From d47907f725947fdb40ad7f39c505cb3cfcc4fd3e Mon Sep 17 00:00:00 2001 From: "A. Wilcox" Date: Mon, 11 Jun 2018 00:40:08 -0500 Subject: system/musl: more complex fixes --- system/musl/complex-math2.patch | 79 +++++++++++++++++++++++++++++++++++++++++ 1 file changed, 79 insertions(+) create mode 100644 system/musl/complex-math2.patch (limited to 'system/musl/complex-math2.patch') diff --git a/system/musl/complex-math2.patch b/system/musl/complex-math2.patch new file mode 100644 index 000000000..49f5c3acd --- /dev/null +++ b/system/musl/complex-math2.patch @@ -0,0 +1,79 @@ +From 10e4bd3780050e75b72aac5d85c31816419bb17d Mon Sep 17 00:00:00 2001 +From: Rich Felker +Date: Wed, 11 Apr 2018 15:05:22 -0400 +Subject: [PATCH 1/2] fix incorrect results for catan with some inputs +MIME-Version: 1.0 +Content-Type: text/plain; charset=UTF-8 +Content-Transfer-Encoding: 8bit + +the catan implementation from OpenBSD includes a FIXME-annotated +"overflow" branch that produces a meaningless and incorrect +large-magnitude result. it was reachable via three paths, +corresponding to gotos removed by this commit, in order: + +1. pure imaginary argument with imaginary component greater than 1 in + magnitude. this case does not seem at all exceptional and is + handled (at least with the quality currently expected from our + complex math functions) by the existing non-exceptional code path. + +2. arguments on the unit circle, including the pure-real argument 1.0. + these are not exceptional except for ±i, which should produce + results with infinite imaginary component and which lead to + computation of atan2(±0,0) in the existing non-exceptional code + path. such calls to atan2() however are well-defined by POSIX. + +3. the specific argument +i. this route should be unreachable due to + the above (2), but subtle rounding effects might have made it + possible in rare cases. continuing on the non-exceptional code path + in this case would lead to computing the (real) log of an infinite + argument, then producing a NAN when multiplying it by I. + +for now, remove the exceptional code paths entirely. replace the +multiplication by I with construction of a complex number using the +CMPLX macro so that the NAN issue (3) prevented cannot arise. + +with these changes, catan should give reasonably correct results for +real arguments, and should no longer give completely-wrong results for +pure-imaginary arguments outside the interval (-i,+i). +--- + src/complex/catan.c | 14 +------------- + 1 file changed, 1 insertion(+), 13 deletions(-) + +diff --git a/src/complex/catan.c b/src/complex/catan.c +index 39ce6cf2..7dc2afeb 100644 +--- a/src/complex/catan.c ++++ b/src/complex/catan.c +@@ -91,29 +91,17 @@ double complex catan(double complex z) + x = creal(z); + y = cimag(z); + +- if (x == 0.0 && y > 1.0) +- goto ovrf; +- + x2 = x * x; + a = 1.0 - x2 - (y * y); +- if (a == 0.0) +- goto ovrf; + + t = 0.5 * atan2(2.0 * x, a); + w = _redupi(t); + + t = y - 1.0; + a = x2 + (t * t); +- if (a == 0.0) +- goto ovrf; + + t = y + 1.0; + a = (x2 + t * t)/a; +- w = w + (0.25 * log(a)) * I; +- return w; +- +-ovrf: +- // FIXME +- w = MAXNUM + MAXNUM * I; ++ w = CMPLX(w, 0.25 * log(a)); + return w; + } +-- +2.15.0 + -- cgit v1.2.3-70-g09d2