Rewrite parser and add support for Nix 2.3 #27
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The Gradle plugin for Grammar-Kit does not support the usage of Util classes. See the following issues from the Gradle plugin. JetBrains/gradle-grammar-kit-plugin#23 JetBrains/gradle-grammar-kit-plugin#3
The test cases are mostly based on the following resources: https://nixos.org/guides/nix-pills/basics-of-language.html https://nixos.org/manual/nix/stable/#ch-expression-language Note that the outcome expected by the tests is not based on the behavior described in the mentioned resources but on the current behavior of the parser. The expected outcome will be corrected in upcoming commits when the related bugs of the parser are fixed. This approach shall make bugfixes on the parser easier to review by showing the old and new behavior in the diff.
The tests are based on src/libexpr/parser.y of commit 75efa421340 in https://github.com/NixOS/nix.
Commit 75efa421340b of https://github.com/NixOS/nix is used as reference.
IDEA might start lexing from the middle of a file for syntax highlighting. In such case, IDEA calls Lexer.start (which calls FlexLexer.reset) to restore the state of the lexer. However, the implementation previous to this commit did only restore parts of the state.
Remove 'NixTokenType.' prefix from expected tokens to improve
readability of error messages. For example the error message
NixTokenType.., NixTokenType.; or NixTokenType.or expected, got '}'
becomes
'.', ';' or 'or' expected, got '}'
JojOatXGME
force-pushed
the
parser_rewrite
branch
from
0756697 to
9baff42
Compare
This was
linked to
issues
Apr 5, 2021
|
I have used the plugin with the updated parser for a few days on a personal project. I did not notice any serious problem. I found IDEA-267912, but it was not a big issue. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This pull request rewrites the whole lexer and parser. I will try to make a field test before I merge the pull request. The new implementation should
The third point is not visible to the user but shall provide the foundation for more advanced features that need to analyze the content of a file. The new implementation no longer has special tokens for certain built-ins like
trueorimport. As a result, built-ins are no longer highlighted. I think the right way to implement the highlighting of built-ins is to use an annotator. I may look into that later.I decided to rewrite the whole parser and lexer because many issues I noticed while working on the test coverage (#24) were not easily fixable without major changes on the previous implementation. While it was a quite steep learning curve, I think I finally finished my implementation. I am aware that this is a huge pull request.
PS: There are still some issues with the new implementation which I was not able to fix using Grammar-Kit. This might include strange error messages in some (hopefully rare) cases. For more details, you might look at the links below. It might also be possible that the lexer, parser, brace matcher, and maybe code folding (once we implement it) get out of sync in invalid files since they all have their own implementation to analyze the structure of the file. I think trying to ensure they are all always in sync (even in invalid files) is not maintainable.