That would be changed from now on as it's too much open ended and
troublesome to support in other subsystems such as JMX beans... Note that
"-" is still in the list, other symbols would not be allowed like "+" or
"@" - Which are very odd and should be very rare.

On Tue, Mar 1, 2016 at 2:52 PM, Anilkumar Gingade <[email protected]>
wrote:

> From the existing javadoc it looks like, application can have any chars in
> region name except the "/".
>
> Changing this one may have implication on existing GemFire customers....
>
> In the past we had to make changes in OQL to support chars like "+", "-",
> "@"...(since customers used this in their region names)...
>
> -Anil.
>
>
> On Tue, Mar 1, 2016 at 2:44 PM, Udo Kohlmeyer <[email protected]>
> wrote:
>
> > +1
> >
> > On 2/03/2016 9:35 am, William Markito wrote:
> >
> >> Folks, it doesn't look like we have actually finished this thread...
> >>
> >> What do you guys think about the following pattern:
> >> "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz123456-_"  ?
> >>
> >> I'm not specifying a regexp to avoid problems with unicode and to keep
> it
> >> only ASCII-only...  Stackoveflow has some suggestions like
> >> "^\\p{ASCII}*$" but
> >> I'd be careful and try to keep it strict to the ones specified in the
> list
> >> above.
> >>
> >> Thanks
> >>
> >>
> >> On Thu, Feb 18, 2016 at 11:19 AM, Darrel Schneider <
> [email protected]
> >> >
> >> wrote:
> >>
> >> The public javadocs on Region#getName say:
> >>> Returns the name of this region. A region's name
> >>>     * can be any non-empty String providing it does not
> >>>     * contain the name separator, a forward slash (/).
> >>>
> >>> Here is the code from LocalRegion that validates the name:
> >>>    static void validateRegionName(String name)
> >>>    {
> >>>      if (name == null) {
> >>>        throw new
> >>>
> >>>
> >>>
> IllegalArgumentException(LocalizedStrings.LocalRegion_NAME_CANNOT_BE_NULL.toLocalizedString());
> >>>      }
> >>>      if (name.length() == 0) {
> >>>        throw new
> >>>
> >>>
> >>>
> IllegalArgumentException(LocalizedStrings.LocalRegion_NAME_CANNOT_BE_EMPTY.toLocalizedString());
> >>>      }
> >>>      if (name.indexOf(SEPARATOR) >= 0) {
> >>>        throw new
> >>>
> >>>
> >>>
> IllegalArgumentException(LocalizedStrings.LocalRegion_NAME_CANNOT_CONTAIN_THE_SEPARATOR_0.toLocalizedString(SEPARATOR));
> >>>      }
> >>>    }
> >>>
> >>>
> >>> On Thu, Feb 18, 2016 at 11:09 AM, William Markito <[email protected]
> >
> >>> wrote:
> >>>
> >>> I don't think we should allow non-alphanumeric region names...   And
> >>>>
> >>> would
> >>>
> >>>> be really nice to have a list or a pattern documenting what's valid.
> >>>>
> >>>>
> >>>> On Thu, Feb 18, 2016 at 11:04 AM Kirk Lund <[email protected]> wrote:
> >>>>
> >>>> I was just looking into a ticket filed because /= was used as the
> >>>>>
> >>>> Region
> >>>
> >>>> name and this caused problems in JMX ObjectNames. I have two
> questions:
> >>>>>
> >>>> 1)
> >>>>
> >>>>> do we really want /= to be a valid Region name? 2) do we have a
> >>>>>
> >>>> complete
> >>>
> >>>> list somewhere of all the non-alphanumeric characters that are usable
> >>>>>
> >>>> in
> >>>
> >>>> Region names?
> >>>>>
> >>>>> -Kirk
> >>>>>
> >>>>> --
> >>>> ~/William
> >>>>
> >>>>
> >>
> >>
> >
>



-- 

~/William

Reply via email to