You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Follow-ups from #547: one test still disabled, two bugs its tests pin #560
An XML ACK with an unknown code is treated as success.
HL7v2ResponseValidatorTest.testXmlUnknownAckCodeLeavesStatusUnchanged pins this. On a running engine (same commit): a destination whose response is an XML ACK with MSA.1 = XX ends SENT, while the same code in an ER7 ACK ends in error with "Invalid HL7 v2.x acknowledgement received." AA and AE behave the same on both paths, so only the unknown-code case differs. The XML branch in HL7v2ResponseValidator has no fallthrough after the success-code check.
Digest auth accepts a request that omits the algorithm and qop directives, whatever the listener allows.
DigestAuthenticatorTest.testNonceWithoutCountCanBeReused pins the nonce half. On a running engine with an HTTP Listener restricted to MD5-sess and auth-int: a request with no algorithm and no qop directive and the right password gets 200, handled as MD5 with no qop. Stating algorithm=MD5 or qop=auth gets 401, wrong password gets 401, MD5-sess with auth-int gets 200. Cause: DigestAuthenticator checks algorithm == null and qop == null before it checks the enabled sets. Effect for someone who has the password: no body integrity, and the header replays for the nonce's 60 s life with no nonce count. Needs the password, so not a break-in; still, the restriction doesn't restrict.
I would like to take item 2 (an XML ACK with an unknown code is treated as success).
What the change does:
In HL7v2ResponseValidator, the XML branch gets the fallthrough that the ER7 branch already has. An ACK code that is not in the successful, error or rejected lists is now invalid: status QUEUED with "Invalid HL7 v2.x acknowledgement received."
An XML response with no MSA.1 value gets the same result, because an empty code matches none of the lists. This matches the ER7 path, where a missing MSA segment is already invalid. If you prefer to leave that case as it is, tell me and I will narrow the change.
testXmlUnknownAckCodeLeavesStatusUnchanged becomes testXmlUnknownAckCodeIsInvalid, and there is one new test for the missing MSA case.
Items 1 and 3 are not part of this change. I have the patch ready locally and will open a draft PR.
Activity
Hisham-InfoGleam commentedon Nov 13, 2026
I would like to take item 2 (an XML ACK with an unknown code is treated as success).
What the change does:
HL7v2ResponseValidator, the XML branch gets the fallthrough that the ER7 branch already has. An ACK code that is not in the successful, error or rejected lists is now invalid: status QUEUED with "Invalid HL7 v2.x acknowledgement received."MSA.1value gets the same result, because an empty code matches none of the lists. This matches the ER7 path, where a missing MSA segment is already invalid. If you prefer to leave that case as it is, tell me and I will narrow the change.testXmlUnknownAckCodeLeavesStatusUnchangedbecomestestXmlUnknownAckCodeIsInvalid, and there is one new test for the missing MSA case.Items 1 and 3 are not part of this change. I have the patch ready locally and will open a draft PR.
Hisham-InfoGleam commentedon Nov 13, 2026
@pacmano1 you wrote that you are happy to open item 1. If you have not started it, I can take it. I have the rename ready locally:
TemplateValueReplacerTestsbecomesTemplateValueReplacerTest(file and class). Nothing else changes../gradlew :server:testrun (850 tests, 0 failures).If that is fine with you, I will open a PR. If you prefer to do it yourself, no problem, I will drop mine.
Item 3 is not part of this.