This is intended to address currently flaky timeouts that are likely caused by the size of the prelude. I'm addressing a performance bottleneck in AnalyzeProgram with trace output. Trying to omit prelude traces reduces most trace output significantly, and I think it'll scale better as the prelude size increases.
The basic mechanics here are:
- In order to consistently track whether tracing is on, I've added a TraceStream class, explorer/interpreter/trace_stream.h.
- The AST now has a num_prelude_declarations field, so that it's provided where the boundary is.
- In order to mark where we try to skip prelude output, I've added calls to set_in_prelude in type_checker.
- In exec_program, I just use num_prelude_declarations directly to skip over.
- Everywhere checks TraceStream::is_enabled before printing, similar to the std::optional check that was previously used.
This does add some timing output in order to better diagnose where slowness is coming from, when tracing. It also adds "verbose" targets to make it easier to get the trace output.
So for example, here's a timing for zero.carbon:
```
Timings:
- Parse: 13ms
- AddPrelude: 25ms
- AnalyzeProgram: 116ms
- ExecProgram: 12ms
```
If I make a small change to just not set skipping_prelude (essentially getting back to current output):
```
- Parse: 13ms
- AddPrelude: 25ms
- AnalyzeProgram: 2359ms
- ExecProgram: 57ms
```
Thus in this trivial example, I'm eliminating about 95% of the execution time.
Note this approach could still be refined in a few ways:
- We could add a flag to allow overriding in_prelude. It should be a small amount of work after this change. But it's a little consistent with how parser_debug works, that it won't print prelude output by default (unless there's an error).
- Execution could skip messages involving initialization of globals declared in the prelude. This is a little noisy right now, but I don't think it's significant for performance because ExecProgram is tiny.
- Once files are more separated, we should be able to change the num_prelude_declarations/set_in_prelude approach.
Co-authored-by: Richard Smith <richard@metafoo.co.uk>
The design of choice types expects the declaration of an alternative to match the usage: if an alternative is declared as `None`, then it should be used as `None` not `None()`, and if it is declared as `None()` then it should be used as `None()` not `None`. Update explorer to match.
Also clean up the handling of choice types and alternative values a little in general, by moving away from identifying choice types and alternatives as strings and towards identifying them symbolically.
Closes#2422
Add support for user-defined assignment, as well as compound assignment and increment, following the design direction in pending proposal #2511.
Some of this isn't fully testable yet: because explorer doesn't properly support `impl` specialization, the blanket `impl`s in the prelude prevent types from customizing assignment.
Also make minor updates to the skeletal design in
docs/design/name_lookup.md following #2113, as there are no longer any prelude names that are made available to unqualified name lookup by default.
Add `type` to the keyword list in
docs/design/lexical_conventions/words.md, following #2360.
In particular, this means that a type can implement `ImplicitAs(Type)` and have values of that type behave like types.
This implies that `()` and `{}` are no longer types. They are now values whose type is the result of converting `()` or `{}` to type `Type`, as has been discussed recently and seems to be the supported direction. This fixes various cases where these types were previously mishandled.
Adds a simple script to merge stdout/stderr and put on labels. This is hidden to the RUN line using lit.cfg.py.
This is my solution to addressing how errors printed by the toolchain break sorting of stdout output; just put stdout first. We could also have toggles for interleaving output or such, which might help test whether we do it properly.
This also moves some previous-distributed replacement logic into lit_autoupdate_base.py: I think having that adjacent to lit.cfg.py is probably the better choice, and it reduces duplication in toolchain scripts. It happens here because I need to change the resulting commands to include the merge.
On #2224 @zygoloid pointed out we needed --implicit-check-not to ensure we were correctly matching output. This is the standard way we're writing explorer tests, so I was looking at unifying our lit approaches.
This is one take on it, making more use of substitutions to bring various testing into alignment, as well as symlinks to avoid config skew (maybe I'll eventually figure out a better solution than symlinks).
Makes a couple small fixes in explorer to remove end-of-line whitespace on output.
* Multiplication and division have the same priority.
* A new builtin interface DivWith is added.
* In some tests expecting a compilation error (syntax error), the error
message now says it is expecting SLASH or binary *.
The most significant change here is that explorer now uses the chosen spelling
rather than the old `Bool` spelling. Also update a few documentation examples
and some skeletal design docs to use the chosen spelling.
- An introprocedural forward analysis that checks the may-be-formed states on local variables.
- Returns compilation error on usage of must-be-unformed variables.
- Implemented as a pass of `ASTNode` traversal.
- Currently supports detection of: function parameter, return expression and rhs of assign.
Following #1191, add initial support for bitwise operators. Support for both
integer operations and for operator overloading via the interfaces specified in
#1191 is provided.
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
This allows us to combine multiple Errors together without repeating the prefix
information. Also fixes several cases where two "COMPILATION ERROR" prefixes
would be prepended to the same message when errors with prefixes and locations
were produced by the lexer and parser.
Allows unformed state for local variables. Reports a run-time error when an unformed local variable is used.
- Added declaration without initialization for local variables in the parser.
- Made the init expression of VariableDefinition optional.
- Expanded (alive, dead) to (uninitialized, alive, dead) in the Heap.
Co-authored-by: Jon Ross-Perkins <jperkins@google.com>
We don't have overload resolution, so I've simply added PrintInt for now. But it's really tempting to turn __intrinsic_print into a total kludge function.
* initial implementation of new and delete
* remove requirement for empty heap from trace.carbon
* more tests
* fixed substitution for generic function types
* added documentation for tracing output, changed the printing of scopes to be part of the stack, made other minor changes hoping to enhance readability
* reformat a list in the README
* minor edit
* fix the lists
* updates to the interpreter/README.md
Support is added for all of the non-type expression contexts where we currently accept built-in conversions, such as reordering the fields in a struct.
- Makes `Self` a keyword
- Inside `class Foo { `... `}`, `Self` means `Foo`
- Inside `class Foo(T:! Type, U:! Type) {` ...` }`, `Self` means `Foo(T, U)`
- `impl as Bar` means `impl Self as Bar`
- Inside `external impl Foo as Bar {`...`}`, not in a scope already defining `Self`, `Self` means `Foo`
Implemented by introducing a new kind of declaration, a `SelfDeclaration`, that is automatically added to class and impl declarations.
Co-authored-by: Jon Meow <jperkins@google.com>
Use FileCheck's `[[@LINE+n]]` mechanism to refer to the next line.
This is made awkward by a couple of things:
* We want to keep the `CHECK` lines in the original order.
* Errors are sometimes more than one line long.
The approach we use is to interleave the original lines and the check lines, putting each check line as early as possible subject to two rules:
1) Check lines never precede the 'AUTOUPDATE' line
2) Except when required by rule (1), a check line that refers to a source line by line number is never placed earlier than a source line that precedes that source line.
The actual `[[@LINE+n]]` annotations are created in a second pass after we've interleaved the lines so that we can work out the correct offsets.
Co-authored-by: Jon Meow <jperkins@google.com>
Jon Ross-PerkinsandRichard Smith
Junhee Cho
pk19604014
Zenong Zhang
Jeremy G. Siek
josh11bandJon Meow