No since spending the GitHub action minutes (or waiting to merge) on
Bazel when only changing markdown or other files that aren't part of the
build and test.
This document tries to lay out a high-level roadmap for Carbon in 2021
following the [roadmap process](/docs/project/roadmap_process.md).
Co-authored-by: Dave Abrahams <dabrahams@google.com>
`clang-format-11` formats this multiple-ternary-operator expression differently than `clang-format-10` and `clang-format` at (current) trunk. Adding an extra set of parentheses brings them all in agreement, avoiding oscillation as folks check the file locally then submit to CI.
PR #280 committed the decision for #199. #199 had the proposal and no decision
and modified `README.md`. #280 had a decision but no proposal, and *didn't*
update `README.md`. However, in a tree with both the proposal *and* its
decision, another modification is necessary.
Since only one PR modified `README.md`, there wasn't a merge conflict that
required a rebase. And since there was no rebase, each PR ran its own CI
blissfully unaware of the other.
For now, committing the results of `pre-commit run --all-files` on trunk.
Later, will look into ways to make the proposal+decision workflow less likely
to break trunk.
We're now one step away from eliminating bare pointers in semantic actions.
There are plenty of other cleanups and modernizations that can be made in the
parser, but the elimination of bare pointers is the one that has the highest
impact for the codebase.
Slowly bringing this into line with Bison's C++ example parser
so we can use strong semantic values for symbols rather than
leaking pointers. First step is to thread a `ParseAndLexContext`
object through the whole syntactic analysis state, like the
example has. In the example, it's called `driver`.
Co-authored-by: Geoff Romer <gromer@google.com>
This is primarily using the configuration matrix facilities of GitHub
actions to consolidate the overall configuration. Beyond avoiding
duplication, this also allows the default and release builds to run in
parallel. The checkout time is duplicated between these, but the rest of
the time is nicely parallelized. This has the most dramatic effect on
macOS builds. Overall, this should reduce the latency on testing from
22-28 minutes to 15-20 minutes from what I've seen which seems
worthwhile.
This does reduce the detail provided in the names of the different
configurations. However, the Bazel build mode is preserved. That seems
like the most critical pieces of information.
Much of this started with me just trying to learn more about GitHub
actions, but once understanding how the job matrix worked, it seemed
worthwhile to send out as an actual change.
Distinguishes parts that come from the parser and lexer. It used to be that all
the files were called "syntax*", but lexing and parsing are distinct phases that
are easier to keep track of when distinguished. syntax.yy.cpp being the source
file generated by flex, containing the lexer was particularly confusing, because
the yy tends to indicate it is a yacc/Bison product, and the ".tab." substring,
indicating "tables" is not really useful to the developer.
These names also match up with what Bison's C++ example uses, which will make
the transition easier.
We switched to the released version to minimize rebuildds of the
toolchain, but with downloading the toolchain this isn't a significant
issue any more. And without this, Bazel doesn't run on ARM macOS.
This may hit issues with broken Bazel builds and have to be reverted.
Added a TODO to reverse this as soon as we can anyways.
Also cleans up the old `.bazelversion` in favor of just using
`.bazeliskrc`. They both work, but the RC file comes first in the
sequence so happy to prefer it here.
* add command-line flag to enable/disable tracing output
* adding missing exit for pattern variable in wrong context and a test case for it (#324)
* Update executable_semantics/interpreter/interpreter.cpp
comment on separate line as code
Co-authored-by: Jon Meow <46229924+jonmeow@users.noreply.github.com>
* fix interpreter's handling of optional else of if statement (#323)
* Update pattern_variable_fail.golden due to error (#334)
* Use llvm's CommandLine for parsing (#332)
* GitHub testing action (#331)
Co-authored-by: Chandler Carruth <chandlerc@gmail.com>
* add copyright
* Create a Dictionary abstraction over the raw Cons list. (#327)
* Create a Dictionary abstraction over the raw Cons list.
* renamed Cons and some methods of Dictionary, various other cleanup
* Update executable_semantics/tracing_flag.cpp
added namespace comment
Co-authored-by: Jon Meow <46229924+jonmeow@users.noreply.github.com>
* added a comment to cpp file
Co-authored-by: Jon Meow <46229924+jonmeow@users.noreply.github.com>
Co-authored-by: Dave Abrahams <dabrahams@google.com>
Co-authored-by: Chandler Carruth <chandlerc@gmail.com>
* Make `field_list` always a list.
* Create separate `paren_expression` and `tuple` nonterminals
* Rename expression_or_field_list.* to field_list.*
Co-authored-by: Jeremy G. Siek <jsiek@indiana.edu>
* improve abstraction for AssocList, fix bug in optional else
* add flag for tracing output, clean up code for output
* turned off tracing by default, updated goldens, removed two examples that use pointers, shouldn't have been there yet
This change creates an unpleasant amount of boilerplate where `Declaration` is declared, in exchange for being able to—very pleasantly—treat it as a simple value that composes with other values everywhere it is used. Applying this technique broadly will pay off in code comprehensibility; once it has been done for all things being new'd, pointers disappear and references are only needed as an idiomatic approximation of inout. The unpleasant code grows only when new polymorphic operations are added, and then only a bit, and is an idiom whose details can readily be ignored once in place. The pleasant code pervades the codebase.
Too bad we don't have existential types in C++ ;-)
* Add assertion to detect UB in interpreter.cpp
This change, applied to 6e5070d ("Adapting jsiek's executable semantics tooling
for commit. (#237)"), causes the if2.6c test to segfault. Becase the assertion
fires only when `stmt == nullptr` and no code has permission to change `stmt`
(it is `const`) before it is dereferenced in the `switch`, and nothiing in
`PrintStatement` is supposed to exit the program, the assertion is a valid
change that detects a bug.
The crash was originally manifest in 29a5994 ("Executable Semantics: 1st-class
stacks (#296)").
* Temporarily disable the if2 test pending #311
See https://github.com/carbon-language/carbon-lang/issues/311
* [executable semantics] Record exit code on expected error.
This will prevent a final segfault from sneaking by, detected as a passing test.
A more principled follow-up commit would bottleneck detected error exit
reporting and have it write something to std::cerr that can be recognized.
* [executable semantics] class-ify Declaration
NFC (no functional change).
Proof of concept that we can simplify code by replacing unions with safer, more
regular types. Hand-rolled existentials (type-erasing CoW wrappers) are a follow-on
step that will further simplify usage.
Began adding `const` where possible, and replacing `std::string*` with
`std::string`. Most `const`s can disappear as we replace reference semantics
with value semantics, but in the meantime it's an important step in the right
direction.
Chandler Carruth
Matthew Riley
Jon Meow
Richard Smith
