vjekob / al-pragma-explorer Goto Github PK

Hey guys,

inspired by something that Vjeko mentioned during one of the pauses, I tried my hand at writing a state machine "lexer" to parse the files.
My results are mixed, to say the least ;-)

(I had no idea what I was doing here, this might be ridiculously wrong and if you laugh your head off, please let me parttake in your amusement :D)

My guess is that this code creates overhead which the JScript interpreter doesn't have to deal with in the current parsing code.

It actually works, but it seems to perform a bit worse than the original code in my tests.

• I wonder if not instantiating an object of the LexerParser class and getting rid of a few if-statements could help
• The step() code seems dirty compared to what I've seen, if it's correct this way at all
• There's some unecessary variables and assignments in there

Lexer Code:
src/parser/Lexer.ts
https://pastebin.com/vJnHqv8z

A simple way to implement it is to add this to src/parser/parse.ts :

const lexerParser = new LexerParser(uri); const results = lexerParser.readFile(); return results;

(Above the current parsing code, comment the rest out.)

Thanks for having a look, kind regards,
Wolfram

Here's some stopwatch code in case you want to measure results:

var startTime: Date, endTime: Date;

function start() {
    startTime = new Date();
}

function end() {
    endTime = new Date();
    var timeDiff = endTime.getMilliseconds() - startTime.getMilliseconds(); //in ms
    console.log(timeDiff + ' milliseconds');
}

Hi all,

I hope you are doing well. Currently, we have implemented to have the pragma in app.json only. According to the docs: (Compiler Directives), we can also have them using #DEFINE "MYPRAGMA".

The main differences are: These pragmas are file scoped.

Would love to try that

Hi Vjeko,

I really like your new extension, thx for supporting our community.
Sometimes I use if not pragma in my projects, at the moment these are presented twice in the pragma-explorer.
Maybe you can change this in one of the next releases.

Thanks, Martin.