Re: lynx-dev patch - "JUSTIFY"

[mattack]

On Wed, 14 Jul 1999, T.E.Dickey wrote:

>Date: Wed, 14 Jul 1999 14:44:54 -0400 (EDT)
>From: "T.E.Dickey" <address@hidden>
>Reply-To: address@hidden
>To: address@hidden
>Subject: Re: lynx-dev patch - "JUSTIFY"
>
>> I realize this is a religious issue, but what I found most difficult about 
>> reading the Lynx code is the brace style. 
>>  
>> There are many different ways of implementing the following (and I realize I 
>> do it differently from others, but will conform to slightly different 
>> styles when dealing with other people's code), but simply matching up 
>> paired curly brackets would help readability a lot. 
>>  
>>   if (LYNoCore) { 
>>       exit(-1); 
>>   } 
>
>where _don't_ they match?  I have some simple checks that I run periodically
>to find stuff like

As I said, they don't match vertically..

if (LYNoCore) 
{ 
        exit(-1); 
} 

(that's no exactly how I would indent if I were writing something from scratch,
but I'm trying to show the least common denominator in the style I mean.)

Even on a 24 line terminal, I'd rather waste the line to have it readable at
a glance rather than having to figure out/use a software feature to match
up the braces.

if (this)
  {
                if (that)
                        {
                                so and so
                        }
                else
                        {
                                such and such
                        }
        }

is readable at a glance, moreso than putting the curly brackets on the same 
line as the if.  (That above is my "weird" style of indenting where even the
brackets get indented..  It's basically a C-ized version of the Pascal indenting
I learned senior year of high school, in 1986-1987.)

Aaagh, I didn't want to get into this, I just wanted to mention that this
indenting style, as well the lack of comments, are what makes lynx lynx hard
to read.

And even if the switches-within-switches were broken into their own functions,
simply for readability purposes (even though they add a tiny bit of runtime
overhead), would be an improvement.

ohwell