On 9. 7. 2026 15:13, [email protected] wrote:
Author: brane
Date: Thu Jul 9 13:13:06 2026
New Revision: 1936025
Log:
* branches/1.15.x/STATUS: Approve r1935776
Modified:
subversion/branches/1.15.x/STATUS
Modified: subversion/branches/1.15.x/STATUS
==============================================================================
--- subversion/branches/1.15.x/STATUS Thu Jul 9 13:09:19 2026
(r1936024)
+++ subversion/branches/1.15.x/STATUS Thu Jul 9 13:13:06 2026
(r1936025)
@@ -47,15 +47,6 @@ Candidate changes:
+1: ivan, kotkov
+0: brane (review only, can't test)
- * r1935776
- Rework XML output validation in tests.
- Justification:
- Support testing with embeddable Python that doesn't provide 'venv'.
- Branch:
- 1.15.x-r1935776
- Votes:
- +1: ivan, kotkov
-
* r1935107
Fix a bug in JavaHL where native code could modify a final attribute.
Votes:
@@ -66,3 +57,12 @@ Veto-blocked changes:
Approved changes:
=================
+
+ * r1935776
+ Rework XML output validation in tests.
+ Justification:
+ Support testing with embeddable Python that doesn't provide 'venv'.
+ Branch:
+ 1.15.x-r1935776
+ Votes:
+ +1: ivan, kotkov, brane
Works for me, with and without schema validation.
One thing I noticed is that with XML schema validation turned off, there
is no indication in the test summary of the failed tests other than this
one line:
XML: Module lxml.etree not found
as well as half a dozen test failures with no additional explanation.
That's hardly descriptive and doesn't tell the casual tester what went
wrong. We can live with that in 1.15 but previously, on trunk, the test
driver knew in advance that the required modules were not available and
would print appropriate error messages for context.
I'd prefer the test run to just fail up front if schema validation was
enabled but the required modules weren't available.
-- Brane
