Christian Morgner

Hi, would it be possible to publish a 0.9.28 release based on this patch on Maven Central?

Thank you very much! The most recent version fixes all warnings and errors in my test case, and all the rules are imported successfully. Would it be possible to build a release and publish it on maven central, so we can officially use the new version without the SNAPSHOT suffix?

Here is another patch that adds support for the "rem" dimension. The patch deprecates the previous one, as it contains the stackoverflow fix and some more test cases. The only thing left in the tailwind.ui CSS is the empty custom property declaration: var(--my-value, ). I added a (failing) test that demonstrates the error; sadly I was not able to find a solution. With "rem" support, the number of import errors is down to about 10, which is really awesome because we are very close to have the full...

Hi, thanks again and sorry for keeping you busy.. :( It seems that the merge accidentially removed the @Test annotation of the var() tests. That alone wouldn't be so bad, but your changes to the patch for the grammar seem to introduce an endless recursion that results in a Stack Overflow, and the tests do not catch it because they are disabled by accident: Exception occurred in target VM: java.lang.StackOverflowError: at java.lang.Math.min(Math.java:1528) at java.util.Arrays.copyOf(Arrays.java:3748)...

Great, thank you for the update again! I tested the current version against my CSS test case (which is minified Tailwind UI CSS (https://tailwindui.com)). I found some empty rules where I expected custom properties and var() expressions, so I added grammar rules for that. The grammar is much simpler, and I think I was more successful this time. (I added more test cases as well.) You can find the changes in the attached diff, which should add support for custom properties and the var() function. Please...

Thank you very much for the update! I have attached a diff with the test cases I was trying to resolve. Some of them still don't work, but I haven't managed to find a solution :( For example, there is a problem with calc(100% / 3 - 40px) where the / 3 part is parsed correctly, but doesn't appear in the final result. I wasn't able to resolve the problem, but it seems that the nextLexicalUnits "/" and "3" are overwritten when adding "-" after exiting calcProduct(). I also included a test case for CSS...

Hi, I just saw that you already merged my initial patch proposal and improved the grammar a lot more than I was able to. Thanks a lot! :) I will happily provide some more test cases, probably later today / this evening.

Hi, I'm working on a better grammar that is closer the the specification and supports parentheses. I also added test cases, but I have not managed to get the grammar working yet, so I need some more time. I'm not sure what you mean by "have spend some time to rewrite the calc parser based on the spec."... do you have spent some time, or should I spend some more time (as I'm currently doing)?