Incorrect Parsing of Floating Point Literals

[SethPoulsen]

Currently, the floating point literal

double f = 4.5;

is parsed to

Decl: f, [], [], []
  TypeDecl: f, []
    IdentifierType: ['double']
  Constant: float, 4.5

when really, the type attribute of the Constant node should be "double". Likewise,

double f = 4.5L;

is parsed to

Decl: f, [], [], []
  TypeDecl: f, []
    IdentifierType: ['double']
  Constant: float, 4.5L

when the type attribute of the Constant node should be "long double". Note that the declared type is not what matters, it is the parsing of the floating point literal.

[eliben]

Good point! Feel free to submit a PR

[eliben]

Fixed by #277

[jam1garner]

Thanks for properly naming the commits/PR @robbert-harms, it made it much easier to find what caused my compiler to break upon installing on another machine (and therefore getting a new version of pycparser)!

[Sparika]

It seems that the case for long int is not handled either.