Since I have a lot of formats created for the SDLang format and I have yet to see a format that would be a good replacement for its place (especially for my own usecases), I was thinking on writing my own parser with a better API (allowing for keeping the comments at their place, etc.), cleaned up type support (reducing the number of integer types to one, maybe even the floating point numbers too), more formatting options (like hexadecimal numbers) etc. I will also try to look into some automatic serialization methods.

I'm currently in the process of doing some research on what kind of gripe people have with the current SDLang implementation, both API and and textual representation level. Then I'll design a class structure for the parsers (Yes, I'll also create a more SAX-like parser too in conjunction of the DOM parser).

On Monday, 22 July 2024 at 17:25:37 UTC, solidstate1991 wrote:

I had such gripes today.

• basically sdlang-d comes with dependencies, that then create warnings. Well I never chose to use sdlang in the first place, so it's annoying to get unrelated warnings on top of that :)
• GitHub README has no parsing example to copy/paste
• the code looks a bit less good on screen than eg. parsing JSON with std.json

    result.name = dubFile["name"].str;
    foreach(e; dubFile["versions"].array)
    {
        if (e.str == "flobber")
            result.flobber = true;
    }

vs

    result.name = sdlFile.getTagValue!string("name");
    foreach(e; sdlFile.getTag("versions").values)
    {
        if (e.get!string() == "flobber")
            result.flobber = true;
    }

but I've not tried the 5(!) alternatives already on DUB.

https://github.com/ZILtoid1991/newsdlang

Created the repository, but likely won't start working on it immediately.

It will be a combined SDLang/KDL(an SDLang-inspired declarative language)/XDL(my own modified SDLang implementation) reader and writer.