Re: [Help-bash] append to $@ one by one

[Eli Schwartz]

On 10/20/19 7:29 AM, Peng Yu wrote:
>> If that is too slow compared
>> to using an array, then I must assume that you have so much data that
>> doing whatever you're doing in another language would be preferable.
> 
> Stdin can have an unlimited amount of data. This approach time
> complexity is not linear with respect to the input. That is what I
> mean by too slow.

Reading lines one at a time and mutating "${a[@]}" repeatedly is also
not very fast when dealing with unlimited data, maybe you'd prefer to
use `mapfile -t a` which also avoids the need to set IFS=

But I figure any script dealing with unlimited number of args will also
have to... do something with an unlimited number of args, so that will
be pretty slow too.

>> Also, is the use of an additional array variable not wanted due to
>> purely aesthetic reasons?
> 
> Yes. The name of the variable "a" is nonessential. It is mentioned
> twice. From the succinct perspective of the language, if it can be
> avoided, it would be best.

It would be even nicer if we could avoid other non-succinct things, like

while
IFS=
read
do
done
set

All of which could be gotten rid of if bash just knew what you wanted. I
demand a new feature for bash which reads my mind and does what I want
without requiring me to program it first.

...

Anyway, the Environmental Protection Agency recommends using `unset a`
when done. Does this suffice?

-- 
Eli Schwartz
Arch Linux Bug Wrangler and Trusted User

signature.asc
Description: OpenPGP digital signature