Parsing strings from binary data

[ZevEisenberg]

I'm parsing a portion of Data blob representing a PDF file. I want to grab a portion near the end of the file that looks like this:

...
some decidedly non-textual binary PDF data, so I don't want to just convert the whole thing to a string.
...
trailer
<< /Size 13 /Root 11 0 R /Info 12 0 R /ID [ <00efb601784c299c9d9082f168910a7e>
<00efb601784c299c9d9082f168910a7e> ] >>
startxref
6265391
%%EOF

I'm specifically looking to get the locations of the two 32-character ID strings, which I'm going to partly do with an updated version of this answer: #69 (comment).

But: how can I parse string-like sequences out of a blob of Data, which is a Collection of UInt8? It thought I might be able to use, for example, "trailer".utf8, but that doesn't seem to work. I found the Data parsing example from the benchmark suite, but that covers parsing individual bytes, not readable strings.

[ZevEisenberg]

To give a simpler example, say I had the following text, encoded as UTF-8 Data:

I want <<this text right here>> and nothing more!

If I wanted to make a parser that took Data in and returned the text between the << and >> as Data, without my having to write a bunch of bytes out the old fashioned way, how might I do that?

[mbrandonw]

All of the parsers that come with the library are super generic and can work on almost any kind of collection (usually where C.SubSequence == C, e.g. Substring, UTF8View, etc.). And Data is such a collection, though ArraySlice<UInt8> would probably be a little more efficient to deal with.

And so when you form a parser like this:

PrefixThrough(Data("<<".utf8))

…that represents consuming all of the bytes up to and including the first occurrence of the subsequence of "<<" bytes in the input. And the PrefixUpTo parser does something similar, except it does not consume the delimiter bytes.

So, in your simplified example, you might want something like this:

import Testing
@Test func parser() throws {
  let input = Data("""
    I want <<this text right here>> and nothing more!
    """.utf8)

  let parser: some Parser<Data, Data> = Parse {
    Skip { PrefixThrough(Data("<<".utf8)) }
    PrefixUpTo(Data(">>".utf8))
    Skip { Rest() }
  }

  let output = try parser.parse(input)
  #expect(String(decoding: output, as: UTF8.self) == "this text right here")
}

You can adapt this code a bit to use ArraySlice too:

let parser: some Parser<ArraySlice<UInt8>, ArraySlice<UInt8>> = Parse {
  Skip { PrefixThrough(ArraySlice("<<".utf8)) }
  PrefixUpTo(ArraySlice(">>".utf8))
  Skip { Rest() }
}

[ZevEisenberg]

Thank you! The part I was missing was using PrefixThrough and Data in concert like that.