Path: csiph.com!newsfeed.hal-mli.net!feeder3.hal-mli.net!newsfeed.hal-mli.net!feeder1.hal-mli.net!newsfeed.xs4all.nl!newsfeed3.news.xs4all.nl!xs4all!newsgate.cistron.nl!newsgate.news.xs4all.nl!post.news.xs4all.nl!not-for-mail
MIME-Version: 1.0
In-Reply-To: <roy-3F8596.09385002032014@news.panix.com>
References: <mailman.7568.1393756930.18130.python-list@python.org> <roy-3F8596.09385002032014@news.panix.com>
Date: Sun, 2 Mar 2014 19:08:42 +0000
Subject: Re: why indentation should be part of the syntax
From: Nicholas Cole <nicholas.cole@gmail.com>
To: Roy Smith <roy@panix.com>
Content-Type: text/plain; charset=ISO-8859-1
Cc: Python <python-list@python.org>
Precedence: list
Newsgroups: comp.lang.python
Message-ID: <mailman.7589.1393787329.18130.python-list@python.org>
Lines: 42
NNTP-Posting-Host: 2001:888:2000:d::a6
Xref: csiph.com comp.lang.python:67459

On Sun, Mar 2, 2014 at 2:38 PM, Roy Smith <roy@panix.com> wrote:
> In article <mailman.7568.1393756930.18130.python-list@python.org>,
>  Stefan Behnel <stefan_ml@behnel.de> wrote:
>
>> Haven't seen any mention of it on this list yet, but since it's such an
>> obvious flaw in quite a number of programming languages, here's a good
>> article on the recent security bug in iOS, which was due to accidentally
>> duplicated code not actually being as indented as it looked:
>>
>> https://www.imperialviolet.org/2014/02/22/applebug.html
>>
>> Stefan
>
> Hogwash.  What this looks like is two gotos in a row.  Anybody who
> reviewed this code would have thrown up a red flag when they saw two
> gotos in a row.  If anything, the "incorrect" indentation makes it even
> more obvious.  Any static code analyzer would have also caught this as
> an unreachable statement.
>
> Paraphrasing this into Python, you get:
>
> def bogus():
>     if SSLHashSHA1.update(hashCtx, serverRandom) != 0:
>         raise fail
>     if SSLHashSHA1.update(hashCtx, signedParams) != 0:
>         raise fail
>         raise fail
>     if SSLHashSHA1.final(hashCtx, hashOut) != 0:
>         raise fail
>
> which is syntactically valid (at least, I can import it), but clearly
> not what the author intended.  So how did Python's indentation rules
> save us?

Actually, that's incorrect.  The bug (written in Python) would have been:

if SSLHashSHA1.update(hashCtx, signedParams) != 0:
    raise fail
raise fail # ie. no indent.

If written with the indent, it's a useless line of code, but it
doesn't become a bug.