Path: csiph.com!eternal-september.org!feeder.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail
From: Eric Sosman <esosman@comcast-dot-net.invalid>
Newsgroups: comp.lang.java.programmer
Subject: Re: Generics and downcasts
Date: Mon, 24 Apr 2017 22:08:35 -0400
Organization: A noiseless patient Spider
Lines: 62
Message-ID: <odmat6$i11$1@dont-email.me>
References: <generics-20170424231405@ram.dialup.fu-berlin.de>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Tue, 25 Apr 2017 02:05:26 -0000 (UTC)
Injection-Info: mx02.eternal-september.org; posting-host="ba845162fa6e6248614a87a2f170f2bf"; logging-data="18465"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1/KVvZ5JgkDyPDOaOOB/wu4"
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.0.1
In-Reply-To: <generics-20170424231405@ram.dialup.fu-berlin.de>
Content-Language: en-US
Cancel-Lock: sha1:jWbn2sL4OPsnLLrYqQurKL/5sRA=
Xref: csiph.com comp.lang.java.programmer:36528

On 4/24/2017 6:23 PM, Stefan Ram wrote:
>    Naftalin and Wadler write that
> 
> List<String> words = new ArrayList<String>();
> words.add("Hello ");
> words.add("world!");
> String s = words.get(0)+words.get(1);
> assert s.equals("Hello world!");
> 
>    is compiled to the same Bytecode as
> 
> List words = new ArrayList();
> words.add("Hello ");
> words.add("world!");
> String s = ((String)words.get(0))+((String)words.get(1))
> assert s.equals("Hello world!");
> 
>    and then write:
> 
> Cast-iron guarantee: the implicit casts added by the
> compilation of generics never fail.
> 
>    This, of course, immediatly leads to the question:
> 
>    Since the downcasts added during the compilation of
>    the above generic code can never fail and since they
>    add run-time costs, can't they be omitted?

     No.

>    Maybe the purpose of these downcasts is to add
>    additional safety for the case of some parts of a
>    program not using generics. (I.e., in the presence of
>    "unchecked" warning.)

     Exactly.  Consider the situation when generics were new: It
would be quite common for a List<T> to be manipulated by code that
worked with an unadorned List.  Might that other code insert a
non-T in its List?  Yes, obviously.  So, when the originating code
fetches things from what it thinks is a List<T>, it is in fact
fetching them from a plain old List -- and who knows whether the
things in that List are T's or non-T's?

     The sharing of a List<T> with a method that uses plain List
is surely less common nowadays than it was in the past, yet I am
morally certain that it occurs.  (And if the `checkcast' operation
were omitted, you can be Absolutely Sure there would soon be
malware taking advantage of it, putting an EvilString into a
List<String> and going "mwah-haah-haaaa!" when an unsuspecting
user of the List<String> did list.get(0).hashCode() and gave
away the keys to the kingdom ...).

>    A program might define a /private/ generic class and make
>    sure that /all/ accesses to this class use generics, yet it
>    still has to pay the price of those downcasts!

     It would be instructive if you were to share your measurements
of the "price."  (You've made them of course.  Of course?)

-- 
esosman@comcast-dot-net.invalid
Thirteen hundred sixty-seven days to go.