fix: parse "infinity" literal fully in `float`

[boxdot]

float first parsed the "inf" literal and then "infinity", therefore even though the latter was parsed correctly, the suffix "inity" was returned as remaining input, which is not correct.

[epage]

Not the maintainer but I'm assuming the intent here is to parse the keywords inf, similar to nan. infinity isn't generally a keyword. The reason inity is being left is because nom-style parsers leave all unrecognized content for the next parser to identify or error on.

For a language when an invalid keyword is used often enough, it can make sense to special case the error situation into the grammar (match on infinity but error).

[boxdot]

infinity isn't generally a keyword

I forgot to mention that nom 6 parsed infinity as float infinity without remaining input. The current code also contains a case and tests for parsing infinity completely. So, I assume it is a regression in nom 7.

My motivation is actually this upgrade: https://github.com/Garvys/rustfst/pull/240

[epage]

Huh, didn't realize that FromStr accepts infinity. FromStr and lexical_core were the nom v6 implementations of float. In nom v7, an attempt was made to switch to minimal_lexical (baf5e01) but that had problems (f8633f9). Whats confusing is I'm not seeing how the minimal_lexical version was meant to handle nan, inf, and infinity.

[Geal]

right, the order was wrong here, nice catch

[coveralls]

Pull Request Test Coverage Report for Build 6597520943

• 4 of 6 (66.67%) changed or added relevant lines in 3 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 62.41%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/number/mod.rs	0	2	0.0%
<!--	Total:	4	6	66.67%	-->

Totals
Change from base Build 6597500402:	0.0%
Covered Lines:	1828
Relevant Lines:	2929

---

💛 - Coveralls

[Geal]

thank you!