forked from jaked/ocamljs
-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathTODO
53 lines (40 loc) · 1.38 KB
/
TODO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
gears
deprecate, move stuff to dom matching HTML5 spec
dom
fill in more of API
more examples
lwt-js
examples
Lwt 2.x
jslib
fill in more of parser
bug with locations
check pattern quotations
redo pretty-printing
examples
javascript
go through JS spec and wrap all the standard objects
jscomp
cleanup
cleaner heap rep (arrays, tag in last elem)
can we do better interoperating with exceptions? would be nice to be
able to declare, catch, and match exceptions from JS libs.
better checking in inline Javascript? easy to refer to unbound
vars. but sometimes useful or necessary.
can we avoid wrapping a function around for-loops when there are no
closures in the body? any other places to avoid extra wrapping?
stdlib
implement String.concat natively
other things to implement natively?
check float format (e.g. float_of_string) JS vs. OCaml
all
docs, wiki
bad ideas:
can we identify options to generate js val or null instead of
boxing? need to be careful about "if (val)" tests, where val could
be unboxed Javascript false value (0, "", etc.); must special case
tests to "if (val != null)".
turns out no. we can do the above (see null-options branch) but it
breaks for nested options, since we can't distinguish e.g. None and
Some None. a scheme to use different reps for different types must
work with polymorphic functions.