RegEx "negative lookahead" within XPath in scorecard

[roederXC]

The regex "negative lookahead" does not work within an XPath in the scorecard-config.JSON

For example, if you use this expression as a rule selector:
"sum(//transactions/transaction[matches(name, '^(?!.*Crawler).+$')]/count)"
you will get this Error in the Report/Scorecard:
Error evaluating rule 'visitsWithoutCrawlerCheck': [Check #0] Syntax error at char 2 in regular expression: No expression before quantifier. Found while atomizing the first argument of fn:sum() in {docOrder(((root/descendant::transactions)/transaction[fn:matches(...)])/count)} on line 1

To better reproduce and handle the issue, I am attaching my debug-scorecard-config.json.
debug-scorecard-config.json

[h-arlt]

Regular expressions in XPath do not support lookahead or lookbehind constructs. That's why it is reported as being erroneous.

See here for details on what regular expression syntax in XPath 3.1.

[h-arlt]

As workaround, you may simply substract the sum of all transaction counts whose names contains TCrawler from the total sum of all transactions: sum(//transactions/transaction/count)-sum(//transactions/transaction[contains(name, 'Crawler')]/count)

[rschwietzke]

With #508 we will get a way to deal with that more easily because we label things upfront and make the queries for the scorecard simpler.

[roederXC]

@h-arlt
Maybe the better workauround is to use only XPath:
`"sum(//transactions/transaction[not(contains(name, 'Crawler'))]/count)"``

But maybe it is still be an known or ongoing issue to us a "?" in ExPath here.

[rschwietzke]

This page explocitly says, that negative lookahead is purposly not part of the spec:
https://www.web3d.org/specifications/X3dRegularExpressions.html

Negative Lookahead filters have potential to disqualify attributes that contain illegal values.

W3C Recommendation for XML Schema (XSD): Regular Expressions does not support this construct, perhaps to avoid the possibility of computational denial-of-service (DoS) attacks.