digitalmars.D.bugs - [Issue 5812] New: Power expression optimisation: constant fold (x^^0) = 1
- d-bugmail puremagic.com (45/45) Apr 03 2011 http://d.puremagic.com/issues/show_bug.cgi?id=5812
- d-bugmail puremagic.com (11/11) Apr 04 2011 http://d.puremagic.com/issues/show_bug.cgi?id=5812
- d-bugmail puremagic.com (7/7) Apr 05 2011 http://d.puremagic.com/issues/show_bug.cgi?id=5812
- d-bugmail puremagic.com (12/12) Apr 05 2011 http://d.puremagic.com/issues/show_bug.cgi?id=5812
- d-bugmail puremagic.com (12/12) Apr 19 2011 http://d.puremagic.com/issues/show_bug.cgi?id=5812
http://d.puremagic.com/issues/show_bug.cgi?id=5812
Summary: Power expression optimisation: constant fold (x^^0) =
1
Product: D
Version: D2
Platform: All
OS/Version: All
Status: NEW
Severity: normal
Priority: P2
Component: DMD
AssignedTo: nobody puremagic.com
ReportedBy: ibuclaw ubuntu.com
--- Comment #0 from Iain Buclaw <ibuclaw ubuntu.com> 2011-04-03 18:06:47 PDT ---
CTFE testcases (that currently fail)
const x = 3;
static assert((7 ^^ 0) == 1);
static assert((7 ^^ 0.0) == 1);
static assert((x ^^ 0) == 1);
static assert((x ^^ 0.0) == 1);
Patch:
--- dmd.orig/expression.c 2011-02-18 01:15:38.000000000 +0000
+++ dmd/expression.c 2011-04-04 02:05:38.631620650 +0100
-10302,6 +10302,15 Expression *PowExp::semantic(Scope *sc)
e = e->semantic(sc);
return e;
}
+ // Replace x ^^ 0 or x^^0.0 by (x, 1)
+ if ((e2->op == TOKint64 && e2->toInteger() == 0) ||
+ (e2->op == TOKfloat64 && e2->toReal() == 0.0))
+ {
+ typeCombine(sc);
+ e = new CommaExp(loc, e1, new IntegerExp(loc, 1, Type::tint32));
+ e = e->semantic(sc);
+ return e;
+ }
// Replace -1 ^^ x by (x&1) ? -1 : 1, where x is integral
if (e2->type->isintegral() && e1->op == TOKint64 &&
(sinteger_t)e1->toInteger() == -1L)
{
Feel free to improve it.
Regards
--
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Apr 03 2011
d-bugmail puremagic.com http://d.puremagic.com/issues/show_bug.cgi?id=5812 --- Comment #1 from Iain Buclaw <ibuclaw ubuntu.com> 2011-04-04 11:32:56 PDT --- You can also optimize: x^^1 => (x) x^^-1 => (1/x) And that will remove two other scenarios from requiring std.math to compute the result. Regards -- Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email ------- You are receiving this mail because: -------
Apr 04 2011
http://d.puremagic.com/issues/show_bug.cgi?id=5812 --- Comment #2 from Iain Buclaw <ibuclaw ubuntu.com> 2011-04-05 16:00:43 PDT --- Created an attachment (id=937) implement noted foldable code in this report -- Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email ------- You are receiving this mail because: -------
Apr 05 2011
http://d.puremagic.com/issues/show_bug.cgi?id=5812
Iain Buclaw <ibuclaw ubuntu.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Attachment #937 is|0 |1
obsolete| |
--- Comment #3 from Iain Buclaw <ibuclaw ubuntu.com> 2011-04-05 16:08:49 PDT ---
Created an attachment (id=938)
just noticed previous patch allows to bypass an error - fixed
--
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Apr 05 2011
http://d.puremagic.com/issues/show_bug.cgi?id=5812
Walter Bright <bugzilla digitalmars.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
CC| |bugzilla digitalmars.com
Resolution| |FIXED
--- Comment #4 from Walter Bright <bugzilla digitalmars.com> 2011-04-19
02:21:06 PDT ---
https://github.com/D-Programming-Language/dmd/commit/c64122e84c1ad304136a39acfd3912caec70fc1d
--
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Apr 19 2011