Downloading a file with a trivial name without overriding a possible existing file with the same name

Question

In CentOS with Bash I download a file with a very trivial file name, install.sh.

Although extremely unlikely in my scenario, this could still, in theory, overwrite some existing file with the same name.

How would you solve this problem?

---

If I understood correctly, one way to deal with this is:

cd DOWNLOADS &&
source <( wget --quiet -O - 'URL/install.sh' )

But as a non professional sysadmin I seek a more "basic" or "simple" way because I am having hard time working with the <(...) syntax, let along with downloading to stdout with various options.

I was thinking about:

• Download a file with a randomly created long file name
• Do action with the file (such as cp and mv to change name)
• delete the downloaded file

Perhaps something like this pseudocode:

random_name=(wget -O kjghfkjdhsgkj.sh URL) # How to create random name by the computer?
destination=(cp "${random_name}" DESTINATION)
mv "${destination}" NEW_DESTINATION
rm "${random_name}"

Answer 1

mktemp is exactly here to let us create a file with a "semi-random" templated filename. See man mktemp; and for cleaning part, see trap.