shell - What is the internal processing of $* and $@ -

actually, understand use of $* , $@.

for example, if run script using: my_script *

then, one_file in $@ each file one_file processing.
there space(s) in filenames, one_file correct filename.

if, however, using one_file in $*, story different.
think guys understand difference , not need go further.

now, interested how.
how kornshell (ksh) interprets my_scrpt *
, pass filenames $@ correctly
, pass filenames $*.

for example, when ksh see my_script *
, put filenames 1 one array,
, put array[1][2][3]... $@ processing ?
and, when seeing $*, concat
filename1+space+filename2+space+... ?

i know may relate more internal coding of ksh.

any insight?

for example, when korn shell see my_script *, put filenames 1 one array, , put array[1][2][3]... $@ processing ? and, when seeing $*, concat filename1+space+filename2+space+... ?

yes, pretty much.

one important thing realize here there 2 separate processes involved: calling shell, expands my_script * myscript arg1 arg2 arg3 , subprocess gets array of arguments. if subprocess ksh (or sh, bash, or similar shell) $* syntax "get argv[1...] concatenated string" while $@ syntax "get argv[1...] array". it's important note $* , $@ meaningfully different when quoted.

in bash, cousin of ksh, idea of array variables has been generalized. (perhaps ksh has done -- haven't used ksh decade...)

in bash, creates array of 3 elements called a. note first element contains space:

[~]$ a=("foo bar" baz quux) [~]$ in ${a[*]}; echo $i; done foo bar baz quux [~]$ in ${a[@]}; echo $i; done foo bar baz quux [~]$ in "${a[*]}"; echo $i; done foo bar baz quux [~]$ in "${a[@]}"; echo $i; done foo bar baz quux 

as can see example, "${a[*]}" concatenates array while "${a[@]}" gives individual elements.