Fix confusing example in paste(1)

[jocki84]

Paste's man page contains an example for a reimplementation of nl(1). This example uses the command line

sed = myfile | paste -s -d '\t\n' - -

in order to concatenate consecutive lines with an intervening tab.

However, the way the example uses the switches -s and -d and two - input files is redundant. There are in fact two equivalent but simpler ways to achieve the desired result:

sed = myfile | paste -s -d '\t\n' -

uses the same style as the previous example, while

sed = myfile | paste - -

is arguably even simpler and illustrates the final sentence of the DESCRIPTION.

I therefore suggest to change the example to the latter style.

[ngie-eign]

Hmmm... it's not exactly equivalent from what I can tell, but I guess it's close..?

$ cat myfile 
a=b
c: d
'e'^1
$ sed = myfile | paste - -                                                                                                                                            
1       a=b
2       c: d
3       'e'^1
$ nl myfile 
     1  a=b
     2  c: d
     3  'e'^1

I don't see why this shouldn't be merged (.Dd will just need to be bumped for the change).

CC: @allanjude

[jocki84]

@ngie-eign yes, it's only nl(1) »in spirit« – the text does mention that the result is »similar« to the output of nl(1).

I think it would be a good idea to use paste - - in this example to demonstrate a technique that only works with stdin. For physical files, paste -s -d '\t\n' has to be used to achieve the effect, which is what the previous example is about.

[bsdimp]

Closing... pushed into head as 8790fe3. The example is simpler and easier to follow. It doesn't have to be perfectly the same as nl(1).