support for declaring and computing secret attributes

[ghost]

Reported by @cristianbogdan on 12 Dec 2002 02:58 UTC this bug is connected to bug 7 (where we might decide something totally different like e.g. dropping secret attributes completely) and bug 40 (which deals with attribute declaration in the page, which would be boring for secret attributes, which are used everywhere)

secret (session) attributes are attributes that keep the user identity, and at present are one of the big design problems of makumba, due to using the same attribute with differnet senses in different parts of the application (bug 256)

the idea here is to have a central place in the BL where secret attributes are declared, with name and type. the name of that file (MDD, Java or properties, not sure yet) can be delcared in MakumbaController.properties, or can be a separate file, MakumbaSession.properties, etc

once secret attributes are declared

• it's impossible to get to a situation like bug 256, since an attr can be of only one type. that means that some parts of karamba will have to be fixed
• the taglib can check if they are used with the correct type, and if not, explain the mak:programmer what's the problem
• the JSPx source viewer can display secret attributes with a different colour, show type info on request and link to the place secret attributes are declared and computed

alternatives for representing secret attributes:

• an MDD file would contain the names and types of the secret attributes, in the normal name=type MDD format (student=ptr best.Student)
• a Java file can do even more: besides declaring the types in methods like String getStudentType(){ return "ptr best.Student"; } it can also implement methods like findStudent(Database db, Attributes a). that would ensure that the attribute is computed in the same way (e.g. that HTTP attributes that it is read from have the same name) throghout the application, which may be a good thing...

Migrated-From: http://trac.makumba.org/ticket/257

[ghost]

Comment by @cristianbogdan on 22 May 2003 13:24 UTC besides bug 256, i think that bug 34 is another bug that can benefit from this

connections with bug 40 and bug 41 are interesting for the idea of static page analysis

[ghost]

Modified by @manuelbernhardt on 28 Jun 2010 07:22 UTC

[ghost]

Modified by @manuelbernhardt on 28 Jun 2010 07:36 UTC