This is again a step closer to the book, but there are some notable
differences:
* Only constants encountered by the compiler are interned, all other
string operations (well, concatenation) happen with heap objects.
* OpReturn will always ensure that a returned string value is newly
heap allocated and does not reference the interner.
Change-Id: If4f04309446e01b8ff2db51094e9710d465dbc50
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2582
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This is based on this matklad post:
https://matklad.github.io/2020/03/22/fast-simple-rust-interner.html
It's modified slightly to provide a safer interface and slightly more
readable implementation:
* interned string IDs are wrapped in a newtype that is not publicly
constructible
* unsafe block is reduced to only the small scope in which it is
needed
* lookup lifetime is pinned explicitly to make the intent clearer when
reading this code
Change-Id: Ia3dae988f33f8e5e7d8dc0c1a9216914a945b036
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2578
Tested-by: BuildkiteCI
Reviewed-by: tazjin <mail@tazj.in>
... including concatenation.
This diverges significantly from the book, as I'm using std::String
instead of implementing the book's whole heap object management
system.
It's possible that Lox in Rust actually doesn't need a GC and the
ownership model works just fine.
Change-Id: I374a0461d627cfafc26b2b54bfefac8b7c574d00
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2577
Tested-by: BuildkiteCI
Reviewed-by: tazjin <mail@tazj.in>
This makes it possible to specify the input & output types of the
binary_op macro. If only one type is specified, it is assumed that the
input and output types are the same.
Change-Id: Idfcc9ba462db3976b69379b6693d091e1a525a3b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2573
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Adds support for true, false & nil. These each come with a new
separate opcode and are pushed directly on the stack.
Change-Id: I405b5b09496dcf99d514d3411c083e0834377167
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2571
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Introduces a new enum which represents the different types of possible
values, and modifies the rest of the existing code to wrap/unwrap
these enum variants correctly.
Notably in the vm module, a new macro has been introduced that makes
it possible to encode a type expectation and return a runtime error in
case of a type mismatch.
Change-Id: I325b5e31e395c62d8819ab2af6d398e1277333c0
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2570
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
If I was adding any dependencies, this might be a good one for a
property-based test thing, but I'm not going to.
Change-Id: Ia801d041479d1a88c59ef9e0fe1460b3640382e3
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2569
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Without this fix we would keep parsing in the same precedence level
and get weird things like:
10 - -10 + 10
=> 10
Change-Id: If2bed4569fbf566027011037165a9b3c09b7427c
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2567
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This should clean up everything in the way of actually running this
end-to-end.
Change-Id: Ie89d82472a458256a251a4fddc1c36d88d21f5f2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2563
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Defines a new precedence levels enum which can be used to restrict the
parser precedence in any given location. As an example, unary
expressions and grouping are implemented, as these have a different
precedence from e.g. expression()
Change-Id: I91f299fc77530f76c3aba717f638985428104ee5
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2558
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This lets us suppress reporting of additional errors from the compiler
until a synchronisation point is reached.
Change-Id: Iacf90949f868fbdb4349750065b5e458cf74d32a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2557
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This one necessarily has to diverge more from the book than the
treewalk interpreter did, so some of this is expected to change, but
I'm happy with the rough shape.
Since we're reusing the old scanner, the compiler/parser struct owns
an iterator over all tokens with which the pull-scanner from the
bytecode chapters is simulated.
Change-Id: Icfa0bd4729d9df786e08f7e49a25cba1b9989a91
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2556
Tested-by: BuildkiteCI
Reviewed-by: tazjin <mail@tazj.in>
This makes it easier to transition between the single/multi error
functions via ?
Change-Id: Ie027f4700da463a549be6f0d4a0022a9b8dc0d61
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2555
Tested-by: BuildkiteCI
Reviewed-by: tazjin <mail@tazj.in>
In the book, the clox interpreter has its own scanner which uses a
pull-based model for a single pass compiler.
I can't be bothered to write another scanner, or amend this one into
pull-mode to work with the treewalk interpreter, so instead I will
just reuse it and pull from a vector of tokens.
The tokens are shared between both interpreters and the scanner is not
what I'm interested in here.
Change-Id: Ib07e89127fce2b047f9b3e1ff7e9908d798b3b2b
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2420
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
It's unclear if the second part of the book can reuse anything from
the first part (I'm guessing probably the scanner, but I'll move that
back if it turns out to be the case).
Change-Id: I9411355929e31ac6e953599e51665406b1f48d55
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2415
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This is significantly simplified from the version in the book, since
I'm using Rust's Vec and not implementing dynamic arrays manually.
We'll see if I run into issues with that ...
Change-Id: Ie3446ac3884b850f3ba73a4b1a6ca14e68054188
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2413
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Right now this introduces a simple mechanism to flip between the
interpreters.
Change-Id: I92ee920c53d76ab6b664ac671993a6d6426af61a
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2412
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This makes the interpreter API a bit cleaner and allows for tighter
integration between the two parts (e.g. for static globals, which are
unhandled in the resolver right now).
Change-Id: I363714dc2e13cefa7731b54326573e0b871295d6
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2407
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
Implements the first part of the resolver from
https://craftinginterpreters.com/resolving-and-binding.html
This is wired up to the execution paths in main, but not yet in the
tests. The resolved depth is also not actually used for variable
lookups (yet).
Change-Id: I3a8615252b7b9b12d5a290c5ddf85988f61b9184
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2403
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
In the book this is implemented via exceptions as control flow, and
I'm sticking somewhat closely to that by doing it via an error
variant.
Change-Id: I9c7b84d6bb28265ab94021ea681df84f16a53da2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2395
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This removes the runtime dependency on a borrow into the program
source code.
It's not yet ideal because there are a lot of tokens where we really
don't care about the lexeme, but this is what the book does and I
am not going to change that.
Change-Id: I888e18f98597766d6f725cbf9241e8eb2bd839e2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2394
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
... with this, functions now work.
Note that this bubbled up another weird code structure nit: The
parser::Function type should probably not carry its name directly.
However this doesn't matter much and I don't care right now.
Change-Id: If8e3b23f07033260433b9acd45f37c0e61fd2ff8
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2393
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This slightly jiggles around interpret_block to let callers pass in an
environment.
Change-Id: I03112a38be0e8696242d8eae8d41da8c2cc66b48
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2392
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
In order to store a function in the interpreter's representation of a
callable, the lifetimes used throughout rlox need to be threaded
through properly.
This is currently not optimal, for two reasons:
* following the design of the book's scanner, the source code slice
needs to still be available at runtime. Rust makes this explicit,
but it seems unnecessary.
* the interpreter's lifetime is now bounded to be smaller than the
source's, which means that the REPL no longer persists state between
evaluations
Both of these can be fixed eventually by diverging the scanner from
the book slightly, but right now that's not my priority.
Change-Id: Id0bf694541ff59795cfdea3c64a965384a49bfe2
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2391
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
This is much easier to read & write. It's been annoying me all the way
through.
Change-Id: Ia91756d3111a2ce3f74e1c14bccc210118d221dd
Reviewed-on: https://cl.tvl.fyi/c/depot/+/2387
Reviewed-by: tazjin <mail@tazj.in>
Tested-by: BuildkiteCI
