> On 25 Dec 2015, at 23:27, stepharo <[email protected]> wrote: > > Thanks sven. > PS: the flight did not take off... we are still in Togo. >
Merry christmas! > Le 25/12/15 20:44, Sven Van Caekenberghe a écrit : >>> On 25 Dec 2015, at 13:35, stepharo <[email protected]> wrote: >>> >>> Hi sven >>> >>> I missed the scenario. >> I guess some people sometimes store a class in some instance variable of a >> model class, maybe for configuration. In that case STON would not work >> before, now it does. >> >> In theory, STON can serialise any class without further configuration, but >> not classes that contain deeply nested/linked/os system objects. >> >>> Stef >>> >>> Le 23/12/15 23:04, Sven Van Caekenberghe a écrit : >>>> Hi, >>>> >>>> In STON #bleedingEdge I added the ability to write and read Class and >>>> Metaclass instances by using their global name. >>>> >>>> So, Point will be written as Class[#Point] while Point class will be >>>> written as Metaclass[#Point]. This is of course a short-circuit solution >>>> for the complex structure that a class actually is, but a reasonable >>>> compromise. It is in any case better than not handling these types of >>>> instances, which was the case before. >>>> >>>> === >>>> Name: STON-Core-SvenVanCaekenberghe.64 >>>> Author: SvenVanCaekenberghe >>>> Time: 23 December 2015, 10:50:14.195912 pm >>>> UUID: b7ced20a-a7a4-4b0f-8f27-db73231eeecb >>>> Ancestors: STON-Core-SvenVanCaekenberghe.63 >>>> >>>> Added the ability to write and read Class and Metaclass instances using >>>> their global name >>>> === >>>> Name: STON-Tests-SvenVanCaekenberghe.59 >>>> Author: SvenVanCaekenberghe >>>> Time: 23 December 2015, 10:51:06.830762 pm >>>> UUID: 9f684bc1-45e1-4bfa-885b-c56ea178bfca >>>> Ancestors: STON-Tests-SvenVanCaekenberghe.58 >>>> >>>> Added unit test for the ability to write and read Class and Metaclass >>>> instances using their global name >>>> === >>>> >>>> Thanks to Norbert Hartl for requesting the feature and for test driving it. >>>> >>>> Sven >>>> >>>> >>>> >>> >> >> > >
