/include/d/std/range/package.d(5371):        std.range.iota(B, E, S)(B begin, E end, S step) if ((isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E))) && isIntegral!S)
/usr/lib/ldc/x86_64-linux-gnu/include/d/std/range/package.d(5481):        std.range.iota(B, E)(B begin, E end) if (isFloatingPoint!(CommonType!(B, E)))
/usr/lib/ldc/x86_64-linux-gnu/include/d/std/range/package.d(5488):        std.range.iota(B, E)(B begin, E end) if (isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E)))
/usr/lib/ldc/x86_64-linux-gnu/include/d/std/range/package.d(5558):        std.range.iota(E)(E end) if (is(typeof(iota(E(0), end))))
/usr/lib/ldc/x86_64-linux-gnu/include/d/std/range/package.d(5567):        std.range.iota(B, E, S)(B begin, E end, S step) if (isFloatingPoint!(CommonType!(B, E, S)))
test_b.d(4):        ... (1 more, -v to show) ...
test_b.d(5): Error: template std.range.iota cannot deduce function from argument types !()(BigInt, BigInt, int), candidates are:
/usr/lib/ldc/x86_64-linux-gnu/include/d/std/range/package.d(5371):        std.range.iota(B, E, S)(B begin, E end, S step) if ((isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E))) && isIntegral!S)
/usr/lib/ldc/x86_64-linux-gnu/include/d/std/range/package.d(5481):        std.range.iota(B, E)(B begin, E end) if (isFloatingPoint!(CommonType!(B, E)))
/usr/lib/ldc/x86_64-linux-gnu/include/d/std/range/package.d(5488):        std.range.iota(B, E)(B begin, E end) if (isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E)))
/usr/lib/ldc/x86_64-linux-gnu/include/d/std/range/package.d(5558):        std.range.iota(E)(E end) if (is(typeof(iota(E(0), end))))
/usr/lib/ldc/x86_64-linux-gnu/include/d/std/range/package.d(5567):        std.range.iota(B, E, S)(B begin, E end, S step) if (isFloatingPoint!(CommonType!(B, E, S)))
test_b.d(5):        ... (1 more, -v to show) ...


Using dmd from d-apt the first code works fine and the second code results in:

test_b.d(4): Error: template `std.range.iota` cannot deduce function from argument types `!()(BigInt, BigInt, BigInt)`, candidates are:
/usr/include/dmd/phobos/std/range/package.d(5890):        `std.range.iota(B, E, S)(B begin, E end, S step) if ((isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E))) && isIntegral!S)`
/usr/include/dmd/phobos/std/range/package.d(6000):        `std.range.iota(B, E)(B begin, E end) if (isFloatingPoint!(CommonType!(B, E)))`
/usr/include/dmd/phobos/std/range/package.d(6007):        `std.range.iota(B, E)(B begin, E end) if (isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E)))`
/usr/include/dmd/phobos/std/range/package.d(6077):        `std.range.iota(E)(E end) if (is(typeof(iota(E(0), end))))`
/usr/include/dmd/phobos/std/range/package.d(6086):        `std.range.iota(B, E, S)(B begin, E end, S step) if (isFloatingPoint!(CommonType!(B, E, S)))`
test_b.d(4):        ... (1 more, -v to show) ...
test_b.d(5): Error: template `std.range.iota` cannot deduce function from argument types `!()(BigInt, BigInt, int)`, candidates are:
/usr/include/dmd/phobos/std/range/package.d(5890):        `std.range.iota(B, E, S)(B begin, E end, S step) if ((isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E))) && isIntegral!S)`
/usr/include/dmd/phobos/std/range/package.d(6000):        `std.range.iota(B, E)(B begin, E end) if (isFloatingPoint!(CommonType!(B, E)))`
/usr/include/dmd/phobos/std/range/package.d(6007):        `std.range.iota(B, E)(B begin, E end) if (isIntegral!(CommonType!(B, E)) || isPointer!(CommonType!(B, E)))`
/usr/include/dmd/phobos/std/range/package.d(6077):        `std.range.iota(E)(E end) if (is(typeof(iota(E(0), end))))`
/usr/include/dmd/phobos/std/range/package.d(6086):        `std.range.iota(B, E, S)(B begin, E end, S step) if (isFloatingPoint!(CommonType!(B, E, S)))`
test_b.d(5):        ... (1 more, -v to show) ...

So I think this problem is not fixed.
Comment 18 dlangBugzillaToGithub 2024-12-01 16:14:18 UTC
THIS ISSUE HAS BEEN MOVED TO GITHUB

https://github.com/dlang/phobos/issues/9909

DO NOT COMMENT HERE ANYMORE, NOBODY WILL SEE IT, THIS ISSUE HAS BEEN MOVED TO GITHUB