On 09/11/2017 03:28 PM, Guido van Rossum wrote:

Oddly I don't like the enum (flag names get too long that way), but I do agree 
with everything else Barry said (it
should be a trivalue flag and please don't name it cmp).

Hmmm, named constants are one of the motivating factors for having an Enum type. It's easy to keep the name a reasonable length, however: export them into the module-level namespace. re is an excellent example; the existing flags were moved into a FlagEnum, and then (for backwards compatibility) aliased back to the module level:

    class RegexFlag(enum.IntFlag):
        ASCII = sre_compile.SRE_FLAG_ASCII # assume ascii "locale"
        IGNORECASE = sre_compile.SRE_FLAG_IGNORECASE # ignore case
        LOCALE = sre_compile.SRE_FLAG_LOCALE # assume current 8-bit locale
        UNICODE = sre_compile.SRE_FLAG_UNICODE # assume unicode "locale"
        MULTILINE = sre_compile.SRE_FLAG_MULTILINE # make anchors look for 
newline
        DOTALL = sre_compile.SRE_FLAG_DOTALL # make dot match newline
        VERBOSE = sre_compile.SRE_FLAG_VERBOSE # ignore whitespace and comments
        A = ASCII
        I = IGNORECASE
        L = LOCALE
        U = UNICODE
        M = MULTILINE
        S = DOTALL
        X = VERBOSE
        # sre extensions (experimental, don't rely on these)
        TEMPLATE = sre_compile.SRE_FLAG_TEMPLATE # disable backtracking
        T = TEMPLATE
        DEBUG = sre_compile.SRE_FLAG_DEBUG # dump pattern after compilation
    globals().update(RegexFlag.__members__)

So we can still do re.I instead of re.RegexFlag.I.

Likewise, if we had:

    class Compare(enum.Enum):
        NONE = 'each instance is an island'
        EQUAL = 'instances can be equal to each other'
        ORDERED = 'instances can be ordered and/or equal'
    globals().update(Compare.__members__)

then we can still use, for example, EQUAL, but get the more informative repr 
and str when we need to.

--
~Ethan~
_______________________________________________
Python-Dev mailing list
Python-Dev@python.org
https://mail.python.org/mailman/listinfo/python-dev
Unsubscribe: 
https://mail.python.org/mailman/options/python-dev/archive%40mail-archive.com

Reply via email to