stefansjs / flatbuffers-intellij-plugin Goto Github PK

field declarations may currently use any type. When the reference finder finds nothing, there is no warning highlighting. There should be warning underline for unknown types.

this is not valid, this attribute must be specified on declaration of Extension Point only (https://plugins.jetbrains.com/docs/intellij/plugin-extension-points.html#dynamic-extension-points), not actual usages

The 0.1.1 version of the plugin is rather slow with parsing and syntax highlighting. I'll want to do some performance tuning using features of the grammarkit bnf grammar. Additionally, a lot of syntax highlighting fails when some statements are incomplete.

Tuning Subtasks

• Implement recoverWhile tuning for better error highlighting
• Cache reference finding
• Serialize named element to stub tree

References

• https://www.jetbrains.org/intellij/sdk/docs/reference_guide/performance/performance.html
• https://www.jetbrains.org/intellij/sdk/docs/basics/indexing_and_psi_stubs/stub_indexes.html
• https://github.com/JetBrains/Grammar-Kit/blob/master/HOWTO.md
• https://github.com/JetBrains/Grammar-Kit/blob/master/README.md#error-recovery-and-reporting

GrammarKit example

https://github.com/JetBrains/Grammar-Kit/blob/master/grammars/Grammar.bnf?source=c#L53

The parser doesn't tolerate the use of native_include in the schema and all subsequent parsing breaks.

In most languages cmd+/ will comment/uncomment a line. Figure out how to implement this for flatbuffers.

GitHub Actions has made many updates since the last build of this plugin. Investigate and update GitHub Actions for building the plugin:

Node.js 12 actions are deprecated. For more information see: https://github.blog/changelog/2022-09-22-github-actions-all-actions-will-begin-running-on-node16-instead-of-node12/. Please update the following actions to use Node.js 16: actions/checkout@v2, actions/setup-java@v1, actions/cache@v2, actions/upload-artifact@v2

The set-output command is deprecated and will be disabled soon. Please upgrade to using Environment Files. For more information see: https://github.blog/changelog/2022-10-11-github-actions-deprecating-save-state-and-set-output-commands/

The save-state command is deprecated and will be disabled soon. Please upgrade to using Environment Files. For more information see: https://github.blog/changelog/2022-10-11-github-actions-deprecating-save-state-and-set-output-commands/

When placing the cursor over a declaration name, the reference isn't highlighted. However when placing the cursor over the keyword (e.g. enum or table) the highlighting does occur.

When one field fails to parse the rest of the fields in the table tend to fail.

Currently the parser errors when parsing multi-line comments inside of flatbuffer files e.g. something like:

namespace blah;

/**
 * Test
 *
 * Describes test in more detail
 */
table Test {
  /**
   * The ID of this connection.
   */
  a:string;
}

My first thought is to actually make a namespace part of the PSI hierarchy. I think this is how the flatbuffers compiler parses namespaces. I will need to investigate how reasonable this is

https://www.jetbrains.org/intellij/sdk/docs/basics/plugin_structure/dynamic_plugins.html

With the following Flatbuffer schema (which flatc is happy to consume):

attribute "MyAttribute";

union MyOneUnion (MyAttribute) {
  MyOneUnionFirst,
  MyOneUnionSecond,
}
table MyOneUnionFirst {
}
table MyOneUnionSecond {
}

I get an error reported by the plugin:

flatbuffers.LCURLY expected, got '('

Just want to say that it's a very helpful plugin! Thank you for the effort you've put into it. :)

i.e

struct Blob {
    data: [uint8:42];
}

This seems pretty important

https://plugins.jetbrains.com/docs/intellij/code-style-settings.html

If a namespace is provided as part of an enum value, the parser fails to identify it. This should use the existing type reference.

First of all, thumbs up for the plugin.

I've just noticed the parser fails when a reserved keyword is used as a field name. This is perfectly valid flatbuffers schema:

table Typeful {
	id           : uint64	;
	int          : int		;
	int8         : int8		;
	int16        : int16	;
	int32        : int32	;
	int64        : int64	;
	uint         : uint		;
	uint8        : uint8	;
	uint16       : uint16	;
	uint32       : uint32	;
	uint64       : uint64	;
	bool         : bool		;
	string       : string	;
	stringvector : [string]	;
	byte         : byte		;
	ubyte        : ubyte	;
	bytevector   : [byte]	;
	ubytevector  : [ubyte]	;
	float32      : float32	;
	float64      : float64	;
	float        : float	;
	double       : double	;
}

Some languages allow you to collapse the function/class block in the IDE. I don't yet know what this involves, but probably involves changing the parser to identify where the curly braces start/stop

When the parser/lexer fails it tend to fail catastrophically. Use the pin and recoverWhile attributes of the flatbuffers grammar to make better error-handling when parsing.

• use recoverWhile to parse until the } of table/enum/struct declarations
• use recoverWhile to parse until the ; of field declarations

recoverWhile depends on the proper use of the pin attribute. This will take quite a bit of testing and unit testing the results may be brittle to each individual change. We'll need unit tests that have partial parsing failures to ensure that recovery happens in the way people would reasonably expect

Hi,
Consider such schema

struct F
{
  f: [int:16];
}

which is fixed size array is allowed. However, the plugins highlight this part as an syntax error.

Rename refactoring doesn't work when clicking on a referenced type. Rename refactoring is not available for declarations.

fix sized array grammar causes syntax error

doc