Regex not working in bash

Question

My regex:

 (<property\sname=\"uri\"\s[value=\"htttp:\/\/]+(\d+\.)+\d+)+

Text sample:

 <bean id="journeyWSClient" parent="abstractClient" class=" lib.JourneyWSClient">
        <property name="uri" value="http://192.24.342.432:20010/some/path/1_0_0"/>
        <!-- Fortuna -->        
        <!-- property name="uri" value="http://164.7.19.11:20010/some/path/1_0_0"/ -->

It works on http://regexr.com/, however when I put the regex in bash script, it doesn't work. Are there some characters I need to escape? Ideas?

Bonus cookie for extracting the IP with only one regex.

Certainly you need to replace `\d` with `[0-9]`/`[[:digit:]]` and `\s` with `[[:space:]]`. And replace `[value=\"htttp:\/\/]+` with `value=\"htttp:\/\/`, and replace `(\d+\.)+\d+` with `([0-9]+(\.[0-9]+)+)` to get the IP into a capture group 2. — Wiktor Stribiżew, May 13 '16 at 12:06
@choroba I'm not using regex as HTML parser, just to compare IPs from a XML file to one in ANT build .properties. — matijasx, May 13 '16 at 12:29
@choroba Any commandline tools? I'd like to keep everything in one script, this is only a small part of much larger validation (that doesn't include XML). — matijasx, May 13 '16 at 13:14

score 2 · Accepted Answer · answered May 13 '16 at 12:33

2

Replace the \d with [0-9] and \s with [[:space:]], and adjust the IP matching part as ([0-9]+(\.[0-9]+)+) (or simplify it to ([0-9.]+)) so as to be able to get its value with ${BASH_REMATCH[1]}:

text='<property name="uri" value="http://192.49.200.142:20010/some/path/1_0_0"/>'
regex='<property[[:space:]]name="uri"[[:space:]]value="http://([0-9]+(\.[0-9]+)+)'
if [[ $text =~ $regex ]]; then
    echo ${BASH_REMATCH[1]};
fi

See the IDEONE demo

answered May 13 '16 at 12:33

Wiktor Stribiżew

607,720
39
448
563

why replacing `\d` with `[0-9]`? – JobaDiniz Apr 26 '23 at 11:45
@JobaDiniz POSIX regex flavor does not support `\d`. – Wiktor Stribiżew Apr 26 '23 at 11:46
I'm trying to convert this _semver_ regex to shell script: `^((0|[1-9][0-9]*)\.(0|[1-9][0-9]*)\.(0|[1-9][0-9]*))(?:-((?:0|[1-9][0-9]*|[0-9]*[a-zA-Z-][0-9a-zA-Z-]*)(?:\.(?:0|[1-9][0-9]*|[0-9]*[a-zA-Z-][0-9a-zA-Z-]*))*))?(?:\+([0-9a-zA-Z-]+(?:\.[0-9a-zA-Z-]+)*))?$` and it works on regex101 but not on shell :( – JobaDiniz Apr 26 '23 at 11:59
1

@JobaDiniz Sure, POSIX regex flavor does not support non-capturing groups, replace all `(?:` with just `(`. – Wiktor Stribiżew Apr 26 '23 at 12:36

score 0 · Answer 2 · answered May 13 '16 at 12:34

0

See if this works for you..

value="http:\/\/(([0-9].?)+):

This worked for me if you have IP only with http pattern

grep http TEXT_SAMPLE_FILE |sed -E    's;.*value="http://(([0-9].?)+):.*;\1;g'

answered May 13 '16 at 12:34

Sanjeev

107
6

Regex not working in bash

2 Answers2