How exactly does sed view meta characters when you replace or delete text?

[null_reflections]

I'm trying to figure out how to use sed without erasing extra things that i want to keep in the document. So far, i've realized that sed is largely user-friendly in terms of how it does away with strings and numbers:

sed -i '/cow/d' <file>

sed -i '/10/d' <file>

in both of these examples, sed will only get rid of entire instances of "10" and "cow", and will not erase every 1,0,c,o,w. However, it doesn't work this way with metacharacters. Let's say i had "**" in a document that i wanted to erase, how do i do that without getting rid of every asterisk? For some reason, it erases all the test in the document when i do

sed -i '/**/d' <file>

, is it just reverting to the bash wildcard? Why does it do that?

Just to be clear, this works just fine if i wanted to get rid of all the asterisks for some reason:

sed -i '/*/d' <file>

 

[null_reflections]

I've answered part of this: you can edit separate strings in sed like this, and it works if you have two ** in a file but only want to get rid of that particular string:

sed -i '/\*\*/d' <file>

This will get rid of every instance where there are two asterisks next to each other, but not instances where there's a lone asterisk. Sed tends to find exact matches unless you use meta characters in regular expressions to broaden things

However, i still wonder why sed erases all the text in my file when i have the two asterisks next to each other.