Mailing-List: contact dev-help@directory.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Apache Directory Developers List" <dev@directory.apache.org>
Received-SPF: pass (nike.apache.org: domain of akarasulu@gmail.com designates
 209.85.161.50 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:sender:in-reply-to:references:date
         :x-google-sender-auth:message-id:subject:from:to:content-type;
        b=o1Fl3QoYIHGQssiiZem3fFk97DovjT0KukD/NSa2ocEY6e5SCIkZNo1pObKGTG4bzt
         XGW7RUOEhDgBwVaO8Ym01w3GjeiJGiNf4L83JtHwP4egJGC+JYc3zh8Lsse/eiz7mNEO
         cg54wPbRogV6jKpk0FRnN9qwbHgPqPeJdMp3I=
MIME-Version: 1.0
Sender: akarasulu@gmail.com
In-Reply-To: <609A6086-AA46-463E-A5F9-7D63AFAA7A92@toolazydogs.com>
References: <4C08D23D.8060803@gmail.com>
	<4C08ED1D.4080403@apache.org>
	<609A6086-AA46-463E-A5F9-7D63AFAA7A92@toolazydogs.com>
Date: Wed, 9 Jun 2010 11:32:48 +0300
Message-ID: <AANLkTikBco2ohEYKhjixjLQYYNLPoG8MFOTCjznB-MtT@mail.gmail.com>
Subject: Re: Should we use assert ?
From: Alex Karasulu <akarasulu@apache.org>
To: Apache Directory Developers List <dev@directory.apache.org>
Content-Type: multipart/alternative; boundary=001485f20c543f6a25048894c083

--001485f20c543f6a25048894c083
Content-Type: text/plain; charset=ISO-8859-1

On Fri, Jun 4, 2010 at 4:08 PM, Alan D. Cabrera <list@toolazydogs.com>wrote:

>
> On Jun 4, 2010, at 5:10 AM, Stefan Seelmann wrote:
>
> > Emmanuel Lecharny wrote:
> >> Hi guys,
> >>
> >> currently (and probably because nobody uses them, due to some Java 1.3
> >> habits we have), we don't use asserts to do simple things like checking
> >> methods parameters (pre-conditions).
> >>
> >> Should we start using them ?
> >
> > I never used them before.
> >
> > For me assert is a bit magic because they are disabled at runtime and
> > the assert statement isn't evaluated unless they are enabled. I prever
> > to throw a IllegalArgumentException instead.
>
> I like asserts because they have the dual function of performing useful
> sanity checking when turned on and also provide implicit documentation to
> other developers as to what my fundamental assumptions are when I designed
> my code.  Take for example
>
>        assert Thread.holdsLock(foo) : "I should be protected before you
> call me";
>
> This is a nice bit of code that explains my assumptions to other developers
> when they use/extend my code with the added benefit that it gets turned off
> once the code is deployed in production.
>
> Slavish use of asserts can lead to trouble; they are not a panacea.
>  Frontline argument and state checking at the client API end should never
> use asserts for obvious reasons.  However, performing the same kind of heavy
> weight checking in my internal classes where I have complete control is a
> bit of overkill; here asserts fit nicely.  They also serve to warn the
> developer that they are now troweling inside the bowels of a codebase.
>
> Just my 2 cents.
>
>
Alan you make very good points. I don't think we should be using asserts for
API semantic checks nor going wild with it but your subtle usage to show
correctness to readers of the code is very important.

Seelmann's points are also solid as well. I would rather use the proper IAE
in conjunction with the proper parameter checks. This falls in line with
Alan's comments.

I'd like to see Alan's approach while writing code to help readers
understand. Plus it has value when assertion checking is enabled during
tests.

My 2 cents.

-- 
Alex Karasulu
My Blog :: http://www.jroller.com/akarasulu/
Apache Directory Server :: http://directory.apache.org
Apache MINA :: http://mina.apache.org
To set up a meeting with me: http://tungle.me/AlexKarasulu

--001485f20c543f6a25048894c083
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<div class=3D"gmail_quote">On Fri, Jun 4, 2010 at 4:08 PM, Alan D. Cabrera =
<span dir=3D"ltr">&lt;<a href=3D"mailto:list@toolazydogs.com">list@toolazyd=
ogs.com</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" style=3D=
"margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div><div></div><div class=3D"h5"><br>
On Jun 4, 2010, at 5:10 AM, Stefan Seelmann wrote:<br>
<br>
&gt; Emmanuel Lecharny wrote:<br>
&gt;&gt; Hi guys,<br>
&gt;&gt;<br>
&gt;&gt; currently (and probably because nobody uses them, due to some Java=
 1.3<br>
&gt;&gt; habits we have), we don&#39;t use asserts to do simple things like=
 checking<br>
&gt;&gt; methods parameters (pre-conditions).<br>
&gt;&gt;<br>
&gt;&gt; Should we start using them ?<br>
&gt;<br>
&gt; I never used them before.<br>
&gt;<br>
&gt; For me assert is a bit magic because they are disabled at runtime and<=
br>
&gt; the assert statement isn&#39;t evaluated unless they are enabled. I pr=
ever<br>
&gt; to throw a IllegalArgumentException instead.<br>
<br>
</div></div>I like asserts because they have the dual function of performin=
g useful sanity checking when turned on and also provide implicit documenta=
tion to other developers as to what my fundamental assumptions are when I d=
esigned my code. =A0Take for example<br>

<br>
 =A0 =A0 =A0 =A0assert Thread.holdsLock(foo) : &quot;I should be protected =
before you call me&quot;;<br>
<br>
This is a nice bit of code that explains my assumptions to other developers=
 when they use/extend my code with the added benefit that it gets turned of=
f once the code is deployed in production.<br>
<br>
Slavish use of asserts can lead to trouble; they are not a panacea. =A0Fron=
tline argument and state checking at the client API end should never use as=
serts for obvious reasons. =A0However, performing the same kind of heavy we=
ight checking in my internal classes where I have complete control is a bit=
 of overkill; here asserts fit nicely. =A0They also serve to warn the devel=
oper that they are now troweling inside the bowels of a codebase.<br>

<br>
Just my 2 cents.<br><br></blockquote><div><br></div><div>Alan you make very=
 good points. I don&#39;t think we should be using asserts for API semantic=
 checks nor going wild with it but your subtle usage to show correctness to=
 readers of the code is very important.</div>
<div><br></div><div>Seelmann&#39;s points are also solid as well. I would r=
ather use the proper IAE in conjunction with the proper parameter checks. T=
his falls in line with Alan&#39;s comments.</div><div><br></div><div>I&#39;=
d like to see Alan&#39;s approach while writing code to help readers unders=
tand. Plus it has value when assertion checking is enabled during tests.=A0=
</div>
<div><br></div><div>My 2 cents.</div><div>=A0</div></div>-- <br>Alex Karasu=
lu<br>My Blog :: <a href=3D"http://www.jroller.com/akarasulu/">http://www.j=
roller.com/akarasulu/</a><br>Apache Directory Server :: <a href=3D"http://d=
irectory.apache.org">http://directory.apache.org</a><br>
Apache MINA :: <a href=3D"http://mina.apache.org">http://mina.apache.org</a=
><br>To set up a meeting with me: <a href=3D"http://tungle.me/AlexKarasulu"=
>http://tungle.me/AlexKarasulu</a><br>

--001485f20c543f6a25048894c083--