Mike Gerwitz
7f71f3f09f
This simply detects whether a value will need to be further parsed for interpolation; it does not yet perform the parsing itself, which will happen during desugaring. This introduces a performance regression, for an interesting reason. I found that introducing a single new variant to `SugaredNir` (with a `(SymbolId, Span)` pair), was causing the width of the `NirParseState` type to increase just enough to cause Rust to be unable to optimize away a significant number of memcpys related to `Parser` moves, and consequently reducing performance by nearly 50% for `tamec`. Yikes. I suspected this would be a problem, and indeed have tried in all other cases to avoid aggregation until the ASG---the problem is that I had wanted to aggregate attributes for NIR so that the IR could actually make some progress toward simplifying the stream (and therefore working with the data), and be able to validate against a grammar defined in a single place. The problem is that the `NirParseState` type contains a sum type for every attribute parser, and is therefore as wide as the largest one. That is what Rust is having trouble optimizing memcpy away for. Indeed, reducing the number of attributes improves the situation drastically. However, it doesn't make it go away entirely. If you look at a callgrind profile for `tameld` (or a dissassembly), you'll notice that I put quite a bit of effort into ensuring that the hot code path for the lowering pipeline contains _no_ memcpys for the parsers. But that is not the case with `tamec`---I had to move on. But I do still have the same escape hatch that I introduced for `tameld`, which is the mutable `Context`. It seems that may be the solution there too, but I want to get a bit further along first to see how these data end up propagating before I go through that somewhat significant effort. DEV-13156 |
||
|---|---|---|
| .. | ||
| benches | ||
| build-aux | ||
| src | ||
| .gitignore | ||
| Cargo.lock | ||
| Cargo.toml | ||
| Makefile.am | ||
| README.md | ||
| autogen.sh | ||
| bootstrap | ||
| configure.ac | ||
| rustfmt.toml | ||
README.md
TAME in Rust (TAMER)
TAME was written to help tame the complexity of developing comparative insurance rating systems. This project aims to tame the complexity and performance issues of TAME itself. TAMER is therefore more tame than TAME.
TAME was originally written in XSLT. For more information about the
project, see the parent README.md.
Building
To bootstrap from the source repository, run ./bootstrap.
To configure the build for your system, run ./configure. To build, run
make. To run tests, run make check.
You may also invoke cargo directly, which make will do for you using
options provided to configure.
Note that the default development build results in terrible runtime performance! See [#Build Flags][] below for instructions on how to generate a release binary.
Build Flags
The environment variable CARGO_BUILD_FLAGS can be used to provide
additional arguments to cargo build when invoked via make. This can be
provided optionally during configure and can be overridden when invoking
make. For example:
# release build
$ ./configure && make CARGO_BUILD_FLAGS=--release
$ ./configure CARGO_BUILD_FLAGS=--release && make
# dev build
$ ./configure && make
$ ./configure CARGO_BUILD_FLAGS=--release && make CARGO_BUILD_FLAGS=
Hacking
This section contains advice for those developing TAMER.
Running Tests
Developers should be using test-driven development (TDD). make check will
run all necessary tests.
Code Format
Rust provides rustfmt that can automatically format code for you. This
project mandates its use and therefore eliminates personal preference in
code style (for better or worse).
Formatting checks are run during make check and, on failure, will output
the diff that would be applied if you ran make fmt (or make fix); this
will run cargo fmt for you (and will use the binaries configured via
configure).
Since developers should be doing test-driven development (TDD) and therefore
should be running make check frequently, the hope is that frequent
feedback on formatting issues will allow developers to quickly adjust their
habits to avoid triggering formatting errors at all.
If you want to automatically fix formatting errors and then run tests:
$ make fmt check
Benchmarking
Benchmarks serve two purposes: external integration tests (which are subject
to module visibility constraints) and actual benchmarking. To run
benchmarks, invoke make bench.
Note that link-time optimizations (LTO) are performed on the binary for benchmarking so that its performance reflects release builds that will be used in production.
The configure script will automatically detect whether the test feature
is unstable (as it was as of the time of writing) and, if so, will
automatically fall back to invoking nightly (by running cargo +nightly bench).
If you do not have nightly, run you install it via rustup install nightly.