0
1
mirror of https://github.com/golang/go synced 2025-11-05 01:08:43 +00:00
Files
go/test/convert5.out
David Chase 0e466a8d1d cmd/compile: modify float-to-[u]int so that amd64 and arm64 match
Eventual goal is that all the architectures agree, and are
sensible.  The test will be build-tagged to exclude
not-yet-handled platforms.

This change also bisects the conversion change in case of bugs.
(`bisect -compile=convert ...`)

Change-Id: I98528666b0a3fde17cbe8d69b612d01da18dce85
Reviewed-on: https://go-review.googlesource.com/c/go/+/691135
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Keith Randall <khr@google.com>
2025-10-09 08:23:21 -07:00

38 lines
634 B
Plaintext

Below this are 'golden' results to check for consistency across platforms. Overflow behavior is not necessarily what we want
uint8 conversions
minus1_32 255
minus1_64 255
p32_plus4k_plus1 255
p64_plus4k_plus1 255
n32_minus4k 0
n64_minus4k 0
inf_32 255
inf_64 255
ninf_32 0
ninf_64 0
u64_plus4k_plus1_64 255
u64_plust44_plus1_32 255
u8plus2 1
p8minus1 126
n8plus2 130
n8minusone 127
int8 conversions
minus1_32 -1
minus1_64 -1
p32_plus4k_plus1 -1
p64_plus4k_plus1 -1
n32_minus4k 0
n64_minus4k 0
inf_32 -1
inf_64 -1
ninf_32 0
ninf_64 0
u64_plus4k_plus1_64 -1
u64_plust44_plus1_32 -1
u8plus2 1
p8minus1 126
n8plus2 -126
n8minusone 127