Posted on 1998-02-06
I have serialized an object to disk. The disk file starts out with eight bytes of gobbledygook (let's call it XXXXXXXX), followed by project.object, followed by eight more bytes of gobbledygook (let's call it YYYYYYYY), followed by the bulk of the serialized object (let's call it BBB...B). Herre are the results of an experiment:
Object with method: XXXXXXXXproject.objectYYYYYYYYBBB...B
Object without method: XXXXXXXXproject.objectZZZZZZZZBBB...B
and YYYYYYYY is not at all similar to ZZZZZZZZ.
I have only one question, but I have to ask it in three pieces: (1) Why is the method-induced information embedded in the serialized object? (2) The method is certainly not stored inside the serialized object, is it? It would have increased the byte count! (3) How the $#!@ am I supposed to make serialized objects publicly available, while the uppity little jerks demand to be reused by only the project that created them? Note: It does not help to put the objects into database storage with public permission, because the byte code within the retrieved object still cites itself as belonging to the project that created it.