Re: [Pan-users] Pan-users Digest, Vol 106, Issue 10

[Graham Lawrence]

> Date: Wed, 9 Nov 2011 01:40:12 +0000 (UTC)
Excerpted From: Duncan <address@hidden>

>
> LF="
> "
>
Yes, I use this form a lot also, but only with echo and bash
concatenation, because it is utterly reliable there, and I use the
ANSI-C hex form only with IFS.  Its a result of googling for answers,
get a solution that works in a given context, and one tends always to
use it in that context.

> MS (or old-Mac) line-terminations (CRLF and CR, respectively,
> Unix is of course LF), \e escape, \b backspace, octal, hex and
> unicode value entries (you used the hex above), and even \cX
> for control-X notation.
>
> For many, either the literal characters within quotes, or the \X
> letter notation will be easier to read than the \hHH or similar
> hex/octal/unicode notation, so it could be preferred (where there
> is such a version, of course) where others may be reading your code.
>
The whole business of quoting and whitespace is incredibly fouled up.
Bash has at least 4 separate methods of quoting things, and there are
divers methods of representing whitespace, and there is no consistency
as to which methods work in which contexts.

>
> None-the-less, getting in the habit of thinking about efficiency
> when it really doesn't matter will help you when it does, too. =:^)
>
It pains me to write clumsy code, though unavoidable at times.  One
way I can detect it when trying a new command, is if there is a
perceptible delay before the command prompt returns.  My stuff almost
always involves just a single file, and usually rather small, so if
the prompt return is not instantaneous, its a fairly good bet there is
a better way of doing it.