On Wed Mar 19, 2025 at 9:11 PM UTC, Rae Moar wrote:
On Mon, Mar 17, 2025 at 12:13 PM Brendan Jackman jackmanb@google.com wrote:
On Fri, 14 Mar 2025 at 06:37, David Gow davidgow@google.com wrote:
On Fri, 14 Mar 2025 at 03:27, Rae Moar rmoar@google.com wrote:
A bug was identified where the KTAP below caused an infinite loop:
TAP version 13 ok 4 test_case 1..4
The infinite loop was caused by the parser not parsing a test plan if following a test result line.
Fix this bug by parsing test plan line to avoid the infinite loop.
Hi Rae,
With this change and this input:
https://gist.githubusercontent.com/bjackman/220265699f346e16161c6534b115019b...
The infinite loop is gone, but it's still hallucinating a [CRASHED] result:
[16:07:15] # SUMMARY: PASS=17 SKIP=0 FAIL=1 [16:07:15] [CRASHED] ... [16:07:15] Testing complete. Ran 19 tests: passed: 17, failed: 1, crashed: 1
Hi! Thanks for the response. This is an interesting problem. Should a test plan at the bottom cause a crash because no tests were found after? Again with KTAP, a crash would make sense. I feel this example demonstrates why there is a need for a general parser that can parse kselftest output as well as KUnit. I'll see how difficult it would be to change the parser to accommodate removing the crash in a new version. Thanks!
Yeah, if this is difficult, it would be fine for the KUnit tool to just say "sorry this is a KTAP parser not a general TAP parser"
With your new patch:
https://lore.kernel.org/linux-kernel/20250319223351.1517262-1-rmoar@google.c...
I get this:
[10:00:41] # SUMMARY: PASS=17 SKIP=0 FAIL=1 [10:00:41] [ERROR] Test: : No more test results! [10:00:41] [NO TESTS RUN]
Which I think is basically fine.
We do want something that can parse this stuff properly but I tihnk that's a separate little project.
Also, vanilla non-K TAP... is honestly a terrible format. A flat structure for the test results (i.e. no nesting) is a bit of a joke - even for just run_vmtests.sh which has a pretty tractable amount of tests, it's already a pain to mentally keep track of which results come from where. Maybe if I want to be able to parse the output of tests I should just switch them to using KTAP.