Replace multiple lines in HTML file which is marked by a start and end pattern.

Question

I'm trying to make an automated build and use Perl in it to update some paths in a file.

Specifically, in an html file, I want to take the block shown below

<!-- BEGIN: -->
<script src="js/a.js"></script>
<script src="js/b.js"></script>
<script src="js/c.js"></script>
<!-- END: -->

and replace it with

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

I have tried a few regexes like:

perl -i -pe  's/<--BEGIN:(.|\n|\r)*:END-->/stuff/g' file.html

or just starting with:

perl -i -pe  's/BEGIN:(.|\n|\r)*/stuff/g' file.html

But I can't seem to get past the first line. Any ideas?

Community · Accepted Answer · 2017-05-23 11:50:01Z

2

perl -i -pe 's/<--BEGIN:(.|\n|\r)*:END-->/stuff/g' file.html

This is so close.

Now just match with the /s modifier, this allows . to match any char, including newlines.
Most importantly, you want to start the match with <!--, note the !.
Also, you want a non-greedy match like .*?, in case you have multiple END markers.
Your example input shows that there may be extra spaces.

This would lead to the following substitution:

s/<!--\s*BEGIN:.*?END:\s*-->/stuff/sg

As @plusplus pointed out, the -p iterates over each line. Let's change Perl's concept of a “line” to “the whole file at once”:

BEGIN { $/ = undef }

or use the -0 command line switch, without a numeric argument.

edited May 23, 2017 at 11:50

CommunityBot

11 silver badge

answered Jun 18, 2013 at 10:53

amon

57.8k2 gold badges93 silver badges152 bronze badges

Sign up to request clarification or add additional context in comments.

5 Comments

plusplus Over a year ago

does that work for you? perl -p is running over the file line by line (while (<>) etc) so it seems unlikely to match like this...?

PeterDev Over a year ago

Thanks! The above regex works beautifully if I put all the above html on one line - all the text is replaced with "stuff". If I separate the html into multi lines, it fails to match. I also tried removing the "-p". Hmmm now I remember why I do my utmost to avoid regex :)

plusplus Over a year ago

see the edit above - amon's original post was untested and didn't work.

PeterDev Over a year ago

Oow sounds fun so would this look something like: <code>perl -i -e 'BEGIN { $/ = undef } s//stuff/sg' file.html</code> must confess I haven't touched Perl in ages.

plusplus Over a year ago

perl -0 -i -pe 's//stuff/sg' file.html (PS - it's backticks to highlight as code in comments)

Collectives™ on Stack Overflow

Replace multiple lines in HTML file which is marked by a start and end pattern.

1 Answer 1

5 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

5 Comments

Your Answer

Sign up or log in

Post as a guest

Related