Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:67806
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.212.178 as permitted sender)
Date: Tue, 25 Jun 2013 15:15:30 +0200
To: Florin Patan <florinpatan@gmail.com>
Cc:  <internals@lists.php.net>
Message-ID: <etPan.51c997f2.836c40e.5757@MacBook-Air-van-Joost-Koehoorn.local>
In-Reply-To: <CAA9esyRFP58dV0LraX5YjA31xoeOqRe5pe2Zd2zK9tVs_9gqPg@mail.gmail.com>
References: <etPan.51c8ce38.54e49eb4.5757@MacBook-Air-van-Joost-Koehoorn.local>
 <CAA9esyRFP58dV0LraX5YjA31xoeOqRe5pe2Zd2zK9tVs_9gqPg@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Subject: =?UTF-8?Q?Re=3A_Announcing_RFC_'Anonymous_Catches'?=
From: joost.koehoorn@gmail.com (Joost Koehoorn)

On Tue, Jun 25, 2013 at 12:54 AM, Joost Koehoorn=C2=A0
<joost.koehoorn=40gmail.com> wrote:=C2=A0
> Hi all,=C2=A0
>=C2=A0
> I just published an R=46C that proposes to add catch-statement without =
needing to specify a variable, and support for fully anonymous catches.=C2=
=A0
>=C2=A0
> Details can be found at:=C2=A0
>=C2=A0
> https://wiki.php.net/rfc/anonymous=5Fcatch=C2=A0
>=C2=A0
> Regards,=C2=A0
>=C2=A0
> Joost Koehoorn=C2=A0
>=C2=A0
> --=C2=A0
> PHP Internals - PHP Runtime Development Mailing List=C2=A0
> To unsubscribe, visit: http://www.php.net/unsub.php=C2=A0
>=C2=A0

Hi,=C2=A0


=46irst of, considering that you would only save three characters by=C2=A0=

having the ' =24e' removed, I don't see how it would make things easier=C2=
=A0
to learn. I mean if you can't learn that maybe you should look into=C2=A0=

something else that's not related to programming :)
I don't recall mentioning that this makes PHP easier to learn. It's reall=
y about avoiding introducing unused variables, which is currently require=
d because the language simply requires and identifier there.

Having this: =22* Avoid indication of a specific exception type=22 could=C2=
=A0
be considered as a possible bad practice, just as you were saying that=C2=
=A0
catch(=5CException =24e) is. Also, it you can add the other ten character=
s=C2=A0
that you save when typing =5CException but does it worth it=3F
I don't really care about saving a few characters, that's not what this i=
s about. I can see that a catch-all can be considered bad practice, but p=
erform a Github search for =60catch (Exception =24e)=60 and find very muc=
h instances of this usage. In all of these situations it seems that only =
specific exceptions are caught, but that's not actually the case. We curr=
ently see that requiring an Exception type doesn't really enforce good pr=
ogramming practice, most instances simply use =60(=5CException =24e)=60 a=
nd have the same bad practice.

How do you plan to handle this: * Note that there may only be one=C2=A0
anonymous catch-statement, and it has to be the last one. Throw a=C2=A0
fatal error=3F If not then what, nothing=3F
In my current implementation, it's simply a syntax error and thus will ne=
ver result in compiled code.

What are the performance downsides of having this check in place=3F I do=C2=
=A0
understand that you mentioned: =22* Runtime needs to perform less=C2=A0
checks. No need to update the symboltable, no need to check exception=C2=A0=

inheritance. =22 but wouldn't that be replaced by what I'm asking about=3F=
=C2=A0
If I could read/understand C better I wouldn't ask for this.
As it's directly enforced by the language grammer, Bison (the parser gene=
rator) won't accept invalid use of the anonymous catch, and thus no addit=
ional runtime checks are necessary.

Could you also please elaborate this: =22* Better possibilities for=C2=A0=

static analysis tools=22=3F=C2=A0
This is about unused variables. When you use an IDE or tools such as PHPM=
D, they will warn you about unused variables in your code. Exception vari=
ables are simply required by the language, even though you may not actual=
ly use them, thus generating unused variables warnings (or the analyser d=
oesn't flag exception variables, in which case you miss unused variables)=
 which you cannot solve.=C2=A0

As for: * People from other languages such as C=23, Python and Ruby=C2=A0=

expect this to work I think it is often pointed out when requesting=C2=A0=

features such as function/method return type hinting (I'm planning a=C2=A0=

R=46C for that), named parameters, threads and other stuff that this=C2=A0=

ain't C=23, Java, Python, Ruby or whatnot so.. what's the official=C2=A0
position on this=3F Do we want PHP like the other languages or not (and=C2=
=A0
I won't say more on this here/now)=3F
I can agree with you here, PHP has its own things and is different from o=
ther languages for a reason. I'm mentioning this because I can see how Py=
thon/Ruby developers working with PHP may find this a flaw in PHP, giving=
 them another reason to bitch about PHP.

Also, the code sample is not that good. If you want to reflect a good=C2=A0=

quality scenario, it should be more like:=C2=A0

while (true) =7B=C2=A0
try=C2=A0
=7B=C2=A0
=24this->connectToServer();=C2=A0
break;=C2=A0
=7D=C2=A0
catch (Connection=46ailedException)=C2=A0
=7B=C2=A0
sleep(3);=C2=A0
=7D=C2=A0
=7D=C2=A0

And I think the R=46C should also add a: Disadvantages section so that=C2=
=A0
it could help people looking on it and proving that you've done all=C2=A0=

the research correctly.
I couldn't really come up with disadvantages. Anonymous catch-statements =
may be considered bad coding practice, but look at the amount of develope=
rs simply using =60Exception=60 as the type which is the same bad practic=
e, only with the false indication that a specific type is expected. By al=
lowing for not specifying the type at all, it's immediately clear that th=
e intention was to not care about the type at all.


Thanks=C2=A0
----=C2=A0
=46lorin Patan=C2=A0
https://github.com/dlsniper=C2=A0
http://www.linkedin.com/in/florinpatan