mirror of
https://github.com/fluencelabs/asmble
synced 2025-04-25 06:42:22 +00:00
Update README explanation about string const max
This commit is contained in:
parent
a66c05ad4a
commit
1127b61eb5
@ -373,7 +373,9 @@ and the JVM:
|
||||
|
||||
* WebAssembly has a nice data section for byte arrays whereas the JVM does not. Right now we use a single-byte-char
|
||||
string constant (i.e. ISO-8859 charset). This saves class file size, but this means we call `String::getBytes` on
|
||||
init to load bytes from the string constant.
|
||||
init to load bytes from the string constant. Due to the JVM using an unsigned 16-bit int as the string constant
|
||||
length, the maximum length is 65536, so we chunk data sections into as many max-65500-byte lengths we need to load it
|
||||
all.
|
||||
* The JVM makes no guarantees about trailing bits being preserved on NaN floating point representations like WebAssembly
|
||||
does. This causes some mismatch on WebAssembly tests depending on how the JVM "feels" (I haven't dug into why some
|
||||
bit patterns stay and some don't when NaNs are passed through methods).
|
||||
|
Loading…
x
Reference in New Issue
Block a user