diff options
Diffstat (limited to 'lux-jvm/source/luxc/lang/translation/jvm/loop.lux')
| -rw-r--r-- | lux-jvm/source/luxc/lang/translation/jvm/loop.lux | 16 |
1 files changed, 8 insertions, 8 deletions
diff --git a/lux-jvm/source/luxc/lang/translation/jvm/loop.lux b/lux-jvm/source/luxc/lang/translation/jvm/loop.lux index bd1e29d0f..85187815d 100644 --- a/lux-jvm/source/luxc/lang/translation/jvm/loop.lux +++ b/lux-jvm/source/luxc/lang/translation/jvm/loop.lux @@ -44,14 +44,14 @@ list.enumeration (list@map (function (_ [register argument]) [(n.+ start register) argument])))] - ## It may look weird that first I compile the values separately, - ## and then I compile the stores/allocations. - ## It must be done that way in order to avoid a potential bug. - ## Let's say that you'll recur with 2 expressions: X and Y. - ## If Y depends on the value of X, and you don't compile values - ## and stores separately, then by the time Y is evaluated, it - ## will refer to the new value of X, instead of the old value, and - ## shouldn't be the case. + ... It may look weird that first I compile the values separately, + ... and then I compile the stores/allocations. + ... It must be done that way in order to avoid a potential bug. + ... Let's say that you'll recur with 2 expressions: X and Y. + ... If Y depends on the value of X, and you don't compile values + ... and stores separately, then by the time Y is evaluated, it + ... will refer to the new value of X, instead of the old value, and + ... shouldn't be the case. valuesI+ (monad.map @ (function (_ [register argS]) (: (Operation Inst) (if (invariant? register argS) |