Hello, I have a question regarding Unicode characters in Contracts. I'm using expressions like

/.{1,4}/

to just assert that a string has a certain length. The dot also includes unicode characters, which is intentional. Pact-JVM saves Unicode characters in the contract file using a backslash escape notation, like

\uDAA5

for a surrogate, even though the file is UTF-8. However, the Pact CLI Tools seem not to be able to handle some of those sequence. By example, the surrogate escape causes the throw of "incomplete surrogate pair at '\uDAA5". Where lies the problem here? Should Pact-JVM just use the actual values or should the CLI tools be able to handle the escaped sequences?

Good question. I think we may have seen some of these errors in PactFlow (probably the same issues in the Ruby CLI tooling as they use the same underlying libs). I don’t have the answers, but looping in a few people to start a convo cc: @phanindra srungavarapu @rholshausen

@rholshausen mentioned [here](https://github.com/pact-foundation/pact-jvm/issues/1536#issuecomment-2533208036) that pact-jvm intentionally escapes all non-ascii sequences.

Oh, that old chestnut. Yes, we recommend giving an example value (second parameter) to avoid this.

I find this a bit sad, because generating the example value from the regex forces the application to actually support the specified format. By example, if I specify

.{9}

in the contract tests but actually just support

[a-Z0-9]

this wouldn't be caught by pact as long as the example value I provide matches

[a-Z0-9]

. Of course this may still happen with a working generator, as long as I CAN provide an example.