I have some ocaml code that reads in a binary file as a string, and then treats it as an array of bytes. This led to some odd errors when I tried to compile it with jsoo, which I eventually traced down to the fact that javascript strings aren’t really suited for the purpose, and I should be using an ArrayBuffer instead. Is there any nice way to create an interface that will compile to the string code in ocamlc and to Typed_array code if compiled with jsoo?
2 Likes
Maybe you could get more help if you explain what where you trying to do and what were the errors in question . AFAIK, OCaml-strings-as-byte-buffers will work regardless of which compiler is used.
I’m trying to load a binary file as a string, and parse it using bitstring. when i try reading the file as a string in javascript, I see a few extra bytes inserted, which makes me think that when I call ReadAsBinaryString, javascript is using a string encoding that spills some characters into two bytes.
Furthermore, when I was googling to see if any one else had had this issue, I saw that ReadAsBinaryString has been deprecated and that I’m supposed to read the file into an ArrayBuffer instead. I’ll experiment with converting the ArrayBuffer to a string and passing it to the original code, but if that doesn’t work either some sort of #ifdef js_of_ocaml equivalent would be useful.
this stackoverflow comment seems relevant:
The reason behind readAsBinaryString() deprecation is in my opinion the following: the standard for JavaScript strings are DOMString which only accept UTF-8 characters, NOT random binary data. So don’t use readAsBinaryString(), that’s not safe and ECMAScript-compliant at all.
poking around the jsoo source I found that runtime/mlString lets me directly create an ocaml string backed by a byte array, I’m now trying to get that playing nicely with gen_js_api, which I believe will solve my issue.
code snippets, for anyone else facing this issue:
this from the js.mli file (processed by gen_js_api):
module ArrayBuffer: sig
type t
val t_of_js: Ojs.t -> t
val t_to_js: t -> Ojs.t
val byte_length: t -> int
end
module Uint8Array: sig
type t
val t_of_js: Ojs.t -> t
val t_to_js: t -> Ojs.t
val new_uint8_array: ArrayBuffer.t -> t [@@js.new]
val to_string: t -> string [@@js.cast]
end
and this from the ml file:
external str_of_uint8Array: Js.Uint8Array.t -> string = "caml_string_of_array"
let str_of_arraybuffer buf =
let arr = Js.Uint8Array.new_uint8_array buf in
str_of_uint8Array arr