fix: typed const conversion validation #3117
Conversation
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #3117 +/- ##
==========================================
- Coverage 63.79% 63.65% -0.14%
==========================================
Files 549 549
Lines 78819 79068 +249
==========================================
+ Hits 50279 50333 +54
- Misses 25150 25326 +176
- Partials 3390 3409 +19
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. |
petar-dambovaliev
requested review from
mvertes,
zivkovicmilos,
jaekwon,
ltzmaxwell,
ajnavarro,
tbruyelle,
gfanton,
notJoon,
piux2 and
r3v4s
| x := uint(tv.GetInt8()) | ||
| tv.T = t | ||
| tv.SetUint(x) | ||
| case Uint8Kind: | ||
| validate(Int8Kind, Uint8Kind, func() bool { return tv.GetInt8() >= 0 }) |
There was a problem hiding this comment.
When an int type value exceeding 255 comes in, it will exceed the range of uint8. I think overflow error should take precedence over convertable error - how do you think?
| @@ -129,6 +129,136 @@ func TestBuiltinIdentifiersShadowing(t *testing.T) { | |||
| } | |||
| } | |||
|
|
|||
| func TestConvertTo(t *testing.T) { | |||
There was a problem hiding this comment.
Nitpicking: this function should be located in values_conversions_tests.go instead of gno_tests.go.
| @@ -796,6 +796,6 @@ func (m *Machine) doOpFuncLit() { | |||
| func (m *Machine) doOpConvert() { | |||
| xv := m.PopValue() | |||
| t := m.PopValue().GetType() | |||
| ConvertTo(m.Alloc, m.Store, xv, t) | |||
| ConvertTo(m.Alloc, m.Store, xv, t, false) | |||
There was a problem hiding this comment.
Here, we don't know if xv is a constant or not. Forcing false for isConst parameter here is incorrect for constant expressions. For example, the following sample should fail but it doesn't:
println(int(float64(1 + 0.5)))
During preprocessing, validates if typed constants are convertible.
Fixes issue
Typed constants are convertible only in a lossless way.
That means that we can convert floats to integers if the fractional part of the float is 0.