Update puppet grammar #25
Conversation
This commit looks for the beginning and end of a resource and only allowing certain patterns in between such as a title, parameters, and valid parameter values.
This commit adds conditionals to the repository and includes them in the global patterns. This will make it easier to maintain conditional logic over time.
For consistency
Easier to maintain over time
This commit makes functions a repository item and removes old ways of identifying them.
This commit does not match the ending '{' character that signifies the
end of a class metadata declaration, and the begining of the class
content. Further, it treats the class name in the 'inherits' declaration
the same as a class name
This comit creates a 'defineparmaeters' repository item that is used for matching parameters in defined types, classes, and application definitions.
This commit properly parses resource parameters and adds new spec tests for resource tokenization
This commit properly parses class definitions with or without parameter and with or without class inheritance. Tests were also added
This commit replaces the defined type pattern with the class pattern since they're almost perfectly identical patterns
Tokenizing resource titles makes it impossible to have syntax highlighting for strings and variables for titles
This commit tokenizes integers and floats anywhere in the Puppet code. In the future, this should be reduced to only where numbers are allowed - as parameter values or assignment operation values.
|
I noticed a few bugs and fixed those as well as added a few more tests. |
| 'endCaptures': | ||
| '0': | ||
| 'name': 'punctuation.definition.class.begin.puppet' | ||
| 'end': '{' |
There was a problem hiding this comment.
Can you move this end above the endCaptures?
This commit adds scoping for case statements. Case statements are really tricky since their parameter syntax structure is unique, but are easy to parse as hash parameters. This causes problems. Since each parameter can contain any valid Puppet code, and we need to parse each parameter explicitly, all valid Puppet code has been moved to a repository item called '#puppet'. This will give the rest of the parser quite a bit of flexibility as we can use this to fully match conditionals blocks
|
Ah, I did indeed forget case statements. So, those were tricky. The case values look similar enough to hashes that it didn't parse correctly. I had to explicitly capture case values, which can contain any valid Puppet code. To do that, I had to make a repository item called '#puppet' that contains all valid Puppet code. This is a good idea because we can fully match definitions and conditional blocks in the future. |
| ] | ||
| } | ||
| ] | ||
| 'puppet': |
There was a problem hiding this comment.
This name seems ambiguous...what does this group do exactly?
EDIT: I see now - you might want to rename this to code to maintain consistency with other language packages.
|
Nice work @ccaum! I've left some comments for you to look at 😄. |
|
I haven't forgotten about this. I've just been pulled onto other things for a bit. |
|
I've found that this breaking highlighting a lot for me: Before: (Vanilla install): After: (Using this branch) |
This pull request updates the puppet grammar to improve matching of classes, comparisons, functions, numbers, and resource parameters. In addition, it improves the parameter matching for class, defined type, and application definitions. This still needs lots more test coverage and it doesn't yet cover the additional syntaxes added in Puppet 4.
This turned out to be a large PR, so I'm happy to work through it as works best for the maintainers.