pp.White skips preceding whitespace characters even if characters removed from default (from issue #350 resolution)

[cgevans]

From revision 1ccf846 onward (the fix to #350), if the default whitespace characters are removed, for example, if there are no default whitespace characters, pp.White(ws) appears to skip over characters in " \t\n\r" to consume ws even if those characters are not in ws and not in the default whitespace characters.

import pyparsing as pp
pp.ParserElement.setDefaultWhitespaceChars("")

("a" + pp.White(" ") + "c").parseString("a\n c")  # Matches, should not.
#[Out]# ParseResults(['a', ' ', 'c'], {})

("a" + pp.White("\n") + "c").parseString("a\nc")  # Matches, should.
#[Out]# ParseResults(['a', '\n', 'c'], {})

("a" + pp.White("\n") + "c").parseString("a \nc")  # Matches, should not.
#[Out]# ParseResults(['a', '\n', 'c'], {})

("a" + pp.White("\n") + "c").parseString("a \n c")  # Does not match, should not.

("a" + pp.White("\n") + "c").parseString("a\n c")  # Does not match, should not.

(This is with Python 3.10.2, pyparsing 1ccf846 up to 938f59d)

[ptmcg]

Thanks for posting, will take me a bit to sift through. The White class has always been an oddity in pyparsing, given that it has to look for spaces that are ignored by most of the rest of the package.

[SamZhangQingChuan]

Hi @ptmcg. I am wondering if there is any update on this issue? This has cause issues in the pyhocon package. Many thanks!