The term "full-expression" is not a grammar production

[jabelloc]

Replace all references (38) to the term "full-expression" with "full expression", as a full expression is not a grammar production.

[cpplearner]

Is there a rule against using hyphen in a term? Plus, "full expression" suggest that it is a kind of expression, but sometimes it is not.

[jabelloc]

AFAICT this is the sole exception in the standard about what seems to be an informal rule.

[zygoloid]

We have other terms containing hyphens, such as cv-qualifier, parameter-type-list, injected-class-name, and so on. This is not a defect. Non-defining uses of such terms are formatted in upright text, which should visually distinguish them from grammar terms. There is, however, a visual ambiguity between the definition of such terms and grammar productions, but we discussed that in the context of #323 and decided to take no action.

[FrankHB]

I'm not trying to revive the issue (because it has nothing to do with hyphens vs. grammar production) but the current use of "full-expression" does have some problems raised from the hyphen. For instance, the localization (e.g. translation in the localized standards) of the ISO C++ rules is more difficult to be precise due to lack of hyphen in the result. This is not same to the cases like cv-qualifiers because prefixed cv is not an adjective by itself, so the hyphen is preserved naturally.

Also @cpplearner, the fact that a full-expression is not necessarily a kind of expression makes the problem worse in another aspect. Why on earth such a context is named after a phrase of "expression"? (This does not seem to be editorial, though.)