Question

I am executing a shell script where i should replace string in a file. So of course i am using sed for that. The string is something like:

<!--replace-->
<script src="1.js"></script>
<script src="2.js"></script>
<!--/replace-->

This should be replaced with:

<script src="3.js"></script>

So i am trying to do it as follows:

$ sed "s/<\!--replace-->.\*<\!--\/replace-->/<script src=\"3.js\"><\/script>/g" file.html

But this does not do the trick. What am i missing?

UPD: Removing \n from file is not a solution in this case, simply because file should be readable.

Answer 1

The problem is related to newlines, you can change your regex to match newlines, or remove newlines
using tr :

user@machine:/tmp$ cat s.txt 

<!--replace-->
<script src="1.js"></script>
<script src="2.js"></script>
<!--/replace-->

user@machine:/tmp$ tr "\n" " " < s.txt | sed "s#<\!--replace-->.*<\!--/replace-->#<script src="3.js"></script>#g" 

<script src=3.js></script> 

In case you do not want to remove the newlines, you can :

a) Try a solution using SED as mentioned here:
how to tell sed "dot match new line"

b) use perl or python to replace the multiline regex,with s modifier , the newlines will be intact :

user@mmmm:/tmp$ cat s.txt 

<!--replace-->
<script src="1.js"></script>
<script src="2.js"></script>
<!--/replace-->

Dummy content containing
newlines for test.

user@mmmm:/tmp$ perl -i -pe 'BEGIN{undef $/;} s#<\!--replace-->.*<\!--/replace-->#<script src="3.js"></script>#gsm' s.txt 

user@mmmm:/tmp$ cat s.txt 

<script src="3.js"></script>

Dummy content containing
newlines for test.