Regex matches under RegexRx, but not under Xojo?

General
1

this is certainly for @Kem_Tekinay !

I have a regex that matches the pattern under regexrx, but match stays nil under xojo
any idea of what it can come from ?

I copied the sqlsource and rx.searchpattern from the xojo debugger into regexrx.

thanks.

Capture d’écran 2023-10-30 à 21.40.07
Capture d’écran 2023-10-30 à 21.40.071884×304 126 KB

Capture d’écran 2023-10-30 à 21.40.19
Capture d’écran 2023-10-30 à 21.40.191694×1748 273 KB

2

Log the SearchPattern and compare the output to what’s in RegExRX. I’ll bet you’ll spot an unexpected difference. (Keep in mind that there could be invisible like EOL characters.)

3

there is an \n before FOREIGN in the searchpattern, but it is also in the pattern so it IS matched
is there a xojo option to match the newline ?

what do you mean exactly by “log the searchpattern” ? in xojo ? in regexrx ?

4

I meant something like using System.DebugLog, but since there are invisible characters are play here, try this:

var pattern as string = *your search pattern*

var c as new Clipboard
c.Text = pattern
c.Close

rx.SearchPattern = pattern
// Etc

Run that, then paste into a new RegExRX window against the same text. It should be easy to spot the differences at that point.

Also, unless you need that comma, I’d change the pattern to:

\bFOREIGN\b[^(]*\( ...
1 Like
5

so you think there may be some invisible characters in the pattern ?
I will look into that.

6

When my regexes work in a regex test app and not in Xojo it’s either the end-of-line characters or some options like greedy that are the culprits.

7

I played with the different options like greedy
 regexrx always get the match, and xojo don’t.
but it could be the newline that causes this.
anyway I think regexrx has been made with xojo, so the match results should be the same ?

8

the pattern doesn’t seems to have hidden characters in it.

Capture d’écran 2023-10-31 à 09.47.28
Capture d’écran 2023-10-31 à 09.47.281524×252 78.4 KB

9

if I remove the \n from the pattern, the regex now has a match in xojo 

anyway I don’t understand why the result is different between xojo and regexrx ?

10

If there is a hidden character then it’s in the text and not the regex. Try with (\r|\n) instead of \n.

11

it is really a 0x0D newline character in the text

Capture d’écran 2023-10-31 à 10.09.51
Capture d’écran 2023-10-31 à 10.09.511502×364 204 KB

12

seems there was a mix between 0x0a (endofline) and 0x0d

thanks Beatrix.

1 Like
13

FYI, \R matches any eol character.

2 Likes