You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
in iKe, while triggering the bug described in #96, there is a second bug that I noticed having to do with copy-on-write when i tried to make a backup of the original array that was being amended:
assert:{[err;t] $[&/t;1;err[]]}
isFlat:~|/0=@:' / x contains no lists
a: 1 1 1 1
t: `
tick:{
assert[`flat0; isFlat a]
t::a / `t` should be a copy of `a` *before* the assignment,
/ but if you inspect it, `t` is `a` *after* the following assignment:
a[,1]::,0
assert[`flat1; isFlat a]}
draw:{}
Note that isFlat a fails on the second run, so the assignment t::a only executes once. (Either it is constructing a view (rather than doing a global assignment), or there is something wrong with copy-on-write.
The workaround was to say t::(),a instead.
The text was updated successfully, but these errors were encountered:
in iKe, while triggering the bug described in #96, there is a second bug that I noticed having to do with copy-on-write when i tried to make a backup of the original array that was being amended:
Note that
isFlat a
fails on the second run, so the assignmentt::a
only executes once. (Either it is constructing a view (rather than doing a global assignment), or there is something wrong with copy-on-write.The workaround was to say
t::(),a
instead.The text was updated successfully, but these errors were encountered: