Path: csiph.com!eternal-september.org!feeder.eternal-september.org!reader01.eternal-september.org!.POSTED!not-for-mail
From: Andreas Leitgeb <avl@logic.at>
Newsgroups: comp.lang.java.programmer
Subject: Re: DefaultTableModel.getDataVector() and Java 11
Date: Wed, 3 Apr 2019 16:41:17 -0000 (UTC)
Organization: A noiseless patient Spider
Lines: 33
Message-ID: <slrnqa9ohd.cfl.avl@logic.at>
References: <slrnqa9lo4.cfl.avl@logic.at> <q82ngd$u5p$1@newsreader4.netcologne.de>
Reply-To: avl@logic.at
Injection-Date: Wed, 3 Apr 2019 16:41:17 -0000 (UTC)
Injection-Info: reader02.eternal-september.org; posting-host="05711f75fa384a44df8cf21cb0776d58"; logging-data="31693"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1/WyRpf2ISLXhdOifGTPnFA"
User-Agent: slrn/1.0.3 (Linux)
Cancel-Lock: sha1:Sg8sXoCHPcSB9zSge9VIKwZtcAg=
Xref: csiph.com comp.lang.java.programmer:38849

Patrick Roemer <sangamon@netcologne.de> wrote:
> Responding to Andreas Leitgeb:
>> In Java up to 8  DefaultTableModel had a method  getDataVector()
>> that returned a raw Vector.
>> 
>> My application's custom TableModel overrode the method to return
>> a Vector<MyRowType>, where MyRowType provided direct type-safe
>> accessors for each column.
>> 
>> In Java 11 it seems to have changed to returning a Vector of 
>> raw Vectors.
>> 
>> At first glance this seems just goofy to me, because that doesn't
>> really make it type-safe (it still involves a raw type), and it even
>> prevents me from making it type-safe myself.
>> 
>> I hope, someone here can provide me with a second glance that
>> clarifies the rationale behind the change.
>
> API doc for this method in Java 8: "Returns the Vector of Vectors that
> contains the table's data values." There's quite a few occurrences of
> casts such as "(Vector)dataVector.elementAt(row)" in the
> DefaultTableModel code, too.
>
> So it seems that using another element type than Vector with
> DefaultTableModel has never really been a safe option, and the change
> only codifies to types what the contract has always been saying. Custom
> [Abstract]TableModel to the rescue...

Thanks a lot - it now makes sense with that background.

Will see what I really need beyond AbstractTableModel.