calc() breaks when using 3+ variable fallbacks

[dgp1130]

From downstream issue cssnano/cssnano#880 which is from downstream issue angular/angular-cli#16910.

Using a calc() command with 3+ variable fallbacks seems to break the parser.

For example, paste the following code into the CSS Nano playground (which I'm told uses postcss-calc via cssnano-preset-default:

.some-rule-lg {
  --width-lg: 1px;
  --width-md: 2px;
  --width-sm: 3px;
  --offset-lg: 4px;
  --offset-md: 5px;
  --offset-sm: 6px;
  width: calc(var(--width-lg, var(--width-md, var(--width-sm, 0))) + var(--offset-lg, var(--offset-md, var(--offset-sm, 0))));
}

Then look at the JavaScript console and see:

"Parse error on line 1: 
var(--width-lg, var(--width-md, var(--width-sm, 0))) + var(--...
---------------------------------------------------^
Expecting end of input, "ADD", "SUB", "MUL", "DIV", got unexpected "RPAREN""

Work around is to use intermediary variables:

.some-rule-lg {
  --width-lg: 1px;
  --width-md: 2px;
  --width-sm: 3px;
  --offset-lg: 4px;
  --offset-md: 5px;
  --offset-sm: 6px;

  --temp-calc1: var(--width-lg, var(--width-md, var(--width-sm, 0)));
  --temp-calc2: var(--offset-lg, var(--offset-md, var(--offset-sm, 0)));
  width: calc(var(--temp-calc1) + var(--temp-calc2));
}

[Semigradsky]

The same issue: #77

[chrsalbert]

Seems to be still an issue. But the workaround works for me!
Will it get fixed soon?

[dzcpy]

Any updates?

[ludofischer]

Any updates?

Since #137 it should just warn instead of crashing. Does is still crash for you? The underlying problem is that the parser is incomplete, but to fix it we would need to duplicate the code that's already in the existing parsers (postcss-value-parser and postcss-selector-parser) so I would prefer to find a way to refactor so we don't parse multiple times (currently postcss-calc does parse -> stringify -> parse -> stringify)

[maranomynet]

This was also reported in #119 (I'm closing that one as duplicate)