Assessment of attributes

[jam1garner]

Tracking issue for figuring out how attributes will translate into the binread/binwrite merge. The main focus being which binread attributes should "just work", which can be accessible for both but not together, etc.

Key:

• br - works for binread only
• bw - works for binwrite only
• br, bw - works for both, however a different value will need to be provided for each (example: map is a one way operation, so two maps will need to be provided to convert in both directions)
• brw - works for both using a single implementation. One example could be "align", where a single alignment value would allow for both reading and writing. Would also allow for separate implementations of br and bw if desired.

binread attributes

attribute	compatibility	more info
big	brw
little	brw
map	br, bw
magic	brw
assert	br, bw	up for debate, usefulness of brw questionable
pre_assert	br, bw	up for debate, usefulness of brw questionable
import	br, bw	up for debate, usefulness of brw questionable
args	br, bw
default	br	possibly should be removed in favor of ignore
ignore	brw
temp	br, bw
postprocess_now	br	dependent on the design of the binwrite redesign, this may have some usability
deref_now	br	ditto
restore_position	br, bw
try	br	binwrite likely has no use as the reverse should be the default behavior for Option<T> where T: BinWrite
parse_with	br
calc	br, bw
is_little	brw
is_big	brw
offset	br, bw(?)	needs discussion, not sure how this will work
if	br	should "just work" since it reads to an Option<T>
pad_before	brw
pad_after	brw
align_before	brw
align_after	brw
seek_before	brw
pad_size_to	brw
return_all_errors	br(?)	error handling probably needs a rework tbh

binwrite attributes

attribute	compatibility	more info
preprocessor	no	superseded by map
postprocessor	???	probably needs a rework
with	bw	rename to write_with, name pending
cstr	brw	a useful shortcut that probably should've carried over to binread
utf16	brw	ditto
utf16_null	brw	ditto
align	no	renamed to align_before
pad	no	renamed to pad_before

[roblabla]

Ideas for attributes to add/replace:

pre_assert is a bit of a hacky workaround to allow arbitrary placement of magics. It won't work nicely for a binwrite variant. Ideally, it should get replaced with something like magic_for, e.g.:

enum MyEnum {
    #[br(magic = "1u32")] X { /* ... */ },
    #[br(magic = "2u32")] Y { /* ... */ },
}

struct X {
    #[brw(magic_for = "data")]
    magic: u32,
    len: u32,
    data: MyEnum
}

If a magic_for attribute is present, the enum should not read the magic itself. Instead, the deserializer for X will provide it himself.

[jam1garner]

Another addition: #[br(dbg)]

is essentially equivalent to #[br(map = |x| dbg!(x))] for printing out values mid-parse in order to allow for debugging when parsing fails.

[kitlith]

possible suggestion for brw attributes: br,bw versions of them so that if (for some reason) you can override them seperately if needed.

tbh this is probably has dubious merit.

[jam1garner]

possible suggestion for brw attributes: br,bw versions of them so that if (for some reason) you can override them seperately if needed.

tbh this is probably has dubious merit.

Yep @kitlith, that's exactly what I meant by the following part of the key:

Would also allow for separate implementations of br and bw if desired.

[kitlith]

Ah, i missed that.

[jam1garner]

Another addition idea: #[br(take = $expr)], equivalent to "taking" $expr bytes. Which is to say, a reader adapter for limiting it to $expr bytes. Which would pretty much just involve temporarily replacing the reader with https://doc.rust-lang.org/std/io/trait.Read.html#method.take

[kitlith]

@jam1garner take as an attribute is a good idea that I've found myself reaching for in the past, but it's not quite that simple as the standard take adapter doesn't implement Seek (so iirc we'd run into trait constraint issues), and we probably want to allow FilePtr to (optionally?) read outside the taken area. The latter, in particular, probably causes the most issues with an adapter based approach?